COMMUNAUTÉ FRANÇAISE DE BELGIQUE UNIVERSITÉ DE LIÈGE GEMBLOUX AGRO-BIO TECH

Transcription

1 COMMUNAUTÉ FRANÇAISE DE BELGIQUE UNIVERSITÉ DE LIÈGE GEMBLOUX AGRO-BIO TECH CARACTÉRISATION DE LA NICHE ÉDAPHIQUE DES ESPÈCES CUPRO-COBALTICOLES DU KATANGA (R.D. CONGO) EN VUE DE LEUR CONSERVATION ET DE LEUR VALORISATION DANS DES STRATÉGIES DE PHYTOSTABILISATION Sylvain BOISSON Dissertation originale présentée en vue de l obtention du grade de docteur en sciences agronomiques et ingénierie biologique Promoteurs : Grégory MAHY et Mylor NGOY SHUTCHA Année civile : 2016

2

3

4 Au terme de la loi belge du 30 juin 1994 sur le droit d auteur et les droits voisins, toute reproduction du présent document, par quelque procédé que ce soit, ne peut être réalisée qu avec l autorisation de l auteur et de l autorité académiques de Gembloux Agro-Bio Tech (ULg). Le présent document n engage que son auteur.

5 COMMUNAUTÉ FRANÇAISE DE BELGIQUE UNIVERSITÉ DE LIÈGE GEMBLOUX AGRO-BIO TECH CARACTÉRISATION DE LA NICHE ÉDAPHIQUE DES ESPÈCES CUPRO-COBALTICOLES DU KATANGA (R.D. CONGO) EN VUE DE LEUR CONSERVATION ET DE LEUR VALORISATION DANS DES STRATÉGIES DE PHYTOSTABILISATION Sylvain BOISSON Dissertation originale présentée en vue de l obtention du grade de docteur en sciences agronomiques et ingénierie biologique Promoteurs : Grégory MAHY et Mylor NGOY SHUTCHA Année civile : 2016

6 Ce travail a été financé par le Fonds pour la Formation à la la Recherche dans l Industrie et dans l Agriculture (FRS-FNRS)

7 Résumé Les sites métallifères naturels représentent des cas d étude uniques pour la recherche scientifique en géo-pédologie, en biologie et en éco-évolution des métallophytes. Le sud-est de la République Démocratique du Congo (R.D.C.) présente les gisements de cuivre et de cobalt parmi les plus importants au monde et considérés comme des ressources minières de hautes valeurs économiques. Les affleurements sont issus de successions géologiques durant des millions d années. Ils forment aujourd hui une centaine de collines isolées dans une région de 300 km de long et 50 km de large dominée par la forêt claire sèche et nommée l Arc Cuprifère Katangais. Au sommet des collines, la proximité entre la surface du sol et l affleurement entraine un excès de cuivre et cobalt dans la fraction biodiponible de la solution du sol qui est quasi inexistant en bas de pente. Ces teneurs dépassant les limites de tolérance pour la plupart des végétaux constituent un facteur de stress déterminant les communautés et la présence d espèces endémiques. Plus de 550 taxa dont environ 10% identifiés comme endémiques sont répartis dans deux communautés végétales herbacées distinctes identifiées comme des savanes steppiques sur les sols peu minéralisés et des steppes sur les sols fortement minéralisés. L impact des activités minières a aujourd hui des conséquences majeures au niveau environnemental et sociétal menant à la disparition des sites métallifères naturels et à la pollution de sites naturels non-métallifères dont certains sont proches des habitations. Etant donné les menaces sur les espèces végétales endémiques de l Arc Cuprifère Katangais, plusieurs stratégies de conservation in situ et ex situ ont été mises en place au cours de ces 10 dernières années en collaboration avec différentes institutions. En plus, les métallophytes constituent de véritables ressources pour la mise en place de biotechnologies telles que la phytoremédiation en raison de leur capacité de tolérance ou d accumulation. Les sols métallifères anthropogéniques de la région pourraient donc constituer de nouveaux sites conciliant la conservation des métallophytes et la remédiation des sols pollués. Cet enjeu double exige de déterminer les relations entre les espèces et les facteurs environnementaux conditionnant leur distribution dans les sites naturels et leur performance dans les sites anthropogéniques. Cette thèse a pour but de caractériser la niche

8 édaphique des métallophytes de l Arc Cuprifère Katangais afin de proposer des actions associant leur conservation et la phytostabilisation des sols pollués de la région. Dans un premier axe, deux concepts fondamentaux en écologie ont été développés, la théorie des niches écologiques et le concept d endémisme édaphique, afin de caractériser les relations entre les métallophytes et leur environnement édaphique à l échelle des individus, des populations, des espèces et des communautés. Plusieurs espèces endémiques menacées ont été choisies comme modèles d étude afin de déterminer l origine de leur présence quasi-exclusive sur des sols métallifères. Un second axe a testé le succès d introduction de ces espèces dans des stratégies de phytostabilisation en association avec une graminée pionnière Microchloa altera. Une évaluation des autres graminées de cette flore selon les critères de sélection exigés par cette stratégie a été effectuée afin de soumettre des propositions concrètes pour l utilisation des métallophytes de l Arc Cuprifère Katangais dans des stratégies de phytostabilisation des sols pollués. Cette thèse a mis en évidence l existence d une diversité de niches réalisées le long des gradients en cuivre et cobalt au sein des communautés avec une présence de taxa endémiques dans des concentrations en cuivre et cobalt généralement plus élevées que les autres sols de la région. La comparaison avec les niches fondamentales des espèces a montré que certains taxa endémiques seraient dépendants de la présence de cuivre dans le sol, qualifiés d endémiques spécialistes tandis que d autres seraient exclus de la végétation dominante, nommés endémiques refuges. Cette diversité d espèces endémiques et non endémiques présentant des niches écologiques distinctes démontre que ces communautés herbacées recèlent un véritable potentiel pour les conserver dans des habitats pollués. Les essais en conditions contrôlées et en conditions réelles ont d ailleurs démontré la possibilité d associer ces espèces endémiques et non endémiques à des stratégies de phytostabilisation dans la région tout en pointant l importance de leur sélection et des caractéristiques des sites à phytostabiliser. Mots clés : Conservation, Métallophyte, Niche écologique, Endémisme, Phytoremédiation, Sol métallifère, Tolérance

9 Edaphic niche characterization of cuprophyte species from southeastern D.R. Congo for conservation and phytostabilization purposes Natural metalliferous habitats represent unique study cases for scientific research as geopedology, biology and eco-evolution of metallophytes. Southeastern Democratic Republic of Congo hosts among the most important copper and cobalt deposits of the world, having a high economic value. Copper and cobalt outcroups have been formed during billion years by geological successions. More than one hundred hills are now scattered in an area of 300 km long and 50 km large in vegetation matrix consisting on an open forest called Miombo. This area is called Katangan Copperbelt. At the top of hills, the close proximity between the deposit and the soil surface lead to a high concentration of bioavailable copper and cobalt in soil solution exceeding the tolerance threshold of most of plants. More than 550 taxa including 10 % are identified as endemics and compose two distinct herbaceous plant formations: the steppic savanna is found on the lowest minralised soils and the steppe occurs in the highest mineralized soils at the top of hills. Mining activities have dramatic consequences on environement and public health, leading to the destruction of natural meatlliferous habitats and to the creation of anthropogenic polluted sites. Considering the threat on endemic plant species of the Katangan Copperbelt, several in situ and ex situ conservation strategies have been implemented for the last ten years in collaboration with several institutions. Furthermore, metallophytes constitute remarkable resources for the biotechnologies such as phytoremediation due to their metal tolerance or their accumulation ability. Anthropogenic metalliferous sites of the Katangan Copperbelt could be considered for the conservation of metallophyte with a perspective of remediation of polluted soils. Both challenges require the characterization of the environmental factors determining the distribution of plant species in natural sites and their performance in anthropogenic sites. This thesis aims at characterising the edaphic niche of metallophytes from the Katangan Copperbelt in order to propose actions combining the conservation of metallophyte and the phytostabilisation of polluted soils. The first axis focused on the study of two fundamental concepts in ecology, the theory of ecological niche and the concept of edaphic endemism, in order to understand the plant-soil relationships. Threatened endemic species were chosen as study case to determine the cause of the restricted distributions of metallophytes on metalliferous soils. The second axis

10 tested the success of the establishment of endemic metallophytes in phytostabilisation strategies using the grass Microchloa altera. An assessment of the potential of other grasses was performed based on the specific criterions needed in phytostabilisation in order to identify candidate species for conservation phystabilisation strategies. The results of thesis highlight that, among communities, a diversity of realized niches of taxa occurs along the copper and cobalt gradients presenting most of endemic species in the highest copper and cobalt concentrations. Comparison with fundamental niches showed that a part of endemic taxa would be dependant on the presence of copper in soils, called specialist endemics. Other endemic taxa would be excluded from the dominant vegetation, called refuge endemics. The diversity of endemic and non-endemic taxa presenting distinct ecological niches demonstrates that these communities could be conserved on anthropogenic polluted soils. The conservation phystabilisation field tests performed in controlled and in situ conditions pointed out that the species selection and the sites characteritics are critical to ensure the success of a long term strategy. Keywords : Conservation, Metallophyte, Ecological niche, Endemism, Phytoremediation, Metalliferous soil, Tolerance

11 Remerciements C est sans aucun doute un chapitre de la thèse qui retiendra votre attention dès l instant où le document s ouvrira. Bien qu il contienne rarement du contenu scientifique, il est difficile à écrire, à articuler et à structurer. Et c est bien normal puisque son contenu est complexe et se veut complet sans jamais être suffisamment exhaustif. Par respect pour les personnes que je ne remercie pas, je ne les citerai pas. Je commence par remercier les miniers de l Arc Cuprifère Katangais (RDC) pour m avoir laissé un soupçon de populations de cuprophytes afin de les étudier et de les apprécier, sans quoi ma thèse se serait limitée à la page de garde avec, peut-être, des remerciements. Merci pour cette aventure mais s il vous plait laissez-en encore pour les chercheurs suivants et en l honneur de Grégory Mahy, Mylor Ngoy Shutcha, Francois Malaisse, Michel-Pierre Faucon, Maxime Séleck, Soizig Le Stradic, Julie Lebrun, Bastien Lange, Guillaume Delhaye, Wilfried Massengo, Edouard Ilunga, et al.! Je m étais promis de quitter la population endémique de Gembloux Agro-Bio Tech après le mémoire, heureusement que j ai omis de le faire... Deux personnes en sont à l origine: Grégory Mahy et Julie Lebrun. Merci pour cette inception bien orchestrée. Je remercie particulièrement Grégory Mahy pour nous avoir initiés à la vie de Lubumbashi et pour nous avoir fait visiter la ville de Likasi, anciennement Jadotville. Je suis malheureusement dans l obligation de remercier une fois de plus Maxime Séleck, un ardant défenseur de la contradiction. Mais cette fois, je le remercie sincèrement! Merci à Soizig Le Stradic (oui, le prénom est bien Soizig), adhérente du parti opposant de la contraction de Maxime. Les collines sans vous, c est juste bizarre... Le groupe resterait incomplet si Aurélie Hick était vacante, toujours sereine dans la forêt de serpents appelée Miombo. Merci pour ce parcours, ces missions et cette expérience inoubliable en République Démocratique du Congo. Nous devons encore partager nos photos... A ceux à qui je dois des remerciements à la hauteur de leurs personnalités et de leurs travails. Blond crollé et grand brun, ces mémoires étaient faits pour vous. Guindailleurs et travailleurs, vous avez remarquablement relevés les défis. Avec une volonté sans limite, vous avez levé à

12 maintes reprises 25 kg d eau au-dessus des grillages de 3 mètres pour arroser vos plants, parfois après une nuit courte. Les cuprophytes et leurs niches vous doivent beaucoup. Merci à Julien Collignon et Olivier Garin. Vers le sud, en faisant un crochet par Bruxelles, on constate que les populations endémiques, quoi que différentes, ont eu des rôles aussi importants dans cette thèse. Merci à l enthousiasme de Bastien Lange lors de notre unique mission quelque peu fun... J en profite pour saluer Guillaume Delhaye, jamais rencontré à Lubumbashi. Avant de quitter Bruxelles, je souhaite remercier Nathalie Verbruggen, Pietro Salis et les membres du laboratoire de physiologie végétale de l ULB qui m ont chaleureusement accueillis et aidés à maitriser la culture in vitro. Merci à François Malaisse, Gilles Colinet, Michel-Pierre Faucon, Mylor Ngoy Shutcha et Pierre Meerts pour l intérêt qu ils ont apporté à ce travail et à son développement. Au sud, si Lubumbashi et ces environs n ont presque plus de secrets pour moi, c est grâce à Patient Tshihwisa et Mour Inwen. Mes amis, je vous remercie de tout cœur! J espère pouvoir partager ces mêmes moments avec Sylvain 2. Si le Campus Kasapa me parait aujourd hui aussi familier, c est grâce à Henry Muabila, Axel Wetshy, Arsène Mbayo, Serge Langunu, Wilfried Masengo et Jean-Marc. Merci également à toute l équipe de la Faculté d Agronomie dont Mylor Ngoy Shutcha et Edouard Ilunga. Merci à la Faculté de Sociologie de Kasapa pour m avoir autorisé à brancher une balance dans le bureau du secrétaire académique. Merci bien évidemment à l équipe de TFM, Héritier, Jean Jacques, Guylain, Patrick, Michel, Amiral, Rodriguez, Bora et les autres pour votre bonne humeur quotidienne. Merci aux responsables de m avoir autorisé à explorer les collines. La niche écologique des populations endémiques pouvant néanmoins être larges, beaucoup d individus de ces populations n ont pas encore été mentionnés. Je ne ferais pas l affront d oublier Cynthia Thomas. Discrète, tu as largement contribué à la réussite de cette thèse, je t en remercie. Je remercie Mélanie Harzé qui a été la partenaire d une mission pour escalader les collines au péril de... ses chaussures. Merci à mon cher collègue de bureau et ami, William Ortmans de me soutenir pas toujours favorablement dans les périodes de rush. Merci à Carline Pitz d avoir participé au roadtrip copperflora et carrière en Finlande, 30 heures d aventure...pour l aller simple. Merci Marie André pour ton enthousiasme. Je te vois encore présenter le plan de ta thèse en même temps que mon protocole de mémoire. Merci à Thales De Haulleville pour ton bureau, ça nous a vraiment fait plaisir William et moi. Courage à tous

13 pour la fin de vos thèses! Je remercie deux statisticiens l unité. Julien Piqueray qui m a démontré lors de mon mémoire qu une ANOVA 6, ça n existait pas. Je n ai plus fait l erreur. Arnaud Monty pour m avoir expliqué de multiples fois en 10 chapitres et 30 sous-chapitres pourquoi je devais suivre sa logique, que j ai souvent suivie. Merci à Jan Bogaert pour m avoir intégré dans des projets de coopération. Merci à Marc Dufrêne, notre nouveau chef, qui partage la même passion pour les bases de données. Merci à toute l équipe Biodiversité et Paysage et du GP : Florent Bachelar, Laura Maebe, Catherine Montfort, Jessica Delangre, Valentin Gilliaux, William Coos, Emilie Pêcheur, Robin Gailly, Fanny Boraeve, Julie Lebeau et al. La théorie des niches écologiques reste une théorie... Des exceptions demeurent toujours et sont souvent nombreuses. Gembloux, Normandie, Vosges, Baie de Somme, etc. Merci à mes amis de longues dates: Benjamin Delory, François Cuisinier, François Rolin, Maxime Hautôt, Kevin Deronne et vos moitiés respetives. J attends la suite! Benjamin, en réponse à ton message dans tes remerciements, je t attends pour le Kilimanjaro. En attendant, porte toi bien en Allemagne. Merci à ceux qui ont profité avec moi de moments de détentes : Pierre- Yves Barez, Deniz Karadurmus, Thomas Dubrowski, Alexandre Carpentier, Hélène Urbain, Camille Ghesquière, Marie-Noëlle Hamoir, Véronique Breuls, Diego Rios Centeno, Caroline Gérard. Merci à mon ami de très très très longue date Anaël Castelein, le seul français jusque ce paragraphe. Salutations à ma promotion Biotechnologies et Défense des Végétaux Merci à mon entourage proche et très proche pour les services gratuits de relectures, de soutiens, de restaurations durant toutes ces années. Certes, je suis convaincu que sans vous je n aurais jamais tenté l aventure d une thèse. Merci à mes parents Sabine et Ruddy qui m ont toujours poussé à faire mieux... Merci à Eric qui s est toujours intéressé à mon travail. Je remercie ma petite famille, particulièrement Papy Bernard, Mamy Thérèse, Jean-Luc, Mamy Jeannine, Patricia et Papy Roland pour m avoir toujours soutenu dans mes démarches et renforcé dans mes choix. Merci à ma ulbiste de sœur Pauline pour sa motivation régulière et à son compagnon Jules pour les cours de zythologie appliquée. Merci à la famille Bizzotto- Louis dont Ariane et Jean-Louis pour leur disponibilité et leur attention vis-à-vis de mes recherches. Je fais un clin d œil particulier à Guy qui a passé plusieurs jours à relire et à comprendre cette thèse et à Nicole pour son entrain.

14 Ces longs remerciements doivent maintenant se clôturer pour laisser place à la science. Je ne parlerai pas en termes de population mais bel et bien d individu endémique. Je réserve ces derniers remerciements à une personne qui a participé bien plus qu à l élaboration d un projet de thèse. Cette personne a d abord survécu à mon mémoire puis au parcours de thésard entre la Belgique et le Congo, sans oublier les dizaines d imprévus. La synergie entre la difficulté des étapes et mon caractère aurait bien pu basculer la situation à d innombrables moments, mais en vain. Tu es toujours là aoujourd hui. Je te remercie Luivine et ta petite préférée Aloe.

15 Table des matières Résumé... 7 Remerciements Contexte La conservation de la biodiversité, un enjeu international La théorie des niches écologiques et ses controverses Le concept d endémisme édaphique Les éléments traces métalliques (ETM) et les organismes vivants Ecophysiologie des métallophytes : tolérance et accumulation Les sites métallifères naturels, modèles pédologique, écologique et botanique Menaces sur la biodiversité des sites métallifères et opportunités de conservation des metallophytes Références L Arc Cuprifère Katangais, un exemple de site métallifère situé en R. D. Congo Localisation, géologie et phytogéographie Relation plante-sol et endémisme La conservation des cuprophytes Références Objectifs et stratégie scientifique Objectif général et lignes directrices des axes Axe 1 : Niche écologique et endémisme édaphique Axe 2 : Conservation et phytostabilisation Revue de la littérature Article 1. BOISSON Sylvain, NGOY SHUTCHA Mylor & MAHY Grégory (Accepté). Concilier la conservation des métallophytes et la phytoremédiation des sols pollués par les éléments traces métalliques. In : COLINET Gilles, MAHY Grégory & BOGAERT Jan eds. Anthropisation des paysages katangais (R.D.C.). Presses Universitaires de Gembloux Axe 1 Niches écologiques et endémisme édaphique Article 2. BOISSON Sylvain, MONTY Arnaud, SÉLECK Maxime, NGOY SHUTCHA Mylor, FAUCON Michel-Pierre & MAHY Grégory (En préparation) Ecological niche distribution of metallophytes from southeastern D.R.Congo along metal toxicity gradients

16 Article 3. BOISSON Sylvain, MONTY Arnaud, LEBRUN Julie, SÉLECK Maxime & MAHY Grégory (Modifications majeures) Edaphic niches of metallophytes from the Katangan Copperbelt: a key implication for the post mining restoration. Journal for nature conservation Article 4. BOISSON Sylvain, ORTMANS William, MARÉCHAL Justine, MAJÉRUS Marie, MAHY Grégory & MONTY Arnaud (Accepté) No copper required for germination of an endangered endemic species from the Katangan Copperbelt (Katanga, DR Congo): Diplolophium marthozianum. Tropical Ecology Article 5. BOISSON Sylvain, LE STRADIC Soizig, COMMANS Morgane, DUMONT Amandine, LECLERC Natasha, THOMAS Cynthia & MAHY Grégory (2016) Copper tolerance of three Crotalaria species from the Katangan Copperbelt at the early development stage. Biotechnology, Agronomy, Society and Environment Article 6. BOISSON Sylvain, LANGE Bastien, GARIN Olivier, TSHOMBA WETSHY Axel, SÉLECK Maxime, LE STRADIC Soizig, NGOY SHUTCHA Mylor, VERBRUGGEN Nathalie, MAHY Grégory & FAUCON Michel-Pierre (Soumis) Intraspecific variation of copper tolerance of four tropical endemic metallophyte to define new conservation strategies. Plant and Soil Axe 2 Conservation et phytostabilisaton Article 7. BOISSON Sylvain, COLLIGNON Julien, LANGUNU Serge, LEBRUN Julie, NGOY SHUTCHA Mylor & MAHY Grégory (2015) Concilier la phytostabilisation des sols pollués avec la conservation de la flore cupro-cobalticole dans la région de Lubumbashi (R.D. Congo) : une stratégie nouvelle pour valoriser les écosystèmes extrêmes? In : Territoires périurbains : Développement, enjeux et perspectives dans les pays du Sud, eds BOGAERT Jan, HALLEUX Jean-Marie. Presses agronomiques de Gembloux, Gembloux Belgique pp Article 8. BOISSON Sylvain, SÉLECK Maxime, LE STRADIC Soizig, COLLIGNON Julien, GARIN Olivier, NGOY SHUTCHA Mylor & MAHY Grégory (En préparation) Using phytostabilisation as a way to conserve threatened endemic species from the Southeastern D.R. Congo Article 9. BOISSON Sylvain, LE STRADIC Soizig, COLLIGNON Julien, SÉLECK Maxime, MALAISSE François, NGOY SHUTCHA Mylor, FAUCON Michel-Pierre, & MAHY Grégory (2015) Potential of copper tolerant grasses to implement phytostabilisation strategies on polluted soils in South D.R.Congo. Environmental Science and Pollution Research Discussion générale Assemblage des communautés, niches édaphiques et classification des cuprophytes. 237 Une flore spécialiste du cuivre ou généraliste? Conservation et phytostabilisation Références

17 Contexte La conservation de la biodiversité, un enjeu international L attrait des hommes pour l étude de la diversité vivante du monde qui les entoure n est pas nouveau (Darwin, 1859; Grinnell, 1917; Elton, 1927). L étude de la conservation des êtres vivants (cf. Biologie de la conservation, Soulé 1985) est cependant plus récente et anime toujours les débats scientifiques (Soulé, 1985; Sutherland et al., 2009). Le terme biodiversité, venant d une contraction de «biological diversity» (diversité biologique) est apparu il y un peu plus de vingt ans suite à la rédaction d une convention internationale, la Convention pour la Diversité Biologique (CDB, United Nations, 1992). Il désigne la variabilité entre les organismes vivants inclus, entre autres, dans les écosystèmes terrestres, marins et aquatiques et les complexes qui en font parties. Cela comprend la diversité au sein des espèces ainsi qu entre les espèces et les écosystèmes. Les pays ayant ratifié la CDB se sont engagés à reconnaitre la biodiversité comme une valeur intrinsèque, à fournir une évaluation sur son statut, à définir des stratégies de conservation et à l inclure dans des programmes éducatifs et de recherches scientifiques. Trois niveaux hiérarchiques constituent la biodiversité. Premièrement, la diversité d écosystèmes comprend la variété de communautés et de processus écologiques existants. Deuxièmement, la diversité d espèces se réfère au nombre et à la fréquence des espèces. Troisièmement, la diversité génétique 17

18 correspond à des variations dans la composition et la structure des gènes contenus dans toutes les espèces et les individus des espèces sur Terre (United Nations, 1992). Les actions de conservation de la biodiversité sont regroupées au sein de deux types de stratégies complémentaires (Guerrant et al., 2004; Primack, 2014). La conservation in situ vise à préserver la biodiversité en maintenant des communautés ou des populations viables d espèces dans leur environnement naturel. Elle préserve les interactions abiotiques et biotiques ainsi que les processus et fonctions écologiques. La création de zones protégées, qui ont été priorisées au début de la ratification de la CDB, en sont des exemples remarquables (United Nations, 1992). Dans certains cas, la protection des zones ou la restauration de l habitat original des espèces est impossible en raison, par exemple, de la disparition continue de leur habitat. La conservation ex situ prévoit alors de déplacer l espèce en dehors de son habitat naturel, ce qui stoppe les interactions existantes dans l environnement d origine (Engelmann et al., 2002). Cette stratégie exige la création de structures particulières telles que des élevages conservatoires pour les animaux ou des jardins botaniques pour les végétaux. Les végétaux présentent également la particularité de pouvoir être préservés sous forme de graines, de propagules, d organes ou de tissus, ce qui rend cette stratégie particulièrement intéressante pour leur conservation (Engelmann et al., 2002). L implémentation de ces stratégies nécessite la fixation de priorités. L Union Internationale pour la Conservation de la Nature (UICN) a mis en place un système structuré de classification du risque d extinction des espèces (Liste rouge de l'uicn, IUCN, 2015). Ce système est devenu une stratégie internationale reconnue par l Assemblée générale des Nations Unies afin de guider les stratégies de conservation des gouvernements, des organisations non-gouvernementales et des institutions scientifiques. Les catégories attribuées aux espèces sont basées sur des analyses scientifiques rigoureuses de critères quantitatifs et qualitatifs définis par des experts internationaux (IUCN, 2012, 2015). Le contenu scientifique des critères dépend de notions équivalentes à celles étudiées en Biologie de la Conservation: écologie, biogéographie, biologie des populations, etc. (Soulé, 1985; IUCN, 2012). Au sein du groupe menacé (CR : critiquement en danger, EN : en danger et VU : vulnérable), les critères évalués pour identifier la catégorie sont la réduction de la taille de la population (critère A), la répartition géographique (critère B), le nombre d individus des 18

19 petites populations et le déclin (critère C), le nombre d individus des très petites populations (critère D) et l analyse quantitative de la probabilité d extinction dans la nature (critère E). La théorie des niches écologiques et ses controverses L identification et la compréhension des relations entre les espèces et les facteurs biotiques et abiotiques qui déterminent leur distribution sont des questions centrales en écologie et pour la conservation de la biodiversité (Chase et al., 2003). La niche écologique représente la gamme de conditions et de ressources au sein de laquelle un individu ou une espèce peut survivre et se reproduire. Cependant, la théorie des niches écologiques est aussi controversée que centrale en écologie et en Biologie de la conservation. Elle est à la base du courant déterministe de la structure des communautés. Celui-ci considère que les communautés sont des groupes d espèces en interaction dont la présence ou l absence peut être déduite par les règles d assemblage comprenant entre autres les relations interspécifiques et le rôle fonctionnel de chaque espèce (Díaz et al., 1998; Weiher et al., 1998). Chaque membre de la communauté aurait évolué de manière à être le meilleur compétiteur dans sa niche écologique. A l inverse, la théorie neutre de la biodiversité est une approche stochastique qui défend le fait que les communautés sont des groupes d espèces résultant de la dispersion et de la colonisation au hasard sans mécanisme de coexistence. A l échelle de l espèce, Grinnell (1917) et Elton (1927) ont, de manière indépendante, établi deux théories basées sur une approche déterministe pour identifier la relation entre une espèce et son environnement. Grinnell définit la niche (ou niche grinellienne) d une espèce selon ses exigences écologiques et introduit le terme «niche» pour la première fois tandis qu Elton met en évidence le rôle trophique et les fonctions des espèces dans leur environnement (niche etlonienne). Ce dualisme de la niche écologique est resté théorique et exclusivement qualitatif jusqu à la fin des années cinquante. Hutchinson (1957) définit formellement la niche écologique comme un volume à n-dimensions caractérisant l ensemble des n-conditions biotiques et abiotiques requises pour le développement et le maintien d individus viables des populations d une espèce indéfiniment dans le temps (Hutchinson, 1957). Il distingue également deux types de niches écologiques : (1) la niche fondamentale qui représente la gamme de conditions et de ressources dans laquelle des organismes peuvent vivre et se reproduire, appelée également niche physiologique, tandis 19

20 que (2) la niche réalisée est une partie de la niche fondamentale de l espèce en considérant d abord l existence de la compétition (principe d exclusion compétitive, Gause, 1937) et, par extrapolation, l ensemble des facteurs biotiques (p. ex. pathogènes, prédation) (Hutchinson, 1957; Pulliam, 2000; Lau et al., 2008). Chaque facteur environnemental est identifié comme une dimension de l espace, ce qui rend ce concept totalement abstrait à plus de trois dimensions mais étudiable statistiquement et mathématiquement à une, deux ou trois dimensions. Le long d une seule dimension, la représentation de la niche écologique indique la variation de l espèce dans l utilisation d une ressource ou la tolérance à une condition. Graphiquement, elle est représentée par une courbe de réponse de l abondance ou de la performance des individus d une espèce le long d un gradient environnemental (Figure 1). Les paramètres de performance sont généralement des traits liés à la croissance ou la reproduction. En conditions optimales, l abondance ou la performance de l espèce est maximale (optimum de la niche). En conditions de stress, l abondance ou la performance diminue sans être nulle et en conditions létales, le paramètre mesuré devient nul. Cette distribution est caractérisée par une amplitude, qui représente la largeur de la distribution, et par un optimum, point autour duquel l abondance de l espèce est la plus élevée (Figure 1). Cette courbe peut-être gaussienne, tronquée à une des deux extrémités du gradient ou asymétrique (Austin et al., 1994; Lawesson et al., 2002; Coudun et al., 2006). La nature multidimensionnelle de la niche écologique rend sa caractérisation complète difficile, ce qui pousse les auteurs à les étudier selon un facteur à la fois. Plusieurs modèles statistiques ont été proposés et testés pour modéliser les courbes de réponse des espèces le long d un gradient environnemental ou d une combinaison linéaire de facteurs (Rydgren et al., 2003; Coudun et al., 2006; Austin, 2007; Heikkinen et al., 2010). Leur application dépend du nombre de données initiales, du type de données, de l échelle de travail, de l objectif (c.-à-d. modélisation spatiale ou graphiques relationnels) ou du taxon étudié (Austin et al., 1996; Austin, 2007). Le modèle le plus utilisé aujourd hui pour représenter la courbe de réponse d une espèce le long d un gradient environnemental est la régression paramétrique ou non paramétrique (Gégout et al., 1998; Guisan et al., 2000; Austin, 2002, 2007). La détermination de la niche fondamentale d une espèce permet d identifier les besoins physiologiques des individus et leur niveau de tolérance à un facteur de stress afin d estimer le potentiel de survie de l espèce considérée suite à une perturbation des conditions de vie 20

21 d origine (Austin, 1982; Wright et al., 2006; Colles et al., 2009). En général, plus une espèce aura une amplitude de niche fondamentale étroite, plus les individus de cette espèce seront sensibles aux modifications des facteurs abiotiques de leur habitat. Les espèces spécialistes, espèces utilisant une gamme restreinte de ressources, sont actuellement identifiées comme étant plus sujettes au déclin que les espèces généralistes qui ont une large préférence en termes de facteurs écologiques (Colles et al., 2009; Clavel et al., 2011). Cependant, les études montrant une corrélation entre la largeur de niche et l étendue de la distribution sont très peu nombreuses malgré les hypothèses qui la soutiennent (Thompson et al., 2003). Figure 1. Représentation conceptuelle de la niche écologique d une espèce le long d un gradient d un facteur environnemental (courbe de réponse). Trois zones peuvent être identifiées, une zone où les conditions sont optimales dans laquelle l abondance de l espèce ou la performance des individus est maximale en raison (vert), une zone de stress où l abondance de l espèce ou la performance des individus est plus faible car les conditions non-optimales induisent un stress (orange) et une zone où les conditions sont létales dans laquelle peu voire aucun individu n est capable de vivre (rouge). A l échelle des communautés, les concepts sur leur organisation ont évolué et deux théories ont initié le débat. Clements (1916) propose que les espèces composant la communauté soient étroitement associées (concept holistique) et que les limites de distribution de chaque espèce correspondent à la distribution de la communauté dans son ensemble (Figure 2-A). Gleason (1917) indique que les espèces sont distribuées indépendamment les unes des autres dans son concept «individualistique». Les limites de ces communautés ne sont donc pas contraintes par la distribution des espèces qui les composent (Figure 2-B, Gleason, 21

22 1939). Cependant, des observations le long de gradients environnementaux sur de courtes distances dans des écosystèmes montagneux montrent que les communautés sont généralement continues et composées d espèces qui se remplacent graduellement (concept du continuum) au lieu de former des groupes bien distincts (Whittaker, 1956). Le long du gradient de ressource, les espèces occuperaient une niche dans laquelle elles seraient les plus compétitives. Même si les deux premiers concepts perdurent encore, l approche de Gleason (1939) et Whittaker (1956) continue à dominer notre conception des associations d espèces au sein des communautés (Austin, 1985; Oksanen et al., 2002). Cette approche est souvent le point de départ d émission de nouvelles hypothèses à propos de la distribution des niches écologiques des espèces le long d un gradient de ressource (Austin et al., 1989; Oksanen et al., 2002; Cabral et al., 2012). En considérant qu une ressource pourrait être en quantité limitée dans l environnement, les espèces peuvent entrer en compétition pour l exploitation de celle-ci. La théorie du partitionnement des ressources avec ou sans strates a souvent été retenue dans les modèles de distribution des niches écologiques des espèces végétales (Figure 2-C, Austin, 1985; Goldberg et al., 1992; Amarasekare, 2003). Ce type d organisation se base sur le principe d exclusion compétitive de Gause qui, découlant de la loi de Lotka-Volterra, énonce que les espèces partageant des niches écologiques similaires ont tendance à modifier leur niche ou à s exclure compétitivement. En d autres termes, des espèces ayant une niche fondamentale identique pour la même ressource vont entrer en compétition, ce qui mènera à une différenciation de leur niche réalisée ou à la disparition d une des deux espèces dans une communauté stable. Le modèle de partitionnement des ressources est caractérisé par une distribution homogène des niches écologiques avec des courbes gaussiennes qui se succèdent le long du gradient de ressource (Gauch et al., 1972; Austin, 1985). La mesure de l intensité de la compétition peut être réalisée par des mesures de chevauchement ou de densité d optima de niches. Toutefois, l homogénéité, l asymétrie et l organisation des courbes de réponse des niches écologiques des espèces le long d un gradient environnemental sont encore discutées aujourd hui (Austin et al., 1994; Lawesson et al., 2002; Vetaas, 2002; Đorić et al., 2007; Heikkinen et al., 2010). 22

23 Figure 2. Evolution des concepts de l organisation des niches écologiques des espèces au sein d une communauté végétale. A Le concept holistique des communautés, B Le concept individualistique des communautés, C Le concept du partitionnement des ressources. Bien que la théorie des niches écologiques semble pertinente pour comprendre de nombreux cas d étude, elle reste insuffisante pour expliquer la présence de faibles différenciations de niches entre certaines espèces ou les valeurs très élevées de diversité alpha (c.-à-d. nombre d espèces coexistantes) dans certaines communautés végétales. Hubbel (2001) renversa complètement le mode de pensée et construisit la théorie neutre (ou théorie neutre unifiée de biodiversité et de biogéographie) en partant de la théorie de la biogéographie insulaire de MacArthur et Wilson (1963, 1967). Elle supporte que les espèces aient toutes le même niveau de compétition, que la différentiation de niche soit inexistante et que la diversité en espèces soit gouvernée par la dérive. Elle est considérée comme une approche stochastique (c.-à-d. dominée par des processus aléatoires) et est apparentée au modèle neutre de la génétique des populations (Moran, 1958; Kimura, 1968). Cette théorie implique qu il n y aurait pas de structure prévisible dans la distribution des niches de communautés, voire même que le concept de niche est inutile pour comprendre les bases de cette organisation (Chase et al., 2003; Silvertown, 2004; Chase, 2011). Son opposition radicale avec l approche déterministe fait l objet de vifs débats qui scindent en deux groupes bien distincts les chercheurs du domaine (Chase et al., 2003). La théorie neutre semble plutôt adaptée aux 23

24 forêts tropicales ou récifs coralliens tandis que la théorie des niches écologiques concerne plutôt les végétations herbacées (Mikkelson, 2005; Stanley Harpole et al., 2006; Zhou et al., 2008). Pour Rosindell (2011), l approche déterministe semble être à l image de la loi des gaz parfaits et la théorie neutre à celle des gaz réels en thermodynamique. Une étude a récemment montré que ces deux théories pourraient expliquer les étapes de la dynamique des communautés de savanes arbustives. Le processus stochastique serait suffisant pour comprendre la dynamique de dispersion des graines et l établissement des espèces, mais serait progressivement remplacé par un processus déterministe qui expliquerait la structure des espèces et leurs niches écologiques (Stokes et al., 2010). En plus de déterminer la structure des communautés, les niches écologiques des espèces peuvent varier dans l espace (et le temps entre les générations) et au cours du cycle de vie des espèces (Eriksson, 2002; Pearman et al., 2008). Ces variations influencent en conséquence la dynamique des communautés végétales (Turnbull et al., 2007). Au niveau spatio-temporel, des variations de conditions environnementales édaphiques et biotiques comprises dans la niche des espèces peuvent être à l origine d une sélection, conduisant à la formation de génotypes différents résultant d une adaptation locale des populations, appelés écotypes (Wright, Stanton, et al., 2006). Un isolement reproducteur accompagne souvent cette sélection de manière directe ou indirecte, accentue les divergences entre ces écotypes et mène à la formation d une nouvelle espèce (Vekemans et al., 1997; Kay et al., 2010; Sobel et al., 2010). Certains auteurs ont décrit et démontré la possibilité d identifier des variations interpopulations ou la présence d écotypes au sein d une espèce en observant des différenciations de niches écologiques au sein d une espèce (Whittaker et al., 1973; Wright, Davies, et al., 2006; Hájková et al., 2008). A un niveau inférieur, les populations les plus généralistes ont tendance à être composées par des individus ou des groupes d individus spécialistes et non par des individus généralistes (Bolnick et al., 2007; Araújo et al., 2011). A l échelle temporelle, le changement de la niche écologique au cours de la vie des individus d une espèce est nommé changement ontogénétique de la niche. Il consiste en une modification de l amplitude de la niche ou de la position de l optimum ou de la niche le long d un ou plusieurs gradients environnementaux. En général, la caractérisation de ce changement s obtient en analysant la réponse de l espèce à un facteur à plusieurs stades de régénération. Pour certains facteurs comme la luminosité ou l humidité, une expansion de la niche écologique est observée entre le stade germination et le stade adulte tandis qu une contraction de la niche est observée pour les nutriments du sol (Parish et al., 1985; Eriksson, 24

25 2002). Cependant, ces lois ne sont pas universelles et dépendent des espèces étudiées. Deux hypothèses peuvent expliquer le changement ontogénétique de la niche écologique d une espèce. Soit ce changement se produit sur la niche réalisée par une modification de l environnement abiotique ou biotique. Soit les individus de l espèce se sont acclimatés aux conditions abiotiques modifiant la niche fondamentale de l espèce. Cependant, la théorie du conservatisme de niche soutient que la niche fondamentale des espèces reste inchangée dans le temps et dans l espace (Wiens et al., 2005). Le concept d endémisme édaphique Les sols sont les supports des écosystèmes terrestres et leur composition varient dans le paysage. Cinq facteurs principaux déterminent leur formation tels que les roches parents, le climat, la topographie, les organismes vivants et le temps (Barcenas-Arguello et al., 2013). Ils sont constitués de solides minéraux et organiques, de liquides et de gaz (Baize, 2004). Une espèce endémique est définie comme une espèce étant distribuée dans une aire géographique très limitée (Kruckeberg et al., 1985). Lorsque cet endémisme est lié au conditions physico-chimiques du sol, on utilise le terme endémisme édaphique (Mason, 1946). Tandis que le climat est un facteur limitant la distribution des espèces à l échelle de la Terre, les caractéristiques géologiques d une région influencent la diversité d habitat (Kruckeberg, 1985; Brady et al., 1996). Les habitats présentant des sols avec des conditions environnementales extrêmes présentent souvent des espèces endémiques (Rajakaruna, 2004). On entend par conditions extrêmes, les sols étant chimiquement mal balancés se trouvant par exemple sur de la serpentine, la dolomie, le gypse ou les sols métallifères. La caractérisation de cet endémisme est une étape permettant de déterminer les conditions de conservation des espèces et la relation entre l environnement édaphique et les taxa (Ba rcenas-arguello et al., 2013). Au point de vue évolutif, l endémisme édaphique peut être originaire de la dérive génétique, la spéciation allopatrique ou la différenciation génétique d individus de ces populations par sélection et l isolement reproducteur (Kruckeberg et al., 1985; Vekemans et al., 1997; Macnair et al., 1998; Rajakaruna, 2004). Dans les écosystèmes métallifères, les taxa endémiques sont souvent apparentés à des espèces plus communes vivant sur des sols non-métallifères adjacents (Duvigneaud, 1958; Jaffré, 1992; Millie Burrell et al., 2012; Faucon et al., 2015). Le modèle classique entend que ces taxa, originellement sur des sols 25

26 non-métallifères, aient produit des individus préadaptés à des conditions intermédiaires (p. ex. plus large amplitude de niche écologique) et par sélection divergente, une différentiation écotypique s est produite. Un isolement reproducteur fort a ensuite renforcé la différentiation gentypique jusqu à la formation d une espèce (Kruckeberg et al., 1985). Dans la littérature plusieurs études comparent l écologie et les génotypes populations d espèces modèles vivant ou non sur des sols aux conditions chimiquement mal balancées afin de comprendre les mécanismes de spéciation (Meerts et al., 1997; Bert et al., 2000; Dubois et al., 2003; Turisová et al., 2013). Au point de vue écologique, trois modèles existent pour comprendre l endémisme édaphique des taxa. Le premier est que les espèces endémiques sont moins compétitives sur les sols plus abondants aux caractéristiques différentes, ce qui les exclus de la végétation dominante vers de conditions de stress qu elles tolèrent (Gankin et al., 1964). L optimum écologique de la niche réalisée le long de ce facteur ne sera pas identique à celui de la niche fondamentale. Ce modèle correspond aux écosystèmes serpentiniques qui sont des habitats aux sols riches en nickel (Ni), en fer (Fe) et en magnésium (Mg) (Rajakaruna, 2004; Anacker, 2014). Le second modèle défend que les taxa sont spécialisés aux sols chimiquement mal balancés et sont plus compétitifs que les autres espèces sur ceux-ci et qu ils dépendent physiologiquement d un ou plusieurs éléments du sol. L optimum écologique de la niche réalisée le long de ce facteur ne sera pas identique à celui de la niche fondamentale. Ce modèle est appelé spécialiste. Finalement, le dernier modèle est plutôt spécifique des sols métallifères et défend que les taxa endémiques sont restreints aux sols métallifères à cause de la plus faible pression de pathogènes dans les sols métallifères comparativement aux autres sols. Ces taxa auraient perdu leur mécanisme de défense et seraient donc plus sensibles à ceux-ci (Boyd, 2007; Fones et al., 2010). Ce modèle pourrait être assimilé à un modèle refuge plus large que la simple compétition végétale. 26

27 Les éléments traces métalliques (ETM) et les organismes vivants Les éléments traces métalliques (ou ETM) sont des métaux qui sont présents naturellement à l état de trace dans le sol, l eau et l air et dont la teneur est inférieure à 0,1 % dans la croute terrestre (Baize, 1997, 2004). Le fer (Fe), le plomb (Pb), le cuivre (Cu), le mercure (Hg), le nickel (Ni), le zinc (Zn) ou le cobalt (Co) constituent par exemple le groupe des ETM. Dans les sols, les ETM peuvent être d origine naturelle dû à leur présence dans les roches-mères, les apports éoliens, le fonctionnement des cycles bio-géochimiques et les redistributions par pédogenèse ou morphogenèse. Ils peuvent également être d origine anthropique en provenant des activités agricoles, industrielles ou domestiques (Baize, 1997). Contrairement aux molécules organiques, les ETM ne sont pas dégradables, s accumulent souvent à la surface des sols (Baize, 1997; Liénard et al., 2014) et ne se trouvent pas tous sous la même forme dans l espace et dans le temps (Dube et al., 2001; Carrillo González et al., 2006). En général, cinq fractions du pool total d ETM présent dans un sol sont identifiées même si les techniques d extraction sont discutables : la fraction échangeable, la fraction adsorbée ou liée aux carbonates, la fraction liée à la matière organique, les fractions liées aux oxyhydroxydes et la fraction résiduelle (Brown et al., 1984; McLean et al., 1992; Cornu et al., 2000; Kabala et al., 2001; Liénard et al., 2014). Comme son nom l indique, la fraction échangeable est en perpétuelle interaction avec les autres fractions. Elle est constituée de la partie de l ETM libre et complexée avec des matières organiques ou inorganiques dans la solution du sol. Une quantité de cette fraction est susceptible de passer directement dans les organismes vivants. Cet état spécifique est nommé biodisponible. Cependant, la quantité biodisponible en ETM dans le sol est difficile à identifier précisément et dépend des techniques d extraction utilisées (Brown et al., 1984; Liénard et al., 2014). Elle varie selon le ph, la capacité d échange cationique, la teneur en matière organique, la quantité et le type d argile, les oxydes de Fe et Mn et le potentiel redox (Ernst, 1996; Carrillo González et al., 2006; Agbenin et al., 2012). Au sein du règne vivant, les ETM peuvent être classés dans deux catégories : les éléments essentiels de type micronutriments, ou non essentiels (Singh, 2005; Pilon-Smits et al., 2009; Kabata-Pendias, 2010). Si un manque en élément essentiel entraîne une déficience pour les végétaux, un excès de cet élément engendrera un phénomène de toxicité. Il est à noter que le seuil de phytotoxicité sera beaucoup plus faible pour les micronutriments (Fe, Mn, Cu, Ni ou 27

28 Zn) présents à moins de 100 mg kg -1 matière sèche végétal (MSv) que pour les macroéléments (C, H, N, P, K, Mg ou Ca présents à plus de mg kg -1 MSv) essentiels. A l inverse, les ETM non essentiels (Cd, Pb ou Hg) sont phytotoxiques à très faibles concentrations (Pilon-Smits et al., 2009). Le terme «éléments traces potentiellement toxiques» remplacerait le terme «métaux lourds», qui est également utilisé lorsque la densité de métaux ou métalloïdes toxiques est supérieure à 5 g cm -3 (Furini, 2012). Au cours du 20 ème siècle, la demande en métaux pour la construction des technologies nouvelles a considérablement augmenté, multipliant les sources anthropogéniques de pollution (Alloway, 1995; Pacyna et al., 2001; Järup, 2003; Singh et al., 2011). Ces activités mènent à l accumulation d ETM dans des écosystèmes aux sols normalement équilibrés, ce qui conduit à une exposition importante des espèces qui les composent. Lorsqu ils se retrouvent en excès dans les êtres vivants, les ETM perturbent leur physiologie et leur métabolisme et entrent dans la chaîne alimentaire par bioaccumulation (Singh, 2005; Manara, 2012). Le Cd, le Pb et le Hg ont d ailleurs été identifiés comme agents chimiques de préoccupation majeure à l échelle mondiale pour la santé humaine par l Organisation Mondiale de la Santé (OMS) même si l écotoxicologie de tous les ETM n est pas encore complètement connue (Pacyna et al., 2001; World Health Organisation, 2016). Les végétaux sont des organismes sessiles qui absorbent principalement les éléments par voie racinaire. Les éléments diffusent d abord par voie apoplasmique puis le transport peut devenir actif au niveau de l endoderme (voie symplasmique) (Pilon Smits et al., 2009). L excès en ETM essentiels ou en ETM non essentiels dans les cellules végétales entraine une production de formes réactives d oxygène («Reactive oxygen species», ROS) qui altèrent des substrats biologiques, modifiant des domaines fonctionnels de molécules. Ces modifications mènent à un bouleversement de la photosynthèse, de la respiration cellulaire, de l activité enzymatique, de la nutrition minérale ou de l osmose pouvant conduire à la mort cellulaire (Stohs et al., 1995; Ryser et al., 2006; Mijovilovich et al., 2009; Furini, 2012; Manara, 2012; Ali, Farooq, et al., 2013; Li et al., 2013). En conséquence, un stress est identifiable par une modification du phénotype, une diminution de la croissance de l individu ou des organes, une diminution de la production de biomasse ou la mort de l individu. 28

29 Ecophysiologie des métallophytes : tolérance et accumulation La tolérance est définie comme la capacité d une plante à survivre en présence de concentrations en ETM létales ou ayant des effets néfastes sur d autres plantes de la même espèce ou d espèces différentes (Antonovics et al., 1971; Shaw, 1990; Maestri et al., 2012). Selon Baker (1987), les végétaux dotés de mécanismes de tolérance contre les ETM se comportent comme des métallophytes. Ils sont retrouvés naturellement sur des substrats riches en ETM, qualifiés de métallifères (Antonovics et al., 1971; Baker, 1987). La tolérance aux ETM implique une activité différentielle des gènes qui régulent l activité enzymatique des végétaux (Jonak et al., 2004; Yang et al., 2005; Maestri et al., 2012; Millie Burrell et al., 2012). Plusieurs mécanismes physiologiques de résistance ont été identifiés contre l excès en ETM dans les cellules ou l organisme végétal comme la compartimentation qui consiste à séquestrer les ETM dans des tissus moins sensibles, l excrétion hors de la plante dans des cristaux salins et la chélation qui permet de diminuer leur disponibilité dans la cellule (intracellulaire) ou dans le sol (extracellulaire) par liaison avec des molécules végétales (Macnair, 1987; Ernst, 2006; Mehes-Smith et al., 2013; Redondo-Gomez, 2013). Plusieurs modèles végétaux sont utilisés pour déterminer avec précision ces mécanismes mais les processus ne sont pas encore déterminés pour tous les ETM (Marquès et al., 2004; Van Rossum et al., 2004; Hinsinger et al., 2008; Lasat, 2009; Krämer, 2010). Plusieurs processus sont spécifiques à certains métaux comme les métallothionéines impliquées pour la tolérance au Cu de Silene vulgaris et S. paradoxa (Schat et al., 1997; Mengoni et al., 2000) ou encore l activité de translocation du Cd et du Zn des racines vers les feuilles de Noccaea caerulescens (anciennement Thlaspi caerulescens) qui peut différer selon les écotypes (Assunção, Schat, et al., 2003; Marquès et al., 2004). Au contraire, il a été démontré que, chez Arabidopsis halleri, les gènes activés lors de la carence cellulaire en Zn dans les racines codent pour des transporteurs qui peuvent également transporter le Cd (Hanikenne et al., 2008). Trois stratégies de tolérance aux ETM ont été identifiées chez les métallophytes (Baker, 1981) : l exclusion est le mécanisme le plus répandu qui consiste à limiter l entrée des ETM dans les parties aériennes afin de maintenir une faible concentration dans la plante sur un large gradient de concentration en ETM dans le sol ; l indication est un comportement caractérisé par une accumulation des ETM dans les tissus végétaux proportionnellement à 29

30 leur concentrations dans le sol ; l accumulation consiste à accumuler les ETM dans les tissus végétaux à des concentrations plus que proportionnellement à celles du sol (Figure 3, Baker, 1981). Au-delà de 100 mg g -1 à mg g -1 ETM (0,01 % à 1 % de MSv) accumulés dans les partie aériennes, le terme «hyperaccumulateur» peut définir la stratégie du taxon (hyper)tolérant (Verbruggen et al., 2009; Krämer, 2010). A l origine de l hyperaccumulation, les études ont révélé une stimulation du prélèvement d ETM du sol, une augmentation de la translocation via le xylème et des mécanismes de détoxification accrus dans les tissus aériens. Les mécanismes impliqués font intervenir des gènes et ne sont encore pas tous identifiés (Verbruggen et al., 2009; Krämer, 2010). Cela ne concerne qu environ 0,2 % des angiospermes, soit environ 500 taxa (Krämer, 2010).L (hyper)accumulation des ETM dans les tissus peut également contribuer à la défense contre les herbivores ou les pathogènes (Behmer et al., 2005; Poschenrieder et al., 2006; Fones et al., 2010). Figure 3. Stratégies de tolérance des métallophytes selon Baker (1981) en lien avec la concentration en ETM dans les tissus végétaux et la concentration en ETM dans le sol. En général, les métallophytes sont également adaptés à d autres facteurs physicochimiques extrêmes des sols (Baker, 1987) comme la xéricité du sol (Schat et al., 1997), les variations en Ca et la disponibilité en Fe (Lombi et al., 2002). Plusieurs métallophytes présentent des relations symbiotiques avec des (ecto)mycorhizes arbusculaires qui piègent les ETM dans leurs cellules avant qu ils pénètrent dans les cellules végétales (Hall, 2002; Khan, 2005; Alford et al., 2010). Les essais expérimentaux sur le rôle exact des 30

31 ectomycorhizes dans la tolérance sont peu nombreux. Une étude a révélé que les ectomycorhizes présentes sur les racines de l espèce Viola calaminaria pouvaient coloniser les racines de maïs (Zea mays L.) et améliorer le rendement sur des sols pollués en ETM (Kaldorf et al., 1999). La tolérance aux ETM en excès présents dans le sol a un coût métabolique dû aux mécanismes de défense mobilisés, pouvant diminuer la biomasse de la plante et la production de graines (Wilson, 1988; Ernst, 2006). Les sites métallifères naturels, modèles pédologique, écologique et botanique Les sites métallifères naturels présentent des sols formés sur des affleurements de roches métallifères (p. ex. minerais métallifères) qui, en s altérant par érosion, entretiennent les teneurs élevées en ETM des sols (Alexander et al., 2006). Ces lieux réunissent des conditions de stress environnemental élevés pour les êtres vivants. Celles-ci sont liées à la présence de hautes concentrations en ETM présents dans le sol mais sont aussi associés à d autres contraintes écologiques comme la xéricité (Antonovics et al., 1971). Souvent de petites superficies, les sites métallifères se retrouvent généralement isolés dans le paysage entre des écosystèmes aux sols non-métallifères (Baker et al., 2010). L assemblage local des taxa d un site métallifère constitue un sous ensemble plus ou moins caractéristique du pool de taxa régional ayant passé une succession de filtres écologiques (Macnair, 1987; Mengoni et al., 2000; Whiting et al., 2002, 2004; Baker et al., 2010; Harrison et al., 2011). Selon certains traits d histoire de vie (c.-à-d. caractères innés et acquis par un organisme au cours de l évolution), le taxon préadapté passera les filtres de dispersion, les filtres abiotiques et les filtres biotiques pour coloniser le site (Cristofoli et al., 2010). Harrisson et al. (2004) a d ailleurs démontré que la diversité floristique de la région contribue à la diversité des communautés végétales des sites métallifères. Au niveau temporel, les taxa établis sont soumis à des pressions de sélection intenses à cause de la concentration élevée en ETM du sol et à la dérive génétique, pouvant mener à la différenciation d écotypes voire de taxa rares, spécialisés ou endémiques (Kruckeberg et al., 1985; Meerts et al., 1997; Vekemans et al., 1997; Wolf et al., 2000; Mengoni et al., 2000; Assunção, Bookum, et al., 2003; Van Rossum et al., 2004; Safford et al., 2005; Bizoux et al., 2008; Faucon et al., 2010; Kay et al., 2010). Parmi les métallophytes présents dans ces communautés, on distingue les métallophytes absolus qui sont des taxa tolérant certains ETM et dont la distribution est strictement limitée aux sols 31

32 métallifères ainsi que des métallophytes facultatifs étant tolérants à certains ETM et présentant également des populations sur les sols non-métallifères (Baker et al., 2010). Menaces sur la biodiversité des sites métallifères et opportunités de conservation des metallophytes Les sites métallifères naturels et leur biodiversité sont en réduction continue en nombre et en surface en raison de l exploitation des ETM (Shaw, 1990; Cooke et al., 2002; Whiting et al., 2002, 2004). L excavation (dans le cas des mines à ciel ouvert) représente une perturbation majeure pour l écosystème jusqu à un état si différent de l initial que la résilience de l écosystème ne peut être envisagée (Gunderson, 2000). Les actions de conservation doivent donc débuter avant l exploitation afin de cibler les métallophytes de valeur conservatoire et de préparer la restauration de l écosystème. Après exploitation, la recolonisation naturelle des zones dégradées par les espèces du pool régional leur demande de repasser les filtres écologiques. Ce processus est donc lent, surtout si le sol est absent. La restauration d écosystème par reconstruction artificielle peut être envisagée en apportant un sol et des espèces préadaptées aux conditions abiotiques du site (Bradshaw, 1983; Cooke et al., 2002). Ce type de stratégie est à l intersection entre la Biologie de la conservation et l Ecologie de la restauration (Dobson et al., 1997) et nécessite de connaitre l écologie et la biologie des métallophytes (Smith, 1979; Bizoux et al., 2004; Jacobi et al., 2008). En parallèle avec cette situation, des sites métallifères anthropogéniques sont créés par les activités minières (Ernst, 1990; Pacyna et al., 2001; Whiting et al., 2002; Blacksmith Institute, 2007; Baker et al., 2010). Les industries minières rejettent dans l atmosphère et dans l eau des déchets contenant des ETM et d autres polluants qui se dispersent dans l environnement et s accumulent dans les écosystèmes. Dans les sols pollués, les changements physico-chimiques sont plus rapides que dans les sites métallifères naturels, induisant des modifications dans la structure et la fonction des écosystèmes (Dube et al., 2001; Wong, 2003; Kozlov et al., 2006; Ryser et al., 2006; Mahmood et al., 2007; Street et al., 2007; Di Salvatore et al., 2008). Les espèces végétales des sites non métallifères subissent alors un stress dû aux ETM et disparaissent pour la plupart. Les sols nus sont ainsi sensibles à l érosion et émettent des particules métallifères dans l environnement pouvant conduire à 32

33 des problèmes de santé chez l Homme (Vangronsveld et al., 1995; Lasat, 2000; Kozlov et al., 2006; Banza et al., 2009; Escarré et al., 2011; Cheyns et al., 2014). Limiter la disponibilité ou réduire la quantité d ETM dans les sols pollués grâce aux biotechnologies végétales et via les métallophytes est une action répondant aux objectifs de la CDB. Notamment, la phytoremédiation inclut un ensemble de méthodes dont l objectif est de réduire, déplacer, dégrader ou immobiliser les composés toxiques dans les sols pollués grâce à la mise en place d une végétation (Cunningham et al., 1995; Peer et al., 2005). Elle comprend des méthodes adaptées pour les polluants organiques (p.ex. la phytodégradation) ou minéraux. Entre autres, la phytostabilisation consiste à diminuer la mobilité des ETM grâce à des espèces natives, exclueuses et possédant une rhizosphère dense tandis que la phytoextraction exige l utilisation d espèces accumulatrices voire hyperaccumulatrices afin d extraire les ETM, par exemple en chimie verte (Berti et al., 2000; Singh et al., 2007; Mendez et al., 2008; Ali, Khan, et al., 2013; Escande et al., 2014). Pour les sites aux sols les plus pollués, la phytostabilisation est reconnue comme une méthode efficace et aisée à mettre en place pour restaurer les fonctions d un écosystème complètement dégradé (Salt et al., 1995; Dubourguier et al., 2001; Mench et al., 2006). La mise en place d espèces en vue de phytostabiliser les sols exige d étudier les relations interspécifique, et entre les espèces et leur environnement (p. ex. niche écologique) (Frérot et al., 2006; Parraga-Aguado et al., 2013). A long terme, ce système végétal pionnier peut promouvoir l installation d autres espèces ou d écotypes tolérants (Parraga-Aguado et al., 2014). Les taxa candidats doivent être natifs de la région pour tolérer les conditions climatiques locales de l environnement (Mendez et al., 2008). Même si certaines études ont pointé l usage d espèces hyperaccumulatrices en phytoremédiation, peu d entre elles ont testé l association entre des espèces candidates et des métallophytes endémiques ou rares (Lone et al., 2008). Les milieux pollués en ETM peuvent pourtant représenter une opportunité pour la conservation des métallophytes menacées en les associant aux stratégies de phytoremédiation (Whiting et al., 2004; Bizoux et al., 2008). 33

34 Références Agbenin J.O. & Welp G., Bioavailability of copper, cadmium, zinc, and lead in tropical savanna soils assessed by diffusive gradient in thin films (DGT) and ion exchange resin membranes. Environ. Monit. Assess. 184(4), Alexander E.B., Coleman R.G., Keeler-Wolf T. & Harrison S.P., Serpentine Geoecology of Western North America : Geology, Soils, and Vegetation, Serpentine Geoecology of Western North America: Geology, Soils, and Vegetation, Oxford University Press, USA, New York, 528. Alford É.R., Pilon-Smits E.A.H. & Paschke M.W., Metallophytes a view from the rhizosphere. Plant Soil 337(1-2), Ali H., Khan E. & Sajad M.A., Phytoremediation of heavy metals-concepts and applications. Chemosphere 91(7), Ali S., Farooq M.A., Yasmeen T., Hussain S., Arif M.S., Abbas F., Bharwana S.A. & Zhang G., The influence of silicon on barley growth, photosynthesis and ultrastructure under chromium stress. Ecotoxicol. Environ. Saf. 89(2013), Alloway B.J., Heavy Metals in Soils, Blackie Academic & Professional. Amarasekare P., Competitive coexistence in spatially structured environments: a synthesis. Ecol. Lett. 6(12), Anacker B.L., The nature of serpentine endemism. Am. J. Bot. 101(2), Antonovics J., Bradshaw A.D. & Turner R.G., Heavy metal tolerance in plants. Adv. Ecol. Res. 7, Araújo M.S., Bolnick D.I. & Layman C. a, The ecological causes of individual specialisation. Ecol. Lett. 14(9), Assunção a. G.L., Bookum W.M., Nelissen H.J.M., Vooijs R., Schat H. & Ernst W.H.O., Differential metalspecific tolerance and accumulation patterns among Thlaspi caerulescens populations originating from different soil types. New Phytol. 159(2), Assunção A., Schat H. & Aarts M., Thlaspi caerulescens, an attractive model species to study heavy metal hyperaccumulation in plants. New Phytol. (1992), Austin M., Use of a relative physiological performance value in the prediction of performance in multispecies mixtures from monoculture performance. J. Ecol. 70(2), Austin M., Continuum concept, ordination methods, and niche theory. Annu. Rev. Ecol. Syst. 16(1985), Austin M., Species distribution models and ecological theory: A critical assessment and some possible new approaches. Ecol. Modell. 200(1-2), Austin M. & Smith T., A new model for the continuum concept. Vegetatio 83, Austin M.., Spatial prediction of species distribution: an interface between ecological theory and statistical modelling. Ecol. Modell. 157(2-3), Austin M.P. & Gaywood M.J., Current problems of environmental gradients and species response curves in relation to continuum theory. J. Veg. Sci. 5(4), Austin M.P. & Meyers J. a., Current approaches to modelling the environmental niche of eucalypts: implication for management of forest biodiversity. For. Ecol. Manage. 85(1-3), Baize D., Teneurs totales en éléments traces métalliques dans les sols (France), INRA Editions, Un Point sur - Institut national de la recherche agronomique, Quae, 409. Baize D., Petit lexique de pédologie, Dictionnaires, Quae, 271. Baker A.J.M., Accumulators and excluders strategies in the response of plants to heavy metals. J. Plant Nutr. 3(1-4), Baker A.J.M., Metal Tolerance. New Phytol. 106, Baker A.J.M., Ernst W.H.O., Van der Ent A., Malaisse F. & Ginocchio R., Metallophytes: the unique biological resource, its ecology and conservational status in Europe, central Africa and Latin America. In: Batty, L., Hallberg, K. eds. Ecology of Industrial Pollution Banza C.L.N., Nawrot T.S., Haufroid V., Decrée S., De Putter T., Smolders E., Kabyla B.I., Luboya O.N., Ilunga A.N., Mutombo A.M. & Nemery B., High human exposure to cobalt and other metals in Katanga, a mining area of the Democratic Republic of Congo. Environ. Res. 109(6), Barcenas-Arguello M.L., Gutierrez-Castorena M. del C. & Terrazas T., The Role of Soil Properties in Plant Endemism A Revision of Conservation Strategies. In: Soil Processes and Current Trends in Quality Assessment. InTech, Behmer S., Lloyd C., Raubenheimer D., Stewart-Clark J., Knight J., Leighton R., Harper F. & Smith J., Metal hyperaccumulation in plants: mechanisms of defence against insect herbivores. Funct. 19, Bert V., Macnair M.R., De Laguerie P., Saumitou-Laprade P. & Petit D., Zinc tolerance and accumulation in metallicolous and nonmetallicolous populations of Arabidopsis halleri (Brassicaceae). New Phytol. 146(2), Berti W.R. & Cunningham S.D., Phytostabilization of metals. In: Raskin, I., Ensley, B.D. eds. Phytoremediation of Toxic Metals: Using Plants to Cleanup the Environment. John Wiley & Sons, Inc., New York, Bizoux J., Brevers F., Meerts P., Graitson E. & Mahy G., Ecology and conservation of Belgian populations 34

35 of Viola calaminaria, a metallophyte with a restricted geographic distribution. Belgian J. Bot. 137(1), Bizoux J.P., Daïnou K., Raspé O., Lutts S. & Mahy G., Fitness and genetic variation of Viola calaminaria, an endemic metallophyte: Implications of population structure and history. Plant Biol. 10(6), Blacksmith Institute, The World s Worst Polluted Places: The Top Ten of The Dirty Thirty, New York, 74. Bolnick D.I., Svanbäck R., Araújo M.S. & Persson L., Comparative support for the niche variation hypothesis that more generalized populations also are more heterogeneous. Proc. Natl. Acad. Sci. U. S. A. 104(24), Boyd R.S., The defense hypothesis of elemental hyperaccumulation: status, challenges and new directions. Plant Soil 293(1-2), Bradshaw A., The reconstruction of ecosystems. J. Appl. Ecol. 20, Brady N.C. & Weil R.R., The nature and properties of soils, Prentice Hall. Brown R.M., Pickford C.J. & Davison W.L., Speciation of Metals in Soils. Int. J. Environ. Anal. Chem. 18(3), Cabral J.S. & Kreft H., Linking ecological niche, community ecology and biogeography: insights from a mechanistic niche model. J. Biogeogr. n/a n/a. Carrillo González R., Šimůnek J., Sauvé S. & Adriano D., Mechanisms and Pathways of Trace Element Mobility in Soils. Adv. Agron. 91(06), Chase J., Ecological niche theory. In: Scheiner, S.M., Willig, M.R. eds. The Theory of Ecology. University of Chicago Press, Chicago, 404. Chase J.M. & Leibold M.A., Ecological niches: linking classical and contemporary approaches, Biodiversity & Conservation, Interspecific interactions. Thompson, J.N. ed., University of Chicago Press, 212. Cheyns K., Banza Lubaba Nkulu C., Ngombe L.K., Asosa J.N., Haufroid V., De Putter T., Nawrot T., Kimpanga C.M., Numbi O.L., Ilunga B.K., Nemery B. & Smolders E., Pathways of human exposure to cobalt in Katanga, a mining area of the DR Congo. Sci. Total Environ. 490, Clavel J., Julliard R. & Devictor V., Worldwide decline of specialist species: toward a global functional homogenization? Front. Ecol. Environ. 9(4), Colles A., Liow L.H. & Prinzing A., Are specialists at risk under environmental change? Neoecological, paleoecological and phylogenetic approaches. Ecol. Lett. 12(8), Cooke J. a & Johnson M.S., Ecological restoration of land with particular reference to the mining of metals and industrial minerals: A review of theory and practice. Environ. Rev. 10(1), Cornu S. & Clozel B., Extractions séquentielles et spéciation des éléments trace métalliques dans les sols naturels. Analyse critique. Etude Gest. des Sols 7(3), Coudun C. & Gégout J.-C., The derivation of species response curves with Gaussian logistic regression is sensitive to sampling intensity and curve characteristics. Ecol. Modell. 199(2), Cristofoli S. & Mahy G., Restauration écologique: Contexte, contraintes et indicateurs de suivi. Biotechnol. Agron. Soc. Environ. 14(1), Cunningham S.D., Berti W.R. & Huang J.W., Phytoremediation of contaminated soils. Trends Biotechnol. 13, Darwin C., On the origin of species by means of natural selection, or the preservation of favoured races in the struggle for life. Murray, J. ed., John Murray, London, UK, 490. Di Salvatore M., Carafa A. & Carratù G., Assessment of heavy metals phytotoxicity using seed germination and root elongation tests: a comparison of two growth substrates. Chemosphere 73(9), Díaz S., Cabido M. & Casanoves F., Plant functional traits and environmental filters at a regional scale. J. Veg. Sci. 9, Dobson a P., Bradshaw a D. & Baker a J.M., Hopes for the future: Restoration ecology and conservation biology. Science (80-. ). 277(5325), Đorić D., Nikolić-Đorić E., Jevremović V. & Mališić J., On measuring skewness and kurtosis. Qual. Quant. 43(3), Dube A., Zbytniewski R., Kowalkowski T., Cukrowska E. & Buszewski B., Adsorption and Migration of Heavy Metals in Soil. Polish J. Environ. Stud. 10(1), Dubois S., Cheptou P.-O., Petit C., Meerts P., Poncelet M., Vekemans X., Lefebvre C. & Escarré J., Genetic structure and mating systems of metallicolous and nonmetallicolous populations of Thlaspi caerulescens. New Phytol. 157(3), Dubourguier H.-C., Petit D., Deram A. & Logeay C., Le phytomanagement: Eléments de synthèse, Pôle de compétence sites et sédiments pollués, Pôle de compétence sites et sédiments pollués, Lille, 53. Duvigneaud P., Etudes sur la végétation du Katanga et de ses sols métallifères. Communication n 1 La végétation du Katanga et de ses sols métallifères. Bull. Soc. Roy. Bot. Belgique 90(2), Elton S.C., Animal ecology. Huxley, J. ed., London, 207. Engelmann F. & Engels J., Technologies and strategies for ex situ conservation. In: Ramanatha, R. V., Brown, A.H.D., Jackson, M. eds. Managing Plant Genetic Diversity Eriksson O., Ontogenetic niche shifts and their implications for recruitment in three clonal Vaccinium shrubs: Vaccinium myrtillus, Vaccinium vitis-idaea, and Vaccinium oxycoccos. Can. J. Bot. 80, Ernst W.H.O., Mine vegetation in Europe. In: Shaw, J. ed. Heavy Metal Tolerance in Plants: Evolutionary Aspects. CRC Press Inc, New York,

36 Ernst W.H.O., Bioavailability of heavy metals and decontamination of soils by plants. Appl. geochemistry 11, Ernst W.H.O., Evolution of metal tolerance in higher plants. For. Snow Landsc. Res. 80(3), Escande V., Garoux L., Grison C., Thillier Y., Debart F., Vasseur J.-J., Boulanger C. & Grison C., Ecological catalysis and phytoextraction: Symbiosis for future. Appl. Catal. B Environ. 146, Escarré J., Lefèbvre C., Raboyeau S., Dossantos A., Gruber W., Cleyet Marel J.C., Frérot H., Noret N., Mahieu S., Collin C. & Van Oort F., Heavy metal concentration survey in soils and plants of the Les Malines Mining District (southern France): Implications for soil restoration. Water. Air. Soil Pollut. 216(1-4), Faucon M., Le Stradic S., Boisson S., Ilunga wa Ilunga E., Séleck M., Lange B., Delhaye G., Shutcha M.N., Pourret O., Meerts P. & Mahy G., Implication of plant-soil relationships for conservation and restoration of coppercobalt ecosystems. Plant Soil (November 2015). Faucon M.-P., Meersseman A., Shutcha M.N., Mahy G., Luhembwe M.N., Malaisse F. & Meerts P., Copper endemism in the Congolese flora: a database of copper affinity and conservational value of cuprophytes. Plant Ecol. Evol. 143(1), Fones H., Davis C. a R., Rico A., Fang F., Smith J.A.C. & Preston G.M., Metal hyperaccumulation armors plants against disease. PLoS Pathog. 6(9), Frérot H., Lefèbvre C., Gruber W., Collin C., Santos a. Dos & Escarré J., Specific Interactions between Local Metallicolous Plants Improve the Phytostabilization of Mine Soils. Plant Soil 282(1-2), Furini A. (Ed.), Plants and Heavy Metals, SpringerBriefs in Molecular Science, Springer Netherlands, Dordrecht, 86. Gankin R. & Major J., Arctostaphylos myrtifolia, its biology and relationship to the problem of endemism. Ecology 45(4), Gauch H.G. & Whittaker R.H., Coenocline Simulation. Ecology 53(3), Gause G.F., Experimental Populations of Microscopic Organisms. Ecology 18(2), Gégout J.-C. & Pierrat J.-C., L autécologie des espèces végétales : Une approche par régression non paramétrique. Ecologie 29(3), Gleason H.A., The Individualistic Concept of the Plant Association. Am. Midl. Nat. 21(1), Goldberg D.E. & Barton A.M., Patterns and Consequences of Interspecific Competition in Natural Communities: A Review of Field Experiments with Plants. Am. Nat. 139(4), 771. Grinnell J., The niche-relationships of the California Thrasher. Auk 34, Guerrant E.O., Havens K. & Maunder M., Ex Situ Plant Conservation: Supporting Species Survival In The Wild, Society for Ecological Restoration International, Island Press. Guisan A. & Zimmermann N.E., Predictive habitat distribution models in ecology. Ecol. Modell. 135(2-3), Gunderson L.H., Ecological Resilience in Theory. Annu. Rev. Ecol. Syst. 31, Hájková P., Hájek M., Apostolova I., Zelený D. & Dítě D., Shifts in the ecological behaviour of plant species between two distant regions: evidence from the base richness gradient in mires. J. Biogeogr. 35, Hall J.L., Cellular mechanisms for heavy metal detoxification and tolerance. J. Exp. Bot. 53(366), Hanikenne M., Talke I.N., Haydon M.J., Lanz C., Nolte A., Motte P., Kroymann J., Weigel D. & Krämer U., Evolution of metal hyperaccumulation required cisregulatory changes and triplication of HMA4. Nature 453(7193), Harrison S., Safford H. & Wakabayashi J., Does the Age of Exposure of Serpentine Explain Variation in Endemic Plant Diversity in California? Int. Geol. Rev. 46(3), Harrison S.P. & Rajakaruna N., Serpentine: The Evolution and Ecology of a Model System. University of California Press ed., University of California Press, Berkeley, 464. Heikkinen J. & Mäkipää R., Testing hypotheses on shape and distribution of ecological response curves. Ecol. Modell. 221(3), Hinsinger P. & Courchesne F., Biogeochemistry of metals and metalloids at the soil root interface, in biophysico-chemical processes of heavy metals and metalloids in soil environments (eds A. Violante, P. M. Huang and G. M. Gadd), John Wiley & Sons, Inc., Hoboken, NJ, USA Hubbell S.P., The Unified Neutral Theory of Biodiversity and Biogeography, Princeton University Press, 392. Hutchinson G., Concluding remarks. Cold Spring Harb. Symp. Quant. Biol. 22(2), IUCN, IUCN Red List Categories and Criteria. Version 3.1. Second edition, IUCN Bulletin. IUCN ed., Gland, Switzerland and Cambridge, UK, 32. IUCN, The IUCN Red List of Threatened Species. Version (11/05/2015). Jacobi C.M. & do Carmo F.F., The contribution of ironstone outcrops to plant diversity in the Iron Quadrangle, a threatened Brazilian landscape. Ambio 37(4), Jaffré T., Floristic and Ecological Diversity of the Vegetation on Ultramafic Rocks in New Caledonia, Proceedings of the First International Conference on Serpentine Ecology, Hampshire, Järup L., Hazards of heavy metal contamination. Br. Med. Bull. 68, Jonak C., Nakagami H. & Hirt H., Heavy metal stress. Activation of distinct mitogen-activated protein 36

37 kinase pathways by copper and cadmium. Plant Physiol. 136(October), Kabala C. & Singh B.R., Fractionation and mobility of copper, lead, and zinc in soil profiles in the vicinity of a copper smelter. J. Environ. Qual. 30(2), Kabata-Pendias A., Trace Elements in Soils and Plants, Third Edition. Press, C. ed., Taylor & Francis, 432. Kaldorf M., Kuhn A.J., Schröder W.H., Hildebrandt U. & Bothe H., Selective Element Deposits in Maize Colonized by a Heavy Metal Tolerance Conferring Arbuscular Mycorrhizal Fungus. J. Plant Physiol. 154(5-6), Kay K.M., Ward K.L., Watt L.R. & Schemske D.W., Plant Speciation. In: Serpentine: The Evolution and Ecology of a Model System Khan A.G., Role of soil microbes in the rhizospheres of plants growing on trace metal contaminated soils in phytoremediation. J. Trace Elem. Med. Biol. 18(4), Kimura M., Evolutionary Rate at the Molecular Level. Nature 217(5129), Kozlov M. V. & Zvereva E.L., Industrial barrens: extreme habitats created by non-ferrous metallurgy. Rev. Environ. Sci. Bio/Technology 6(1-3), Krämer U., Metal hyperaccumulation in plants. Annu. Rev. Plant Biol. 61, Kruckeberg a R. & Rabinowitz D., Biological Aspects of Endemism in Higher Plants. Annu. Rev. Ecol. Syst. 16(1), Kruckeberg A.R., California Serpentines: Flora, Vegetation, Geology, Soils, and Management Problems, University of California Press, Berkeley, 196. Lasat M.M., Phytoextraction of metals from contaminated soil: A review of plant/soil/ metal interaction and assessment of pertinent agronomic issues. J. Hazard. Subst. Res. 2(5), Lasat M.M., Zinc Hyperaccumulation in Thlaspi caerulescens (703). Lau J. a, McCall A.C., Davies K.F., McKay J.K. & Wright J.W., Herbivores and edaphic factors constrain the realized niche of a native plant. Ecology 89(3), Lawesson J. & Oksanen J., Niche characteristics of Danish woody species as derived from coenoclines. J. Veg. Sci. (Gaston 1994), Li X., Yang Y., Jia L., Chen H. & Wei X., Zinc-induced oxidative damage, antioxidant enzyme response and proline metabolism in roots and leaves of wheat plants. Ecotoxicol. Environ. Saf. 89(2013), Liénard A. & Colinet G., Review of the mobility of trace elements in soils and of the use of physicochemical stabilization as an in situ management strategy for contaminated sites. Biotechnol. Agron. Soc. Environ. 18(2), Lombi E., Tearall K.L., Howarth J.R., Zhao F.-J., Hawkesford M.J. & McGrath S.P., Influence of Iron Status on Cadmium and Zinc Uptake by Different Ecotypes of the Hyperaccumulator Thlaspi caerulescens. PLANT Physiol. 128(4), Lone M.I., He Z., Stoffella P.J. & Yang X., Phytoremediation of heavy metal polluted soils and water: progresses and perspectives. J. Zhejiang Univ. Sci. B 9(3), MacArthur R.H. & Wilson E.O., An Equilibrium Theory of Insular Zoogeography. Evolution (N. Y). 17(4), MacArthur R.H. & Wilson E.O., The Theory of Island Biogeography, Princeton University Press, Princeton, 224. Macnair M., Heavy metal tolerance in plants: a model evolutionary system. Trends Ecol. Evol. 2(12), Macnair M. & Gardner M., The Evolution of Edaphic Endemics. In: Howard, D., Berlocher, S. eds. Endless Forms. Species and Speciation. Oxford University Press, New York, Maestri E. & Marmiroli M., Genetic and Molecular Aspects of Metal Tolerance and Hyperaccumulation. In: Metal Toxicity in Plants: Perception, Signaling and Remediation. Springer Berlin Heidelberg, Berlin, Heidelberg, Mahmood T., Islam K. & Muhammad S., Toxic effects of heavy metals on early growth and tolerance of cereal crops. Pakistan J. Bot. 39(2), Manara A., Plant Responses to Heavy Metal Toxicity. In: Plants and Heavy Metals Marquès L., Cossegal M., Bodin S., Czernic P. & Lebrun M., Heavy metal specificity of cellular tolerance in two hyperaccumulating plants, Arabidopsis halleri and Thlaspi caerulescens. New Phytol. 164(2), Mason H.L., The Edaphic Factor in Narrow Endemism. I. The nature of environmental influences. Madrono 8, McLean J. & Bledsoe B.E., Behavior of metals in soils. Gr. Water Issue. Meerts P. & Van Isacker N., Heavy metal tolerance and accumulation in metallicolous and nonmetallicolous populations of Thlaspi caerulescens from continental Europe. Plant Ecol. 133, Mehes-Smith M., Nkongolo K. & Cholew E., Coping Mechanisms of Plants to Metal Contaminated Soil. In: Environmental Change and Sustainability. InTech, Mench M., Vangronsveld J., Lepp N., Bleeker P., Ruttens A. & Geebelen W., Phytostabilisation of metalcontaminated sites. In: Morel, J.-L., Echevarria, G., Goncharova, N. eds. Phytoremediation of Metal- Contaminated Soils. Springer, Trest, Mendez M.O. & Maier R.M., Phytostabilization of mine tailings in arid and semiarid environments--an emerging remediation technology. Environ. Health Perspect. 116(3), Mengoni A., Gonnelli C., Galardi F., Gabbrielli R. & Bazzicalupo M., Genetic diversity and heavy metal 37

38 tolerance in populations of Silene paradoxa L. (Caryophyllaceae): a random amplified polymorphic DNA analysis. Mol. Ecol. 9(9), Mijovilovich A., Leitenmaier B., Meyer-Klaucke W., Kroneck P.M.H., Götz B. & Küpper H., Complexation and toxicity of copper in higher plants. II. Different mechanisms for copper versus cadmium detoxification in the copper-sensitive cadmium/zinc hyperaccumulator Thlaspi caerulescens (Ganges Ecotype). Plant Physiol. 151(2), Mikkelson G.M., Niche-Based vs. Neutral Models of Ecological Communities. Biol. Philos. 20(2005), Millie Burrell A., Hawkins A.K. & Pepper A.E., Genetic analyses of nickel tolerance in a North American serpentine endemic plant, Caulanthus amplexicaulis var. barbarae (Brassicaceae). Am. J. Bot. 99(11), Moran P.A.P., Random processes in genetics. Math. Proc. Cambridge Philos. Soc. 54(01), 60. Oksanen J. & Minchin P.R., Continuum theory revisited: what shape are species responses along ecological gradients? Ecol. Modell. 157(2-3), Pacyna J.M. & Pacyna E.G., An assessment of global and regional emissions of trace metals to the atmosphere from anthropogenic sources worldwide. Environ. Rev. 9(4), Parish J. & Bazzaz F., Ontogenetic niche shifts in old-field annuals. Ecology 66(4), Parraga-Aguado I., Gonzalez-Alcaraz M.N., Alvarez-Rogel J., Jimenez-Carceles F.J. & Conesa H.M., The importance of edaphic niches and pioneer plant species succession for the phytomanagement of mine tailings. Environ. Pollut. 176, Parraga-Aguado I., Querejeta J.-I., González-Alcaraz M.-N., Jiménez-Cárceles F.J. & Conesa H.M., Usefulness of pioneer vegetation for the phytomanagement of metal(loid)s enriched tailings: grasses vs. shrubs vs. trees. J. Environ. Manage. 133, Pearman P.B., Guisan A., Broennimann O. & Randin C.F., Niche dynamics in space and time. Trends Ecol. Evol. 23(3), Peer W.A., Baxter I.R., Richards E.L., Freeman J.L. & Murphy S., Phytoremediation and hyperaccumulator plants. Mol. Biol. Met. Homeost. Detoxif. 14(August), Pilon-Smits E.A.H., Quinn C.F., Tapken W., Malagoli M. & Schiavon M., Physiological functions of beneficial elements. Curr. Opin. Plant Biol. 12(3), Poschenrieder C., Tolrà R. & Barceló J., Can metals defend plants against biotic stress? Trends Plant Sci. 11(6), Primack R.B., Essentials of Conservation Biology, Sinauer Associates, Incorporated. Pulliam H.R., On the relationship between niche and distribution. Ecol. Lett. 3(4), Rajakaruna N., The Edaphic Factor in the Origin of Plant Species. Int. Geol. Rev. 46, Redondo-Gomez S., Abiotic and Biotic Stress Tolerance in Plants. In: Rout, G.R., Das, A.B. eds. Molecular Stress Physiology of Plants. Springer India, India, Rosindell J., Hubbell S.P. & Etienne R.S., The Unified Neutral Theory of Biodiversity and Biogeography at Age Ten. Trends Ecol. Evol. 26(7), Rydgren K., Økland R.H. & Økland T., Species response curves along environmental gradients. A case study from SE Norwegian swamp forests. J. Veg. Sci. 14(6), Ryser P. & Sauder W.R., Effects of heavy-metalcontaminated soil on growth, phenology and biomass turnover of Hieracium piloselloides. Environ. Pollut. 140(1), Safford H., Viers J. & Harrison S., Serpentine endemism in the California flora: a database of serpentine affinity. Madroño 52, Salt D.E., Blaylock M., Kumar N.P.B.A., Dushenkov V., Ensley B.D., Chet I. & Raskin I., Phytoremediation: A Novel Strategy for the Removal of Toxic Metals from the Environment Using Plants. Biotechnology 13(5), Schat H. & Vooijs R., Multiple tolerance and cotolerance to heavy metals in Silene vulgaris: A cosegregation analysis. New Phytol. 136(3), Shaw J., Heavy Metal Tolerance in Plants: Evolutionary Aspects. Shaw, J. ed., CRC Press Inc, New York, 268. Silvertown J., Plant coexistence and the niche. Trends Ecol. Evol. 19(11), Singh A. & Prasad S.M., Reduction of heavy metal load in food chain: technology assessment. Rev. Environ. Sci. Bio/Technology 10(3), Singh S.N. & Tripathi R.D., Environmental bioremediation technologies, Environmental Bioremediation Technologies, Springer-Verlag Berlin Heidelberg, Singh V.P., Metal Toxicity and Tolerance in Plants and Animals, Sarup & sons. Smith R., The occurrence and need for conservation of metallophytes on mine wastes in Europe. Miner. Environ. 1, Sobel J.M., Chen G.F., Watt L.R. & Schemske D.W., The biology of speciation. Evolution (N. Y). 64(2), Soulé M., What is conservation biology. Bioscience 35(11), Stanley Harpole W. & Tilman D., Non-neutral patterns of species abundance in grassland communities. Ecol. Lett. 9(1), Stohs S.J. & Bagchi D., Oxidative mechanisms in the toxicity of metal ions. Free Radic. Biol. Med. 18(2),

39 Stokes C.J. & Archer S.R., Niche differentiation and neutral theory: An integrated perspective on shrub assemblages in a parkland savanna. Ecology 91(4), Street R.A., Kulkarni M.G., Stirk W.A., Southway C. & Van Staden J., Toxicity of metal elements on germination and seedling growth of widely used medicinal plants belonging to Hyacinthaceae. Bull. Environ. Contam. Toxicol. 79(4), Sutherland W.J., Adams W.M., Aronson R.B., Aveling R., Blackburn T.M., Broad S., Ceballos G., CÔtÉ I.M., Cowling R.M., Da Fonseca G. a B., Dinerstein E., Ferraro P.J., Fleishman E., Gascon C., Hunter M., Hutton J., Kareiva P., Kuria a., MacDonald D.W., MacKinnon K., Madgwick F.J., Mascia M.B., McNeely J., Milner-Gulland E.J., Moon S., Morley C.G., Nelson S., Osborn D., Pai M., Parsons E.C.M., Peck L.S., Possingham H., Prior S. V., Pullin a. S., Rands M.R.W., Ranganathan J., Redford K.H., Rodriguez J.P., Seymour F., Sobel J., Sodhi N.S., Stott a., Vance-Borland K. & Watkinson a. R., One hundred questions of importance to the conservation of global biological diversity. Conserv. Biol. 23(3), Thompson K. & Ceriani R.M., No relationship between range size and germination niche width in the UK herbaceous flora. Funct. Ecol. 17(3), Turisová I., Štrba T., Aschenbrenner Š. & Andráš P., Arabidopsis arenosa (L.) Law. on metalliferous and nonmetalliferous sites in central Slovakia. Bull. Environ. Contam. Toxicol. 91(4), Turnbull L.A., Coomes D.A., Purves D.W. & Rees M., How spatial structure alters population and community dynamics in a natural plant community. J. Ecol. 95(1), United Nations, Convention on biological diversity. Diversity 30. Van Rossum F., Bonnin I., Fenart S., Pauwels M., Petit D. & Saumitou-Laprade P., Spatial genetic structure within a metallicolous population of Arabidopsis halleri, a clonal, self-incompatible and heavy-metal-tolerant species. Mol. Ecol. 13(10), Vangronsveld J., Van Assche F. & Clijsters H., Reclamation of a bare industrial area contaminated by non-ferrous metals: in situ metal immobilization and revegetation. Environ. Pollut. 87(1), Vekemans X. & Lefèbvre C., On the evolution of heavy-metal tolerant populations in Armeria maritima: Evidence from allozyme variation and reproductive barriers. J. Evol. Biol. 10(2), Verbruggen N., Hermans C. & Schat H., Molecular mechanisms of metal hyperaccumulation in plants. New Phytol. 181(4), Vetaas O.R., Realized and potential climate niches: a comparison of four Rhododendron tree species. J. Biogeogr. 29, Weiher E., Clarke G.D.P. & Keddy P. a., Community Assembly Rules, Morphological Dispersion, and the Coexistence of Plant Species. Oikos 81(2), 309. Whiting S.N., Reeves R.D. & Baker A.J.M., Conserving biodiversity: Mining, metallophytes and land reclamation. Min. Environ. Manag. 10(2), Whiting S.N., Reeves R.D., Richards D., Johnson M.S., Cooke J.A., Malaisse F., Johns R., Mcintyre T., Purvis O.W., Salt D.E., Schat H., Zhao F.J. & Baker A.J.M., Research Priorities for Conservation of Metallophyte Biodiversity and their Potential for Restoration and Site Remediation. Restor. Ecol. 12(1), Whittaker R., Levin S. & Root R., Niche, habitat, and ecotype. Am. Nat. 107(955), Whittaker R.H., Vegetation of the great smoky mountains. Ecol. Monogr. 26(1), Wiens J.J. & Graham C.H., NICHE CONSERVATISM: Integrating Evolution, Ecology, and Conservation Biology. Annu. Rev. Ecol. Evol. Syst. 36(1), Wilson J., The cost of heavy-metal tolerance: an example. Evolution (N. Y). 42(2), Wolf A., Howe R. & Hamrick J., Genetic diversity and population structure of the serpentine endemic Calystegia collina (Convolvulaceae) in northern California. Am. J. 87(8), Wong M.H., Ecological restoration of mine degraded soils, with emphasis on metal contaminated soils. Chemosphere 50(6), World Health Organisation, Chemical Safety - Activity Report 2015, Wright J., Davies K., Lau J., McCall A. & McKay J., Experimental verification of ecological niche modeling in a heterogeneous environment. Ecology 87(10), Wright J.W., Stanton M.L. & Scherson R., Local adaptation to serpentine and non-serpentine soils in Collinsia sparsiflora. Evol. Ecol. Res. 8(1), Yang X.E., Jin X.F., Feng Y. & Islam E., Molecular mechanisms and genetic basis of heavy metal tolerance/hyperaccumulation in plants. J. Integr. Plant Biol. 47(9), Zhou S. & Zhang D., Neutral theory in community ecology. Front. Biol. China 3(1),

40

41 L Arc Cuprifère Katangais un exemple de site métallifère situé en R. D. Congo Localisation, géologie et phytogéographie Le sud-est de la République Démocratique du Congo (R.D. Congo) présente les gisements de cuivre (Cu) et de cobalt (Co) parmi les plus importants au monde, associés à d autres métaux tels que le zinc (Zn), le plomb (Pb) et l uranium (U) (Figure 4a, Cailteux et al., 2005; Kampunzu et al., 2009). Au point de vue géologique, les gisements de Cu et Co sont des écailles de la Série des Mines qui affleurent en surface en raison de l épisode tectonique panafricain. Cette Série des Mines a été formée par séries sédimentaires dans une dépression située au croisement de l orogenèse kibarienne et lufilienne (1300 Ma et 500 Ma). Cette dépression est bordée par les massifs kibariens, les plateaux de Kundelungu et de Biano au nord et au sud-ouest. Au niveau stratigraphique, la Série des Mines compose la base de la couche du Katangien tandis que le Kundelungu est au sommet. Cette Série des Mines est constituée de bas en haut, du Roan inférieur (roches silico-détritiques et dolomitiques rouges non minéralisées), du Roan moyen (roches silico-détritiques et dolomitiques grises à noires, souvent minéralisées économiquement), et du Roan supérieur (roches silicodétritiques et dolomitiques grises à blanchâtres, non minéralisées économiquement). 41

42 Normalement, cette couche devrait se situer à plusieurs milliers de mètres de profondeur, sous les sédiments du Kundelungu mais la période tectonique panafricaine ( Ma) l a découpée et a fait affleurer des blocs par charriage ou par extrusion au sud de la République Démocratique du Congo (François, 1973; Okitaudji, 2001; Cailteux et al., 2005; Kampunzu et al., 2009). Ces blocs ont été recouverts de sédiments durant le quaternaire et se présentent aujourd hui sous la forme de collines isolées dans une matrice de forêt claire sèche nommée le Miombo (Duvigneaud et al., 1963; Campbell, 1996). Une centaine de collines aux sols métallifères sont dispersées dans une zone de 300 km de long et 50 km de large entre Kolwezi et Lubumbashi appelée l Arc Cuprifère Katangais. Au sommet des collines, la proximité entre l affleurement et la surface du sol engendrent des teneurs biodisponibles (extraction acétate EDTA 1N ph 4,65) en Cu de 30 à mg kg -1 et en Co 2 à mg kg -1. Ces concentrations sont jusqu à 500 fois plus élevées que celles mesurées dans les sols du Miombo (Hick, communication personnelle). L affleurement étant plus profond à la base, la quantité de particules métallifères présentes à la surface du sol sont plus faibles qu au sommet. Les teneurs biodisponibles sont proches de celles des sols aux conditions différentes et des plus abondants dans la région. Contrairement à d autres sites métallifères comme les sols ultramafiques, aucune déficience en nutriments Mg, P et Ca n a été observée dans ces sols (Kruckeberg, 1985; O Dell et al., 2006). Cette région est inclue dans une zone phytogéographique plus vaste, le Copperbelt allant jusqu au nord de la Zambie (Figure 4a,b, Duvigneaud et al., 1963). Mise en évidence il y a plus de 50 ans, la végétation des collines de l Arc Cuprifère Katangais a rapidement suscité l intérêt des scientifiques pour sa composition et son adaptation à des conditions de stress presque uniques au monde (Duvigneaud, 1958; Duvigneaud et al., 1959, 1963). Plus de 550 taxa spermatophytes sont distribués sur les collines aux sols riches en Cu et Co. Ils sont également appelés taxa cupro-cobalticoles étant donné leur affinité pour ces sols (Duvigneaud et al., 1963; Brooks et al., 1992; Leteinturier & Malaisse, 1999; Leteinturier, 2002). 42

43 Hélène Maurel / RFI 2009 Vétoquinol Soizig Le Stradic Julie Lebrun Maxime Séleck Sylvain Boisson Sylvain Boisson Julie Lebrun Sylvain Boisson Figure 4. Localisation et paysage des collines de l Arc Cuprifère Katangais, des communautés végétales principales et de deux de leurs espèces indicatrices. a. Localisation de la République Démocratique Du Congo, du Katanga et de l Arc Cuprifère Katangais. b. Paysage de l Arc Cuprifère Katangais avec une colline en avant plan entourée de Uapaca robynsii. c1. Savane steppique. c2. Loudetia simplex. c3. Cryptosepalum maraviense. d1. Steppe. d2. Ascolepis metallorum. d3. Xerophyta equisetoïdes. 43

44 Deux formations végétales principales ont été définies selon les concentrations en Cu et Co du sol (Saad et al., 2012; Séleck et al., 2013). Une steppe se développe au sommet de la colline sur des sols minéralisés avec des concentrations s échelonnant entre mg kg -1 Cu et mg kg -1 Co. Cette formation est caractérisée généralement par une végétation d une hauteur moyenne maximale de 30 cm, avec 33% de recouvrement du sol par les graminées, 36% de sol nu et une richesse spécifique de 11 espèces par mètre carré. Une savane steppique colonise des sols peu minéralisés dont les concentrations se distribuent entre mg kg -1 Cu et 0-20 mg kg -1 Co. Elle se distingue de la steppe par une végétation plus haute (moyenne de 62 cm), plus fermée (moyenne de 55% de sol recouvert par les graminées et 13% de sol nu) et a une richesse spécifique plus grande (moyenne de 14 espèces par mètre carré). Ces deux formations végétales se distinguent également par les formes de vie présentes (Delhaye et al., 2016). Notamment, Cryptosepalum maraviense Oliv. (Caesalpiniaceae) est une espèce à xylopode qui structure la savane steppique avec l espèce Loudetia simplex (Nees) C.E.Hubb. (Poaceae) (Figure 4c, Saad et al., 2012; Ilunga wa Ilunga et al., 2013; Séleck et al., 2013; Faucon et al., 2015). Au contraire, des espèces annuelles et celles présentant un système racinaire moins profond dominent la steppe même si des espèces pérennes peuvent y être relevées. L espèce Ascolepis metallorum P.A.Duvign. & G.Léonard (Cyperaceae) et dans certains cas Xerophyta equisetoides Baker var. equisetoides (Velloziaceae) y sont notamment retrouvées et identifiées comme indicatrices (Séleck et al. 2013, Figure 4d). Elle est également caractérisée par la présence d espèce à xylopodes et des espèces Loudetia simplex (Nees) C.E.Hubb. (Poaceae) et Cryptosepalum maraviense Oliv. (Caesalpiniaceae) (Figure 4c, Saad et al., 2012; Ilunga wa Ilunga et al., 2013; Séleck et al., 2013; Faucon et al., 2015). Parmi ces deux formations, une diversité de types de communautés végétales a été identifiée selon, entre autres, les concentrations en Cu et Co. Aussi, les roches et les failles faiblement minéralisées et aux teneurs en Cu inférieures à 900 mg kg -1 sont colonisées par une végétation chasmophytique (Figure 5, Malaisse et al., 1994; Leteinturier, Baker, et al., 1999; Faucon et al., 2015). 44

45 Figure 5. Les huit types de végétation identifiées et décrites par Leteinturier, Baker et al A. Uapaca robynsii - Ceinture de savane arbustive ; B. Savane steppique à Loudetia simplex- Monocymbium ceresiiforme et Acalypha cupricola; C. Steppe enrochée à Xerophyta equisetoides.; D. Végétation sur affleurements rocheux crevassés (steppe); E. Savane boisée à Hymenocardia acida; F. Pelouse sur sol cuprifère remanié à Haumaniastrum robertii (steppe); G. Pelouse à Microchloa altera sur sol compact (steppe); H. Pelouse à Bulbostylis pseudoperennis (steppe). (Malaisse et al., 1994) Relation plante-sol et endémisme Les facteurs Cu et Co ne sont pas les seuls qui déterminent les communautés végétales et les espèces des sites cuprifères de l Arc Cuprifère Katangais (Faucon et al., 2015). Au sein et entre les sites, ces communautés sont formées par différentes mosaïques d espèces conditionnées également par d autres facteurs édaphiques, tel que le ph, inter-dépendants dans le sol (Saad et al., 2012; Ilunga wa Ilunga et al., 2013; Muyumba et al., 2015). Par exemple, les sols des steppes présentent un ph plus grand et des teneurs en calcium (Ca) et phosphore (P) plus élevées ainsi qu un ratio C:N et des teneurs en fer (Fe) plus faibles que dans les savanes steppiques (Séleck et al., 2013). Les variations complexes entre ces facteurs entrainent des modifications de la biodisponibilité du Cu et du Co dans les sols à l échelle spatiale et temporelle qui contribuent indirectement aux variations floristiques et à l hétérogénéité de la végétation au sein et entre les sites (Lange et al., 2014; Faucon et al., 2015; Pourret, Lange, Houben, et al., 2015). Cependant, les variations des teneurs biodisponibles en Cu et Co conditionnent les traits des espèces (Delhaye et al., 2016). Le long des gradients métalliques, le patron de substitution de traits interspécifiques est plus marqué qu une simple variation de traits à l échelle interspécifique. A l échelle des communautés, la variation des traits résulte des différences d abondance entre les formes de vie (Séleck et al., 2013). Les communautés végétales des sites cuprifères sont caractérisées par une proportion de taxa endémiques plus faible que dans d autres végétations métallicoles tels que les écosystèmes serpentiniques (Faucon et al., 2010). Les sols métallifères (ultramafiques) de 45

46 Californie et d Italie présentent respectivement 23% et 29% d espèces endémiques contre 5% sur les sites cupro-cobaltifères de l Arc Cuprifère Katangais. Cette végétation est composée de 32 taxa endémiques stricts (ou métallophytes absolus) dont toutes les populations connues sont exclusivement localisées sur des sols riches en Cu et 25 endémiques larges de l Arc Cuprifère Katangais (ou metallophytes facultatifs et taxa tolérants) dont plus de 75% des populations connues sont localisées sur des sols riches en Cu de l Arc Cuprifère Katangais (Brooks et al., 1985; Leteinturier, 2002; Faucon et al., 2010; Champluvier, 2011). Cette faible proportion d endémisme est justifiée par de plus petits habitats que les sols ultramafiques et l influence de variations paléoclimatiques (Faucon et al., 2015). D ailleurs, Ilunga wa Ilunga (2014, données non publiées) a démontré que la richesse en taxa endémiques au sein de l Arc Cuprifère Katangais est corrélée positivement avec la surface du site. Ensuite, l hypothèse des variations climatiques soutient que l apparition de divergences évolutives pour les populations présentes sur les collines entre les épisodes glaciaires et chauds expliquerait la ressemblance morphologique entre des taxa endémiques (c.-à-d. taxa congénériques) de l Arc Cuprifère Katangais et des taxa plus répandus dans la région (phénomène de vicariance). Finalement, les processus de pression de sélection et d isolement écologique pourraient être surestimés. En effet, la biodisponibilité du le Cu et le Co varie avec les autres éléments édaphiques notamment la matière organique (Faucon et al., 2015). La classification des métallophytes a été adoptée pour les cuprophytes de la végétation de l Arc Cuprifère Katangais. Les eucuprophytes correspondent aux métallophytes absolus, les cuprophytes locaux aux métallophytes facultatifs et les cuprophiles ainsi que les cuprorésistants aux taxa tolérants (Duvigneaud et al., 1963). Cette classification dépend exclusivement de leur distribution sur les sols métallifères et non-métallifères. Plusieurs de ces taxa, tels que Haumanisatrum katangense (S.Moore) P.A.Duvign. & Plancke ou Crepidorhopalon perennis (P.A.Duvign.) Eb.Fisch., ont été identifiés comme des exclueurs et d autres, tel qu Anisopappus davyi S. Moore, comme hyperaccumulateurs de Cu et/ou de Co (Brooks et al., 1987, 1992; Chipeng et al., 2010; Faucon et al., 2012). La présence d espèces endémiques présentant une grande affinité pour le Cu et Co du sol (eucuprophytes ou endémiques strictes et cuprophytes locaux ou endémiques larges) peut également être la conséquence d un besoin constitutif en ces métaux ou de la pression de facteurs biotiques (Tadros, 1957; Remon, 2006; Wu et al., 2009; Kay et al., 2010). Chipeng (2010) a en effet mis en évidence que la biomasse et la fitness de H. katangense augmentait avec les teneurs 46

47 en Cu du substrat (cf. endémisme édaphique spécialiste). Ces espèces pourraient également être moins compétitives dans la végétation dominante (cf. endémisme édaphique refuge) étant donné le coût adaptatif de la tolérance aux ETM (Macnair et al., 2000). La perte de résistance aux pathogènes et la défense contre les herbivores constituent une dernière hypothèse. En effet, les pressions de sélection dans les sols métallifères sont intenses et pourraient limiter la croissance de ces organismes même si certains auteurs ont démontré qu ils étaient capables de s adapter (Kazakou et al., 2008; Lau et al., 2008). D autres études ont d ailleurs démontré que certains microorganismes sont sélectionnés par la pression des ETM présents dans le sol et sont capables de vivre dans des sols métallifères (Becerra-Castro et al., 2012). La conservation des cuprophytes Plusieurs sites hébergeant des cuprophytes ont déjà été détruits par l exploitation minière (Brooks et al., 1992; Faucon et al., 2011). La plupart des espèces endémiques de l Arc Cuprifère Katangais (c.-à-d. 79 %) sont actuellement menacée d extinction selon les lignes directrices fixant les critères de la Liste Rouge de l IUCN (Faucon et al., 2010). La conservation et l étude des espèces endémiques dans une partie de l Arc Cuprifère Katangais font l objet de collaborations internationales entre Gembloux Agro-Bio Tech (Université de Liège, BE), la Faculté des Sciences Agronomiques de Lubumbashi (Université de Lubumbashi, R.D.C.), l Université Libre de Bruxelles (ULB, BE), l Institut Polytechnique LaSalle Beauvais (ISAB, FR) et la société Tenke Fungurume Mining (Filiale de Freeport-McMoran Inc., R.D.C.) depuis plusieurs années (Figure 6; Semereab et al., 2009). Le défi majeur de ces collaborations est de produire les connaissances nécessaires à la conservation des cuprophytes de l Arc Cuprifère Katangais plus rapidement que la disparition des sites. La finalité du projet a pour but de restaurer les communautés métallophytes après exploitation à partir du matériel végétal mis en conservation pendant l exploitation. Dans le contexte minier, les stratégies de conservation in situ comme les zones protégées sont difficilement envisageables. Des stratégies de conservation ex situ sont donc implémentées à deux échelles : communauté et espèce. A l échelle de la communauté, la translocation des communautés vivantes avec leur sol permet de déplacer les steppes dans des zones nonexploitées en conservant la majorité des espèces. Cette méthode est insuffisante pour les savanes steppiques qui ne présentent plus la même composition et la même structure 47

48 végétale après translocation et qui sont rapidement colonisées par des espèces tolérantes envahissantes (Le Stradic et al., 2015). A l échelle des espèces, la conservation ex situ de matériel végétal sous forme de graines à l université de Lubumbashi et au Jardin botanique de Meise est un succès pour 19 taxa endémiques qui montrent un taux de germination minimum de 30 %. Cependant, certains taxa produisent des graines non-viables (Godefroid et al., 2013). Les difficultés rencontrées dans les stratégies de conservation ex situ mettent en avant les particularités des végétations herbacées tropicales anciennes (appelées old growth grassland) qui sont des écosystèmes stables dans le temps comprenant de nombreux chaméphytes et des cryptophytes. Ces écosystèmes sont souvent peu pris en compte dans des programmes de conservation et donc peu étudiés (Veldman et al., 2015). Dans le cas des végétations cupricoles, étant donné la présence des ETM, une contrainte écologique supplémentaire, la connaissance de la distribution in situ des taxa le long du gradient de concentration en Cu (niches édaphiques réalisées) est essentielle (Figure 5). Depuis la description in situ de Duvigneaud et al. (1963), peu d études ont quantifié les niches réalisées de taxa composant les communautés végétales de cuprophytes de l Arc Cuprifère Katangais (Ilunga wa Ilunga et al., 2013). De plus, les informations à propos de la tolérance et l accumulation en Cu ou Co à l échelle des espèces et des populations fournissent des éléments clés pour leur conservation et leur valorisation dans des stratégies de phytoremédiation (Leteinturier, Baker, et al., 1999; Faucon et al., 2007, 2012; Chipeng et al., 2010; Peng et al., 2012; Lange et al., 2014). Les espèces comme Crepidorhopalon tenuis (P.A.Duvign.) Eb.Fisch., Microchloa altera (Rendle) Stapf et Bulbostylis pseudoperennis Goetgh. (IUCN : critiquement en danger), entre autres, sont retrouvées sur des sols métallifères anthropogéniques (Leteinturier, Baker, et al., 1999; Shutcha et al., 2010, 2013; Faucon et al., 2011). Cette démarche exige également de déterminer les interactions entre les éléments édaphiques dont la biodisponibilité des ETM selon d autres facteurs édaphiques (Lange et al., 2014; Pourret, Lange, Bonhoure, et al., 2015; Pourret, Lange, Houben, et al., 2015). 48

49 Figure 6. Structuration des études marquantes réalisées sur la végétation métallicoles de l Arc Cuprifère Katangais et concepts ou tâches associés étudiés durant la thèse de doctorat. Schéma : Le Stradic S. et Boisson S. Références Becerra-Castro C., Monterroso C., Prieto-Fernández a, Rodríguez-Lamas L., Loureiro-Viñas M., Acea M.J. & Kidd P.S., Pseudometallophytes colonising Pb/Zn mine tailings: a description of the plant-microorganismrhizosphere soil system and isolation of metal-tolerant bacteria. J. Hazard. Mater , Brooks R.R., Baker A.J.M. & Malaisse F., Copper flowers. Res. Explor. 8(3), Brooks R.R. & Malaisse F., The Heavy Metal Tolerant Flora of Southcentral Africa: a mutlidisciplinary approach, A.A. Balkema, Rotterdam, 199. Brooks R.R., Naidu S.M., Malaisse F. & Lee J., The elemental content of metallophytes from the copper/cobalt deposits of Central Africa. Bull. Soc. Roy. Bot. Belgique 119(2), Cailteux J.L.H., Kampunzu A.B., Lerouge C., Kaputo A.K. & Milesi J.P., Genesis of sediment-hosted stratiform copper cobalt deposits, central African Copperbelt. J. African Earth Sci. 42(1-5), Campbell B., The Miombo in transition: woodlands and welfare in Africa. Campbell, B. ed., Center for International Forestry Research (CIFOR), Bogor, Indonesia, 266. Champluvier D., New and overlooked Acanthaceae taxa from D.R.Congo, Rwanda and Burundi: (1) the genus Barleria. Plant Ecol. Evol. 144(1), Chipeng F.K., Hermans C., Colinet G., Faucon M.-P., Ngongo M., Meerts P. & Verbruggen N., Copper tolerance in the cuprophyte Haumaniastrum katangense (S. Moore) P.A. Duvign. & Plancke. Plant Soil 328(1-2), Delhaye G., Violle C., Séleck M., Ilunga wa Ilunga E., Daubie I., Mahy G. & Meerts P., Plant traits variation along a copper-cobalt contamination gradient: a community-level perspective exploring the relative importance of species turnover and within species variation. J. Veg. Sci. Online Duvigneaud P., Etudes sur la végétation du Katanga et de ses sols métallifères. Communication n 1 La végétation du Katanga et de ses sols métallifères. Bull. Soc. Roy. Bot. Belgique 90(2), Duvigneaud P. & Denaeyer-De Smet S., Etudes sur la végétation du Katanga et de ses sols métallifères. Communication n 7 Cuivre et végétation au Katanga. Bull. Soc. Roy. Bot. Belgique, Travaux du Centre scientifique et médical de l Université libre de Bruxelles en Afrique centrale 96(2), Duvigneaud P. & Timperman J., Etudes sur la végétation du Katanga et de ses sols métallifères. 49

50 Communication n 3 Etudes sur le genre Crotalaria. Bull. Soc. Roy. Bot. Belgique 91(2), Faucon M., Le Stradic S., Boisson S., Ilunga wa Ilunga E., Séleck M., Lange B., Delhaye G., Shutcha M.N., Pourret O., Meerts P. & Mahy G., Implication of plant-soil relationships for conservation and restoration of coppercobalt ecosystems. Plant Soil (November 2015). Faucon M.-P., Chipeng F., Verbruggen N., Mahy G., Colinet G., Shutcha M., Pourret O. & Meerts P., Copper tolerance and accumulation in two cuprophytes of South Central Africa: Crepidorhopalon perennis and C. tenuis (Linderniaceae). Environ. Exp. Bot. 84, Faucon M.-P., Meersseman A., Shutcha M.N., Mahy G., Luhembwe M.N., Malaisse F. & Meerts P., Copper endemism in the Congolese flora: a database of copper affinity and conservational value of cuprophytes. Plant Ecol. Evol. 143(1), Faucon M.-P., Parmentier I., Colinet G., Mahy G., Ngongo Luhembwe M. & Meerts P., May Rare Metallophytes Benefit from Disturbed Soils Following Mining Activity? The Case of the Crepidorhopalon tenuis in Katanga (D. R. Congo). Restor. Ecol. 19(3), Faucon M.-P., Shutcha M.N. & Meerts P., Revisiting copper and cobalt concentrations in supposed hyperaccumulators from SC Africa: influence of washing and metal concentrations in soil. Plant Soil 301(1-2), François A., L extrémité occidentale de l'arc Cuprifère Shabien, Etude géologique-département de géologie de la Gécamines, Likasi (République du Zaire), 65. Godefroid S., Van de Vyver A., Massengo Kalenga W., Handjila Minengo G., Rose C., Ngongo Luhembwe M., Vanderborght T. & Mahy G., Germination capacity and seed storage behaviour of threatened metallophytes from the Katanga copper belt (DR Congo): implications for ex situ conservation. Plant Ecol. Evol. 146(2), Ilunga wa Ilunga E., Séleck M., Colinet G., Meerts P. & Mahy G., Small-scale diversity of plant communities and distribution of species niches on a copper rock outcrop in Upper Katanga, DR Congo. Plant Ecol. Evol. 146(2), Kampunzu A.B., Cailteux J.L.H., Kamona A.F., Intiomale M.M. & Melcher F., Sediment-hosted Zn Pb Cu deposits in the Central African Copperbelt. Ore Geol. Rev. 35(3-4), Kay K.M., Ward K.L., Watt L.R. & Schemske D.W., Plant Speciation. In: Serpentine: The Evolution and Ecology of a Model System Kazakou E., Dimitrakopoulos P.G., Baker a. J.M., Reeves R.D. & Troumbis a. Y., Hypotheses, mechanisms and trade-offs of tolerance and adaptation to serpentine soils: From species to ecosystem level. Biol. Rev. 83(4), Kruckeberg A.R., California Serpentines: Flora, Vegetation, Geology, Soils, and Management Problems, University of California Press, Berkeley, 196. Lange B., Faucon M.-P., Meerts P., Shutcha M., Mahy G. & Pourret O., Prediction of the edaphic factors influence upon the copper and cobalt accumulation in two metallophytes using copper and cobalt speciation in soils. Plant Soil 379(1-2), Lau J. a, McCall A.C., Davies K.F., McKay J.K. & Wright J.W., Herbivores and edaphic factors constrain the realized niche of a native plant. Ecology 89(3), Le Stradic S., Séleck M., Lebrun J., Boisson S., Handjila Minengo G., Faucon M., Enk T. & Mahy G., Comparison of translocation methods to conserve metallophyte communities in the Katangan Copperbelt, DRC. Environ. Sci. Pollut. Res. Leteinturier B., Evaluation du potential phytocénotique des gisements cupriferes d Afrique centro-australe en vue de la phytoremédiation de sites pollués par l'activité. Fac. des Sci. Agron. Gembloux. Leteinturier B., Baker A.J.M. & Malaisse F., Early stages of natural revegetation of metalliferous mine workings in South Central Africa: a preliminary survey. Biotechnol. Agron. Soc. Environ. 3(1), Leteinturier B. & Malaisse F., The copper flora of Katanga: a phytogeographical analysis. Geo. Eco. Trop. 23, Macnair M., Tilstone G. & Smith S., The Genetics of Metal Tolerance and Accumulation in Higher Plants. In: Terry, N. ed. Phytoremediation of Contaminated Soil and Water. CRC Press Inc, 408. Malaisse F., Brooks R.R. & Baker A.J.M., Diversity of vegetation communities in relation to soil heavy metal content at the Shinkolobwe copper/cobalt/uranium mineralization, upper Shaba, Zaire. Belgian J. Bot. 127(1), Muyumba D.K., Liénard A., Mahy G. & Ngongo M., Caractérisation des systèmes sols-plantes dans les collines de l arc cuprifère du Katanga (synthèse bibliographique). Biotechnol., Agron. Société Environ. 19(2), O Dell R.E. & Claassen V.P., Relative performance of native and exotic grass species in response to amendment of drastically disturbed serpentine substrates. J. Appl. Ecol. 43(5), Okitaudji R., Modele de formation des gisements cuivre-cobalt du Shaba en République Démocratique du Congo. Bull. Acad. Soc. Lorraines. Sci. 40(4), 17. Peng H., Wang-Müller Q., Witt T., Malaisse F. & Küpper H., Differences in copper accumulation and copper stress between eight populations of Haumaniastrum katangense. Environ. Exp. Bot. 79, Pourret O., Lange B., Bonhoure J., Colinet G., Decrée S., Mahy G., Séleck M., Shutcha M. & Faucon M.-P., Assessment of soil metal distribution and environmental impact of mining in Katanga (Democratic Republic of Congo). Appl. Geochemistry Pourret O., Lange B., Houben D., Colinet G., Shutcha M. & Faucon M.-P., Modeling of cobalt and copper speciation in metalliferous soils from Katanga (Democratic Republic of Congo). J. Geochemical Explor. 149,

51 Remon E., Tolérance et accumulation des métaux lourds par la végétation spontanée des friches métallurgiques : vers de nouvelles méthodes de biodépollution. Univ. Jean Monnet 166. Saad L., Parmentier I., Colinet G., Malaisse F., Faucon M.- P., Meerts P. & Mahy G., Investigating the Vegetation-Soil Relationships on the Copper-Cobalt Rock Outcrops of Katanga (D. R. Congo), an Essential Step in a Biodiversity Conservation Plan. Restor. Ecol. 20(3), Séleck M., Bizoux J.-P., Colinet G., Faucon M.-P., Guillaume A., Meerts P., Piqueray J. & Mahy G., Chemical soil factors influencing plant assemblages along coppercobalt gradients: implications for conservation and restoration. Plant Soil 373(1/2), Semereab E., Rensonnet A., Lebrun J., Handjila G., Saad L., Parmentier I., Senterre B., Guillaume A., Weiskopf T., Malaisse F. & Mahy G., Biodiversity Action Plan Tenke Fungurume Mining , Gembloux, 89. Shutcha M., Meerts P., Kazadi S., Faucon M.-P. & Visser M., Low Caryopsis Production of the Toothbrush Grass (Microchloa altera) from Katanga (DR Congo) Could Limit the Revegetation of Trace Metal Contaminated Lands by. Ecol. Restor. 31(3), Shutcha M.N., Mubemba M.M., Faucon M.-P., Luhembwe M.N., Visser M., Colinet G. & Meerts P., Phytostabilisation of copper-contaminated soil in Katanga: an experiment with three native grasses and two amendments. Int. J. Phytoremediation 12(6), Tadros T.T.M., Evidence of the presence of an edapho-biotic factor in the problem of serpentine tolerance. Ecology 38(1), Veldman J.W., Buisson E., Durigan G., Fernandes G.W., Le Stradic S., Mahy G., Negreiros D., Overbeck G.E., Veldman R.G., Zaloumis N.P., Putz F.E. & Bond W.J., Toward an old-growth concept for grasslands, savannas, and woodlands. Front. Ecol. Environ. 13(3), Wu F.Y., Leung H.M., Wu S.C., Ye Z.H. & Wong M.H., Variation in arsenic, lead and zinc tolerance and accumulation in six populations of Pteris vittata L. from China. Environ. Pollut. 157(8-9),

52

53 Objectifs et stratégie scientifique Objectif général et lignes directrices des axes Cette thèse présente une contribution à l amélioration des connaissances sur l écologie des taxa végétaux cupro-cobalticoles de l Arc Cuprifère Katangais dans le but d implémenter des stratégies de conservation les plus pertinentes associées à la phytostabilisation des sols pollués. En particulier, plusieurs questions sont à la base de la démarche scientifique de cette thèse. Le premier axe pose les questions de la spécialisation des taxa endémiques au Cu du sol en débutant par la caractérisation des niches réalisées puis fondamentales des taxa endémiques et non-endémiques le long des gradients de Cu et/ou Co. L étude de ces notions permet de mettre en évidence des lignes directrices pour des actions de conservation. Il se termine par une étude confrontant les deux notions de niches des mêmes espèces afin de déterminer le modèle d endémisme édaphique (spécialiste et refuge) pour des taxa endémiques. Le second axe vise à démontrer la faisabilité de mise en place des stratégies de conservation et de phytostabilisation en associant les taxa cupro-cobalticoles endémiques et non-endémiques dans des sols métallifères anthropogéniques. Il se clôture par un article de sélection de graminées candidates à la phytostabilisation afin d augmenter le choix des espèces à utiliser (Figure 6, Figure 7). 53

54 Les espèces ou les variétés végétales étudiées durant la thèse ont été sélectionnées selon leur distribution le long du gradient de concentration en Cu et Co, leurs caractéristiques dans les communautés (endémique, hyperaccumulatrice ou structurante), les priorités de conservation ainsi que leur forme de vie. Ces taxa annuels ou pérennes se développent dans les steppes ou les savanes steppiques, ce qui permet de couvrir une partie du gradient de concentration en ETM (Tableau 1). Même si certains taxa ont été étudiés dans la majorité des études de cette thèse, la diminution continue de certaines populations d espèces endémiques due à l exploitation a contraint le choix dans certaines étapes. L Article 1 présente une revue de la littérature à propos du lien existant entre la conservation des espèces végétales et l utilisation de celles-ci pour le bien-être de l Homme en décrivant l historique de la conservation et les concepts scientifiques impliqués dans ces démarches. L utilisation des végétaux dans des services écosystémiques est, en effet, un type d action pouvant associer la conservation des métallophytes avec les services de régulation, des services de soutien aux conditions favorables à la vie sur Terre et des services culturels. Ces services sont obtenus par la phytoremédiation. La végétation de l Arc Cuprifère Katangais représente un cas d étude remarquable pour mettre en relation la conservation et une technique de la phytoremédiation, la phytostabilisation. Cette revue de la littérature s inscrit dans un contexte sociétal et environnemental et renforce la démarche poursuivie dans cette thèse de doctorat cherchant à valoriser les métallophytes d intérêt conservatoire dans des techniques de réhabilitation. 54

55 Tableau 1. Synthèse des taxa étudiés dans minimum deux études de la thèse de doctorat. Légende des statuts UICN : EN= en danger, CR= critiquement en danger, NA= Taxon Statut UICN proposé Famille Structurant les communautés végétales Forme de vie Communauté Recherches effectuées au cours de la thèse Caractérisation des niches édaphiques conservation Réalisée Fondamentale phytostabilisation Installation Croissance Andropogon schirensis Hochst. ex A.Rich. NA Poaceae Pérenne Steppe - Savane steppique Article 2 Article 9 Eragrostis racemosa (Thunb.) Steud. NA Poaceae Pérenne Steppe Article 2 Article 9 Hyparrhenia diplandra (Hack.) Stapf var. diplandra NA Poaceae Pérenne Savane steppique Article 2 Article 9 Loudetia simplex (Nees) C.E.Hubb. NA Poaceae Pérenne Steppe - Savane steppique Article 2 Article 9 Monocymbium ceresiiforme (Nees) Stapf NA Poaceae Pérenne Steppe - Savane steppique Article 2 Article 9 Trachypogon spicatus (L.f.) Kuntze NA Poaceae Pérenne Steppe - Savane steppique Article 2 Article 9 Tristachya bequaertii De Wild. NA Poaceae Pérenne Savane steppique Article 2 Article 9 Endémique Anisopappus davyi S.Moore Asteraceae Pérenne Steppe Article 2 Article 7 Crotalaria cobalticola P.A.Duvign. & Plancke EN Fabaceae Annuelle Steppe Article 2 Article 4 Article 6 Article 7 Crotalaria peschiana P.A.Duvign. & Timp. CR Fabaceae Pérenne Savane steppique Article 2 Article 4 Article 8 Diplolophium marthozianum P.A.Duvign. EN Apiaceae Pérenne Steppe - Savane steppique Article 2 Article 6 Article 5 Article 6 Gladiolus ledoctei P.A.Duvign. & Van Bockstal EN Iridaceae Pérenne Steppe Article 2 Article 6 Article 6 Triumfetta welwitschii var. rogersii (N.E.Br.) Brummitt & Seyani NA Malvaceae Pérenne Savane steppique Article 2 Article 6 Article 6 Article 8 non évalué. 55

56 Axe 1 : Niche écologique et endémisme édaphique L étude des concepts de niche écologique et d endémisme édaphique à l échelle des espèces et de l organisation des communautés est essentielle dans le but d assurer la réussite des différentes étapes de la conservation allant de la germination au développement d individus viables. Le concept de niche écologique a été très peu étudié sur les espèces cuprocobalticoles endémiques de l Arc Cuprifère Katangais. Etant donné la complexité des interactions plante-sol, les facteurs déterminant l assemblage des communautés ont été sélectionnés dans ces études. Cinq des neuf articles sont consacrés à la compréhension des notions de niches édaphiques réalisées (Article 2, Article 3, Article 6) et fondamentales (Article 4, Article 5, Article 6) pour les espèces cupro-cobalticoles endémiques et non endémiques, ce qui représente une base scientifique nécessaire à la compréhension de la structure des communautés de cet écosystème (Figure 7). Même si cet axe se focalise sur des aspects conceptuels, les études apportent chacune des réponses pragmatiques et techniques concernant la conservation des espèces cupro-cobalticoles. La végétation cupro-cobalticole de l Arc Cuprifère Katangais se développant le long d un gradient de concentration en ETM offre un cas d étude unique pour caractériser la distribution des niches réalisées des espèces végétales le long d un gradient de toxicité. L Article 2 consiste en une modélisation de la réponse de 80 taxa le long des gradients de concentration en Cu et en Co (ou gradient de toxicité) à partir de données de présence-absence. L utilisation du modèle additif généralisé (ou GAM) a permis de modéliser des courbes de réponse sans forcer un modèle mathématique défini a priori. Cette étude a permis de caractériser la distribution des niches réalisées à l échelle de la communauté et de tester trois hypothèses portant sur les relations entre les optima et les largeurs des niches, identifiées sur d autres gradients non-métalliques. Dans un contexte d exploitation minière, la disparition rapide de la structure et de la composition de ces habitats ne permet pas d utiliser les méthodes classiques nécessitant un nombre conséquent de données. De plus, certains taxa cuprocobalticoles endémiques sont plus rares et forment des populations très localisées, ce qui les a exclus de la base de données de l Article 2. L Article 3 caractérise la niche réalisée le long des gradients de concentration en Cu, Co et du gradient de minéralisation (ratio C:N) de huit taxa cupro-cobalticoles endémiques via la modélisation de courbes de densité d individus et les statistiques descriptives. Une méthode d échantillonnage rapide et pragmatique a été développée sur base de relevés par transects parallèles aux gradients Cu- 56

57 Co afin d agir dans une situation critique de disparition de sites tout en ciblant des espèces rares et peu abondantes. Deux articles ont porté sur la caractérisation de la niche fondamentale le long d un gradient de Cu (ou réponse physiologique). Ces études ont majoritairement mesuré le processus de germination, les traits liés au développement des plants et la survie. L Article 4 compare la réponse au Cu de trois espèces d un même genre (Crotalaria) au stade d installation venant de communautés végétales distinctes le long du gradient de concentration en Cu. Six milieux de culture avec des concentrations en Cu croissantes ont été créés pour mettre en germination et en croissance les trois espèces. La germination de certaines espèces endémiques pouvant être entravée par des contaminations fongiques lors de leur reproduction, une seconde étude en conditions in vitro (ex situ, Article 5) a ciblé le caractère fongicide du Cu et son effet sur la germination de l espèce Diplolophium marthozianum. Trois concentrations en Cu croissantes ont été croisées avec trois traitements stérilisants (eau, javel, éthanol) afin de déterminer leur effet sur la germination et la contamination des graines. Pour finir, la détermination du modèle d endémisme édaphique des espèces cuprocobalticoles composant cette végétation est cruciale pour comprendre leur distribution et implémenter des stratégies de conservation. L Article 6 confronte la niche réalisée et la niche fondamentale de taxa endémiques le long du gradient de Cu. La niche réalisée a été calculée selon le même modèle que l Article 2 tandis que la niche fondamentale représente la réponse des individus (c.-à-d. performance à partir de la graine) à trois concentrations en Cu ajoutées dans un sol arable non-métallifère (0 mg kg -1, 100 mg kg -1 ou mg kg -1 Cu). Etant donné que différentes populations de la même espèce peuvent présenter des variations de performances sur un même substrat, les graines ont été choisies parmi trois populations conservées à la banque de graines de la Faculté des Sciences Agronomiques de l Université de Lubumbashi. Le test a été effectué en condition de climat local au jardin botanique de la faculté. 57

58 Axe 2 : Conservation et phytostabilisation La faisabilité de la mise en place et l optimisation des stratégies de conservation ex situ des espèces cupro-cobalticoles sont présentées dans les trois derniers articles (Article 7, Article 8 et Article 9). L Article 7 et l Article 8 présentent une comparaison de la régénération de taxa endémiques ou d intérêt scientifique à partir de semis sur des parcelles phytosabilisées par Microchloa altera mises en place récemment (Shutcha et al., 2015). La flore cuprocobalticole de l Arc Cuprifère Katangais comprend des espèces aux formes de vie très variées. En conséquence, deux des taxa choisis ont une forme de vie annuelle (ou pseudo-annuelle) et deux sont pérennes. Deux sites ont été expérimentés: le premier présentait un dispositif expérimental en parcelles divisées (couvert par une graminée, non-couvert croisé avec des amendements), le second était un site métallifère anthropogénique pollué notamment par du Cu. Considérant que la mise en place d association d espèces demande une connaissance sur les espèces endémiques mais aussi sur les espèces associées, l Article 9 a proposé, parmi sept graminées de l Arc Cuprifère Katangais, des candidates pour la phytostabilisation des sols pollués dans la région selon des critères basés sur des paramètres végétatifs et reproductifs. 58

59 Figure 7. Synthèse des objectifs des Axes de la thèse de doctorat ainsi que des objectifs et de la méthodologie des neufs articles. 59

60

61 Revue de la littérature 61

62

63 Revue de la littérature -Article 1 Accepté - Concilier la conservation des métallophytes et la phytoremédiation des sols pollués par les éléments traces métalliques BOISSON Sylvain, NGOY SHUTCHA Mylor & MAHY Grégory Résumé La conservation de la biodiversité est devenue un challenge social, politique et économique au niveau mondial. Etant donné que certaines régions du monde sont fortement soumises à l anthropisation et à ses impacts, il est indispensable de modifier notre perception et notre réflexion en ce qui concerne la conservation de la biodiversité et la restauration écologique. Ce chapitre explore les possibilités existantes et les adapte pour associer la conservation des espèces végétales d intérêt avec la phytoremédiation des sols pollués. Cette nouvelle stratégie semble encourageante pour conserver le lien entre la société et l environnement. En utilisant la diversité végétale locale, il est possible de créer de nouvelles communautés végétales afin de réhabiliter des sols pollués par les éléments traces métalliques. Le cas de la ville de Lubumbashi et de la flore de l Arc Cuprifère Katangais en est un exemple marquant qui a été récemment testé avec succès. Cependant, d autres études sur les métallophytes sont essentielles pour les utiliser de manière appropriée. 63

64 La conservation de la biodiversité, de nouveaux objectifs pour 2020 La conservation de la biodiversité est reconnue comme un enjeu majeur à l échelle mondiale depuis plus de 20 années (United Nations, 1992). La prise de conscience de son importance avait déjà émergée dans le monde scientifique sous la forme d une discipline de crise appelée Biologie de la conservation (Soulé, 1985). Cette science constitue une approche pluridisciplinaire (écologie, biogéographie, démographie, génétique, systématique, biologie évolutive, etc.) nécessaire à la mise en place de stratégies de conservation en réponse à un déclin de la biodiversité (Soulé, 1985; Primack, 2010). Les stratégies principales qui en découlent sont divisées entre la conservation in situ et ex situ. Face à la pression continue sur les écosystèmes et la difficulté de mettre en place ces stratégies dans certains cas, l Ecologie de la restauration a vu le jour et poursuit un objectif similaire avec une philosophie différente. Tandis que la Biologie de la Conservation tente de préserver et maintenir les habitats et la biodiversité à long terme, l Ecologie de la restauration assure que le processus de dégradation des habitats soit réversible par des actions intentionnelles ou assistées (Young, 2000; Clewell et al., 2004). Cependant, ces deux disciplines basent leur stratégie en lien avec un écosystème historique à conserver ou à atteindre appelé écosystème de référence (Bullock et al., 2011). Contrairement à la Biologie de la Conservation qui a une identité théorique, les concepts de base de l Ecologie de la restauration, plus appliqués, font encore l objet de discussions (Van Andel et al., 2006) Malgré la complémentarité des deux disciplines écologiques, les mesures stratégiques prises lors de la Convention pour la Diversité Biologique (CDB) en 1992 et en 2002 n ont pas réussi à remplir les objectifs de maintien de la biodiversité (Secretariat of the Convention on Biological Diversity, 2014). Les causes principales de la perte de la biodiversité en sont la diminution et la dégradation des habitats dues à la pression anthropique (Millenium Ecosystem Assessment, 2005; Vié et al., 2008). En colonisant les milieux naturels, l Homme change l occupation et la fonction primaire des écosystèmes pour les adapter à ses besoins : agriculture, élevage, pêche, industries, constructions, voiries, etc. (Vitousek, 1997). Vu l hétérogénéité de développement dans les populations humaines, la difficulté dans la mise en place de telle convention réside dans l importance donnée à la biodiversité par rapport à la société (Adams et al., 2004). Le plan stratégique Vision de cette convention (c.- à-d. Objectifs d Aichi) poursuit 20 objectifs en développant des actions conjointes à toute échelle afin de comprendre, de conserver et d utiliser la biodiversité (UNEP, 2015). Un des 64

65 objectifs spécifiques est d utiliser de manière durable la diversité végétale au travers, entre autres, d innovations afin d appuyer la sécurité alimentaire et la santé (UNEP, 2015). La Vision 2050 suit, dont l objectif final propose que «la diversité biologique soit valorisée, conservée, restaurée et utilisée avec sagesse, en assurant le maintien des services d écosystèmes, en maintenant la planète en bonne santé et en procurant des avantages essentiels à tous les peuples». Services, restauration et traits fonctionnels L utilisation de la diversité biologique comme un outil ou un service à la société est un concept très récent qui est en plein essor ces dernières années (Daily, 1997; UNEP, 2015). Les services écosystémiques peuvent être définis comme les bénéfices que les humains obtiennent à partir des écosystèmes où les services sont le résultat des processus et des fonctions de l écosystème ; le stockage de carbone, les flux d eau et la fertilité du sol en sont les exemples récurrents (Millenium Ecosystem Assessment, 2005). Même si le lien entre la biodiversité et les services d un écosystème est complexe à déterminer, il a été démontré que la richesse spécifique d un écosystème est positivement corrélée à la qualité des services fournis (Balvanera et al., 2006). L état de santé des écosystèmes est déterminant pour la biodiversité. Une méta-analyse récente a d ailleurs montré une corrélation positive entre la restauration d écosystème, la biodiversité et la fourniture de services écosystémiques (Rey Benayas et al., 2009). Cependant, le succès de la restauration de l écosystème et de ces services n est pas systématique et dépend de l écosystème de référence considéré (Bullock et al., 2011). Le choix de la référence est une étape aussi importante que complexe ; elle représente un état historique non dégradé naturel ou seminaturel de l écosystème (Egan et al., 2005). Quand le retour à la composition, aux conditions, aux fonctions et aux services historiques de l écosystème est impossible ou hors d atteinte sans manipulation et gestion très conséquentes et coûteuses, il convient de se détacher des concepts traditionnels. Le concept de «nouvel écosystème» ou «écosystème émergeant» est le résultat d une considération accrue des changements rapides de nos écosystèmes qui amènent à la création d autres écosystèmes. Ils diffèrent en composition et/ou en fonction par rapport aux écosystèmes historiques présents dans les biomes reconnus (Hobbs et al., 2006). De nouveaux écosystèmes peuvent être volontairement créés pour fournir des services 65

66 équivalents ou différents de l écosystème historique (Hobbs et al., 2009). Ce concept permet de ne pas tenir compte de l écosystème de référence ou historique et autorise donc des changements d ampleur variable dans la composition du nouvel écosystème. Dans ce cadre, l idée de préservation ou du maintien de l intégrité des écosystèmes à toutes échelles n est pas assurée, ce qui entraine un conflit avec les objectifs de la conservation de la biodiversité et la Biologie de la Conservation (García-Palacios et al., 2010). Ce concept ne peut donc se présenter comme une solution unique mais doit rester une alternative s approchant de la notion de réhabilitation d un écosystème face à l impossibilité d appliquer les stratégies issues de la Biologie de la Conservation et de l Ecologie de la restauration (Hobbs et al., 2009; Bullock et al., 2011). L évaluation du succès de restauration, de conservation ou de réhabilitation de communautés végétales est une approche complexe qui doit s appuyer sur des paramètres bien définis et quantifiables dans le temps. Une manière classique en écologie est d évaluer les traits fonctionnels de la ou des communauté(s) végétale(s) (Lavorel et al., 1997; Cornelissen et al., 2003). Les traits fonctionnels sont définis comme des paramètres morphologiques, physiologiques ou phénologiques, mesurés au niveau de l individu, qui influent sur sa performance dans l écosystème (Violle et al., 2007). Cette classification relie la forme et la fonction des espèces et est centrée sur l utilisation des traits de vie des plantes. Ils incluent par exemple les caractéristiques morphologiques racinaires et aériennes, la composition biochimique des feuilles et des racines ou les caractéristiques reproductives comme la masse et le nombre de graines (Lavorel et al., 1997, 2002). Certaines espèces semblent avoir des traits fonctionnels qui leur confèrent un rôle plus dominant que d autres dans l écosystème (Tilman, 1997), c est notamment le cas des espèces invasives (Hobbs et al., 2006). L approche par les traits fonctionnels est également utilisée dans les études de réponses des communautés aux perturbations ou aux stress et pour évaluer leur capacité de résilience (Castro et al., 2010; Mabry and Fraterrigo, 2009) (Ilunga wa Ilunga et al., 2015; Delhaye et al., 2016). Bien qu en Ecologie de la restauration, l objectif est de déterminer le rôle fonctionnel des espèces de l écosystème de référence afin de rétablir l intégrité fonctionnelle de l écosystème, l approche est très différente en réhabilitation ou en création de nouveaux écosystèmes (Hobbs et al., 2006). Le rôle fonctionnel des espèces de la nouvelle communauté ne serait plus identifié par rapport à l écosystème de référence mais bien par rapport aux services à fournir (Ilunga wa Ilunga et al., 2015). 66

67 La pollution des sols et les phytotechnologies Une des causes actuelles de la dégradation de la qualité des écosystèmes est la pollution par les éléments traces métalliques (Secretariat of the Convention on Biological Diversity, 2014). Les sources anthropiques des contaminations métalliques en sont principalement l exploitation et la transformation des gisements, la production d énergie, l agriculture et les boues d épuration qui mènent à leur dispersion dans l air, l eau et le sol (Bradshaw et al., 1965; Prasad et al., 2003; Kozlov et al., 2006). Ces éléments sont en production croissante à cause de la hausse de la demande et l émergence de plusieurs pays d Asie (ICMM, 2012). Ces métaux, non dégradables, interagissent physiquement et chimiquement avec d autres facteurs édaphiques qui modifient leur mobilité et leur disponibilité (Dube et al., 2001). Dans le sol, leur biodisponibilité varie selon le ph, la capacité d échange cationique, la teneur en matière organique, la quantité et le type d argile, les oxydes de Fe et Mn et le potentiel redox (Wong, 2003). Une fois présents dans la fraction biodisponible, ils s accumulent dans les tissus des organismes vivants qui sont d abord les plantes et les microorganismes. Ce phénomène de bioaccumulation se répercute le long de la chaine alimentaire et demeure responsable de problèmes de toxicité chez les êtres vivants végétaux et animaux (Järup, 2003; Ali et al., 2013). Chez les plantes, les éléments traces métalliques (Fe, Cu, Cd, Cr, Pb, Hg, Ni, V) catalysent notamment la production de radicaux libres qui bouleversent plusieurs mécanismes physiologiques cellulaires, détruisent l ADN cellulaire et ainsi induisent un stress (Stohs et al., 1995; Furini, 2012). A l extrême, de hautes teneurs en métaux, combinées avec des caractéristiques édaphiques particulières peuvent mener à la disparition complète de la végétation naturelle exposant les zones environnantes aux métaux (Vangronsveld et al., 1995). Ces zones polluées perdent ainsi leur condition, leur composition et leur fonction historique puis deviennent une menace permanente pour l environnement et la santé humaine (Järup, 2003; Banza et al., 2009; Cheyns et al., 2014). Afin de réduire l abondance des éléments traces métalliques dans l environnement, des techniques de remédiation physiques (p. ex. excavation) et chimiques (p. ex. chélation, précipitation) des sols existent (Bolan et al., 2014). Elles restent cependant couteuses et laborieuses à mettre en place à grande échelle (Cunningham et al., 1996). L usage des plantes pour fournir des services à la société est apparu suite aux études sur les interactions plante-sol dans les années 90, la phytoremédiation (Cunningham et al., 1995; van Breemen et al., 1998). Cette technique inclut un ensemble de méthodes dont l objectif est de réduire, 67

68 déplacer, dégrader ou immobiliser les composés toxiques dans les sols pollués grâce à la mise en place d une végétation (Cunningham et al., 1995; Peer et al., 2005). L avantage est de combiner un coût plus faible que les techniques de remédiation classiques, une stratégie durable et un impact paysager plutôt favorable. On y distingue entre autres la phytoextraction qui permet d extraire les particules métalliques du sol vers les tissus végétaux grâce à l utilisation de plantes accumulatrices ; la phytostabilisation qui permet d adsorber, d absorber ou de précipiter les métaux dans les tissus végétaux ou le sol au niveau de la rhizosphère grâce à des plantes tolérantes et la phytodégradation qui permet de transformer les éléments organiques polluants en matériaux non-toxiques dans la rhizosphère (Berti et al., 2000; Ali et al., 2013). La phytoremédiation est, contrairement à la restauration écologique, une méthode qui n a pas toujours pour ambition de restaurer l ensemble des caractéristiques de l écosystème historique. Tandis que la phytoextraction pourrait être envisagée comme un service qui, à long terme, restaure les conditions d origine de l écosystème, la phytostabilisation s oriente plutôt vers la mise en place de nouveaux écosystèmes qui doivent nécessairement perdurer à long terme pour éviter la mobilisation des métaux séquestrés. Ainsi, un nouvel écosystème composé d autres espèces se constitue à court terme. D autres espèces végétales de la matrice environnante peuvent venir enrichir la zone remédiée. Ainsi, le nouvel écosystème fournit par exemple des services de régulation (p. ex. réduction de maladie liée à l intoxication), des services de soutien aux conditions favorables à la vie sur Terre (p. ex. immobilisation des métaux toxiques) et des services culturels (Millenium Ecosystem Assessment, 2005). Les métallophytes, une ressource biologique Plusieurs régions du monde présentent des sols aux teneurs en éléments traces métalliques naturellement élevées appelés sols métallifères. Généralement, ces zones naturelles sont dispersées et isolées dans une matrice contenant des sols non-enrichis (Johnson, 1978). Les communautés végétales uniques de ces milieux sont fortement distinctes de celles de la matrice végétale environnante et contiennent souvent des espèces ou des écotypes, souvent endémiques, tolérants spécifiquement les métaux (Shaw, 1990; Wolf, 2001; Ernst, 2006; Baker et al., 2010). Ces taxa, les métallophytes, ont acquis des mécanismes physiologiques complexes au travers de processus (micro-)évolutifs qui leur permettent de 68

69 tolérer, de survivre et de se reproduire jusqu à un seuil déterminé de toxicité dépendant de l espèce (Antonovics et al., 1971; Ernst, 2006; Alford et al., 2010; Furini, 2012). Selon l affinité des métallophytes pour les sols métallifères, deux comportements existent : les métallophytes absolus sont des espèces qui ont une tolérance très élevée pour les éléments traces métalliques avec une distribution dépendante de ces éléments ; les métallophytes facultatifs sont des écotypes ou des sous-espèces qui ont une tolérance spécifique mais qui sont également retrouvés sur des sols non enrichis en éléments traces métalliques. Dans les communautés végétales contenant des métallophytes, peuvent aussi être retrouvées des espèces associées, issues de la matrice végétale dominante, tolérante (pseudo-métallophytes) ou non-tolérante (Baker et al., 2010). Parmi les métallophytes absolus, trois stratégies de tolérance ont été identifiées : les (hyper)accumulateurs qui concentrent dans leurs tissus aériens les métaux en quantité plus que proportionnelle à celle du sol ; les exclueurs qui limitent le transport et l entrée des métaux vers les parties aériennes et les indicateurs qui accumulent une quantité de métaux proportionnelle à celle du sol (Baker, 1981). Les métallophytes présentent une grande diversité d espèces, même si beaucoup d hyperaccumulateurs connus appartiennent à la famille des Brassicaceae (Lone et al., 2008; Baker et al., 2010; Krämer, 2010). Etant donné leur adaptations physiologiques remarquables à tolérer ou accumuler les éléments traces métalliques, ces communautés ont souvent été pointées pour leur ressources phytogénétiques utilisables dans des stratégies de phytoremédiation de sols pollués (Leteinturier et al., 1999; Whiting et al., 2002; Jacobi et al., 2007; Shutcha et al., 2010; Furini, 2012). Chacune des stratégies constituant une technique de phytoremédiation demande de sélectionner des traits fonctionnels définis. La dépollution de zones contenant des excès en éléments traces métalliques demande idéalement d utiliser des espèces accumulatrices productives, facilement cultivables et repoussant les herbivores afin d exporter les métaux (phytoextraction) rapidement sans aucun danger pour les animaux (Ali et al., 2013). Cependant, l accumulation ou l hyperaccumulation des métaux dans les tissus aériens des plantes sont des stratégies de tolérance spécifiques à certains métallophytes pour certains métaux uniquement (Pollard et al., 2002; Boyd, 2004; Verbruggen et al., 2009). Ali et al. (2013) ont dressé une liste non exhaustive des hyperaccumulateurs de As, Cd, Cu, Mn, Ni, Pb, Zn sur une estimation de 450 existants. Ce mécanisme d accumulation à un coût physiologique qui impacte la production de biomasse et la vitesse de croissance des plantes, 69

70 ce qui rend le processus d extraction très lent (Cunningham et al., 1995; Singh et al., 2007). Des méthodes standards de croisement sont possibles. Aussi, l utilisation de plants transgéniques, quoi que leur commerce soit encore discuté, a récemment été explorée mais ne résout pas encore l ensemble des problèmes comme l accumulation lors des contaminations mixtes (Pilon-Smits et al., 2002; Pollard et al., 2002; Cherian et al., 2005; Stone, 2010; Furini, 2012). La phytostabilisation, plus aisée à mettre en place, demande d utiliser des espèces ou des variétés tolérantes qui ont un couvert souterrain et aérien important en vue de limiter la dispersion des métaux par érosion éolienne, lixiviation ou ruissellement (Neuman et al., 2006). Après la mise en place par semis ou par transplantation du couvert qui peut demander de fertiliser le sol, les plants demandent peu de maintenance (Berti et al., 2000; Córdova et al., 2011). Le choix des espèces est déterminant et les traits fonctionnels sélectionnés doivent assurer la longévité et l efficience de la stratégie. Des espèces natives de la région, avec une rhizosphère dense et volumineuse, tolérantes aux métaux en question et limitant l accumulation dans les tissus aériens sont conseillées (Berti et al., 2000; Mendez et al., 2008). Des graminées tolérantes natives et pérennes ont déjà été utilisées avec succès dans certains pays du monde (O Dell et al., 2006; Shutcha et al., 2010; Parraga-Aguado et al., 2014). Conservation et phytoremédiation Les habitats métallifères, malgré leur composition végétale unique et valorisable, sont propices à l exploitation minière du fait de leur intérêt économique (Whiting et al., 2002). Ces dernières années, les activités minières ont détruit et détruisent encore ces écosystèmes, mettant en danger la survie des métallophytes (Brooks et al., 1992; Ernst, 2000; Whiting et al., 2002, 2004; Bizoux et al., 2004; Jacobi et al., 2007; Faucon et al., 2010; Mendoza- Hernández et al., 2012). Ces espèces sont détruites dès les premières étapes de l exploitation de surface et sont difficilement restaurables après les activités du fait des modifications physico-chimiques des écosystèmes (Ernst, 2000; Whiting et al., 2002). Les actions de conservation sont donc urgentes afin d éviter leur disparition. Cependant, elles sont encore difficilement justifiables et coûteuses vis-à-vis des industries (Whiting et al., 2004). Les habitats secondaires créés par l industrie minière peuvent pourtant être favorables à l installation de certaines espèces métallicoles d intérêt conservatoire (Bizoux et al., 2008). Etant donné la volonté du plan stratégique Vision de la CDB à améliorer notre interaction avec la biodiversité, l ensemble des industries liées aux pays qui 70

71 ont adopté cette convention vont devoir se soumettre à ses exigences (ICMM, 2014; Secretariat of the Convention on Biological Diversity, 2014). L intégration de la conservation des métallophytes retrouvés sur un site minier avec la phytoremédiation des sols pollués répond aux exigences en matière de biodiversité et offre plusieurs services à la société liés à l environnement et la santé publique. Parmi les méthodes, la phytoextraction semble commercialement viable tandis que la phytostabilisation des métaux est considérée comme une méthode de gestion des risques (Whiting et al., 2004; Ali et al., 2013). Au sein d une industrie minière, plusieurs étapes sont nécessaires à partir de l initialisation des démarches pour exploiter une zone hébergeant des métallophytes. Les stratégies de conservation-phytoremédiation doivent être un compromis entre le coût, la complexité technique et l objectif de conservation (Barceló et al., 2003; Whiting et al., 2004). La caractérisation des conditions physico-chimiques du sol du site minier doit être effectuée en premier lieu pour déterminer les concentrations en éléments traces métalliques et en nutriments. En général, des sondages sont effectués par l industrie pour adapter l exploitation et la transformation du gisement. L identification des métallophytes présents sur les sols du site minier et leur écologie est une étape cruciale pour la mise en place des stratégies de conservation et de phytoremédiation. Les métallophytes présentant un intérêt conservatoire peuvent être des hyperaccumulateurs, des espèces ou des écotypes endémiques des sites métallifères ou des espèces présentant un statut de menace officielle de l Union pour la Conservation de la Nature IUCN (Vié et al., 2008). Selon les caractéristiques édaphiques du site minier et les métallophytes présents, un ou plusieurs sites pollués peuvent être choisis selon les métaux présents et leur teneur en éléments traces métalliques. Ces sites pollués peuvent se trouver dans la concession minière (remblais, etc.) ou à proximité d une usine de transformation de minerai (sols dégradés, etc.) tant qu ils soient soumis à des conditions climatiques similaires. Etant donné l hétérogénéité de certaines surfaces polluées, des espèces ayant une plus large distribution écologique parmi les concentrations en métaux sont favorables. Par exemple, pour les sols serpentiniques, certaines espèces présentent des écotypes spécialisés tandis que pour d autres, aucune variation n est constatée au sein de l espèce (O Dell et al., 2011). La priorité dans le choix du site pollué doit être la proximité avec des habitations ou de cours d eau afin de réduire la dispersion des métaux. Un amendement organique ou calcaire peut être nécessaire pour installer la végétation et reste moins coûteux que l ajout d un sol superficiel. Il permet de réduire la biodisponibilité des métaux avant l installation des plantes (Singh et al., 2007; 71

72 Alvarenga et al., 2008; Mendez et al., 2008; Shutcha et al., 2010; Córdova et al., 2011). Lorsque le site est prêt, des propagules ou des individus sont collectés afin de les amener au niveau du site à phytoremédier. En général, la transplantation d individus est plus efficace que le semis mais demande beaucoup plus d investissement (Mendez et al., 2008). Les stratégies de phytoremédiation sont souvent étudiées par espèce. Cependant, dans les écosystèmes naturels, les espèces végétales coexistent et interagissent ensemble au sein des communautés (Bruno et al., 2003; Tirado et al., 2005; Kawai et al., 2007; Verdú et al., 2008). Dans l optique de conserver les espèces à long terme sur les sols pollués pour extraire les métaux ou pour les stabiliser, il semble intéressant d utiliser une approche pluri-espèces. Ainsi, les métallophytes d intérêt conservatoire pourraient être associées à des espèces plus communes des communautés. L écosystème nouvellement créé entre dans une stratégie de conservation ex situ et bénéficie d un intérêt double. Premièrement, il peut être une source de propagules pour enrichir des sites conservés, des pépinières ou des sites à restaurer. Deuxièmement, la couverture végétale limite la dispersion des métaux. De la même manière, les parcelles de phytoremédiation peuvent accueillir de nouveaux individus ou de nouvelles espèces selon leurs dimensions, leurs conditions et l abondance des espèces déjà présentes. En fin d exploitation, le site minier pourra être en partie restauré avec les espèces conservées dans les parcelles. Si les parcelles de phytoremédiation sont conservées à long terme, elles seront une source de propagules pour alimenter d autres sites et d autres stratégies. Enjeux au Katanga L Le sud de la province du Katanga (R.D. Congo) et le nord de la Zambie ont été le siège d activités métallogéniques intenses amenant à la formation de gisements métallifères (Cailteux et al., 2005; Kampunzu et al., 2009). Ces gisements se distinguant de la forêt claire dominante (le Miombo) sous la forme de collines sont essentiellement composés de cuivre (Cu) et de cobalt (Co) associés à d autres métaux tels que le zinc (Zn), le plomb (Pb) et l uranium (U) (Duvigneaud et al., 1963; Campbell, 1996; Cailteux et al., 2005; Kampunzu et al., 2009). Les centaines de collines sont dispersées sous forme d un arc appelé le Copperbelt et figurent parmi les gisements de cuivre et de cobalt les plus importants au monde. La République Démocratique du Congo détient la partie katangaise appelée Arc Cuprifère Katangais et demeure actuellement le premier producteur mondial de cobalt et le cinquième producteur de cuivre (USGS, 2015). Toutefois, ces gisements ont été exploités de 72

73 manière artisanale et non intensive pendant plusieurs siècles. L industrialisation de l exploitation a eu lieu à partir du XX ème siècle due à la colonisation (UMHK, 1956). Plusieurs villes dont Kolwezi, Likasi, Kipushi et Lubumbashi, ont alors dès 1940 vu l installation d usines assurant la transformation du minerais en métal directement exportable par chemin de fer (UMHK, 1956). Particulièrement, la ville de Lubumbashi (Elisabethville) représente une ville minière type avec en son centre l usine de transformation appartenant à la Général Carrière des Mines (GECAMINES, Figure 1-a,b) depuis l indépendance (Prasad, 1989). Pendant plusieurs dizaines d années, les processus d extraction chimiques des métaux ont conduit à la pollution des sols et des eaux environnants par les retombées de particules atmosphériques métallifères et la formation de pluies acides à l ouest de la ville (cône de pollution, Figure 1-c,e,f) dues aux vents sud-est dominants (Mbenza et al., 1989; Mpundu et al., 2011). Les teneurs en cuivre disponibles dépassent les mg.kg sol -1 et atteignent mg.kg sol -1 par endroit, équivalent à 70 fois les teneurs moyennes en cuivre rencontrées en ville (Figure 2) (Mpundu et al., 2011; Narendrula et al., 2012). Le paysage du cône de pollution diffère clairement des zones non polluées avoisinantes par une végétation plus basse et distribuée de manière plus fragmentée due essentiellement à la présence en excès d éléments traces métalliques en surface (Figure 1-e) (Vranken et al., 2013). Cette zone, initialement inhabitée était classée comme zone agricole en 1990 malgré l état de ses sols (Bruneau et al., 1990). Elle est aujourd hui habitée et des projets d aménagement urbanistique sont en cours (Figure 1-e). Cependant, l exposition des habitants aux différents éléments traces métalliques comme le cuivre et le cobalt est très élevée dans la région, ce qui entraine des conséquences graves sur la santé des habitants par leur abondance dans l air, l eau, le sol et in fine leur bioaccumulation dans les cultures maraichères et les tissus animaux (Figure 1-c,d) (Banza et al., 2009; Manda et al., 2010; Mpundu et al., 2013; Cheyns et al., 2014). Une récente étude a d ailleurs mis en évidence que l ajout d amendement organique ou calcaire dans le sol ne suffisait pas pour limiter le transfert des métaux dans les légumes consommés (Mpundu et al., 2014). 73

74 Figure 1. Photographies du centre (usine de la GECAMINES) et de l ouest de la ville Lubumbashi (cône de pollution). a. Terril de Lubumbasi situé au sud-est du cône de pollution. b. Usine d extraction de minerais de la GECAMINES située à proximité du terril et responsable des rejets dans l atmosphère. c. Paysage dans le cône de pollution. A l avant plant, une parcelle cultivée devant une zone nue à proximité des habitations. d. Plant de maïs adulte chlorosé cultivé dans le cône de pollution. e. Vue générale vers l ouest prise dans le cône de pollution, En avant plant une zone colonisée sporadiquement par des Setaria sphacelata et Celosia trigyna avec au loin à droite un projet de construction d habitation. f. Vue dirigée vers le nord prise dans le cône de pollution. Etat de la végétation sous les lignes à haute tension. 74

75 Pour pallier aux problèmes d exposition des habitants de cette zone polluée, les méthodes de remédiation classiques demanderaient d excaver le sol ou d utiliser des méthodes chimiques couteuses pour extraire ou immobiliser les métaux (Berti et al., 2000). Les technologies végétales d extraction (c.-à-d., la phytoextraction) sont également peu efficaces face à des niveaux de concentration en éléments traces métalliques aussi élevés (Lasat, 2000), les espèces (hyper)accumulatrices étant souvent peu productives en terme de biomasse (Cunningham et al., 1995; Ernst, 1996; Singh et al., 2007). Dès lors, la phytostabilisation a été privilégiée dans ce quartier de Lubumbashi. Shutcha (2010) a comparé la régénération de trois graminées indigènes en conditions contrôlées sur des substrats quatre fois moins enrichis en cuivre que la zone en question. Etant donné les taux de survie élevés de Microchloa altera sur les sols pollués non amendés, comparé à Monocymbium ceresiiforme et Cynodon dactylon, cette espèce a été installée avec succès à plus grande échelle dans un champ expérimental situé dans le cône de pollution avec et sans amendement. Même si les individus se maintiennent depuis sa mise en place en 2009, leur capacité d autopropagation sur les substrats pollués semble très limitée à cause d une faible production de graines et d une inhibition totale de la germination à 4000 mg Cu.kg sol -1 (Shutcha et al., 2013). L Arc Cuprifère Katangais héberge plus de 550 espèces végétales dont environ 50 espèces graminées comme Eragrostis racemosa, Andropogon schirensis, Monocymbium ceresiiforme et Loudetia simplex présentant des traits fonctionnels de croissance et reproductifs conformes à la phytostabilisation des sols pollués par les éléments traces métalliques dans la région (Duvigneaud et al., 1963; Leteinturier, 2002; Boisson, Le Stradic, et al., 2015). A partir des espèces recommandées par (Boisson, Le Stradic, et al., 2015), des semis devrait être testés en conditions réelles afin d augmenter la diversité d espèces et ainsi la diversité de traits permettant d immobiliser plus efficacement les métaux. 75

76 Figure 2. Répartition des teneurs en cuivre (mg.kg-1) des sols dans la ville de Lubumbashi et de ces alentours. Le cône de pollution à l ouest de la ville représente une zone aux teneurs en cuivre élevées (Mpundu et al., 2011) La flore de l Arc Cuprifère Katangais recèle d espèces métallophytes (ou cuprophytes) endémiques et hyperaccumulatrices (Chipeng et al., 2010; Faucon et al., 2010; Faucon, Chipeng, et al., 2012; Lange et al., 2014; Boisson et al., 2016), se développant sur des sols ayant des teneurs disponibles en cuivre et en cobalt allant respectivement jusque mg.kg -1 et jusque 1000 mg.kg -1 (Duvigneaud et al., 1963; Leteinturier, 2002; Séleck, Bizoux, et al., 2013). Au total, 57 espèces endémiques ayant des formes de vie et des capacités de tolérance différentes ont été identifiées à ce jour (Faucon et al., 2010; Champluvier, 2011; Faucon, Chipeng, et al., 2012; Séleck, Lebrun, et al., 2013). L intensification des activités minières ces dernières décennies entraine une destruction accrue des collines et des écosystèmes avoisinants. L exploitation du gisement se faisant à ciel ouvert, les impacts sont irrémédiables sur la végétation unique. De plus la construction d infrastructures, de voiries et d installations transforme entièrement l écosystème. Selon une évaluation basée sur les critères IUCN, plus de 65 % des espèces endémiques sont en danger critique d extinction (CR), 3 % en danger (EN), 9 % vulnérables (VU) et 9 % se sont déjà éteintes (Faucon et al., 2010). Plusieurs études ont mis en évidence les relations entre les communautés végétales, les espèces et les éléments du sol (Chipeng et al., 2010; Faucon, Tshilong, et al., 2012; Saad et al., 2012; Ilunga wa Ilunga et al., 2013; Séleck, Bizoux, et al., 2013; Lange et al., 2014). Elles ont contribué significativement à la mise en place de 76

77 stratégies de conservation ex situ telles qu une banque de graines et un jardin botanique (Boisson et al., 2012; Godefroid et al., 2013). Même si cette biodiversité devrait être restaurée après exploitation, peu d études au Katanga ont testé le succès de la restauration des espèces après exploitation ou sur un sol similaire via des graines ou des individus. Une seule étude a évalué, via une approche consistant à mesurer les traits fonctionnels, le succès de recolonisation spontanée sur un sol pollué (Ilunga wa Ilunga et al., 2015). Certains auteurs ont pointé la capacité de certaines espèces endémiques ou non endémiques à coloniser spontanément les zones où les sols ont été perturbés (Malaisse et al., 1982; Faucon et al., 2011). Dans un effort de conservation, il serait pourtant essentiel de valoriser les espèces d intérêt scientifique (c.-à-d. hyperaccumulatrices) ou endémiques menacées dans des technologies végétales telle que la phytostabilisation des sols pollués par les éléments traces métalliques. A ce jour, une étude a testé avec succès une stratégie originale visant à associer une espèce phytostabilisatrice avec des espèces endémiques dans le cône de pollution (Boisson, Collignon, et al., 2015). Cette approche amène à la création de communautés végétales différentes des écosystèmes de référence sur des sites qui présentent une opportunité pour la conservation des espèces métallophytes menacées et qui demeurent un réel fléau pour la santé des habitants. Avec un choix judicieux des espèces, la création de ces communautés originales représentent donc une alternative de choix pour la réhabilitation des sols à long terme dans les zones comportant des sols pollués par les éléments traces métalliques. Références Adams W.M., Aveling R., Brockington D., Dickson B., Elliott J., Hutton J., Roe D., Vira B. & Wolmer W., Biodiversity conservation and the eradication of poverty. Science (80-. ). 306(2004), Alford É.R., Pilon-Smits E.A.H. & Paschke M.W., Metallophytes a view from the rhizosphere. Plant Soil 337(1-2), Ali H., Khan E. & Sajad M.A., Phytoremediation of heavy metals-concepts and applications. Chemosphere 91(7), Alvarenga P., Gonçalves A.P., Fernandes R.M., de Varennes A., Vallini G., Duarte E. & Cunha-Queda A.C., Evaluation of composts and liming materials in the phytostabilization of a mine soil using perennial ryegrass. Sci. Total Environ. 406(1-2), Antonovics J., Bradshaw A.D. & Turner R.G., Heavy metal tolerance in plants. Adv. Ecol. Res. 7, Baker A.J.M., Accumulators and excluders strategies in the response of plants to heavy metals. J. Plant Nutr. 3(1-4), Baker A.J.M., Ernst W.H.O., Van der Ent A., Malaisse F. & Ginocchio R., Metallophytes: the unique biological resource, its ecology and conservational status in Europe, central Africa and Latin America. In: Batty, L., Hallberg, K. eds. Ecology of Industrial Pollution Balvanera P., Pfisterer A. & Buchmann N., Quantifying the evidence for biodiversity effects on ecosystem functioning and services. Ecol. Lett. 9(10), Banza C.L.N., Nawrot T.S., Haufroid V., Decrée S., De Putter T., Smolders E., Kabyla B.I., Luboya O.N., Ilunga A.N., Mutombo A.M. & Nemery B., High human exposure to cobalt and other metals in Katanga, a mining area of the Democratic Republic of Congo. Environ. Res. 109(6),

78 Barceló J. & Poschenrieder C., Phytoremediation : principles and perspectives. Contrib. to Sci. 2(3), Berti W.R. & Cunningham S.D., Phytostabilization of metals. In: Raskin, I., Ensley, B.D. eds. Phytoremediation of Toxic Metals: Using Plants to Cleanup the Environment. John Wiley & Sons, Inc., New York, Bizoux J., Brevers F., Meerts P., Graitson E. & Mahy G., Ecology and conservation of Belgian populations of Viola calaminaria, a metallophyte with a restricted geographic distribution. Belgian J. Bot. 137(1), Bizoux J.P., Daïnou K., Raspé O., Lutts S. & Mahy G., Fitness and genetic variation of Viola calaminaria, an endemic metallophyte: Implications of population structure and history. Plant Biol. 10(6), Boisson S., Collignon J., Langunu S., Lebrun J., Shutcha M.N. & Mahy G., Concilier la phytostabilisation des sols pollués avec la conservation de la flore cuprocobalticole dans la région de Lubumbashi (R.D. Congo) : une stratégie nouvelle pour valoriser les écosystèmes extrêmes? In: Bogart, J., Halleux, J.-M. eds. Territoires Périurbains: Développement, Enjeux et Perspectives Dans Les Pays Du Sud. Presses Agronomiques de Gembloux, Gembloux, Boisson S., Le Stradic S., Collignon J., Séleck M., Malaisse F., Ngoy Shutcha M., Faucon M.-P. & Mahy G., Potential of copper-tolerant grasses to implement phytostabilisation strategies on polluted soils in South D. R. Congo. Environ. Sci. Pollut. Res. Boisson S., Le Stradic S., Commans M., Dumont A., Leclerc N., Thomas C. & Mahy G., Copper tolerance of three Crotalaria species from southeastern D. R. Congo at the early development stage. Biotechnol. Agron. Soc. Environ. 20(2), Boisson S., Lebrun J., Séleck M. & Mahy G., Copperflora Website. (20/05/2013). Bolan N., Kunhikrishnan A., Thangarajan R., Kumpiene J., Park J., Makino T., Kirkham M.B. & Scheckel K., Remediation of heavy metal(loid)s contaminated soils-- to mobilize or to immobilize? J. Hazard. Mater. 266, Boyd R., Ecology of metal hyperaccumulation. New Phytol. 162, Bradshaw A.D., McNeilly T.S. & Gregory R.P.G., Industrialisation, evolution and the development of heavy- metal tolerance in plants. In: British Ecology Society Symposium ed. Ecology and the Industrial Society. Blackwell Scientific Publications, Oxford. Brooks R.R., Baker A.J.M. & Malaisse F., Copper flowers. Res. Explor. 8(3), Bruneau J.-C. & Pain M., Atlas de Lubumbashi. Centre d études géographiques sur l'afrique noire ed., Université Paris X, Nanterre, 133. Bruno J.F., Stachowicz J.J. & Bertness M.D., Inclusion of facilitation into ecological theory. Trends Ecol. Evol. 18(3), Bullock J.M., Aronson J., Newton A.C., Pywell R.F. & Rey- Benayas J.M., Restoration of ecosystem services and biodiversity: Conflicts and opportunities. Trends Ecol. Evol. 26(10), Cailteux J.L.H., Kampunzu A.B., Lerouge C., Kaputo A.K. & Milesi J.P., Genesis of sediment-hosted stratiform copper cobalt deposits, central African Copperbelt. J. African Earth Sci. 42(1-5), Campbell B., The Miombo in transition: woodlands and welfare in Africa. Campbell, B. ed., Center for International Forestry Research (CIFOR), Bogor, Indonesia, 266. Champluvier D., New and overlooked Acanthaceae taxa from D.R.Congo, Rwanda and Burundi: (1) the genus Barleria. Plant Ecol. Evol. 144(1), Cherian S. & Margarida Oliveira M., Transgenic Plants in Phytoremediation : Recent Advances and New Possibilities. Environ. Sci. Technol. 39(24), Cheyns K., Banza Lubaba Nkulu C., Ngombe L.K., Asosa J.N., Haufroid V., De Putter T., Nawrot T., Kimpanga C.M., Numbi O.L., Ilunga B.K., Nemery B. & Smolders E., Pathways of human exposure to cobalt in Katanga, a mining area of the DR Congo. Sci. Total Environ. 490, Chipeng F.K., Hermans C., Colinet G., Faucon M.-P., Ngongo M., Meerts P. & Verbruggen N., Copper tolerance in the cuprophyte Haumaniastrum katangense (S. Moore) P.A. Duvign. & Plancke. Plant Soil 328(1-2), Clewell a, Aronson J. & Winterhalder K., The SER International primer on ecological restoration. Ecol. Restor. 2, Córdova S., Neaman A., González I., Ginocchio R. & Fine P., The effect of lime and compost amendments on the potential for the revegetation of metal-polluted, acidic soils. Geoderma 166(1), Cornelissen J., Lavorel S. & Garnier E., A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. Aust. J. Bot. 51(4), 335. Cunningham S.D., Berti W.R. & Huang J.W., Phytoremediation of contaminated soils. Trends Biotechnol. 13, Cunningham S.D. & Ow D.W., Promises and Prospects of Phytoremediation. Plant Physiol. 110(3), Daily G., Nature s Services: Societal Dependence On Natural Ecosystems, Island Press, 392. Delhaye G., Violle C., Séleck M., Ilunga wa Ilunga E., Daubie I., Mahy G. & Meerts P., Plant traits variation along a copper-cobalt contamination gradient: a community-level perspective exploring the relative importance of species turnover and within species variation. J. Veg. Sci. Dube A., Zbytniewski R., Kowalkowski T., Cukrowska E. & Buszewski B., Adsorption and Migration of Heavy Metals in Soil. Polish J. Environ. Stud. 10(1),

79 Duvigneaud P. & Denaeyer-De Smet S., Etudes sur la végétation du Katanga et de ses sols métallifères. Communication n 7 Cuivre et végétation au Katanga. Bull. Soc. Roy. Bot. Belgique, Travaux du Centre scientifique et médical de l Université libre de Bruxelles en Afrique centrale 96(2), Egan D., Howell E.A. & Meine C., The Historical Ecology Handbook: A Restorationist s Guide to Reference Ecosystems, Science and practice of ecological restoration, Island Press, 488. Ernst W.H.O., Bioavailability of heavy metals and decontamination of soils by plants. Appl. geochemistry 11, Ernst W.H.O., Evolution of metal hyperaccumulation and phytoremediation hype. New Phytol. 146(3), Ernst W.H.O., Evolution of metal tolerance in higher plants. For. Snow Landsc. Res. 80(3), Faucon M.-P., Chipeng F., Verbruggen N., Mahy G., Colinet G., Shutcha M., Pourret O. & Meerts P., Copper tolerance and accumulation in two cuprophytes of South Central Africa: Crepidorhopalon perennis and C. tenuis (Linderniaceae). Environ. Exp. Bot. 84, Faucon M.-P., Meersseman A., Shutcha M.N., Mahy G., Luhembwe M.N., Malaisse F. & Meerts P., Copper endemism in the Congolese flora: a database of copper affinity and conservational value of cuprophytes. Plant Ecol. Evol. 143(1), Faucon M.-P., Parmentier I., Colinet G., Mahy G., Ngongo Luhembwe M. & Meerts P., May Rare Metallophytes Benefit from Disturbed Soils Following Mining Activity? The Case of the Crepidorhopalon tenuis in Katanga (D. R. Congo). Restor. Ecol. 19(3), Faucon M.-P., Tshilong B.M., Van Rossum F., Meerts P., Decocq G. & Mahy G., Ecology and Hybridization Potential of Two Sympatric Metallophytes, the Narrow Endemic Crepidorhopalon perennis ( Linderniaceae ) and its More Widespread Congener. Biotropica 44(4), Furini A. (Ed.), Plants and Heavy Metals, SpringerBriefs in Molecular Science, Springer Netherlands, Dordrecht, 86. García-Palacios P., Soliveres S., Maestre F.T., Escudero A., Castillo-Monroy A.P. & Valladares F., Dominant plant species modulate responses to hydroseeding, irrigation and fertilization during the restoration of semiarid motorway slopes. Ecol. Eng. 36, Godefroid S., Van de Vyver A., Massengo Kalenga W., Handjila Minengo G., Rose C., Ngongo Luhembwe M., Vanderborght T. & Mahy G., Germination capacity and seed storage behaviour of threatened metallophytes from the Katanga copper belt (DR Congo): implications for ex situ conservation. Plant Ecol. Evol. 146(2), Hobbs R.J., Arico S., Aronson J., Baron J.S., Bridgewater P., Cramer V. a., Epstein P.R., Ewel J.J., Klink C. a., Lugo A.E., Norton D., Ojima D., Richardson D.M., Sanderson E.W., Valladares F., Vilà M., Zamora R. & Zobel M., Novel ecosystems: Theoretical and management aspects of the new ecological world order. Glob. Ecol. Biogeogr. 15, 1 7. Hobbs R.J., Higgs E. & Harris J. a, Novel ecosystems: implications for conservation and restoration. Trends Ecol. Evol. 24(11), ICMM, Trends in the mining and metals industry, 10 Principles, 16. ICMM, Biodiversity performance review: Executive summary, London, UK, 8. Ilunga wa Ilunga E., Mahy G., Piqueray J., Séleck M., Shutcha M.N., Meerts P. & Faucon M.-P., Plant functional traits as a promising tool for the ecological restoration of degraded tropical metal-rich habitats and revegetation of metal-rich bare soils: A case study in copper vegetation of Katanga, DRC. Ecol. Eng. 82, Ilunga wa Ilunga E., Séleck M., Colinet G., Meerts P. & Mahy G., Small-scale diversity of plant communities and distribution of species niches on a copper rock outcrop in Upper Katanga, DR Congo. Plant Ecol. Evol. 146(2), Jacobi C.M., Carmo F.F., Vincent R.C. & Stehmann J.R., Plant communities on ironstone outcrops: a diverse and endangered Brazilian ecosystem. Biodivers. Conserv. 16(7), Järup L., Hazards of heavy metal contamination. Br. Med. Bull. 68, Johnson M.S., Land Reclamation and the Botanical Significance of Some Former Mining and Manufacturing Sites in Britain. Environ. Conserv. 5(03), Kampunzu A.B., Cailteux J.L.H., Kamona A.F., Intiomale M.M. & Melcher F., Sediment-hosted Zn Pb Cu deposits in the Central African Copperbelt. Ore Geol. Rev. 35(3-4), Kawai T. & Tokeshi M., Testing the facilitationcompetition paradigm under the stress-gradient hypothesis: decoupling multiple stress factors. Proc. Biol. Sci. 274(1624), Kozlov M. V. & Zvereva E.L., Industrial barrens: extreme habitats created by non-ferrous metallurgy. Rev. Environ. Sci. Bio/Technology 6(1-3), Krämer U., Metal hyperaccumulation in plants. Annu. Rev. Plant Biol. 61, Lange B., Faucon M.-P., Meerts P., Shutcha M., Mahy G. & Pourret O., Prediction of the edaphic factors influence upon the copper and cobalt accumulation in two metallophytes using copper and cobalt speciation in soils. Plant Soil 379(1-2), Lasat M.M., Phytoextraction of metals from contaminated soil: A review of plant/soil/ metal interaction and assessment of pertinent agronomic issues. J. Hazard. Subst. Res. 2(5), Lavorel S., Díaz S., Cornelissen J.H.C., Garnier E., Harrison S.P., Mcintyre S., Pausas J.G., Catherine N.P. & Carlos R., Plant Functional Types : Are We Getting Any Closer to the Holy Grail? In: Terrestrial Ecosystems in a Changing World

80 Lavorel S., McIntyre S., Landsberg J. & Forbes T.D., Plant functional classifications: from general groups to specific groups based on response to disturbance. Trends Ecol. Evol. 12(12), Leteinturier B., Evaluation du potential phytocénotique des gisements cupriferes d Afrique centro-australe en vue de la phytoremédiation de sites pollués par l'activité. Fac. des Sci. Agron. Gembloux. Leteinturier B., Baker A.J.M. & Malaisse F., Early stages of natural revegetation of metalliferous mine workings in South Central Africa: a preliminary survey. Biotechnol. Agron. Soc. Environ. 3(1), Lone M.I., He Z., Stoffella P.J. & Yang X., Phytoremediation of heavy metal polluted soils and water: progresses and perspectives. J. Zhejiang Univ. Sci. B 9(3), Malaisse F., Brooks R.R. & North P., Colonisation of modified metalliferous environments in Zaire by the copper flower Haumaniastrum katangense. Plant Soil 64, Manda B., Colinet G. & André L., Evaluation de la contamination de la chaîne trophique par les éléments traces (Cu, Co, Zn, Pb, Cd, U, V et As) dans le bassin de la Lufira supérieure (Katanga/. Tropicultura Mbenza M., Aloni K. & Muteb M., Quelques considérations sur la pollution de l air à Lubumbashi (Shaba, Zaïre). Geo. Eco. Trop. 13(1-4), Mendez M.O. & Maier R.M., Phytostabilization of mine tailings in arid and semiarid environments--an emerging remediation technology. Environ. Health Perspect. 116(3), Mendoza-Hernández P.E., Orozco-Segovia A., Meave J. a., Valverde T. & Martínez-Ramos M., Vegetation recovery and plant facilitation in a human-disturbed lava field in a megacity: searching tools for ecosystem restoration. Plant Ecol. 214(1), Millenium Ecosystem Assessment, Ecosystems and Human Well-being: Biodiversity Synthesis, World Resources Institute, Washington, DC, 100. Mpundu M.M., Sikuzani Y.U., Kimuni L.N. & Colinet G., Effets d amendements carbonatés et organiques sur la culture de deux légumes sur sol contaminé à Lubumbashi ( RD Congo ). Biotechnol. Agron. Société Environ. 18(3), Mpundu M.M.M., Mobinzo P., Bindelle J., Vandenput S., Ngongo Luhembwe M. & Colinet G., Soil, plant and goat faeces can be used to map trace elements contaminations in the town of Lubumbashi (DRCongo). In: The 6th International Conference on Soils of Urban, Industrial, Traffic, Mining and Military Areas. Marrakech, 1. Mpundu M.M.M., Useni Sikuzani Y., Ntumba Ndaye F., Muyambo Musaya E., Kapalanga Kamina P., Mwansa M., Ilunga K. & Nyembo Kimuni L., Évaluation des teneurs en éléments traces métalliques dans les légumes feuilles vendus dans les différents marchés de la zone minière de Lubumbashi. J. Appl. Biosci. 66, Narendrula R., Nkongolo K. & Beckett P., Comparative Soil Metal Analyses in Sudbury (Ontario, Canada) and Lubumbashi (Katanga, DR-Congo). Bull. Environ. Contam. Toxicol. 88(2), Neuman D. & Ford K.L., Phytostabilization as a remediation alternative at mining sites, Technical note 420, 48. O Dell R. & Classen V., Restoration and Revegetation of Harsh Soils. In: Serpentine : The Evolution and Ecology of a Model System O Dell R.E. & Claassen V.P., Relative performance of native and exotic grass species in response to amendment of drastically disturbed serpentine substrates. J. Appl. Ecol. 43(5), Parraga-Aguado I., Querejeta J.-I., González-Alcaraz M.-N., Jiménez-Cárceles F.J. & Conesa H.M., Usefulness of pioneer vegetation for the phytomanagement of metal(loid)s enriched tailings: grasses vs. shrubs vs. trees. J. Environ. Manage. 133, Peer W.A., Baxter I.R., Richards E.L., Freeman J.L. & Murphy S., Phytoremediation and hyperaccumulator plants. Mol. Biol. Met. Homeost. Detoxif. 14(August), Pilon-Smits E.A.H. & Pilon M., Phytoremediation of Metals Using Transgenic Plants. CRC. Crit. Rev. Plant Sci. 21(5), Pollard A.J., Powell K.D., Harper F.A. & Smith J.A.C., The Genetic Basis of Metal Hyperaccumulation in Plants. CRC. Crit. Rev. Plant Sci. 21(6), Prasad M.N. V & De Oliveira Freitas H.M., Metal hyperaccumulation in plants - Biodiversity prospecting forphytoremediation technology. Electron. J. Biotechnol. 6(3), Prasad M.S., Production of copper and cobalt at Gecamines, Zaire. Miner. Eng. 2(4), Primack R.B., Essentials of Conservation Biology, Sinauer Associates, Incorporated, 601. Rey Benayas J.M., Newton A.C., Diaz A. & Bullock J.M., Enhancement of Biodiversity and Ecosystem Services by Ecological Restoration: A Meta-Analysis. Science (80-. ). 325(5944), Saad L., Parmentier I., Colinet G., Malaisse F., Faucon M.- P., Meerts P. & Mahy G., Investigating the Vegetation-Soil Relationships on the Copper-Cobalt Rock Outcrops of Katanga (D. R. Congo), an Essential Step in a Biodiversity Conservation Plan. Restor. Ecol. 20(3), Secretariat of the Convention on Biological Diversity, Global Biodiversity Outlook 4, Montréal, 155. Séleck M., Bizoux J.-P., Colinet G., Faucon M.-P., Guillaume A., Meerts P., Piqueray J. & Mahy G., Chemical soil factors influencing plant assemblages along coppercobalt gradients: implications for conservation and restoration. Plant Soil 373(1/2), Séleck M., Lebrun J., Boisson S., Le Stradic S., Masengo Kalenga W. & Mahy G., Guide to Cu-Co species of concern regeneration - V2.2, Gembloux,

81 Shaw J., Heavy Metal Tolerance in Plants: Evolutionary Aspects. Shaw, J. ed., CRC Press Inc, New York, 268. Shutcha M., Meerts P., Kazadi S., Faucon M.-P. & Visser M., Low Caryopsis Production of the Toothbrush Grass (Microchloa altera) from Katanga (DR Congo) Could Limit the Revegetation of Trace Metal Contaminated Lands by. Ecol. Restor. 31(3), Shutcha M.N., Mubemba M.M., Faucon M.-P., Luhembwe M.N., Visser M., Colinet G. & Meerts P., Phytostabilisation of copper-contaminated soil in Katanga: an experiment with three native grasses and two amendments. Int. J. Phytoremediation 12(6), Singh S.N. & Tripathi R.D., Environmental bioremediation technologies, Environmental Bioremediation Technologies, Springer-Verlag Berlin Heidelberg, Soulé M., What is conservation biology. Bioscience 35(11), Stohs S.J. & Bagchi D., Oxidative mechanisms in the toxicity of metal ions. Free Radic. Biol. Med. 18(2), Stone G.D., The Anthropology of Genetically Modified Crops. Annu. Rev. Anthropol. 39(1), Tilman D., Community invasibility, recruitment limitation, and grassland biodiversity. Ecology 78(1), Tirado R. & I. Pugnaire F., Community structure and positive interactions in constraining environments. Oikos 111(3), UMHK, Union Minière du Haut Katanga , Bruxelles, 283. UNEP, Global Strategy for Plant Conservation. Conv. Biol. Divers. (25/08/2015). United Nations, Convention on biological diversity. Diversity 30. USGS, Mineral Commodity Summaries 2015, U.S. Geological Survey, 196. Van Andel J. & Aronson J., Restoration ecology: the new frontier, Restoration Ecology, Blackwell Publishing Ltd, 319. van Breemen N. & Finzi A., Plant-soil Interactions: Ecological Aspects and Evolutionary Implications. Biogeochemistry 42(1-2), Vangronsveld J., Van Assche F. & Clijsters H., Reclamation of a bare industrial area contaminated by non-ferrous metals: in situ metal immobilization and revegetation. Environ. Pollut. 87(1), Verbruggen N., Hermans C. & Schat H., Molecular mechanisms of metal hyperaccumulation in plants. New Phytol. 181(4), Verdú M. & Valiente-Banuet A., The nested assembly of plant facilitation networks prevents species extinctions. Am. Nat. 172(6), Vié J.-C., Hilton-taylor C. & Stuart S.N., Wildlife in a changing world, IUCN Gland Switzerland, 180. Violle C., Navas M.-L., Vile D., Kazakou E., Fortunel C., Hummel I. & Garnier E., Let the concept of trait be functional! Oikos 116(5), Vitousek P.M., Human Domination of Earth s Ecosystems. Science (80-. ). 277(5325), Vranken I., Amisi Y.M., Kankumbi F.M., Bamba I., Veroustraete F., Visser M. & Bogaert J., The Spatial Footprint of the Non-ferrous Mining Industry in Lubumbashi. Tropicultura 31(1), Whiting S.N., Reeves R.D. & Baker A.J.M., Conserving biodiversity: Mining, metallophytes and land reclamation. Min. Environ. Manag. 10(2), Whiting S.N., Reeves R.D., Richards D., Johnson M.S., Cooke J.A., Malaisse F., Johns R., Mcintyre T., Purvis O.W., Salt D.E., Schat H., Zhao F.J. & Baker A.J.M., Research Priorities for Conservation of Metallophyte Biodiversity and their Potential for Restoration and Site Remediation. Restor. Ecol. 12(1), Wolf A., Conservation of endemic plants in serpentine landscapes. Biol. Conserv. 100(1), Wong M.H., Ecological restoration of mine degraded soils, with emphasis on metal contaminated soils. Chemosphere 50(6), Young T.P., Restoration ecology and conservation biology. Biol. Conserv. 92,

82

83 Axe 1 Niches écologiques et endémisme édaphique

84

85 Axe 1 - Article 2 En preparation - Ecological niche distribution of metallophytes from southeastern D.R.Congo along metal toxicity gradients BOISSON Sylvain, MONTY Arnaud, SÉLECK Maxime, NGOY SHUTCHA Mylor, FAUCON Michel- Pierre & MAHY Grégory Abstract Ecological niche modelling helps to understand the spatial assembly of species in heterogeneous environment. Along macronutrient gradients, three patterns have been highlighted: (1) species optima are unevenly distributed, with a higher frequency in mesic conditions; (2) species response curves are narrower where optima density is higher, and (3) species with optima at the extremes of the gradient have skewed response curves with a longer tail toward mesic conditions. This study aims at testing these patterns on threatened vegetation occurring in metalliferous soils comprising copper and cobalt in southeastern D.R. Congo. Realized niches of 80 taxa were modeled using generalized additive models and the niche optima and the niche width were determined for each taxon. Copper and cobalt gradients ranged from 29 mg kg -1 to mg kg -1 and cobalt from 2 mg kg -1 to 927 mg kg -1. Results showed that three groups have been highlighted according to the niche optima location along the gradients. Two groups corresponded to taxa having niche optima equal to the lower or the upper limit. One group contained taxa having well-identified niche optima. The patterns found along macronutrient gradients were transposable along a micronutrient gradient. This diversity of realized niche had consequence on the plant conservation at communities and species scale. 85

86 Introduction Vegetation ecology has long been concerned with describing the distribution of species along edaphic gradients (Austin, 1985, 1987, 2007). Ecological niche theory and modelling help to understand the spatial assembly of species in heterogeneous environment (Chase et al., 2003; Higgins et al., 2012). Since Hutchinson s definition (1957), ecological niches of species along one or several environmental factors is addressed with statistical and mathematical tools (Yee et al., 1991; Guisan et al., 2002; Austin, 2007). The fundamental niche is related to the physiological tolerances of species, whereas the realized niche is defined as the range of conditions and resources that permit to an organism to survive and to reproduce given the presence of biotic interactions (Hutchinson, 1957). Species response curves along one environmental factor, obtained from field occurrence or abundance data, is the classical representation of the distribution of realized niches (Austin et al., 1996; Rydgren et al., 2003; Coudun et al., 2006; Austin, 2007). Along edaphic factors, species response curves are typically characterized by three patterns : (1) species optima are unevenly distributed, with a higher frequency in mesic conditions (Minchin, 1989; Lawesson et al., 2002); (2) species response curves are narrower when optima density is higher (Minchin, 1989; Lawesson et al., 2002). ; and (3) species with optima at the extremes of the gradient present skewed response curves with a longer tail toward mesic conditions (Austin et al., 1994; Chase, 2011). These patterns were observed for distinct vegetation types along macronutrient gradients such as calcium (Hájková et al., 2008), magnesium (Wright et al., 2006), phosphorus (Austin et al., 1996) or carbon-nitrogen ratio (Heikkinen et al., 2010). However, the existence of such distribution patterns along micronutrient gradients has not been addressed so far, despite soils of some areas in the world present unbalanced micronutrients concentrations that constraint vegetation type and species distributions (Duvigneaud et al., 1963; Kruckeberg, 1992; Whiting et al., 2002; Bizoux et al., 2004; Brady et al., 2005; Baker et al., 2010). As micronutrients (including beneficial elements) are needed in very small quantities by plants, range in their available concentrations in soil may differ from macronutrient gradients. First, contrarily to macronutrients, the deficiency in micronutrients is not common and generally observed on a small range of concentrations (Pilon-Smits et al., 2009). Second, many micronutrients are toxic at a lower threshold than macronutrients by definition (Ernst, 86

87 2006; Pilon-Smits et al., 2009), which can profoundly affect the distribution of species. At high toxicity levels, like in metalliferous soils, the capacity to tolerate or accumulate micronutrients can be determinant for the occurrence of a particular species. Such capacity could have originated from a strategy to minimize biotic interactions with other plants species, herbivores and pathogens (Antonovics et al., 1971; Macnair, 1987; Pollard, 2000; Hall, 2002; Boyd, 2004; Ernst, 2006). Some metallophyte communities constitute interesting examples of vegetation having species or populations occurring along large gradients of micronutrients being trace metals (Duvigneaud et al., 1963a; Baker et al., 2010; Escarré et al., 2013). Notably, the metallicolous plant communities of the southeastern D.R. Congo represent remarkable grassland of which the communities assemblage are dependent on copper and cobalt concentrations in soils (Duvigneaud et al., 1963a; Saad et al., 2012; Séleck, Bizoux, et al., 2013). In theory, such vegetation can be considered as an opportunity to test the three presented patterns. In practice, the characterization of the realized niches of taxa along the copper and the cobalt gradients appears to be particularly valuable in a context of plant communities and taxa conservation. The present study aims at describing the realized niche distribution of plant species occurring in the threatened vegetation growing along strong copper and cobalt gradients in the southeastern D.R. Congo, in the Katangan Copperbelt. Specifically, we tested the existence of the three patterns along the two micronutrient gradients being copper and cobalt, i.e., (1) the uneven distribution of optima along the gradients; (2) the existence of a negative correlation between the niche optima density and the niche widths; and (3) the existence of a negative correlation between the skewness coefficient and metal concentrations. Finally, we discussed how the pattern of the characterized distribution affects the conservation strategies. Method Study area and vegetation This study was performed in the Katangan Copperbelt (D.R. Congo). The annual average precipitations about 1174 mm are concentrated in wet season (November March). Average daily temperatures during wet season are 20.2 C and temperatures ranges from C, at the beginning of dry season (Kazadi et al., 1996). More than one hundred copper-cobalt 87

88 outcrops are scattered in an open deciduous forest over more than 300 km, forming copper hills with metalliferous soils (Duvigneaud et al., 1963). On each copper hill, ammonium acetate-edta extractable copper and cobalt concentrations (i.e., available concentrations) are dependent on the topographical gradient with concentrations of 29 mg kg -1 Cu and 2 mg kg -1 Co at the bottom and can reach 10,000 mg kg -1 Cu in soil and 1,000 mg kg -1 Co at the top (Figure 1a, Séleck et al., 2013). Small spatial variations are also identified within sites, influencing the plant communities and species distribution (Ilunga wa Ilunga et al., 2013). More than 550 plant species occur on the copper hills (Duvigneaud et al., 1963a; Leteinturier & Malaisse, 1999). Among vegetation, two plant communities have been identified according to the copper and the cobalt concentrations: the steppic savannah, on the bases and the slopes of the hills (less than 3,500 mg kg -1 of available Cu in soil and 50 mg kg -1 of available Co), and, the steppe, on soils highly enriched in Cu and Co (from 3,500 mg kg -1 to 10,000 mg kg -1 of available Cu in soils and from 50 mg kg -1 to 1,000 mg kg -1 of available Co in soils) and located on the top of a Cu hill (Séleck et al., 2013). Figure 1. Predicted copper and cobalt concentrations on one copper hill, Fungurume V (a) and the location of the performed sampling plots (b). 88

89 Vegetation sampling and soil analysis During March 2008 and 2009, 172 square meter plots (1m 2 ) were established in a diversity of topographical points on three natural sites in the area between Tenke and Fungurume towns: Fungurume V, Kavifwafwaulu [1] and Kazinyanga (Table 1). For each site, sampling was performed according to a systematic grid following six to eight parallel transects stretched from the bottom to the top of the hill. Each plot was located at node between both transect and level curb all the 5 m altitude (Figure 1b). For each plot, all taxa were identified and one composite sample was created from four soil subsamples collected at the four plot corners at a depth of 15 cm. For unidentified species, herbarium voucher sample were collected and deposited at Gembloux Agro-Bio Tech, University of Liège, Belgium. Identification of specimen was based on the African Plant Database (Conservatoire et Jardin Botanique de la ville de Genève and South African National Biodiversity Institute Pretoria 2013). Due to the lack of recent taxonomic revision, some individuals could not be identified at the species level. They were individualized at putative species level. Each composite sample soil was air-dried and sieved to 2 mm prior to physic-chemical analyses. Soil extractable concentrations of copper (Cu) and cobalt (Co) were measured using 0.5N ammonium acetate-edta (CH 3 COONH 4 EDTA, (Lakanen et al., 1971). The extraction soil/solution ratio was fixed to 1:5 with a buffered at 4.65 (Kucak et al., 1998; Faucon et al., 2009; Saad et al., 2012). Finally, flame atomic absorption spectroscopy (VARIAN 220) of the filtrate (used filter: S&S-595-½) measured metal concentrations after stirring during 30 minutes. Table 1. Copper hill ID, geographical coordinates (utm), sampled area (ha), number of sampling plots and minimum and maximum available copper and cobalt concentrations (mg kg -1 ) of the three propected copper hills (Fungurume V, Kavifwafwaulu and Kazinyanga). Copper hill ID Geographical coordinates (utm) Sampled area (ha) Number of plots [Cu] (mg kg -1 ) [Co] (mg kg -1 ) X Y Min Max Min Max Fungurume V Kavifwafwaulu [1] Kazinyanga

90 Species response curve modelling Taxa present in a minimum of 7 plots were chosen to model their species response curves. Many statistical methods exist to model species response curves and provide models for one or more environmental variables (Austin et al., 1996; Guisan et al., 2000, 2002; Oksanen et al., 2002; Austin, 2007; Meynard et al., 2007). We selected the generalized additive models GAM (GAM, Hastie et al., 1990; Yee et al., 1991) with binomial likelihoods to model the probability of occurrence of each selected species along the two gradients: available copper and cobalt concentrations. GAMs do not require any assumption regarding the shape of the response curves to environmental factors (Oksanen et al., 2002). For each taxon, the best fitted model among 3 degrees of smoothing (3, 4 or 5) was selected with the Aikaike Information Criterion (AIC). Niche optimum of each species along the two gradients was estimated by the projected value of the highest probability of occurrence on the x-axis. The borders of niche width (amplitude) were determined by the projection of lower and upper limits (LL and UL) of 80% area under curve centered on the optimum on the x-axis (Gégout et al., 1998). Niche width was calculated by the difference between UL and LL. Statistical modelling and values extraction were performed via a function GAM.summary (Appendix C) operating with mgcv library (Wood, 2006) in the R statistical software to model response curves with generalized additive models (GAM). This function is able to generate the response curves with the parameters of the model and append the curve of each factor included in the dataset. The significance of the relation between copper and cobalt optima location was tested using a Pearson s correlation. To test for an uneven distribution of niche optima along the copper-cobalt gradients, we used a kernel density function. Density function uses the Fourier transformation and applies a linear approximation to model the curve. To test relationship between the optima density and the niche width a Pearson s correlation between niche width and optima density were performed. To test the relationship between the skewness and the optima location, the skewness of species response curves was determined using Fisher s coefficient (Fisher, 1930): 90

91 , where is the third central moment and is the standard deviation. Each response curve can be assimilate to probability density function with an area under curve (A) higher than 1. Mathematical expectations were thus defined as ( ) ( ) ( ), where m is the number of rectangles, is and ( ) is the probability of occurrence of species and the sum of rectangles areas. For the calculation: m=1000 and = maximum value of measured copper or cobalt. Mathematical expectations had permit to estimate V(x), the variance (2 th central moment) of each response curve and the n th central moment. Finally, linear regression was used to determinate the relationship between skewness coefficients and optima values along gradients using R. Results Data ranges and metals gradients Copper concentrations ranged from 29 mg kg -1 (Kavifwafwaulu [1]) to 10,136 mg kg -1 (Kazinyanga) and cobalt from 2 mg kg -1 (Kavifwafwaulu [1] and Kazinyanga) to 927 mg kg -1 (Fungurume V) (Table 1). Spearman correlation coefficient between copper and cobalt concentrations was 0.61 (p value < 0.05). Species distribution overview On a total of 184 taxa in dataset, 80 occurred in at least 7 plots (Appendix A). Along copper and cobalt gradients, three groups of taxa, corresponding to 3 types of distribution, presented the same structure based on the niche optima and the response curves. Two groups had taxa for which response curves along gradients were highly truncated with tails toward the highest concentrations (Group 1) and inversely, with tails toward the lowest concentrations (Group 3). Group 2 contained taxa with well-determined niche optima (i.e., niche optimum unequal to LL or UL) along gradients. Along copper gradient, 40 taxa presented niche optima at the lowest value (29 mg kg -1 Cu, Group 1, Figure 2). 31 taxa had optima ranging from 453 to 7,467 mg kg -1 Cu (Group 2, Figure 91

92 2). Group 3 included 9 taxa with optima at the lowest value of the gradient (10,136 mg kg -1 Cu, Group 3, Figure 2, Appendix B). Along cobalt gradient, 45 taxa included their niche optima at the lowest value (i.e., 2 mg kg -1 Co, Group 1, Figure 3, Appendix C). 27 taxa had been identified in the Group 2, having their niche optima ranged from 26 to 664 mg kg -1 Co. 7 taxa had niche optima at the highest value of the gradient (927 mg kg -1 Co, Group 3, Figure 3). The correlation coefficient between the copper and the cobalt optima values was 0.38 (p value < 0.001). A total of 61% of taxa (49 taxa) was in the same copper and cobalt groups, matching at 75 % between Groups 1, 55% between Groups 2, 22% between Groups 3. 92

93 Figure 2. Modelled realized niches along available copper gradient (29 to 10;136 mg kg -1 ) by generalized additive models (GAM) for 80 metallophyte taxa of the Katanga Copperbelt sorted by niche optima. Plain dots, squares and diamonds are the niche optima of taxa determined by the gradient value of the highest probability of occurrence for respectively Group 1, 2 and 3. Straight lines are the niche widths calculated by the difference between the borders of area of 80 % under GAM-curve. Asterisk (*) and double asterisks (**) mean that the taxon is respectively strict or broad endemic (Faucon et al. 2010). 93

94 Figure 3. Modelled realized niches along available cobalt gradient (2 to 927 mg kg -1 ) by generalized additive models (GAM) for 80 metallophyte taxa of the Katanga Copperbelt sorted by niche optima. Empty dots, squares and diamonds are the niche optima of taxa determined by the gradient value of the highest probability of occurrence for respectively Group 1, 2 and 3. Straight lines are the niche widths calculated by the difference between the borders of area of 80 % under GAM-curve. Asterisk (*) and double asterisks (**) mean that the taxon is respectively strict or broad endemic (Faucon et al. 2010). 94

95 Pattern 1: Uneven distribution of optima along the gradients The density of optima was the highest in low metal concentrations (43 taxa between 0-1,000 mg kg -1 Cu and 49 taxa between mg kg -1 Co) and decreased with increasing metals concentrations (Figure 4). Along cobalt gradient only, density of niche optima approached zero (lower than ) at concentrations from 530 to 780 mg kg -1 (Figure 4b). Figure 4. Density of taxa optima along copper and cobalt gradient (mg kg -1 ) calculated by kernel density estimation. Pattern 2: Negative correlation between niche optima density and niche width The niche widths of taxa were significantly correlated with the density of niche optima along the copper and the cobalt gradients. Along the copper gradient, the niche widths of taxa decreased with increasing density of niche optima for all taxa (r = -0.75; p value < 0.001) and also for Group 2 (r = -0.68; p value < 0.001). Along cobalt gradient, the same relationship was observed for all taxa (r = -0.61; p value < 0.001) and also for Group Co-2 (r = -0.69; p value < 0.001). The niche widths of taxa were the narrowest in the lowest concentrations whatever the metal gradients. Pattern 3: Negative correlation between the skewness coefficient and the niche optima. Skewness coefficient of niches was significantly negatively correlated with copper optima locations for all taxa (R 2 = 0.66; p value < 0.001) and for the Group 2 (R 2 = 0.52; p value < 0.001). Realized niches were more symmetric when optima ranged between and mg kg -1 (Figure 5a). Along cobalt gradient, skewness coefficients had the same relationship, with the niche widths for all taxa (R 2 = 0.48; p value < 0.001) and for the Group 2 only (R 2 = 0.29; p value < 0.001). Symmetry of species responses curves was 95

96 encountered at 150 mg kg -1 for Group 2 and at 500 mg kg -1 for all taxa (Figure 5b). Along these two metal gradients, taxa whose niche optimum located at the lowest metal concentrations had response curves with a tail towards highest concentrations, contrasting with the taxa having niche optima in the highest metal concentrations. Figure 5. Skewness coefficients relationship with the copper and cobalt optima positions for the 80 identified taxa from the three groups. Plain or empty dots, squares and diamonds are the niche optima for respectively Group 1, 2 and 3 along copper/cobalt gradient. (a) Linear regression of skewness coefficients along copper niche optima. (b) Linear regression of the skewness coefficients along cobalt niche optima. Discussion General patterns of the realized niches distribution along metal gradients Our results demonstrate that the patterns of plant niche distributions identified along macronutrient gradients were also found along a toxicity gradient copper and cobalt gradient. The most taxa were found in the metal concentrations lower than mg kg -1 Cu and 100 mg kg -1 Co and had the narrowest niche width. Close to the limits of the gradients, the response curves were skewed with a longer tail into the opposite part of the metal gradients. Even if the correlation between copper and cobalt optima location was significant, the coefficient was low, indicating that a taxon occurred not necessarily in the same range of the relative scale of concentrations along the gradients. For example, Crotalaria cobalticola P.A.Duvign. & Plancke (Fabaceae) was located in Group 3 for copper, signifying that the species optimum occurs in the highest concentrations and in Group 1 for cobalt, containing the species optimum occurring in the lowest concentrations. 96

97 Origins of patterns and consequences The highest optima density found in the lowest metal concentrations suggest that competition should a determinant factor leading to narrow niche width by competitive exclusion principle (Hardin, 1960; Grime, 1973). Competition has been recently identified as higher in the steppic savannah than in the steppe (Delhaye et al., In press). This variation of the niche width has already been highlighted along the copper gradient for taxa occurring on a copper hill near Lubumbashi (Ilunga wa Ilunga et al., 2013). However, some taxa like the grass Loudetia simplex (Nees) C.E.Hubb. having the niche optima equal to the lower limit, had a broad niche width, that confirms the presence of taxa in both plant communities (Boisson et al., In press). This suggests that the species present a wide range of tolerance to metals compared to taxa with narrow realized niches in the same group, for example, Aeschynomene pygmaea J.Léonard. A part of taxa occurring in the Group 1 along both copper and cobalt could be considered as colonizers presenting higher tolerance threshold than other species from non metalliferous soils. The broad realized niches of some taxa could results from the high between-population variations in realized niches, considering that metals exert a strong selective pressure on taxa (Antonovics et al., 1971; Shaw, 1990; Whiting et al., 2002; Baker et al., 2010) and the metal concentrations varied according to sites. Adaptation and differentiation are also promoted by extreme dispersal limit, that occurs on the plant taxa of the copper hills (Ackerly, 2003). The taxa present difference in realized niches between individuals. Individuals could be specialized to distinct ranges of concentrations along the gradient (Bolnick et al., 2007). Determination of the fundamental niche of distinct individual from populations among a taxon could help to understand the origin of the broad realized niches. A recent study on one tolerant species from metalliferous soils suggested that niches may evolve relatively quickly in serpentines soils between ecotypes Collinisia sparsiflora (Wright et al., 2006). The skewed response curves are more difficult to interpret in ecological theory even if it was often observed in studies about ecological niches (Bio, 1998; Rydgren et al., 2003). Skewed curve of realized niche could be from a progressive decrease in physiological advantages for an extreme-adapted species until it is outperformed by the more productive species (Austin et al., 1989, 1994). The relationship between the skewness and the optima location was not identical for both metals, that was expected because copper and cobalt have not the same 97

98 physiological role in plants. Copper is known as an essential redox-active micronutrient while cobalt is identified as a non-essential but beneficial element (Yruela, 2005; Pilon-Smits et al., 2009). Cobalt should be thus toxic at lower concentrations than copper. In Katanga (D.R. Congo), mesic conditions are found between 0-65 mg kg -1 and 0-6 mg kg -1 of available copper and cobalt (Anderson et al., 1973; McLean et al., 1992; Ma et al., 2010; Faucon et al., 2011; Colinet, 2013). The results match with a previous study showing the indicator species of two plant communities in the Katangan Copperbelt (Saad et al., 2012; Séleck, Bizoux, et al., 2013). For each community, three indicator species has been proposed. Cryptosepalum maraviense (Caesalpiniaceae), Loudetia simplex (Poaceae), Scleria bulbifera (Cyperaceae) characterize steppic savannahs. In this study, these species occur in Group 1 along both copper and cobalt gradient (excepted C. maraviense). The taxa Xerophyta equisetoides (Velloziaceae), Anisopappus davyi (Asteraceae), Ascolepis metallorum (Cyperaceae) distinguish the steppes and are present in concentrations over than 5000 mg kg -1 Cu and 200 mg kg -1 Co (Group 2). Niche concept is complex and species responses curves along available metals depend on other edaphic factors (Whittaker et al., 1973; Austin, 2007; Soberón et al., 2009). Soil factors which have an effect on metal bioavailability include the total metal present in the soil, ph, clay and hydrous oxide content, organic matter and redox conditions (Reichman, 2002). Also, acetate-edta method extracts metals at a constant ph (4.65) but in the rhizosphere, plants have a significant effect on the local bioavailability of nutrient by changing local ph (Marschner et al., 1986). Implication for conservation and restoration The diversity of realized niche of taxa along large copper and cobalt gradients has consequence on the conservation strategies. The recreation of these extreme conditions is thus required in order to restore the plant communities along the metal gradients. At species scale, the species having realized niche on highest metal concentrations should be considered firstly because of their potential dependence to metal in soils. The study of their fundamental niches, as suggested for the determination of population variations, could help to understand the relationship between these taxa and the metal. The species with the niche optima in the highest metal concentration and a broad niche width represent an opportunity 98

99 to test them in rehabilitation strategies like Crepidorhopalon perennis, a rare highly-tolerant taxa to Cu in the Katangan Copperbelt (Faucon et al., 2011). Acknowledgments Both travels realized for the present work was made possible thanks to the financial intervention of the Fonds de la Recherche dans l Industrie et dans l Agriculture (FRIA) of the FNRS (Fonds National de la Recherche Scientifique), Belgium. Tenke Fungurume Mining S.a.r.l. provided all logistic support necessary to perform this study. References Austin M., Continuum concept, ordination methods, and niche theory. Annu. Rev. Ecol. Syst. 16(1985), Austin M., Models for the analysis of species response to environmental gradients. Vegetatio 69, Austin M., Species distribution models and ecological theory: A critical assessment and some possible new approaches. Ecol. Modell. 200(1-2), Higgins S.I., O Hara R.B. & Römermann C., A niche for biology in species distribution models. J. Biogeogr. 39(12), Chase J.M. & Leibold M.A., Ecological niches: linking classical and contemporary approaches, Biodiversity & Conservation, Interspecific interactions. University of Chicago Press. Hutchinson G., Concluding remarks. Cold Spring Harb. Symp. Quant. Biol. 22(2), Yee T.W. & Mitchell N.D., Generalized additive models in plant ecology. J. Veg. Sci. 2, Guisan A., Edwards T.C. & Hastie T., Generalized linear and generalized additive models in studies of species distributions: setting the scene. Ecol. Modell. 157(2-3), Rydgren K., Økland R.H. & Økland T., Species response curves along environmental gradients. A case study from SE Norwegian swamp forests. J. Veg. Sci. 14(6), Coudun C. & Gégout J.-C., The derivation of species response curves with Gaussian logistic regression is sensitive to sampling intensity and curve characteristics. Ecol. Modell. 199(2), Austin M.P. & Meyers J. a., Current approaches to modelling the environmental niche of eucalypts: implication for management of forest biodiversity. For. Ecol. Manage. 85(1-3), Lawesson J. & Oksanen J., Niche characteristics of Danish woody species as derived from coenoclines. J. Veg. Sci. (Gaston 1994), Minchin P.R., Montane vegetation of the Mt. Field massif, Tasmania: a test of some hypotheses about properties of community patterns. Vegetatio 83(1-2), Austin M.P. & Gaywood M.J., Current problems of environmental gradients and species response curves in relation to continuum theory. J. Veg. Sci. 5(4), Chase J., Ecological niche theory. In: Scheiner, S.M., Willig, M.R. eds. The Theory of Ecology. University of Chicago Press, Chicago, 404. Hájková P., Hájek M., Apostolova I., Zelený D. & Dítě D., Shifts in the ecological behaviour of plant species between two distant regions: evidence from the base richness gradient in mires. J. Biogeogr. 35, Wright J., Davies K., Lau J., McCall A. & McKay J., Experimental verification of ecological niche modeling in a heterogeneous environment. Ecology 87(10), Heikkinen J. & Mäkipää R., Testing hypotheses on shape and distribution of ecological response curves. Ecol. Modell. 221(3), Brady K.U., Kruckeberg A.R. & Bradshaw Jr. H.D., Evolutionary Ecology of Plant Adaptation To Serpentine Soils. Annu. Rev. Ecol. Evol. Syst. 36(1), Bizoux J., Brevers F. & Meerts P., Ecology and conservation of Belgian populations of Viola calaminaria, a metallophyte with a restricted geographic distribution. Belgian J. Bot. 137(1), Baker A.J.M., Ernst W.H.O., Van der Ent A., Malaisse F. & Ginocchio R., Metallophytes: the unique biological resource, its ecology and conservational status in Europe, central Africa and Latin America. In: Batty, L., Hallberg, K. eds. Ecology of Industrial Pollution Whiting S.N., Reeves R.D. & Baker A.J.M., Conserving biodiversity: Mining, metallophytes and land reclamation. Min. Environ. Manag. 10(2), Kruckeberg A.R., Plant life of western North American ultramafics. In: Roberts, B.A., Proctor, J. eds. 99

100 The Ecology of Areas with Serpentinized Rocks: A World View. Kluwer Academic Publishers, Dordrecht, Pilon-Smits E.A.H., Quinn C.F., Tapken W., Malagoli M. & Schiavon M., Physiological functions of beneficial elements. Curr. Opin. Plant Biol. 12(3), Ernst W.H.O., Evolution of metal tolerance in higher plants. For. Snow Landsc. Res. 80(3), Macnair M., Heavy metal tolerance in plants: a model evolutionary system. Trends Ecol. Evol. 2(12), Antonovics J., Bradshaw A.D. & Turner R.G., Heavy metal tolerance in plants. Adv. Ecol. Res. 7, Pollard A., Metal hyperaccumulation: a model system for coevolutionary studies. New Phytol. 146, Hall J.L., Cellular mechanisms for heavy metal detoxification and tolerance. J. Exp. Bot. 53(366), Boyd R., Ecology of metal hyperaccumulation. New Phytol. 162, Escarré J., Lefèbvre C., Frérot H., Mahieu S. & Noret N., Metal concentration and metal mass of metallicolous, non metallicolous and serpentine Noccaea caerulescens populations, cultivated in different growth media. Plant Soil 370(1-2), Duvigneaud P. & Denaeyer-De Smet S., Etudes sur la végétation du Katanga et de ses sols métallifères. Communication n 7 Cuivre et végétation au Katanga. Bull. Soc. Roy. Bot. Belgique 96(2), Kazadi S. & Kaoru F., Interannual and long-term climate variability over the Zaire River Basin during the last 30 years. J. Geophys. Res. 101(21), Ilunga wa Ilunga E., Séleck M., Colinet G., Meerts P. & Mahy G., Small-scale diversity of plant communities and distribution of species niches on a copper rock outcrop in Upper Katanga, DR Congo. Plant Ecol. Evol. 146(2), Leteinturier B. & Malaisse F., The copper flora of Katanga: a phytogeographical analysis. Geo. Eco. Trop. 23, Séleck M., Bizoux J.-P., Colinet G., Faucon M.-P., Guillaume A., Meerts P., Piqueray J. & Mahy G., Chemical soil factors influencing plant assemblages along coppercobalt gradients: implications for conservation and restoration. Plant Soil 373(1/2), Lakanen E. & Erviö R., A comparison of eight extractants for the determination of plant available micronutrients in soil. Acta Agral. Fenn. 123, Saad L., Parmentier I., Colinet G., Malaisse F., Faucon M.- P., Meerts P. & Mahy G., Investigating the Vegetation-Soil Relationships on the Copper-Cobalt Rock Outcrops of Katanga (D. R. Congo), an Essential Step in a Biodiversity Conservation Plan. Restor. Ecol. 20(3), Faucon M.-P., Colinet G., Mahy G., Ngongo Luhembwe M., Verbruggen N. & Meerts P., Soil influence on Cu and Co uptake and plant size in the cuprophytes Crepidorhopalon perennis and C. tenuis (Scrophulariaceae) in SC Africa. Plant Soil 317(1-2), Kucak A. & Blanuša M., Comparison of two extraction procedures for determination of trace metals in soil by atomic absorption spectrometry. Arh. Hig. Rada Toksikol. 49(October), Oksanen J. & Minchin P.R., Continuum theory revisited: what shape are species responses along ecological gradients? Ecol. Modell. 157(2-3), Meynard C.N. & Quinn J.F., Predicting species distributions: a critical comparison of the most common statistical models using artificial species. J. Biogeogr. 34(8), Guisan A. & Zimmermann N.E., Predictive habitat distribution models in ecology. Ecol. Modell. 135(2-3), Hastie T.J. & Tibshirani R.J., Generalized Additive Models, Chapman and Hall/CRC Monographs on Statistics and Applied Probability Series. Chapman & Hall. Gégout J.-C. & Pierrat J.-C., L autécologie des espèces végétales : Une approche par régression non paramétrique. Ecologie 29(3), Wood S., Generalized additive models: an introduction with R. Chapman and Hall/CRC Press. Fisher R., The moments of the distribution for normal samples of measures of departure from normality. Proc. R. Soc.. Grime J.P., Competitive Exclusion in Herbaceous Vegetation. Nature 242(5396), Hardin G., The competitive exclusion principle. Science (80-. ). 131(3409), Delhaye G., Violle C., Séleck M., Ilunga wa Ilunga E., Daubie I., Mahy G. & Meerts P., n.d. Plant traits variation along a copper-cobalt contamination gradient: a community-level perspective exploring the relative importance of species turnover and within species variation. J. Veg. Sci. Boisson S., Le Stradic S., Collignon J., Séleck M., Malaisse F., Ngoy Shutcha M., Faucon M.-P. & Mahy G., Potential of copper-tolerant grasses to implement phytostabilisation strategies on polluted soils in South D. R. Congo. Environ. Sci. Pollut. Res. Shaw J., Heavy Metal Tolerance in Plants: Evolutionary Aspects. CRC Press Inc, New York. Ackerly D., Community assembly, niche conservatism, and adaptive evolution in changing environments. Int. J. Plant Sci. 164(May 2003), Bolnick D.I., Svanbäck R., Araújo M.S. & Persson L., Comparative support for the niche variation hypothesis that more generalized populations also are more heterogeneous. Proc. Natl. Acad. Sci. U. S. A. 104(24), Bio A.M.F., Determining alternative models for vegetation response analysis: a non- parametric approach. J. Veg. Sci. 9(1),

101 Austin M. & Smith T., A new model for the continuum concept. Vegetatio 83, Yruela I., Copper in plants. Brazilian J. Plant Physiol. 17(1), McLean J. & Bledsoe B.E., Behavior of metals in soils. Gr. Water Issue. Anderson A.J., Meyer D.R. & Mayer F.K., Heavy metal toxicities: levels of nickel, cobalt and chromium in the soil and plants associated with visual symptoms and variation in growth of an oat crop. Aust. J. Agric. Res. 24(4), Ma Y. & Hooda P.S., Chromium, Nickel and Cobalt. In: Trace Elements in Soils. John Wiley & Sons, Ltd, Faucon M.-P., Parmentier I., Colinet G., Mahy G., Ngongo Luhembwe M. & Meerts P., May Rare Metallophytes Benefit from Disturbed Soils Following Mining Activity? The Case of the Crepidorhopalon tenuis in Katanga (D. R. Congo). Restor. Ecol. 19(3), Colinet G., Average available copper and cobalt concentrations in Katanga. Whittaker R., Levin S. & Root R., Niche, habitat, and ecotope. Am. Nat. 107(955), Soberón J. & Nakamura M., Niches and distributional areas: concepts, methods, and assumptions. Proc. Natl. Acad. Sci. 106(2), Reichman S., The Responses of Plants to Metal Toxicity: A Review Focusing on Copper, Manganese & Zinc. Australian Minerals & Energy Environment Foundation. Marschner H., Römheld V., Horst W.J. & Martin P., Root-induced changes in the rhizosphere: Importance for the mineral nutrition of plants. Zeitschrift für Pflanzenernährung und Bodenkd. 149(4),

102 Appendix A. Name, family, number of presence among the three sites and ID of the 80 studied taxa Taxon Family Number of presence ID Acalypha cupricola Euphorbiaceae 31 Acalcupr Aeschynomene katangensis Fabaceae 21 Aesckata Aeschynomene pygmaea Fabaceae 10 Aescpygm Andropogon schirensis Poaceae 131 Andrschi Anisopappus davyi Asteraceae 37 Anisdavy Ascolepis metallorum Cyperaceae 38 Ascometa Brachiaria serrata Poaceae 30 Bracserr Buchnera randii Orobanchaceae 25 Buchquad Buchnera trilobata Orobanchaceae 14 Buchtril Bulbostylis cupricola Cyperaceae 27 Bulbcupr Bulbostylis filamentosa Cyperaceae 68 Bulbfila Bulbostylis macra Cyperaceae 24 Bulbfusi Bulbostylis pseudoperennis Cyperaceae 11 Bulbpseu Chamaecrista pratensis Caesalpiniaceae 15 Chammimo Chloris sp. Poaceae 37 Chlosp Chlorophytum subpetiolatum Anthericaceae 18 Chlosubp Commelina sp.1 Commelinaceae 13 Commsp1 Commelina sp.3 Commelinaceae 11 Commsp2 Commelina sp.4 Commelinaceae 8 Commsp3 Crotalaria argenteotomentosa Fabaceae 41 Crotarge Crotalaria cobalticola Fabaceae 8 Crotcoba Crotalaria sp.1 Fabaceae 11 Crotsp1 Cryptosepalum maraviense Caesalpiniaceae 105 Crypmara Cyanotis aff. cupricola Commelinaceae 44 Cyanafcu Cyanotis longifolia Commelinaceae 42 Cyanlong Cyperus cf.kibweanus 1 Cyperaceae 9 Cypecfk1 Cyperus margaritaceus Cyperaceae 10 Cypemarg Digitaria cf uniglumis Poaceae 35 Digicfun Digitaria sp. Poaceae 11 Digisp Diheteropogon grandiflorus Poaceae 20 Dihegran Diplolophium marthozianum Apiaceae 21 Diplmart Droogmansia pteropus Fabaceae 17 Drooptpt Endostemon dissitifolius Lamiaceae 29 Endodiss Eragrostis racemosa Poaceae 29 Eragrace Eriosema englerianum Fabaceae 36 Erioabys Eriosema shirense Fabaceae 11 Erioengl Eriospermum abyssinicum Eriospermaceae 8 Erioshir Euphorbia zambesiana Euphorbiaceae 9 Euphzamb Gladiolus gregarius Iridaceae 14 Gladgreg 102

103 Gladiolus ledoctei Iridaceae 26 Gladledo Haumaniastrum praealtum Lamiaceae 57 Haumprae Haumaniastrum robertii Lamiaceae 25 Haumrobe Haumaniastrum rosulatum Lamiaceae 39 Haumrosu Trachypogon spicatus Poaceae 64 Hetecont Hibiscus rhodanthus Malvaceae 20 Hibirhod Hyparrhenia diplandra Poaceae 49 Hypadipl Hyparrhenia sp. Poaceae 8 Hypasp Indigofera sp.2 Fabaceae 8 Indisp2 Ipomoea recta Convolvulaceae 28 Ipomlino Iridaceae sp.2 Iridaceae 10 Iridsp2 Justicia elegantula Acanthaceae 31 Justele Loudetia simplex Poaceae 131 Loudsimp Loudetia superba Poaceae 23 Loudsupe Monocymbium ceresiiforme Poaceae 71 Monocere Murdannia simplex Commelinaceae 36 Murdsimp Ocimum sp.1 Lamiaceae 26 Ocimsp1 Ocimum vanderystii Lamiaceae 16 Ocimvand Olax obtusifolia Olacaceae 10 Olaxobtu Oxalis obliquifolia Oxalidaceae 8 Oxalobl Oxalis semiloba Oxalidaceae 8 Oxalsemi Pandiaka carsonii Amaranthaceae 31 Pandcars Pimpinella acutidentata Apiaceae 39 Pimpacut Poaceae sp.2 Poaceae 53 Poacsp1 Polygala petitiana Polygalaceae 12 Polypeti Microchloa altera Poaceae 14 Rendalte Rhytachne rottboellioides Poaceae 35 Rhytrott Sacciolepis transbarbata Poaceae 11 Sacctran Schizachyrium brevifolium Poaceae 20 Schibrev Scleria bulbifera Cyperaceae 69 Sclebulb Sphenostylis sp. Fabaceae 10 Sphesp Sporobolus cf mildbreadii Poaceae 10 Sporcfmi Sporobolus congoensis Poaceae 26 Sporcong Striga asiatica Scrophulariaceae 9 Striasia Thesium quarrei Santalaceae 14 Thesquar Tristachya bequaertii Poaceae 48 Trisbequ Tristachya superba Poaceae 17 Trissupe Triumfetta welwitschii Malvaceae 44 Triuwe_w Vernonia aff.turbinella Asteraceae 8 Vernaffa Vernonia suprafastigiata Asteraceae 49 Vernsupr Xerophyta equisetoides Velloziaceae 39 Xeroequi 103

104 104 Appendix B. Species response curves along the available copper gradient (mg kg -1 ) using GAM. Green dots represent the copper concentration where taxa are found. Grey dashes represent the sampled concentrations. Legend: pres = number of presence, AdRsqrd = Adjusted R 2

105 105

106 106

107 107

108 108 Appendix C. Species response curves along the available cobalt gradient (mg kg -1 ) using GAM. Green dots represent the cobalt concentration where taxa are found. Grey dashes represent the sampled concentrations. Legend: pres = number of presence, AdRsqrd = Adjusted R 2.

109 109

110 110

111 111

112 Appendix D. GAM.summary function created to generate the species response curve ################################################################################## # FUNCTION GAM.summary.r # # Copyright Biodiversity and Landscape Unit - Gembloux Agro-Bio Tech - ULg # ################################################################################## #BEFORE USE GAM.summary ##WARNING: Function use mgcv package (S. Wood). ##WARNING: Column names must be registered in col.names. It works automatically with Rstudio, Rcmdr (for example). #TO PREPARE DATASET ##Put species and environmental data in column. ##Put sites or plot in line with presence/absence values (1/0) or abundance (%). First column must be contain the site or plot names. #FUNCTION PARAMETERS DESCRIPTION ##nsp: Number of species in dataset. MANDATORY ##minpres: minimal threshold to take species in account (>=). DEFAULT is 8. ##fcol: State of first column : TRUE if site names are present, FALSE if the first column contains species data. ##distribution: Type of distribution : poisson, gaussian, Gamma, binomial. BETWEEN QUOTATION. DEFAULT is BINOMIAL ##data: Name of dataset containing species and environmental data after importation in R object. MANDATORY ##nproj: Name of project and output files. BETWEEN QUOTATION ##graph: Export PDF files with species response curves X environmental parameters. One factsheet is generated by species (TRUE/FALSE). DEFAULT is TRUE #FUNCTION CODE GAM.summary<function (nsp, minpres, fcol, distribution, kmin, kmax, data, nproj, graph) { library(mgcv) library(tcltk) ##Data check if (missing(data)) { "Error with data: please attribute a dataset : data=?" } else { if (missing(nsp)) { "Error with number of species : please specify the number of species in data." } else { if (missing(minpres)) { "Threshold of presence for analysis is 8 by default." minpres<-8 } else {minpres<-minpres} if (missing(fcol)) { fcol<-false } else {fcol<-fcol} if (missing(distribution)) { "Species distribution family is binomial by default." distribution<-'binomial' } else {distribution<-distribution} if (missing(kmin)) { kmin<-3 } else {kmin<-kmin} if (missing(kmax)) { kmax<-5 } else {kmax<-kmax} if (missing(nproj)) { nproj<-"my_project" } else {nproj<-nproj} if (missing(graph)) { graph<-true } else {graph<-graph} progressbar <- tkprogressbar(title = "We calculate for you!", min = 0, max = 100, width = 300) 112

113 #SCRIPT ##Begin, variables initialization data.gam <- data if (fcol == TRUE) {colsp=2} else {colsp=1} nli<-dim(data.gam)[1] nco<-dim(data.gam)[2] nomco<-names(data.gam) nomli<-row.names(data.gam) newnomco <- NULL mpval <- NULL mndeg <- NULL mrsq <- NULL medf <- NULL moptval <- NULL newnsp <- 0 newnomli <- NULL npres <- NULL ndeg <-NULL ##Species loop for (i in colsp:(nsp+colsp-1)) { ###variables and vectors (re)initialization pval<-null rsq <- NULL edf <- NULL ndeg <- NULL optval <- NULL amplvalmin<- NULL amplvalmax<- NULL allaic<-null allgam<-null vaic<-null newnco <- 0 sp <- as.vector(data.gam[,i]) ###Which distribution? n_method <- "GCV.Cp" #By default, models were validated with CV method if (distribution == "poisson") {#if you choose Poisson family, we convert data values on integer test_int<-sp == round(sp) nb_dec<-test_int[test_int==f] if ((length(nb_dec)) >0) { #You choose Poisson method but non integer values were detected. No problem, we use quasipoisson family distribution <-"quasipoisson" n_method ="REML" paste("it seems that your data contain non integer values, quasipoisson distribution will be use in place of poisson.") paste("default method for quasi- distribution is RM/ML. This validation method method will be used now.") } } else { if (distribution == "quasipoisson") { n_method ="REML"#if you choose quasipoisson family, REML validation method is used(like CANOCO) paste("default method for quasi- distribution is RM/ML. This validation method method will be used now.") }} nbpres <- length(sp[sp>0]) ###We display the progressbar Sys.sleep(0.1) settkprogressbar(progressbar, round((i+colsp-1+1)/(nsp)*100, 0), title=paste(round((i+colsp- 1)/(nsp)*100, 0), "% done")) ###Are the species present in more than 'minpres' plot? if (nbpres >= minpres) { newnsp <- newnsp + 1 #Number of species analyzed spact <- nomco[i] newnomli <- c(newnomli, spact) #We add the name of the species in vector newnomli 113

114 continue name in vector family=distribution, method=n_method)) #Model formula distinct object cannot model the data, we save a very big AIC vector vector best model according to AIC value type='response') min(fit.gam.def)[1], length = 100)) >= p.auc) {last.l <- l; break ()} npres <- c(npres, nbpres) ####Graph is TRUE or FALSE? if (graph == TRUE) { e<-c(1, 21, 41, 61, 81, 101, 121) # Title is display all 20 charts maximum pdf(paste(nproj,"-gam-",spact,".pdf"), width=10, height=14, title=spact) #PDF opening par(mfrow=c(5,4),oma=c(2,3,5,3)) # Page is divided in 20 parts to display charts } ####Variable loop for (j in (colsp+nsp):nco) { var<-as.vector(data.gam[,j]) #We take the variable in dataset fact<-nlevels(var) #It is a factor? 1 = yes, 0= no if (fact == 0) { #If variable is not a factor, we can be analyzed it! varf<-factor(var) #Number of distinct value in variable: Step 1 nterms<-nlevels(varf) #Number of distinct value in variable: Step 2 if (nterms > 2) { #There are more than 2 distinct values? Ok, we colact <- nomco[j] #Names of variable newnco <- newnco + 1 #We add a new column if (newnsp == 1) { #If it is a new species, we add the column newnomco <- c(newnomco, colact) } #####Modelling with GAM nbaic <-1 ######Degree of smoothing loop for (l in kmin:kmax) { if (nterms > l-1) { assign(paste("gam",l,sep="."),gam(sp~s(var, k=l), a<-get(paste("gam",l,sep=".")) #We stock the models in ######According to familly, we use REML or GCV/UBRE if (n_method == "REML") { assign(paste("aic",l,sep="."), a$gcv.ubre[1]) } else { assign(paste("aic",l,sep="."), extractaic(a)[2]) } } else {assign(paste("aic",l,sep="."), )} #if we allaic <-c(allaic, get(paste("aic",l,sep=".")))#we create a AIC allgam<-c(allgam, paste("gam",l,sep=".")) #We create a gam vaic <- c(vaic, nbaic) nbaic <- nbaic +1 } gam.def<-get(allgam[allaic==min(allaic)]) #We select the GAM.deg <- l #Number of degree of freedom is saved ######Species response curve div.var<-seq(range(var)[1], range(var)[2], len=1001) fit.gam.def<-predict(gam.def, list(var=div.var), ######Optimum opt<-div.var[fit.gam.def==max(fit.gam.def)][1] ######Niche width 80% auc<-80 #We save auc value w.auc <- sum (fit.gam.def) p.auc <- w.auc * (auc/100) for (l in seq(max(fit.gam.def)[1], } prob.ampl <- last.l { if (sum (fit.gam.def[fit.gam.def >= l]) 114

115 prob.ampl]) minmax <- range(div.var[fit.gam.def >= amplmin<-minmax[1] amplmax<-minmax[2] ######Vector creation or update pval.unit <- summary(gam.def)$s.pv rsq.unit<-summary(gam.def)$r.sq amplvalmin<- c(amplvalmin, amplmin) amplvalmax <- c(amplvalmax, amplmax) optval<- c(optval, opt) ndeg<- c(ndeg, GAM.deg) pval<- c(pval, pval.unit) rsq<- c(rsq, rsq.unit) edf<- c(edf, summary(gam.def)$edf) if (graph == TRUE) { ###### If graph=true, we create the chart pdf vpage <- e == newnco u.pval<-round(pval.unit, digit=3) u.rsq<-round(rsq.unit, digit=2) sp.pres<-var[sp>0] if (distribution == "binomial") { xleg <- "Probability of occurence" #Probabily of occurence was display if model family is binomial } else {xleg <- "Relative abundance"} #Relative abundance was display if model family is quasipoisson/poisson if (pval.unit<=0.05) { plot(div.var, fit.gam.def, lty=1, type="l", col="blue", xlab=list(colact, cex = 1), ylab=list(xleg, cex = 1), cex.axis=0.8) points(var,y=rep(min(fit.gam.def),nli), pch=" ", col="gray51", cex=0.8) points(sp.pres,y=rep(max(fit.gam.def),nbpres), pch=".", col="green4", cex=1.5) mtext(paste("p-value=",u.pval), side=3, outer=f, cex=0.5, adj=0) mtext(paste("adrsqrd=",u.rsq), side=3, outer=f, cex=0.5, adj=1) } else { plot(div.var, fit.gam.def, lty=2, type="l", col="blue", xlab=list(colact, cex = 1), ylab=list(xleg, cex = 1), cex.axis=0.8) points(var,y=rep(min(fit.gam.def),nli), pch=" ", col="gray51", cex=0.8) points(sp.pres,y=rep(max(fit.gam.def),nbpres), pch=".", col="green4", cex=1.5) mtext(paste("p-value=",u.pval), side=3, outer=f, cex=0.5, adj=0) mtext(paste("adrsqrd=",u.rsq), side=3, outer=f, cex=0.5, adj=1) } if (max(vpage) == 1) { mtext(paste(spact,"(pres:",nbpres,")"), side=3, outer=t) } } }#Distinct values in environmental variable =< 2, this is not a gradient, sorry! Next variable -> } #Environmental variable is factor. Next variable -> else { var<- NULL newnco <- newnco } } #Variable Loop end if (graph == TRUE) {dev.off()} #we close de pdf file for species i 115

116 ####Result matrix creation if (newnsp == 1) { mpval <- pval #p-value mrsq <- rsq #R square medf <- edf #Degree of freedom mndeg <- ndeg #Degree of smoothing moptval <- optval #Optima mamplvalmin <- amplvalmin #Lower limit of ecological niche mamplvalmax <- amplvalmax #Higher limit of ecological niche names(mpval) <- newnomco names(mrsq) <- newnomco names(medf) <- newnomco names(mndeg) <- newnomco names(moptval) <- newnomco names(mamplvalmin) <- newnomco names(mamplvalmax) <- newnomco } else { mpval <- rbind(mpval, pval) mrsq <- rbind(mrsq, rsq) medf <- rbind(medf, edf) mndeg <- rbind(mndeg, ndeg) moptval <- rbind(moptval, optval) mamplvalmin <- rbind(mamplvalmin, amplvalmin) mamplvalmax <- rbind(mamplvalmax, amplvalmax) } } } #Species loop end #Analysis end #Result Table creation sp <- newnomli mpval <- cbind(sp, mpval) mrsq <- cbind(sp, mrsq) medf <- cbind(sp, medf) mndeg <- cbind(sp, mndeg) moptval <- cbind(sp, moptval) mamplvalmin <- cbind(sp, mamplvalmin) mamplvalmax <- cbind(sp, mamplvalmax) names(npres) <- sp #Exportation write.table(mpval, paste(nproj,"-gam_pval.csv"), dec=".", sep=";", row.names=f) write.table(mrsq, paste(nproj,"-gam_rsq.csv"), dec=".", sep=";", row.names=f) write.table(medf, paste(nproj,"-gam_edf.csv"), dec=".", sep=";", row.names=f) write.table(mndeg, paste(nproj,"-gam_degs.csv"), sep=";", row.names=f) write.table(moptval, paste(nproj,"-gam_optval.csv"), dec=".", sep=";", row.names=f) write.table(mamplvalmin, paste(nproj,"-gam_amplvalmin.csv"), dec=".", sep=";", row.names=f) write.table(mamplvalmax, paste(nproj,"-gam_amplvalmax.csv"), dec=".", sep=";", row.names=f) write.table(npres, paste(nproj,"-gam_nbpres.csv"), sep=";") } } } close(progressbar)#progressbar stop dos<- getwd() paste("files save in: ", dos) 116

117 Axe 1 -Article 3 Modifications majeures - Edaphic niches of copper endemics from southeastern D. R. Congo: implications for post-mining restoration BOISSON Sylvain, MONTY Arnaud, LEBRUN Julie, SÉLECK Maxime & MAHY Grégory Abstract Southeastern D. R. Congo, about 550 metallophytes grow on soils with high copper and cobalt concentrations with among which 57 are endemics to these metalliferous environments. About 70% of those endemics are considered threatened by mining activities because of habitat destruction during exploitation phase. To provide guidelines for future restoration programs, the edaphic ecological niches of eight copper endemics were studied using a pragmatic sampling method adapted for urgent conservation needs. Niches were modelled using violin plot along Cu, Co and C:N gradients representing the two main independent edaphic gradients among nine edaphic variables (C, N, C:N, K, P, ph, Co, Cu, and Mn). Copper endemics presented distinct edaphic niches along the copper and cobalt gradients, but differentiation was lower along the C:N gradient. In addition, edaphic elements presented covariations among them and metalliferous soils had higher nutrient content and elements presented compared to the non-metalliferous soils of the region dominated by the Miombo woodland. The complexity of the soil composition and the edaphic niches of copper endemics revealed an important challenge in the implementation of the species conservation and the habitat restoration strategies of post-mining sites. 117

118 Introduction Metal-rich outcrops in many parts of the world are recognised as major foci of biodiversity (Whiting et al., 2004). These ecosystems host highly distinctive plant communities due to their ecological isolation and strong environmental selective pressures (Baker, Ernst, Van der Ent, Malaisse, & Ginocchio, 2010; Bizoux, Brevers, Meerts, Graitson, & Mahy, 2004). These plants occurring on heavy metal-enriched soils are called metallophyte and present various physiological mechanisms to tolerate the excess of metal in soils (Hall, 2002; Shaw, 1990). Such plants communities often include rare and endemic taxa (Whiting et al., 2004) and benefit of high conservation value. Southeastern Democratic Republic of Congo hosts one of the most important metalliferous areas in the world. The Katangan Copperbelt includes some 160 outcrops (i.e., copper hills) scattered over more than 300 km (W-E) in a Miombo woodland matrix (Brooks & Malaisse, 1985; Duvigneaud & Denaeyer-De Smet, 1963). Mineralization of the parent rocks promoted a copper and cobalt enrichment of the soils along the slope of hills with concentrations of copper from 20 to 10,000 mg kg -1 and cobalt from 2 to 1,000 mg kg -1 toward the top of the copper hills (Séleck et al., 2013). The Katangan Copperbelt is recognised as a hotspot for metallophytes diversity (Brooks & Malaisse, 1985). Two main vegetation types are found on the copper hills: a steppic savanna on the lowest and intermediate metal concentrations and a steppe on the highest metal concentrations (Saad et al., 2012; Séleck et al., 2013). These plant communities host more than 550 species tolerant to high cobalt and copper concentrations (Chipeng et al., 2009; Faucon, Shutcha, & Meerts, 2007; Lange et al., 2014; Leteinturier, 2002). Among those tolerant species, 57 have been identified as endemic from the Katangan copper hills (i.e., copper endemics). Two levels of endemism have been described: the broad copper endemics having more than 75% of population occurrence on copper hills in Katanga and the strict endemics occurring solely on the copper hills (Champluvier, 2011; Faucon et al., 2010). Due to a recent revival of mining activities, most copper hills in Katanga have now been allotted to mining companies and will be exploited in the coming years and decades. Mining activities have already totally or partially destroyed several sites (Faucon et al., 2010). Although some copper metallophytes may actually benefit from secondary habitats created by the mineral industry (Faucon et al., 2011), there is no doubt that the copper endemics 118

119 could be considered highly threatened. A recent assessment of their conservational status suggested that according to IUCN criteria, 80% of copper endemics would be considered threatened (i.e., critically endangered, endangered, vulnerable) (Faucon et al., 2010). The best strategy from a biodiversity conservation point of view includes preservation of a proportion of copper endemic populations in protected areas (International Council on Mining and Minerals (ICMM), 2006). However, this may turn out to be a limited and economically unrealistic option considering the increasing demand for copper and cobalt in the world and the high contribution of the copper-cobalt market to the economy of D. R. Congo. In mining sites, the most realistic way to ensure the preservation of threatened metallophytes may be to promote their ex situ conservation during mining activity associated with their use in ecological restoration at mine closure (Faucon et al., 2011; Saad et al., 2012; Whiting et al., 2004). However, designing ex situ conservation programs and restoring ecosystems that can support viable populations of copper endemics require sufficient knowledge about the ecology of those species (Whiting et al., 2004). The characterisation of the edaphic niche of metallophytes is particularly critical (Ellstrand & Elam, 1993; Whiting et al., 2004). Metallophytes present a broad variation of tolerance to heavy metals resulting in specific distribution in relationship to soil properties. Absolute metallophytes are restricted to metalliferous soils, whereas facultative metallophytes show various proportions of metallicolous and non-metallicolous populations (Baker et al., 2010). Even among absolute metallophytes, tolerance to heavy metals may vary. Recent studies point out that other edaphic factors may also be important to explaining the distribution of plant species on copper hills (Saad et al., 2012; Faucon et al., 2012; Séleck et al., 2013). In contrast to most metalliferous soils worldwide, metalliferous soils from Katangan Copperbelt present high concentrations of nutrients such as P, Ca and K. Saad et al. (2012) suggested that nutrient status may also be of prime importance in explaining plant community variability. The complexity of plant-soil relationships in the copper hills requires the development of quantitative approaches to characterise the ecological niches of copper endemics. For a long time, botanists have relied on the seminal work of Duvigneaud for taxonomy, ecology and distribution of Katangan metallophytes (Duvigneaud & Denaeyer-De Smet, 1963; Duvigneaud & Timperman, 1959). Those studies suggest that copper endemics may present 119

120 different niche optimum and width in relationship to copper and cobalt. However, these studies were based on qualitative approaches and did not provide an understanding of the fine scale variation of edaphic niches (B Leteinturier & Malaisse, 1999; Saad et al., 2012). The classical methods of ecological niche modelling along edaphic or climatic factors implies the collection of large amount of species presence (or abundance) and absence data (Heikkinen & Mäkipää, 2010; Vetaas, 2002; Wiser, Peet, & White, 1998; Yee & Mitchell, 1991). Presence-only data can be used at large scale provided pseudo-absence data are generated (Guisan, Edwards, & Hastie, 2002; Zaniewski, Lehmann, & Overton, 2002). However, in critical situations like the Katangan Copperbelt, where the exponential development of mining activities may rapidly overcome conservation efforts, it is crucial to develop pragmatic, conservation-oriented methods to survey the endemic taxa and to characterise their relationship with edaphic factors. In this study, a conservation-oriented method using transects was selected to survey eight endemic metallophytes of the Katangan Copperbelt and to characterise their edaphic niches on copper hills in order to understand their relationship with determinant soil factors and to derive implications for conservation at species scale including these taxa in future habitat restoration strategies. Methods Study site and plant materials This study was performed in the south-eastern Democratic Republic of Congo between Tenke (10.61 S; E) and Fungurume (10.62 S; E). This region has a humid subtropical climate (Köppen-Geiger: CWa) tempered by the relatively high elevation (c m a.s.l.). Rainy season extends from November to the end of March, and dry season is from May to September. Eight copper endemics with putative threatened IUCN status sensu Faucon 2010 (Table 1) were considered in the present study: Barleria lobelioides Champl. (Acanthaceae), Basananthe kisimbae Malaisse & Bamps (Passifloraceae) endangered; Commelina zigzag P. A. Duvign. & Dewit (Commelinaceae) critically endangered; Euphorbia cupricola (Malaisse & Lecron) Bruyns (Euphorbiaceae) critically endangered; Lopholaena deltombei P. A. Duvign. (Asteraceae) critically endangered; Sopubia neptunii P. A. Duvign. & Van Bockstal 120

121 (Orobanchaceae) vulnerable; Tinnea coerulea var. obovata (Robyns & Lebrun) Vollesen (Lamiaceae) vulnerable; and Triumfetta likasiensis De Wild (Malvaceae) vulnerable. Identifications of taxa were based on Bamps ( ), Board of Trustees Kew Royal Botanic Gardens ( ) and Flora of Tropical East Africa (Kew Royal Botanic Gardens, ) completed with more recently published taxonomic literature for particular genera and species (Leteinturier, 2002; Séleck et al., 2013). Table 1. Proposed IUCN status of Faucon et al. (2010), mean density of individuals among sites and sites (i.e. copper hills) where the two meters transects were performed for species data collection. Copper hills (i.e., presenting metalliferous soils) are located between the towns of Tenke and Fungurume. A cross (X) indicated that the endemic metallophyte was surveyed on the considered site. No value means that the copper endemic was not sampled on the considered site. Legend of IUCN status: CR= Critically endangered, EN= Endangered, VU= Vulnerable, NA= not assessed. Fungurume Goma Kabwelunono I Kakavilundo West Orthodoxe II Mean Taxon IUCN status II III IV V X II III density of individual B. kisimbae EN X X X 2.9±4.8 B. lobelioides - X X X 2.5±1.9 C. zigzag CR X X X 3.8±2.7 E. cupricola CR X X X 4.1±4.7 L. deltombei CR X X X 1.0±0.0 S. neptunii VU X X X 3.7±3.5 T. coerulea VU X X X 1.3±0.8 T. likasiensis VU X X X 1.5±0.6 Field data collection To assess the edaphic niches of each studied species, a sampling method requiring the collection of substantially less data than the classical presence-(pseudo-)absence samplings was used. This approach is applied in botany and plant ecology and is based on the characterisation of population density and soil parameters at key positions along transect performed on the heavy metal gradients found on the copper hills. The survey was performed during the growing season 2012 (Table 1). The first step consisted of characterising the distribution of the population along the environmental gradients (Figure 1). For each species, three copper hills where the species occurred were sampled for a total of 10 copper hills. Topography is a good proxy of the copper-cobalt gradient on copper hills (Séleck et al., 2013). On each hill, a two-meter wide transect was materialised from the bottom to the top of the hill. All individuals of the studied metallophytes were numbered based on their position in transect. In a second step, the rank 121

122 distribution of individuals was used to select five key positions: the two individuals at the extremities of the population, the individual with median rank, one individual at the 1 st quartile rank and one individual at the 3 rd quartile rank. In the third step, a one-square-meter plot was centered on each key position. In the plot, the species density was recorded, and four 10 cm-deep soil samples were randomly collected and pooled. A total of 15 plots was sampled per species (3 hills*5 key positions). To compare the nutrient and metal concentration of the copper hills with the nonmetalliferous soils of the region, soil samples were taken in the non-perturbed areas of the plain of Lubumbashi located in the Miombo woodland. For this sampling, 145 (for the analysis of C, N, K, ph and Mn) and 53 plots (for the analysis of Cu) were performed. Figure 1. Scheme of the field data collection on a copper hill (i.e., having metalliferous soil). Copper endemic population sampling was performed along a topographic transect from the Miombo (excluded in transect) to the top of hill on each site. Each individual was numbered. Five ranks were chosen in order to establish five plots. Soil analysis Soil samples were sieved to less than 2 mm. The ph was measured in 1N KCl with a glass electrode in a 2:5 soil to solution ratio after 2h equilibration time. Total organic and nitrogen content were measured in soil after dry combustion. Carbon and nitrogen were oxidized to CO 2, NO x and N 2 by heating the soil to at least 900 C in a flow of oxygen-containing gas. The amount of CO 2 released was then measured by gas chromatography (GC) method. Nitrogen 122

123 content was measured by means of thermal conductivity detection (Margesin et al., 2005). Extractable fractions of Co, Cu, K, P, and Mn were determined after extraction with 0.5N CH3COONH 4 -EDTA for 30 min (Lakanen et al., 1971). The soil/solution ratio was 1:5, and the ph of the extraction solution was buffered at The supernatant was filtered through a S&S 595 folded filter and analysed using a flame atomic absorption spectrometer (Varian 220). Phosphorus content was determined with a Shimadzu UV-1205 spectrophotometer (Shimadzu Corporation) at 430 nm after the development of blue coloration. Extractable fractions of the elements are the concentrations that are theoretically available for plant uptake at the time scale of one growing season. In addition to heavy metals, C, N, P and K were included in the analysis because they contribute to the characterisation of the soil nutrient status and to the level of mineralisation. Data analysis A Kruskal-Wallis rank sum test was performed to compare the nutrient and metal concentrations between the soils from the copper hills and the Miombo woodland (nonmetalliferous soil). To explore interrelationships among the considered edaphic factors and to select independent chemical edaphic parameters for niches modelling, a principal component analysis (PCA) using soil data was performed from all plots. For each species, the one-dimensional ecological niches along copper, cobalt and independent edaphic factor(s) identified from the PCA were modelled for each species using violin plots (Hintze & Nelson, 1998). The violin plot combines the box and density trace into one diagram. The density trace was modelled with the kernel density method and was plotted symmetrically on each side of the box plot. The density curve was the species response curve including extreme values. Niche optimum was defined by the median of the data. The central part of the niche (i.e., central niche width) was the interquartile range (Q3-Q1) in the box plot in which 50 % of individuals were present. All analyses were performed using the R statistical software with stats and vioplot package (R Development Core Team, 2010). 123

124 Results and discussion Chemical edaphic variables variations As expected, copper concentrations were more than 100 times higher in the soils of the copper hills compared to the soils of the Miombo woodland that represents the nonmetalliferous soils of the region. As observed by Faucon (2011), soil from the copper hills had significantly higher nutrients concentrations (i.e., C, N and P) and ph value than the nonmetalliferous soils, except for potassium (Table 2). In metalliferous soils, the C:N ratio value was higher than 15, suggesting that carbon mineralisation was regular. Mean carbon, nitrogen and phosphorus content of metalliferous soils were respectively 26.2 ± 13.3 g kg -1, 1.36 ± 0.74% and 1.25 ± 1.55 mg 100g -1 with a ph of 4.9 ± 0.5. The metal concentrations were 26.2 ± 13.3 mg kg -1 of Co and 1,506 ± 2,438 kg -1 of Cu, respectively. These results confirmed that these communities differ from those founded on serpentine soils, where deficiency of N, P, K and Ca has been suggested (Faucon et al., 2011; O Dell & Claassen, 2006). Table 2. Means, standard deviations and ranges [min-max] of the prospected sites for the species sampling. Soils data of the copper hills were calculated from 128 plots and the soil data of the non-metalliferous soils (Miombo woodland) were calculated from 145 plots (C, N, K, ph and Mn) and 53 plots (Cu) located in the woodland around Lubumbashi (D.R.C.). Soil analyses were performed with the same method that the present study. Copper Hills Non-metalliferous soils Mean±SD Range [min-max] Mean±SD Range [min-max] X 2 p-value C (g.kg -1 ) 26.2 ± 13.3 [ ] 13.7 ± 4.8 [ ] < N (%) 1.36 ± 0.74 [ ] 0.10 ± 0.03 [ ] < C:N ratio 17.6 ± 2.6 [ ] 13.0 ± 1.4 [ ] < K (mg.100g -1 ) 6.8 ± 3.1 [ ] 12.7 ± 5.3 [ ] < P (mg.100g -1 ) 1.25 ± 1.55 [LQ-10.77] 0.24 ± 0.07 [ ] < ph KCl 4.9 ± 0.5 [ ] 4.2 ± 0.2 [ ] < Co (mg.kg -1 ) 26.2 ± 13.3 [ ] NA NA - - Cu (mg.kg -1 ) 1,506 ± 2,438 [53-11,013] 12 ± 9 [1-59] < Mn (mg.kg -1 ) 94.5 ± [ ] 49.1 ± 50.3 [ ] < LQ means that the value was lower than the limit of quantification In most cases, the soil of the copper hills and the non-metalliferous soil shared a same part of edaphic element variations, but the range of variation of soil elements was largest in the soil of the copper hills. The first two axes of the PCA for nine edaphic variables explained 42.8% of the total variation (Figure 2). Two main groups of correlated edaphic factors could be 124

125 highlighted. The variables most correlated with the first axis were Co, ph, K, Mn, P, Cu. The elements C, N and the C:N ratio relating the dynamic of mineralisation were more correlated with the second axis. As a result, the edaphic niches were characterised along Cu and Co that correspond to the most relevant factors structuring the vegetation (Saad et al., 2012; Séleck et al., 2013) and the C:N ratio that appeared as a factor independent from metals. Figure 2. Principal component analysis of the edaphic variables from the square meter plots being in metalliferous soils. Dots represent the sampling plots. Arrows represent edaphic factors. Axis 1 and Axis 2 represent respectively 21.9 % and 20.9 of the edaphic variability. Copper endemics niche The response curves of copper endemics along C:N gradient overlapped greatly with density maxima and medians within a range from 15 to 19. In contrast, response curves along heavy metals concentrations spanned a wide range with obvious interspecific variability in optimum and central niche width. Copper endemics were distributed along a copper gradient ranging from 50 to 11,000 mg Cu.kg -1. Along the cobalt gradient, copper endemics occurred in concentrations ranging from 4 to 1,000 mg Co.kg

126 Three groups of taxa could be identified according to their optima location and the central niche width along the two metal gradients. The first group was composed of five species having optima in concentrations lower than 300 mg Cu.kg -1 and 55 mg Co.kg -1 : Tinnea coerulea var. obovata; Barleria lobelioides; Triumfetta likasiensis; Basanathe kisimbae; and Lopholaena deltombei. All the species of this group have been reported as absolute metallophyte (strict endemic sensu Faucon et al., 2010) except for Tinnea coerulea var. obovata (facultative metallophyte) (Champluvier, 2011; Paul Duvigneaud & Denaeyer-De Smet, 1963; Faucon et al., 2010). Except for B. kisimbae, the response curves of this group appeared truncated in the lower limits of the copper-cobalt ranges that correspond to Cu, Co concentrations in non-metalliferous soils of the region (Table 2). Copper and cobalt response curves overlapped greatly for endemics of this group. The species B. kisimbae appeared to be the most specialised copper endemic along the copper gradient in the three studied sites considering its central niche width. Ilunga et al. (2013) observed a similar pattern for nonendemic metallophytes on copper hills. A second group was composed of two species with a large central niche width for both copper and cobalt: Commelina zigzag (5,302 to 8,600 mg Cu.kg -1 ) and Sopubia neptunii (1,810 to 7,170 mg Cu.kg -1 ). For copper, both species had a niche optimum in high copper concentrations (C. zigzag: 7,570 mg Cu.kg -1, S. neptunii: 5,325 mg Cu.kg -1 ). For cobalt, despite the two species displaying the largest niche, their response curves were truncated in the lower limit of the cobalt range. However, the two species presented a different behaviour in relationship to cobalt concentration. S. neptunii had its optimum in the lowest part of the cobalt range (33 mg Co.kg -1 ), while C. zigzag presented the highest optimum for cobalt concentrations (162 mg Co.kg -1 ). Both species occur preferentially in the steppic vegetation of copper hills (Saad et al., 2012). C. zigzag has been reported to be an absolute metallophyte having very restricted distribution in the Katangan Copperbelt. In contrast, S. neptunii has been considered a facultative metallophyte species (Brooks & Malaisse, 1985; Saad et al., 2012) due to its occurrence outside the copper hills (Faucon et al., 2010). It was identified as a hemiparasitic species and thus may depend on the edaphic niches of its hosts (Poaceae) (Morawetz J. J., personal communication). Euphorbia cupricola presented its optimum (higher than 1,200 mg Cu.kg -1 ) and its central niche width (332 to 1,313 mg Cu.kg -1 ) in the intermediate values of the copper concentrations. Along cobalt concentrations, the species had its optimum in the lowest 126

127 concentrations (53 mg Co.kg -1 ), ranging from 45 to 90 mg Co.kg -1. This taxon was previously identified as an absolute metallophyte and is one of the rarest copper endemics (Faucon et al., 2010). 127

128 128 Figure 3. Species response curves of eight copper endemics along available copper and cobalt concentrations (mg kg -1 ) and the C:N ratio. Used functions were created from vioplot package (R-software). Legend: Plain black dots, squares or diamonds are the medians along respectively copper, cobalt and C:N gradients, the gray rectangles is a boxplot (the lower and upper limits are the first and third quartile) and it represents the central niche width. The light gray surfaces are the density of individuals (i.e. response curve) along considered gradients taking into account 100 % individuals.

129 Implication of niche characteristics for copper endemics conservation Our results point to a general relationship between the reported metallophyte status of studied endemics and their observed ecological niche. The three facultative metallophytes and non- hemiparasitic species B. kisimbae, T. coerulea var. obovata and T. likasiensis were found in the lowest copper concentrations corresponding to non-metalliferous soils of the region. In contrast, the four absolute metallophytes, B. lobelioides, C. zigzag, E. cupricola and L. deltombei, presented different central niche widths along the copper gradient. Barleria lobelioides occurred in the lowest copper concentrations compared to C. zizag, which had the largest edaphic niche in the highest copper concentrations. The pattern is less clear along the cobalt gradient because except for C. zigzag, the copper endemics had similar cobalt edaphic niches. The diversity of niche patterns observed for copper metallophyte may result from complex inter-relationships of heavy metals with other factors. Species response may be influenced by limiting biotic factors such as microorganisms. Copper endemics appear specifically to be susceptible to soil fungi (Chipeng et al., 2009; Faucon et al., 2012; Malaisse, Brooks, & North, 1982) due to reduced pathogenic pressure on metal-rich soils (Tadros, 1957). Narrow ranges of some copper endemics can be also related to a low competitive ability linked to their nature as stress-tolerant plants (Devictor et al., 2010; Palacio et al., 2007). In the case of facultative metallophyte, taxa are described as genotype or ecotype (or subspecies) of common species with a specific tolerance to metals (Baker et al., 2010). Surprisingly, most of the studied copper endemics have their optima and central niche width in the lowest metal concentrations with truncated response curves, particularly along the cobalt gradient. These observations call into question the absolute metallophyte status of those species. This suggests that these copper endemics may be able to grow in nonmetalliferous similar vegetation (steppic savanna or herbaceous vegetation). Targeted explorations for the copper endemics are needed to confirm that those species do not occur in non-metalliferous herbaceous communities. Those results also suggest that those copper endemics may be maintained on non-metalliferous soils in ex situ conservation programs. It has been recently demonstrated that copper endemics are able to germinate in non-metal enriched substrate (Godefroid et al., 2013). Thus, further regeneration tests in nonmetalliferous soils should be implemented to provide guidelines for the ex situ conservation of copper endemics. 129

130 Successful restoration requires fundamental knowledge about both biotic and abiotic factors and also about ecological processes on several scales (Clewell, Rieger, & Munro, 2005). The diversity of the modelled edaphic niches implies that conservation has to be planned at the species scale in order to restore successfully the habitats and the communities. The reproduction of the species habitats' conditions in a post-mining restoration project is an economic and environmental challenge that requires financial and human resources. Furthermore, the restoration of harsh edaphic conditions in ex-situ areas as metalliferous soils is contrary to the principles of soil rehabilitation and reclamation of polluted sites. Compared to the non-metalliferous soils of the region, the soils of studied sites contained higher concentrations of total nitrogen, phosphorus and organic carbon and had a higher ph value. Concentrations of potassium and manganese were similar (Table 2). This highest level of nutrient status has been already shown by Faucon et al. (2011). Furthermore, these edaphic compounds such as ph, organic matter and oxyhydroxides are correlated to the heavy metal bioavailable concentrations in soils of the Katangan Copperbelt (Saad et al., 2012; Séleck et al., 2013; Pourret et al., 2015). In this context, the habitat restoration through the species transplantation or sowing has to consider the particular edaphic conditions of the metalliferous soils and the variations between their elements in order to create the most favorable growing environments. Acknowledgements The present work was made possible thanks to the financial intervention of the Fonds de la Recherche dans l Industrie et dans l Agriculture (FRIA) and the FNRS (Fonds National de la Recherche Scientifique), Belgium. We thank Tenke Fungurume Mining S.a.r.l., which provided logistical support in the field. References Baker, A. J. M., Ernst, W. H. O., Van der Ent, A., Malaisse, F., & Ginocchio, R. (2010). Metallophytes: the unique biological resource, its ecology and conservational status in Europe, central Africa and Latin America. In L. Batty & K. Hallberg (Eds.), Ecology of Industrial Pollution (pp. 7 40). Bamps, P. (1993). Flore d Afrique centrale (Zaïre Rwanda Burundi). Jardin Botanique National de Belgique, Meise, Belgique. Bizoux, J., Brevers, F., Meerts, P., Graitson, E., & Mahy, G. (2004). Ecology and conservation of Belgian populations of Viola calaminaria, a metallophyte with a restricted geographic distribution. Belgian Journal of Botany, 137(1), Retrieved from Board of trustees Kew Royal Botanic Gardens. (2010). Flora Zambesiaca. Royal Botanic Gardens Kew, Richmond. Brooks, R. R., & Malaisse, F. (1985). The Heavy Metal Tolerant Flora of Southcentral Africa: a mutlidisciplinary approach. Rotterdam: A.A. Balkema. Retrieved from Champluvier, D. (2011). New and overlooked Acanthaceae taxa from D.R.Congo, Rwanda and Burundi: 130

131 (1) the genus Barleria. Plant Ecology and Evolution, 144(1), doi: /plecevo Chipeng, F. K., Hermans, C., Colinet, G., Faucon, M.-P., Ngongo, M., Meerts, P., & Verbruggen, N. (2010). Copper tolerance in the cuprophyte Haumaniastrum katangense (S. Moore) P.A. Duvign. & Plancke. Plant and Soil, 328(1-2), doi: /s z Clewell, A., Rieger, J., & Munro, J. (2005). Guidelines for Developing and Managing Ecological Restoration Projects. Society for Ecological Restoration International, 1(December), Retrieved from Devictor, V., Clavel, J., Julliard, R., Lavergne, S., Mouillot, D., Thuiller, W., Mouquet, N. (2010). Defining and measuring ecological specialization. Journal of Applied Ecology, 47(1), doi: /j x Duvigneaud, P., & Denaeyer-De Smet, S. (1963). Etudes sur la végétation du Katanga et de ses sols métallifères. Communication n 7 Cuivre et végétation au Katanga. Bulletin de La Société Royale de Botanique de Belgique, 96(2), Duvigneaud, P., & Timperman, J. (1959). Etudes sur la végétation du Katanga et de ses sols métallifères. Communication n 3 Etudes sur le genre Crotalaria. Bulletin de La Société Royale de Botanique de Belgique, 91(2), Ellstrand, N. N. C., & Elam, D. D. R. (1993). Population genetic consequences of small population size: implications for plant conservation. Annual Review of Ecology and Systematics, 24(1993), Faucon, M.-P., Chipeng, F., Verbruggen, N., Mahy, G., Colinet, G., Shutcha, M., Meerts, P. (2012). Copper tolerance and accumulation in two cuprophytes of South Central Africa: Crepidorhopalon perennis and C. tenuis (Linderniaceae). Environmental and Experimental Botany, 84, doi: /j.envexpbot Faucon, M.-P., Meersseman, A., Shutcha, M. N., Mahy, G., Luhembwe, M. N., Malaisse, F., & Meerts, P. (2010). Copper endemism in the Congolese flora: a database of copper affinity and conservational value of cuprophytes. Plant Ecology and Evolution, 143(1), doi: /plecevo Faucon, M.-P., Parmentier, I., Colinet, G., Mahy, G., Ngongo Luhembwe, M., & Meerts, P. (2011). May Rare Metallophytes Benefit from Disturbed Soils Following Mining Activity? The Case of the Crepidorhopalon tenuis in Katanga (D. R. Congo). Restoration Ecology, 19(3), doi: /j x x Faucon, M.-P., Shutcha, M. N., & Meerts, P. (2007). Revisiting copper and cobalt concentrations in supposed hyperaccumulators from SC Africa: influence of washing and metal concentrations in soil. Plant and Soil, 301(1-2), doi: /s Godefroid, S., Van de Vyver, A., Massengo Kalenga, W., Handjila Minengo, G., Rose, C., Ngongo Luhembwe, M., Mahy, G. (2013). Germination capacity and seed storage behaviour of threatened metallophytes from the Katanga copper belt (DR Congo): implications for ex situ conservation. Plant Ecology and Evolution, 146(2), doi: /plecevo Guisan, A., Edwards, T. C., & Hastie, T. (2002). Generalized linear and generalized additive models in studies of species distributions: setting the scene. Ecological Modelling, 157(2-3), doi: /s (02) Hall, J. L. (2002). Cellular mechanisms for heavy metal detoxification and tolerance. Journal of Experimental Botany, 53(366), Retrieved from Heikkinen, J., & Mäkipää, R. (2010). Testing hypotheses on shape and distribution of ecological response curves. Ecological Modelling, 221(3), doi: /j.ecolmodel Hintze, J., & Nelson, R. (1998). Violin plots: a box plotdensity trace synergism. The American Statistician, 52(2), Ilunga wa Ilunga, E., Séleck, M., Colinet, G., Meerts, P., & Mahy, G. (2013). Small-scale diversity of plant communities and distribution of species niches on a copper rock outcrop in Upper Katanga, DR Congo. Plant Ecology and Evolution, 146(2), doi: /plecevo International Council on Mining and Minerals (ICMM). (2006). Good Practice Guidance for Mining and Biodiversity. London: ICMM. Kew Royal Botanic Gardens. (2008). Flora of tropical East Africa. In R. M. Polhill (Ed.), Flora of tropical East Africa. Kew, UK: Royal Botanic Gardens. Lakanen, E., & Erviö, R. (1971). A comparison of eight extractants for the determination of plant available micronutrients in soil. Acta Agralia Fennica, 123, Lange, B., Faucon, M.-P., Meerts, P., Shutcha, M., Mahy, G., & Pourret, O. (2014). Prediction of the edaphic factors influence upon the copper and cobalt accumulation in two metallophytes using copper and cobalt speciation in soils. Plant and Soil, 379(1-2), doi: /s y Leteinturier, B. (2002). Evaluation du potential phytocénotique des gisements cupriferes d Afrique centro-australe en vue de la phytoremédiation de sites pollués par l'activité. Faculté des Sciences Agronomiques de Gembloux. Leteinturier, B., & Malaisse, F. (1999). The copper flora of Katanga: a phytogeographical analysis. Geo-Eco-Trop, 23, Malaisse, F., Brooks, R. R., & North, P. (1982). Colonisation of modified metalliferous environments in Zaire by the copper flower Haumaniastrum katangense. Plant and Soil, 64, Margesin, R., & Schinner, F. (2005). Manual for Soil Analysis - Monitoring and Assessing Soil Bioremediation: Monitoring and Assessing Soil Bioremediation. (Springer Science & Business Media, Ed.). Springer. O Dell, R. E., & Claassen, V. P. (2006). Relative performance of native and exotic grass species in response to amendment of drastically disturbed 131

132 serpentine substrates. Journal of Applied Ecology, 43(5), doi: /j x Palacio, S., Escudero, A., Montserrat-Martí, G., Maestro, M., Milla, R., & Albert, M. J. (2007). Plants living on gypsum: beyond the specialist model. Annals of Botany, 99(2), doi: /aob/mcl263 Pourret, O., Lange, B., Houben, D., Colinet, G., Shutcha, M., & Faucon, M.-P. (2015). Modeling of cobalt and copper speciation in metalliferous soils from Katanga (Democratic Republic of Congo). Journal of Geochemical Exploration, 149, doi: /j.gexplo R Development Core Team. (2010). A language and environment for statistical computing. (R Foundation for Statistical Computing, Ed.). Vienna (Austria). Saad, L., Parmentier, I., Colinet, G., Malaisse, F., Faucon, M.-P., Meerts, P., & Mahy, G. (2012). Investigating the Vegetation-Soil Relationships on the Copper-Cobalt Rock Outcrops of Katanga (D. R. Congo), an Essential Step in a Biodiversity Conservation Plan. Restoration Ecology, 20(3), doi: /j x x Séleck, M., Bizoux, J.-P., Colinet, G., Faucon, M.-P., Guillaume, A., Meerts, P., Mahy, G. (2013). Chemical soil factors influencing plant assemblages along coppercobalt gradients: implications for conservation and restoration. Plant & Soil, 373(1/2), doi: /s Séleck, M., Lebrun, J., Boisson, S., Le Stradic, S., Masengo Kalenga, W., & Mahy, G. (2013). Guide to Cu-Co species of concern regeneration - V2.2. Gembloux. Shaw, J. (1990). Heavy Metal Tolerance in Plants: Evolutionary Aspects. (J. Shaw, Ed.). New York: CRC Press Inc. Tadros, T. T. M. (1957). Evidence of the presence of an edapho-biotic factor in the problem of serpentine tolerance. Ecology, 38(1), Vetaas, O. R. (2002). Realized and potential climate niches: a comparison of four Rhododendron tree species. Journal of Biogeography, 29, Whiting, S. N., Reeves, R. D., Richards, D., Johnson, M. S., Cooke, J. A., Malaisse, F., Baker, A. J. M. (2004). Research Priorities for Conservation of Metallophyte Biodiversity and their Potential for Restoration and Site Remediation. Restoration Ecology, 12(1), Wiser, S. K., Peet, R. K., & White, P. S. (1998). Prediction of rare-plant occurence: A Southern appalachian example. Ecological Applications, 8(4), Yee, T. W., & Mitchell, N. D. (1991). Generalized additive models in plant ecology. Journal of Vegetation Science, 2, Zaniewski, A. E., Lehmann, A., & Overton, J. M. (2002). Predicting species spatial distributions using presenceonly data: a case study of native New Zealand ferns. Ecological Modelling, 157(2-3), doi: /s (02)

133 Axe 1 -Article 4 Publié - Copper tolerance of three Crotalaria species from the Katangan Copperbelt at the early development stage BOISSON Sylvain, LE STRADIC Soizig, COMMANS Morgane, DUMONT Amandine, LECLERC Natasha, THOMAS Cynthia & MAHY Grégory Abstract Description of the subject. Microevolutionary processes in metallophytes established on copper enriched soils can lead to a diversity of plant species showing distinct tolerance capacities among genus. Researches about the relationship between these endangered plants and soil copper concentrations are critical in order to improve ex situ conservation methods in Southeastern Democratic Republic of Congo (Katangan Copperbelt). Objectives. The aim of the study was to test the effect of copper on the germination and root elongation of three Crotalaria species naturally occurring along a natural copper gradient. The hypothesis is that copper concentration has different effect on germination and root elongation according to the species of Crotalaria genus. Method. Three species were selected: Crotalaria cobalticola, Crotalaria peschiana and Crotalaria cornetii, occurring on soils with the highest to the lowest copper concentrations respectively. Germination and root elongation tests were performed in vitro (MS vitaminenriched medium) in six copper mediums ranging from 0 to 125 µm Cu 2+. Results. No significant differences in germination percentage were observed according to the copper concentrations. Crotalaria cornetii had the lowest germination percentage. Root elongation of Crotalaria peschiana did not differ with copper concentration, but root elongation of C. cobalticola was higher at the greatest copper concentration (125 µm Cu 2+ ). Conclusions. Even if C. cobalticola presented better growth at highest Cu concentrations, it appeared that C. cobalticola and C. peschiana do not require copper for their early stages of development and could thus be conserved in non-contaminated substrate. C. cornetii seemed to present a physical seed dormancy. 133

134 Introduction Metalliferous areas present high concentrations of trace metals associated with other environmental constraints including xericity and poor depauperate nutrient content (Ernst, 1974; Kruckeberg, 1985). These harsh abiotic conditions are strong drivers for local adaptation of plant species (Antonovics et al., 1971; Brady et al., 2005) that results in the formation of highly distinctive plant communities (Duvigneaud et al., 1963a; Ernst, 1974; Rajakaruna, 2004; Saad et al., 2012). Such ecosystems are found in many places around the world (e.g. Arizona, Peru, Chili, Australia or Democratic Republic of Congo). They are usually recognized as major hotspot of biodiversity (Whiting et al., 2004; Veldman, Overbeck, et al., 2015), offering outstanding examples of microevolution and speciation processes (Antonovics et al., 1971; Faucon et al., 2010; Harrison et al., 2011). Ecosystems established on soils naturally enriched with copper are rare, except in South Central Africa where occurs a large number of copper-cobalt mineralized areas (Duvigneaud et al., 1963a; Baker et al., 2010; Chipeng et al., 2010). The Katangan Copperbelt, in the Katanga province (South East of Democratic Republic of Congo), extends over 300 km long between Kolwezi and Lubumbashi with more than 150 copper- and cobalt-rich outcrops (François, 1988; Cailteux et al., 2005). These outcrops appear as isolated hills scattered in the Miombo woodland matrix, and available copper concentrations often exceed 10,000 mg Cu.kg -1 at the outcrop summit. Those copper hills support more than 550 metallophyte taxa (Leteinturier, 2002). Recent studies show that trace metals and related nutrient concentrations play a significant role to structure plant communities along the metal gradient (Saad et al., 2012; Ilunga wa Ilunga et al., 2013; Séleck, Bizoux, et al., 2013). In Katanga, copper hills represent a biodiversity hotspot (Malaisse, 1983; Brooks et al., 1985; Leteinturier et al., 2002; Faucon et al., 2010), including 56 endemic species in which 32 are exclusively restricted to copper enriched soils (absolute metallophytes) (Faucon et al., 2010). Some authors have shown that the edaphic endemism could result of both abiotic and biotic interactions (Meyer, 1986; Palacio et al., 2007). First, taxa may be restricted to a limited area because of their specialization to particular physicochemical soil conditions making them more competitive than in regular soils (Meyer, 1986). Second, taxa can be stress-tolerant, but not specifically adapted, to adverse physicochemical conditions, and take refuge in the stressful area where competition with other plant species is lowest 134

135 (Gankin et al., 1964). Plants of the Katanga (D.R.Congo) occurring on copper-rich soils remained poorly studied (Faucon, Chipeng, et al., 2012). To focus on endemic and nonendemic species within a genus appears interesting in order to understand the relationship between plant and copper concentrations in soil. The copper hills have been severely disturbed for more than one century through the mining of natural resources leading in some cases to their degradation or total destructions (Brooks et al., 1992; Faucon et al., 2010). The determination of the in vitro multiplication ability of some endemics appears useful in a purpose of ex situ conservation. In this paper, we aimed to test the effect of copper on the germination and the root elongation of Crotalaria cobalticola, Crotalaria peschiana and Crotalaria cornetii. The results also contribute to discuss the relationship between the endemic status of the three species and their affinity for copper. Material and methods We investigated both the germination and the root elongation in the first weeks according to different copper concentrations for three species of the genus Crotalaria (Fabaceae) from the Katangan Copperbelt. These species, Crotalaria cobalticola, Crotalaria peschiana and Crotalaria cornetii, were chosen according to their supposed copper tolerances (i.e. present in their habitat) and their endemism status in natural ecosystem; Crotalaria cobalticola, being the most tolerant species to strong copper concentrations and Crotalaria cornetii the less tolerant species. Table 1. General characteristics of the studied species. Habitat descriptions and life forms are from Duvigneaud et al. (1959), Duvigneaud et al. (1963) and Saad et al. (2012). Conservation status and proposed IUCN status are from Faucon et al. (2010). Fruiting time, fruit type, seed sizes are from field observations. Taxon Life form Habitat in copper hills Fruiting time Fruit type Seed size (mm) Proposed IUCN status Conservation status C. cobalticola Annual Highly Cu/Coenriched steppe April-May Dehiscent pod 3-4 EN Strict endemic C. cornetii Perennial Low Cu/Coenriched steppic savannah September- October Dehiscent pod 4-5 NA Non-endemic C. peschiana Perennial Cu/Co-enriched slope August- September (April-May) Dehiscent pod 3-4 CR Strict endemic CR: Critically endangered, EN: Endangered, NA: Not assessed 135

136 Plant species Crotalaria is a genus of the fabaceae family. The genus Crotalaria appears to be mainly adapted to the seed dispersal by wind, water and gravity (Le Roux et al., 2011). Crotalaria cobalticola P.A. Duvign. & Plancke is a strict endemic and a threatened species (proposed IUCN status: EN) present in the natural highest Cu-enriched steppes of the copper hills or in disturbed copper and cobalt soils in the Katangan Copperbelt (Figure 1, Table 1). This annual forb have a natural density of about 5 to 10 individual m -2 and was identified as an hyper accumulator of cobalt in concentrations higher than 3,000 mg kg -1 (Duvigneaud et al., 1959; Brooks et al., 1981). Crotalaria cornetii Taub. & Dewevre is a non-endemic species presenting a large distribution in the Southeastern D.R. Congo with a density of about 0.5 to 1 individual m -2. It occurs on the low-enriched substrate (Figure 1, Duvigneaud et al., 1963), located in the steppic savannahs of the copper hills (Table 1, Duvigneaud et al., 1959). To date, its IUCN status has not been assessed. Crotalaria peschiana P.A. Duvign. & Timp. is a strict endemic and a critically endangered species (proposed IUCN status: CR) occurring in the slopes of steppic savannah of the natural copper hills in the Katangan Copperbelt (Duvigneaud et al., 1959, Faucon et al., 2010). Its natural density is about 1 to 2 individuals m -2 (Figure 1, Table 1). Figure 1. Pictures of adult individuals of Crotalaria cobalticola, Crotalaria cornetii and Crotalaria peschiana - Fabaceae ( 136

137 Seed collection Seeds were harvested in populations occurring in the Southeastern D.R. Congo, in the Copperbelt, between the cities of Tenke (10.61 S; E) and Fungurume (10.62 S; E) during the year This region presents 40 isolated copper hills between 1300 m to 1400 m. The climate is subtropical with a rainy season from November to the end of March and a dry season from May to September. October and April are the transition months between the two seasons. Annual rainfalls are ca mm, most of which occurs during the rainy season. Temperatures range from 15 C to 17 C at the beginning of the dry season and from 31 C to 33 C between September and October. Dry season minima can reach 5 C at night (Saad et al., 2012). Seeds were collected on 72 individuals scattered in one population of C. cobalticola (ca individuals), on 75 individuals in one population of C. cornetii (ca individuals) and on 75 individuals in one population of C. peschiana (ca individuals). Among population, mature and well-developed individuals were targeted in order to obtain mature seeds. For the three species, fruits are dehiscent pods and seed sizes comprised between 3 and 5 mm. C. cornetii has been largest seeds (4-5 mm) (Table 1). For each species, seeds from the different individuals were pooled and seed samples were then placed in conservation. Considering that the three species have different conservation purposes (i.e. short-term conservation for C. cornetii and C.peschiana and long-term conservation for C cobalticola) they were stored in different sites during a maximum two years. Crotalaria cobalticola was stored in the Botanic Garden Meise (Belgium) at -20 C and 15% RH in order to apply long term conservation strategy (Godefroid et al., 2013). The two other species (C. cornetii and C. peschiana) were conserved at 5 C in the nursery of Tenke Fungurume Mining Sarl (D.R.Congo) in order to use their seeds in propagation programs (Table 2). Table 2. Location of harvested sites with collecting and conservation data of the seeds of the three studied species. Collecting data Conservation data Taxon Date Site Sampled ind. Date Method Site C. cobalticola 07/05/2013 Fungurume II 72 28/05/ C - 15% HR BGM C. cornetii 03/01/2013 Kavifwafwaulu I 75 02/01/ C - ambient HR TFM C. peschiana 24/09/2013 Kazinyanga 75 24/09/ C - ambient HR TFM BGM: Botanic Garden Meise, TFM: Tenke Fungurume Mining sarl 137

138 Germination tests Germination tests were performed in controlled conditions between October and December Mediums were prepared with agar (8 g l -1 ), Murashige and Skoog (MS)-vitamins (2.2 g l - 1 ), sucrose (10 g l -1 ), MES-buffer (0.5 g l -1 ) and copper sulfate solution (CuSO 4.5H 2 O 0.1M). Six copper mediums were prepared: 0 (the control), 25, 50, 75, 100 and 125 μm (i.e. μmol l -1 ) copper added, based on the results of Chipeng et al. (2010) and two other studies about nontolerant legumes (Di Salvatore et al., 2008; Lin, 2011). The ph was adjusted at 5.7 with HCl 0.5 M or NaOH 0.1 M. 50 ml of medium was put in each Petri dishes; each medium was previously sterilized in autoclave during 90 minutes. In order to limit the pathogen development, seeds were first treated with a bleach water solution (5%) during five minutes and rinsed three times with distilled water following by a five minutes soak in an ethanol solution (70%) (Boisson et al., In press). Seeds were then rinsed again three times in distilled water. Then, for each species and for each copper concentration, 90 seeds were placed in three Petri dishes (i.e. 30 seeds by petri dishes according to a systematic grid following six columns and five rows). Petri dishes were placed horizontally in a tropical growth chamber with photoperiod 12/12 at 25 C and 70% relative humidity at 450 μe light and 15 C and 50% relative humidity in the dark during 38 days (23 October 01 December). Germination was recorded when the radicle was at least 2 mm long. Germination was monitored twice a week during three weeks and once a week as soon as one germinated seed maximum was recorded by Petri dish. Root elongation tests In vitro elongation tests were performed for C. cobalticola and C. peschiana; the number of C. cornetii seedlings was too low in order to test the effect of copper on root elongation. The root elongation test mediums were the same than the germination mediums; 0 (the control), 25, 50, 75, 100 and 125 μm (i.e. μmol l -1 ) copper added. Seedlings were germinated on MS, and transplanted when cotyledons appeared after a treatment with bleach water (10%) during five minutes following by five rinsing baths in distilled water in order to keep it in sterile conditions. For each species and for each copper concentration, ten seedlings were transplanted in two Petri dishes (i.e. five seedlings by petri dishes) and dishes were installed vertically in the same conditions than for the germination tests during four weeks. The 138

139 distance (mm) between the crown and the apical point of the roots was measured at the end of the test. A first test demonstrated the independence of seedlings root elongation among the Petri dishes. For this purpose, seedlings were germinated in the non-enriched MS. Five seedlings were transplanted respectively in five Petri dishes while five seedlings were transplanted in one Petri dishes. The final distance (mm) between the crown and the apical point of the roots was also measured. Data analysis Germination percentages were calculated as the ratio between the number of germinated seeds and total number of seeds in a Petri dish. Cumulative germination curves were created in order to evaluate the global dynamic of the germination process. The mean germination time (MGT) was calculated as a proxy of the germination speed:, where was the beginning of the test (in days); was the time at test closure (in days); was the number of germinated seeds at the time ; was the period (in days) between and ; was the total number of germinated seeds (Ellis et al., 1981). The effect of copper concentration on both the final germination percentages and mean germination time (MGT) were analyzed by a one-way ANOVA. Effect of copper concentration on root elongation was assessed using one-way ANOVA. The independence test had confirmed the non-effect of the Petri dishes on the root elongation (one-way ANOVA C. cobalticola: F = 1.03, P-value > 0.05; one-way ANOVA C. peschiana: F = 3.59, P-value > 0.05). So, ten replications were considered for the analysis. However, some seedlings had died during the test. The number of replication was lower than 10 in two treatments for C. cobalticola (25 µm, n = 9; 75 µm, n = 5) and for one treatment for C. peschiana (50 µm, n = 9). For all analysis, the normality and heteroscedasticity was verified. Root elongation data for C. cobalticola were square-root transformed. All significant differences in analyses were followed by Tukey's range test (HSD) to structure the means. Statistical analyses were performed using the R statistical software (R Development Core Team, 2010). 139

140 Results Effect of copper on germination The final germination percentages were 34.3 ± 2.4% for C. cobalticola, 46.0 ± 2.8% for C. peschiana and 3.0 ± 0.5% for C. cornetii. Although no significant differences of the final germination percentage were highlighted between copper concentrations (Figure 2, Figure 3), some trends were observed on the cumulative germination curves. The cumulative germination percentage of C. cobalticola at 25 µm Cu 2+ stayed the lowest (22.0 ± 4.8% at the end of the experiment), while the cumulative germination percentage at 75 µm Cu 2+ was the highest from the day 6 (42.0 ± 6.8% at the end of the experiment). Crotalaria peschiana had the highest cumulative germination percentage at 25 µm Cu 2+ (63.3±6.7% at the end of the experiment). Crotalaria cornetii, which had overall low final germination percentages, had a highest final germination percentage (4.4 ± 1.1%) at low copper concentrations (0, 25 and 50 µm Cu 2+ ). However, no well-defined trend on the cumulative germination curves was observed (Figure 2). C. cobalticola C. cornetii 0 C. peschiana Copper concentrations (µm) Figure 2. Cumulative germination curves of the percentage of germination (%, n=3) of Crotalaria cobalticola, C. cornetii and C. peschiana after sowing of seeds in Petri dishes containing six concentrations of copper.0 µm (black squares), 25 µm (black dots), 50 µm (black triangles), 75 µm (white squares), 100 µm (white dots) and 125 µm (white triangles). Error bars correspond to standard error (SE). 140

141 No significant differences of the MGT were highlighted between copper concentrations for C. cobalticola, with a MGT of 7.3 ± 0.3 days (Figure 3) and for C. peschiana with a MGT equal to 10.8 ± 0.3 days (Figure 3). For C. cornetii, the MGT differed according to copper concentrations (F = 5.36, P < 0.01), seeds tended to germinate more rapidly when copper concentrations increased. The highest MGT was observed for 25 µm of copper (35.6 ± 2.3 days), although it was not significantly different from that at 0 and 50 µm Cu 2+ (27.8 ± 5.3 days). Its lowest MGT was at 75 µm Cu 2+ (3.3 ± 3.3 days) that was significantly different than the MGT at 25 µm Cu 2+ (Figure 3). Figure 3. Germination percentages (%, black dots) and mean germination time (days, white squares) of Crotalaria cobalticola, C. peschiana and C. cornetii after 38 days of experiment according to six copper concentration (µm). Mean germination time was calculated by MGT =, where defines the begin of the test; is the time that closes the test; is the number of germinated seeds at the time ; is the period (in days) between and ; is the total number of germinated seeds. Error bars correspond to standard error (SE). The different letters indicate significant (P < 0.05) different means after ANOVA and Tukey s range test (error 5%). Effect of copper on root elongation Crotalaria cobalticola had a higher root elongation at 125 µm Cu 2+ (150 ± 41mm) compared to the others copper concentrations (F = 6.31, P < 0.001, Figure 4). Crotalaria peschiana presented an equal elongation root uneven the copper concentrations (25 ± 1mm, Figure 4). 141

142 Figure 4. Copper concentrations effects on the root elongation of Crotalaria cobalticola and C. peschiana after 28 days of exposition to solutions containing 0, 25, 50, 75, 100 and 125 µm Cu 2+. Data are presented as the mean root elongation (mm ± SE) of 10 seedlings for each concentration but 25 µm (n = 9) and 75 µm (n = 5) for C. cobalticola and 50 µm (n = 9) for C. peschiana. Dead seedlings were removed from the analysis. Error bars correspond to standard error (SE). Letters above dots indicate significant (P < 0.05) differences after ANOVA and Tukey s range test. Means sharing the Discussion Germination and early establishment The three tropical metallophyte species, C. cobalticola, C. peschiana and C. cornetii, germinated under all copper concentrations, without a clear effect of copper concentrations on the germination (i.e., germination percentage and mean germination time). This pattern was already pointed out for another copper endemic species Diplolophium marthozianum (Apiaceae) (Boisson et al., In press). Chipeng et al. (2010) also showed that copper addition did not significantly improve the germination of the high copper tolerant Haumaniastrum katangense (Lamiaceae). On the other hand, Crotalaria cornetii seems more sensitive to higher copper concentration, since this species showed lower mean germination time at the highest copper concentrations. Seed coats can constitute a barrier between the embryo and its direct environment or adsorb metals as reported by previous studies (Mayer et al., 1989; Araújo et al., 2005; Di Salvatore et al., 2008). Seeds of Fabaceae are known for their physical dormancy, where an impermeable seed-coat prevents the entry of water to the embryo (Bewley et al., 1994). The low germination percentages recorded for C. cornetii (< 5%) may indicate a dormancy of seeds; scarification with sandpaper or soaking in hot water permitted to increase the percentage of germination at 56% (Le Stradic, personal communication). 142

143 The genus Crotalaria permitted to study two endemics that have distinct life histories. Crotalaria cobalticola is an annual species while C. peschiana is a perennial. Due to the shorter life cycle, annual have to growth rapidly in order to reproduce (Garnier, 1992a). While the percentages of germination were higher for the perennial C. peschiana, the MGT was also longer and could be partially the consequence of the perennial behavior of the species. Crotalaria cobalticola germinates in November and have to produce seeds before the dry season (May). This could explain that the root elongation was higher for C. cobalticola. These two endemic species from the Katangan Copperbelt are classified in the same clade Dispermae contrary to C. cornetii which is in Glaucae (Le Roux, personal communication). This phylogenetic proximity supports the higher supposed copper tolerance of C. cobalticola and C. peschiana compared to C. cornetii. As expected, C. cobalticola, which is a strict endemic of copper hills, was the most copper tolerant species at the early stages of growth (Duvigneaud & Timperman 1959). Root growth for C. cobalticola was stimulated by highest copper concentration, favoring a rapid development of the individuals which was not the case for C. peschiana at least for the copper concentrations used. Root growth was already reported to be more sensitive than germination to metal toxicity (Jeliazkova et al., 1998; Araújo et al., 2005). Growth stimulation by elevated concentrations of copper has rarely been reported (Faucon, Chipeng, et al., 2012). Rapid development of individuals confers an advantage in terms of competition in the highest copper-enriched areas. This high tolerance capacity could also lead to a physiological cost, reducing the fitness and competitive ability of individuals (Cox et al., 1981; Bradshaw, 1984). These assumptions support the hypothesis that this species could not grow in regular soils of the savannahs surrounding the hills where copper content is low ( 100 mg kg -1 Cu) (Duvigneaud et al., 1959; Saad et al., 2012). Additional studies should focus on a broader range of metal concentrations in order to obtain the total copper response curve of the species. Considering that Duvigneaud & Timperman (1959) describe it as a species highly dependent on cobalt in soils, the identification of its response to cobalt would also be interesting in terms of comparison with copper response. Crotalaria peschiana was able to grow in all copper concentrations without expressing any quantified stress. Contrary to C. cobalticola, copper seems not to be associated with a physiological need for its growth in the early growth stage; it did not limit its growth either. Crotalaria peschiana is not recorded in non-metaliferous substrate in the Katangan province 143

144 (Faucon et al., 2010), which imply that this stress-tolerant endemic species might not be limited by edaphic conditions, but is likely excluded from the dominant vegetation by biotic interactions as competition (Gankin et al., 1964) or herbivory. The less productive ability in regular soils lead this species in refuge areas (i.e. metalliferous substrates) in which competition is lower (McNeilly et al., 1968). Copper is also known as a broad-spectrum fungicide. Copper-rich soils may have lower pathogen content in comparison to regular soils, conducting to a loss of pathogens defense (Tadros, 1957; Chipeng et al., 2010). Some studies has pointed out the loss of resistance of some copper tolerant species to pathogens (Malaisse et al., 1982; Paton et al., 1996; Chipeng et al., 2010; Faucon, Chipeng, et al., 2012). According to the elemental defense hypothesis, metal accumulation in the apoplast can inhibit growth of some bacterial pathogens (Fones et al., 2010). On the copper hills, C. peschiana are absent in the steppe where copper concentrations exceed 4,000 mg kg soil -1 Cu, suggesting that this species tolerate lower copper concentration in soil than C. cobalticola (Séleck, Bizoux, et al., 2013). Additional studies are required to evaluate the growth response of the C. peschiana and C. cobalticola in higher copper concentrations than 125 µm in order to highlight their upper tolerance limits. Conservation of copper endemics and revegetation In a conservation effort, the two endemic and threatened species C. cobalticola and C. peschiana could be put in germination in non-contaminated mediums, like C. cornetii. The ex situ conservation should allow to export and preserve the plant material outside its natural habitat. It is interesting to germinate seeds in vitro before transplanting it in the nursery to keep a maximum of individuals. Then, a non-copper enriched medium would be sufficient to grow these two species during their installation stages which should facilitate the multiplication of such species. However, more studies should be performed to determine the other factors (i.e. origin of population, biotic interactions or dormancy) which might influence their growth. As the response could be different between populations due to local adaptations, conservation actions should focus on several sites (Faucon, Chipeng, et al., 2012; Godefroid et al., 2013; Ilunga wa Ilunga et al., 2013). In addition, physiological plant response to available metals is related to other with all chemical edaphic conditions (Austin, 2007; Soberón et al., 2009), which are not reproduced in our experiments. Similarly, bioavailability of copper and cobalt depends on the total edaphic metal content, ph, clay and hydrous oxide content, organic matter and redox 144

145 conditions (Reichman, 2002; Saad et al., 2012; Séleck, Bizoux, et al., 2013). So, to define a clear relationship between the availability of copper in agar and soils is still difficult (McLean et al., 1992). The determination of the copper and cobalt bioavailability by mathematical models is currently performed in order to study the ecological response to metals of some endemic species and are necessary to improve conservation strategies (Lange et al., 2014). These results could also contribute to the implementation of other strategies, as revegetation or phytoremediation. The knowledge about ecology of metallophyte endemics is the basis to associate phytotechnologies and ex situ conservation strategies. Secondary habitats created by the mining exploitation are opportunities to preserve part of the threatened metallophyte and endemic species (Whiting et al., 2004; Faucon et al., 2011). These secondary metalliferous habitats present often high physico-chemical heterogeneity in metal concentrations and species having a broad response to metals should be preferred. While restoration of mining sites has to be performed after exploitation, ex situ conservation of these metallophytes and endemic species in favorable sites should be a priority, and these secondary habitats created by the mining can be an opportunity to preserve this biological resource. References Antonovics J., Bradshaw A.D. & Turner R.G., Heavy metal tolerance in plants. Adv. Ecol. Res. 7, Araújo A.S.F. & Monteiro R.T.R., Plant bioassays to assess toxicity of textile sludge compost. Sci. Agric. 62(3), Austin M., Species distribution models and ecological theory: A critical assessment and some possible new approaches. Ecol. Modell. 200(1-2), Baker A.J.M., Ernst W.H.O., Van der Ent A., Malaisse F. & Ginocchio R., Metallophytes: the unique biological resource, its ecology and conservational status in Europe, central Africa and Latin America. In: Batty, L., Hallberg, K. eds. Ecology of Industrial Pollution Bewley J.D. & Black M., Seeds: physiology of development and germination, Language of science, Plenum Press, New York, 445. Boisson S., Ortmans W., Maréchal J., Majerus M. & Mahy G., No copper required for germination of an endangered endemic species from the Katangan Copperbelt (Katanga, DR Congo): Diplolophium marthozianum. Trop. Ecol. Bradshaw A.D., Adaptation of plants to soils containing toxic metals - a test for conceit. In: Evered, D., Collins, G.M. eds. Origins and Development of Adaptation, CIBA Foundation Symposium 102. Pitman, London, Brady K.U., Kruckeberg A.R. & Bradshaw Jr. H.D., Evolutionary Ecology of Plant Adaptation To Serpentine Soils. Annu. Rev. Ecol. Evol. Syst. 36(1), Brooks A., Collins J.C. & Thurman D.A., The mechanism of zinc tolerance in grasses. J. Plant Nutr. 3(1-4), Brooks R.R., Baker A.J.M. & Malaisse F., Copper flowers. Res. Explor. 8(3), Brooks R.R. & Malaisse F., The Heavy Metal Tolerant Flora of Southcentral Africa: a mutlidisciplinary approach, A.A. Balkema, Rotterdam, 199. Cailteux J.L.H., Kampunzu A.B., Lerouge C., Kaputo A.K. & Milesi J.P., Genesis of sediment-hosted stratiform copper cobalt deposits, central African Copperbelt. J. African Earth Sci. 42(1-5), Chipeng F.K., Hermans C., Colinet G., Faucon M.-P., Ngongo M., Meerts P. & Verbruggen N., Copper tolerance in the cuprophyte Haumaniastrum katangense (S. Moore) P.A. Duvign. & Plancke. Plant Soil 328(1-2), Cox R.M. & Hutchinson T.C., Multiple and cotolerance to metals in the grass Deschampsia cespitosa: 145

146 adaptation, preadaptation and cost. J. Plant Nutr. 3, Di Salvatore M., Carafa A. & Carratù G., Assessment of heavy metals phytotoxicity using seed germination and root elongation tests: a comparison of two growth substrates. Chemosphere 73(9), Duvigneaud P. & Denaeyer-De Smet S., Etudes sur la végétation du Katanga et de ses sols métallifères. Communication n 7 Cuivre et végétation au Katanga. Bull. Soc. Roy. Bot. Belgique 96(2), Duvigneaud P. & Timperman J., Etudes sur la végétation du Katanga et de ses sols métallifères. Communication n 3 Etudes sur le genre Crotalaria. Bull. Soc. Roy. Bot. Belgique 91(2), Ellis R.H. & Roberts E.H., The quantification of ageing and survival in orthodox seeds. Seed Sci. Technol. 9, Ernst W.H.O., Schwermetallvegetation der Erde. G. Fisher ed., Stuttgart, 194. Faucon M.-P., Chipeng F., Verbruggen N., Mahy G., Colinet G., Shutcha M., Pourret O. & Meerts P., Copper tolerance and accumulation in two cuprophytes of South Central Africa: Crepidorhopalon perennis and C. tenuis (Linderniaceae). Environ. Exp. Bot. 84, Faucon M.-P., Meersseman A., Shutcha M.N., Mahy G., Luhembwe M.N., Malaisse F. & Meerts P., Copper endemism in the Congolese flora: a database of copper affinity and conservational value of cuprophytes. Plant Ecol. Evol. 143(1), Faucon M.-P., Parmentier I., Colinet G., Mahy G., Ngongo Luhembwe M. & Meerts P., May Rare Metallophytes Benefit from Disturbed Soils Following Mining Activity? The Case of the Crepidorhopalon tenuis in Katanga (D. R. Congo). Restor. Ecol. 19(3), Fones H., Davis C. a R., Rico A., Fang F., Smith J.A.C. & Preston G.M., Metal hyperaccumulation armors plants against disease. PLoS Pathog. 6(9), François A., Synthèse géologique sur l Arc cuprifère du Shaba (Rép.du Zaïre). Société belge de Géologie Centenaire(1987), Gankin R. & Major J., Arctostaphylos myrtifolia, its biology and relationship to the problem of endemism. Ecology 45(4), Garnier E., Growth analysis of congeneric annual and perennial grass species. J. Ecol. 80(4), Godefroid S., Van de Vyver A., Massengo Kalenga W., Handjila Minengo G., Rose C., Ngongo Luhembwe M., Vanderborght T. & Mahy G., Germination capacity and seed storage behaviour of threatened metallophytes from the Katanga copper belt (DR Congo): implications for ex situ conservation. Plant Ecol. Evol. 146(2), Harrison S.P. & Rajakaruna N., Serpentine: The Evolution and Ecology of a Model System. University of California Press ed., University of California Press, Berkeley, 464. Ilunga wa Ilunga E., Séleck M., Colinet G., Meerts P. & Mahy G., Small-scale diversity of plant communities and distribution of species niches on a copper rock outcrop in Upper Katanga, DR Congo. Plant Ecol. Evol. 146(2), Jeliazkova E.A., Jeliazkov V.D., Craker L.E. & Xing B., Heavy Metals and Seed Germination in Medicinal and Aromatic Plants. HortScience 33(2), 206. Kruckeberg A.R., California Serpentines: Flora, Vegetation, Geology, Soils, and Management Problems, University of California Press, Berkeley, 196. Lange B., Faucon M.-P., Meerts P., Shutcha M., Mahy G. & Pourret O., Prediction of the edaphic factors influence upon the copper and cobalt accumulation in two metallophytes using copper and cobalt speciation in soils. Plant Soil 379(1-2), Le Roux M.M., Van Wyk B.E., Boatwright J.S. & Tilney P.M., The systematic significance of morphological and anatomical variation in fruits of Crotalaria and related genera of tribe Crotalarieae (Fabaceae). Bot. J. Linn. Soc. 165(1), Leteinturier B., Evaluation du potential phytocénotique des gisements cupriferes d Afrique centro-australe en vue de la phytoremédiation de sites pollués par l'activité. Fac. des Sci. Agron. Gembloux. Leteinturier B. & Malaisse F., On the tracks of botanical collectors on copper outcrops of South Central Africa (in French). Syst. Geogr. Plants 71, Lin Y., Effects of Copper Ion on Seeds Germination of Cichorium intybus L.Determined by Agar Plate Method. J. Hebei Agric. Sci. 10. Malaisse F., Phytogeography of the copper and cobalt flora of Upper Shaba (Zaïre), with emphasis on its endemism, origin and evolution mechanisms. Bothalia 14, Malaisse F., Brooks R.R. & North P., Colonisation of modified metalliferous environments in Zaire by the copper flower Haumaniastrum katangense. Plant Soil 64, Mayer A.M. & Poljakoff-Mayber A., The Germination of Seeds. Pergamon, Press eds., New York. McLean J. & Bledsoe B.E., Behavior of metals in soils. Gr. Water Issue. McNeilly T. & Bradshaw A.D., Evolutionary processes in populations of copper tolerant Agrostis tenuis Sibth. Evolution (N. Y). 22(1), Meyer S., The ecology of gypsophile endemism in the eastern Mojave Desert. Ecology 67(5), Palacio S., Escudero A., Montserrat-Martí G., Maestro M., Milla R. & Albert M.J., Plants living on gypsum: beyond the specialist model. Ann. Bot. 99(2), Paton A. & Brooks R.R., A re-evaluation of Haumaniastrum species as geobotanical indicators of copper and cobalt. J. Geochemical Explor. 56, R Development Core Team, A language and environment for statistical computing. R Foundation for Statistical Computing ed., Vienna (Austria). Rajakaruna N., The edaphic factor in the origin of species. Int. Geol. Rev. 46,

147 Reichman S., The Responses of Plants to Metal Toxicity: A Review Focusing on Copper, Manganese & Zinc, Australian Minerals & Energy Environment Foundation, Saad L., Parmentier I., Colinet G., Malaisse F., Faucon M.- P., Meerts P. & Mahy G., Investigating the Vegetation-Soil Relationships on the Copper-Cobalt Rock Outcrops of Katanga (D. R. Congo), an Essential Step in a Biodiversity Conservation Plan. Restor. Ecol. 20(3), Séleck M., Bizoux J.-P., Colinet G., Faucon M.-P., Guillaume A., Meerts P., Piqueray J. & Mahy G., Chemical soil factors influencing plant assemblages along coppercobalt gradients: implications for conservation and restoration. Plant Soil 373(1/2), Soberón J. & Nakamura M., Niches and distributional areas: concepts, methods, and assumptions. Proc. Natl. Acad. Sci. 106(2), Tadros T.T.M., Evidence of the presence of an edapho-biotic factor in the problem of serpentine tolerance. Ecology 38(1), Veldman J.W., Overbeck G.E., Negreiros D., Mahy G., Le Stradic S., Fernandes G.W., Durigan G., Buisson E., Putz F.E. & Bond W.J., Tyranny of trees in grassy biomes. Science (80-. ). 347(6221), Whiting S.N., Reeves R.D., Richards D., Johnson M.S., Cooke J.A., Malaisse F., Johns R., Mcintyre T., Purvis O.W., Salt D.E., Schat H., Zhao F.J. & Baker A.J.M., Research Priorities for Conservation of Metallophyte Biodiversity and their Potential for Restoration and Site Remediation. Restor. Ecol. 12(1),

148

149 Axe 1 -Article 5 Accepté - No copper required for germination of an endangered endemic species from the Katangan Copperbelt (Katanga, DR Congo): Diplolophium marthozianum BOISSON Sylvain, ORTMANS William, MARÉCHAL Justine, MAJÉRUS Marie, MAHY Grégory & MONTY Arnaud Abstract Two hypotheses were tested with respect to the germination of Diplolophium marthozianum, an endemic plant species of the copper-cobalt outcrops in Katanga, Democratic Republic of Congo: (1) germination of copper-endemics is limited by fungal infection in the absence of pathogen control and (2) the germination success of this copper-endemic species is copperdependent. Seed lots of twenty seeds were weighed, soaked in different disinfection treatments and then placed in a germination medium containing four distinct copper concentrations for 30 days. Seed viability was measured at the beginning and at the end of the experiment by a cut test. Final germination percentage (15.2 ± 8.2%) and seed viability (24.2±10.3%) were not affected by copper concentration or disinfection treatments. D. marthozianum is able to germinate in a substrate without added copper, despite pervasive fungal infection. However, seed mass had a significant positive effect on seed germination suggesting that selecting the largest seeds may ensure the highest germination success in ex situ conservation programs. 149

150 Introduction The Global Strategy for Plant Conservation aims to halt the continuing loss of plant diversity. Understanding, documenting and developing conservation strategies for plant species are key issues that need to be urgently addressed as international responsibilities (UNEP, 2007). Biodiversity conservation priorities should focus especially on tropical regions, which include 90% of extant species (Myers et al., 2000). Rare endemic taxa are considered priorities for conservation programs when threats are identified (Schemske et al., 1994; Primack, 2010). Metalliferous areas (i.e. soils with elevated concentrations of trace metals) often host rare ecologically endemic taxa due to the severe selection pressures induced by trace metals, founder effects and genetic drift induced by the insularity of these habitats (Baker et al., 2010). Plants that are distributed only on metal-rich soils are called absolute metallophytes (Pollard et al., 2002). Many of these species are threatened by mining, which leads to degradation or even total destruction of their habitats, increasing extinction risks (Whiting et al., 2002; Saad et al., 2012). The Katangan Copperbelt (Southeastern Democratic Republic of Congo) is formed by more than 150 copper-cobalt outcrops, which are globally unique (Duvigneaud et al., 1963a). These hills are geographically and ecologically isolated and have a unique flora that includes over 600 species, including 32 taxa that are strict endemic species (i.e. occurring solely on Cu-rich soils) or absolute cuprophytes (e.g. Crepidorhopalon perennis), and 24 taxa that are broad endemic species (i.e. more than 75% occurrence on Curich soils) (Faucon et al., 2010). Many copper-endemic species are highly threatened by habitat destruction caused by industrial mining (Faucon et al., 2010). Many factors influence the success of current ex situ conservation efforts focused on reproduction (Di Salvatore et al., 2008; Godefroid et al., 2013). Chipeng et al. (2010) showed that copper concentrations or the disinfection of seeds prior to germination trials may improve germination percentage for the absolute cuprophyte Haumaniastrum katangense (S.Moore) P.A.Duvign. & Plancke. The regeneration of metallophytes poses particular problems because of their limitation to highly metal-enriched substrates. This limitation may be due to various factors. First, these species may have higher copper physiological needs than other species (Faucon, Chipeng, et al., 2012; Peng et al., 2012). Second, copper may limit the development of pathogens, e.g. by acting as a natural fungicide (Antonovics et al., 150

151 1971); thus, metal hyperaccumulation can provide an effective form of defence (called elemental defence) against a wide range of pathogens (Boyd, 2007; Fones et al., 2010). From an evolutionary point of view, the metallophytes may have therefore lost their pathogen-resistance mechanisms (Tadros, 1957). Third, these species may also be limited to metalliferous soils because of inter-specific competition; e.g., endemics could be less competitive in other environments but able to grow on non-metal-enriched soils in the absence of competition (i.e. edaphic endemism, Gankin & Major, 1964). In order to develop a rapid conservation strategy and maintain maximum genetic diversity, the scientific community seeks to reproduce copper-endemics from seed collections rather than by cloning methods. However, there is still no standardized protocol for managing seed lots of these species from the wild, and it is unclear which disinfection and other specific treatments have to be applied to seeds in order to enhance their germination. The aim of this study is to provide guidelines for ex situ conservation of endemic species from copper-rich ecosystems. The first critical process of a plant propagation strategy is seed germination; therefore, this study focuses on germination. Two hypotheses regarding seed germination of an endemic species from copper ecosystems, D. marthozianum, were tested. First, germination of this copper-endemic species is limited by fungal attack in the absence of pathogen control by copper. Second, the germination success of this species is copperdependent. In addition, the effect of seed mass on germination was also examined to determine if seeds should be sorted by size prior to germination in conservation programs. Material and methods Diplolophium marthozianum P.A. Duvign. (Apiaceae) is a perennial cuprophyte and a broad endemic species with a high affinity for Cu Co rich substrates in Katanga (Faucon et al., 2010; Saad et al., 2012). Its proposed classification for the IUCN Red List of Threatened Species is endangered (EN; Faucon et al., 2010). Like many plants of the Apiaceae family, seed teguments have microcavities on their surface, where fungal pathogens can shelter (Bulajic et al., 2009). The achenes of D. marthozianum (henceforth referred to as seeds) used in the present experiments were collected on three sites in the Katangan Copperbelt between Tenke (10.61 S; E) and Fungurume (10.62 S; E) towns; one collection was made in July 2011 at Mwinansefu, another in July 2011 at Fungurume IV and the third 151

152 seed lot was collected in September 2011 at Kavifwafwaulu. Seeds were stored in paper bags at room temperature, in Gembloux, Belgium, for 5 to 15 weeks before germination tests. Germination tests were performed in Petri dishes under sterile conditions on a culture medium consisting of agar (0.8%) (Kumari et al., 2000). Copper in the form of copper (II) sulphate (CuSO 4 ) was used. Previous studies on seed germination of non-metallophyte species have shown that the phytotoxicity of copper appears between 2.5 and 10 mg Cu.l -1, whereas the absolute cuprophyte H. katangense (which occurs in the Katangan Copperbelt) tolerates more than 10 mg Cu.l -1 (Di Salvatore et al., 2008; Chipeng et al., 2010). Based on these previous findings, four Cu concentrations were added to the medium: 0 mg l -1 Cu (control), 5 mg l -1 Cu, 10 mg l -1 Cu and 20 mg l -1 Cu. After the copper (II) sulphate addition, the ph of the medium across treatments was 5.4 ± 0.1. Due to the small population size and the IUCN status of the species, the three collected populations (ca seeds per population) were pooled before the experiment. Seed lots of 20 seeds were weighed to the nearest 0.1 mg and received one of three disinfection treatments: bleach, fungicide or sterile water. The first was based on a widely used protocol of seed surface disinfection using 1.6% sodium hypochlorite (Sauer et al., 1986). The second used a broad-spectrum anti-fungal agent (ROVRAL SC, 2 % - BASF), and the third was the control (no disinfection treatment). Seeds were soaked for 5 minutes in either disinfection treatment followed by three rinses of 5 seconds in three different vessels of sterile water. Finally, seed lots were placed in Petri dishes having distinct copper concentrations. Ten replications were performed for each treatment combination (copper concentrations disinfection) resulting in 120 dishes and 2400 seeds. The dishes were wrapped with Parafilm and placed in an incubator at 25 C with a 12 hour light/12 hour dark photoperiod. This corresponds to the natural conditions at the beginning of the growing season in Katanga. Photosynthetic photon flux density (PPFD) was set at 30 µmol m -2 s -1 using cool white fluorescent lamps (Philips TL 8W/33-640). This standard intensity used for germination tests is lower than the intensity in natural tropical areas. Germination was recorded and Petri dishes were randomized three times a week. Seeds were considered germinated when the radicle emerged. Seed viability was measured with two cut tests. An initial cut test was performed on 65 seeds from the initial seed sample. A second cut test was performed on a subset of 10 non- 152

153 germinated seeds randomly chosen from each treatment combination (4 copper concentrations 3 disinfection treatments, for a total of 120 seeds) at the end of the experiment. The experiment duration was 30 days, based on mean germination time (MGT) observed by Godefroid et al. (2013) for this copper-endemic species. A two-way ANCOVA was performed to compare the germination percentage between copper concentrations and disinfection treatments (as fixed factors) using seed mass (i.e. mass of lots of 20 seeds) as a covariate. To compare the percentage of seed viability, two separate one-way ANOVAs were performed among the copper concentrations or the disinfection treatments. Statistical analyses were performed using the statistical software R (R Development Core Team, 2010). Results An average final germination percentage of 15.2 ± 8.2% (Mean % ± SD) was obtained. No significant effect of either fixed factor on final germination percentage was found (Fig. 1). The cumulative germination curves showed a similar germination rate over time among copper concentrations and disinfection treatments (Fig. 2). Fig. 1. Final percentage of germination (mean standard deviation, %) of the species D. marthozianum under different copper concentration and disinfection treatments. 153

154 The statistical analyses showed a significant effect of the covariate (seed mass) on final germination percentage (F 1, 113 = 23.99, P < 0.001), and a positive correlation between these two variables (r = 0.422, P < 0.001, N = 120). Heavier seed lots showed higher germination percentages. During the experiment, infection by fungi was observed on the surface of seeds, and more than 90% of seeds were affected regardless of the disinfection treatment or the copper concentration. The cut test performed prior to germination showed that approximately half of the seeds (46.1%) contained a viable embryo. The second cut test carried out on non-germinated seeds showed that the viability of seeds was unaffected by disinfection treatments or copper concentrations (average viability of non-germinated seeds: 24.2 ± 10.8%). Fig. 2. Cumulative germination percentage of the species D. marthozianum under different copper concentration and disinfection treatments during 30 days. Discussion A previous study on D. marthozianum showed that the percentage of germination was around 37.4 ± 35.0% (Mean % ± SD) in non-copper-enriched substrate at 22 C and under similar light intensity (Godefroid et al., 2013). However, this experiment used only viable seeds, whereas we did not sort viable seeds from seed lots. If we consider the seed lot viability of 46.1%, the germination percentage of fully developed seeds would be 32.9 ± 17.6%. In both cases, the germination percentage was similar and presented a high standard deviation that could be due to population variation, since Godefroid et al. (2013) also used three populations from different sites but found lower seed viability. 154

155 Our results did not support the first hypothesis that the germination of D. marthozianum is limited by fungal attack in the absence of pathogen control by copper. The addition of copper in the medium did not hinder the growth of fungi, and fungi did not affect the germination percentage of D. marthozianum. In contrast, Haumanisatrum katangense has been identified as highly susceptible to pathogen attacks in non-copper-enriched soils (Chipeng et al., 2010). As copper is known as a mineral anti-fungal agent (Carlile et al., 2001), the presence of high levels of cupric ions in the soil may promote the selection of fungi possessing genetic determinants for copper resistance, mainly controlled by the cell wall (Cervantes et al., 1994). The presence of pathogens in all media suggests that they were from the seed surface and microcavities of D. marthozianum. This form of fungi could be classified as seedborne fungi that are dependent on seeds as a host for their dispersal and development (Carlile et al., 2001). The hypothesis that germination of D. marthozianum requires high copper concentrations was not supported, since no significant differences in germination percentage were found among copper concentrations. However, either high copper requirements or low tolerance could occur at other developmental stages. Seedlings have been suspected to be more sensitive to metals than mature plants for some species (Street et al., 2007). Also, the exclusion of copper endemics from the dominant vegetation in soils that are not enriched in copper- (i.e. edaphic endemism) could be expected. This species response to copper in more advanced growth stages should be studied. Seed mass was an important source of phenotypic variation among D. marthozianum that proportionally affected the germination independently of the treatments (i.e. copper concentrations or disinfection). This result suggests that maternal effects can play a significant role in germination dynamics and subsequent seedling traits, as already documented for other species (Roach et al., 1987; Monty et al., 2013). These results have direct implications for conservation strategies of this species regarding the selection of seeds for propagation in a nursery. From a conservation perspective, our results indicate that the copper endemic D. marthozianum is not copper-dependent and that, although it is susceptible to fungal attack, this does not affect seed germination. Thus, disinfection of seeds does not seem to be a crucial step for the propagation of this species. In contrast, the selection of the largest seeds 155

156 may ensure the highest germination success in a nursery. Apart from germination, future research should consider the growth and the survival of seedlings of D. marthozianum, which may be more sensitive to copper and to pathogens than seeds. A better understanding of the effect of copper concentration on the germination and development of copper endemic species is crucial for their ex situ conservation. Acknowledgements The authors would like to thank Julie Lebrun and Maxime Séleck, who collected the seeds used in the present study, and Cynthia Thomas for her help during the experimental manipulations. Sandrine Godefroid contributed to the implementation of the experiment and provided information about the species. This study was funded by the Fonds de la Recherche dans l Industrie et dans l Agriculture (FRIA) of the FNRS (Fonds National de la Recherche Scientifique, Belgium) and under FNRS Grant n References Antonovics J., Bradshaw A.D. & Turner R.G., Heavy metal tolerance in plants. Advances in ecological research 7: Baker A.J.M., Ernst W.H.O., Van der ent A., Malaisse F. & Ginocchio R., Metallophytes: the unique biological resource, its ecology and conservational status in Europe, central Africa and Latin America. In: Batty, L., Hallberg, K. eds. Ecology of Industrial Pollution Boyd R.S., The defense hypothesis of elemental hyperaccumulation: status, challenges and new directions. Plant and Soil 293: Bulajic A., Djekic I., Lakic N. & Krstic B., The presence of Alternaria spp. on the seed of Apiaceae plants and their influence on seed emergence. Archives of Biological Sciences 61: Carlile M.J., Watkinson S.C. & Gooday G.W., The Fungi. Academic Press. Cervantes C. & Gutiérrez-Corona F., Copper resistance mechanism in bacteria and fungi. FEMS Microbiol. Rev. 14: Chipeng F.K., Hermans C., Colinet G., Faucon M.-P., Ngongo M., Meerts P. & Verbruggen N., Copper tolerance in the cuprophyte Haumaniastrum katangense (S. Moore) P.A. Duvign. & Plancke. Plant and Soil 328: Di Salvatore M., Carafa A. & Carratù G., Assessment of heavy metals phytotoxicity using seed germination and root elongation tests: a comparison of two growth substrates. Chemosphere 73: Duvigneaud P. & Denaeyer-De Smet S., Etudes sur la végétation du Katanga et de ses sols métallifères. Communication n 7 Cuivre et végétation au Katanga. Bulletin de la Société Royale de Botanique de Belgique 96: Faucon M.-P., Chipeng F., Verbruggen N., Mahy G., Colinet G., Shutcha M., Pourret O. & Meerts P., Copper tolerance and accumulation in two cuprophytes of South Central Africa: Crepidorhopalon perennis and C. tenuis (Linderniaceae). Environmental and Experimental Botany 84: Faucon M.-P., Meersseman A., Shutcha M.N., Mahy G., Luhembwe M.N., Malaisse F. & Meerts P., Copper endemism in the Congolese flora: a database of copper affinity and conservational value of cuprophytes. Plant Ecology and Evolution 143: Fones H., Davis C. a R., Rico A., Fang F., Smith J.A.C. & Preston G.M., Metal hyperaccumulation armors plants against disease. PLoS pathogens 6: Gankin R. & Major J., Arctostaphylos myrtifolia, its biology and relationship to the problem of endemism. Ecology 45: Godefroid S., Van de Vyver A., Massengo Kalenga W., Handjila Minengo G., Rose C., Ngongo Luhembwe M., Vanderborght T. & Mahy G., Germination capacity and seed storage behaviour of threatened metallophytes from the Katanga copper belt (DR Congo): implications for ex situ conservation. Plant Ecology and Evolution 146: Kumari S. & Ichhpujani R., Guidelines on standard operating procedures for Microbiology. World Health Organization. 156

157 Monty A., Bizoux J.P., Escarré J. & Mahy G., Rapid Plant Invasion in Distinct Climates Involves Different Sources of Phenotypic Variation. PLoS ONE 8. Myers N., Mittermeier R. a, Mittermeier C.G., da Fonseca G. a & Kent J., Biodiversity hotspots for conservation priorities. Nature 403: Peng H., Wang-Müller Q., Witt T., Malaisse F. & Küpper H., Differences in copper accumulation and copper stress between eight populations of Haumaniastrum katangense. Environmental and Experimental Botany 79: Pollard A.J., Powell K.D., Harper F.A. & Smith J.A.C., The Genetic Basis of Metal Hyperaccumulation in Plants. Critical Reviews in Plant Sciences 21: Primack R.B., Essentials of Conservation Biology. Sinauer Associates, Incorporated. R Development Core Team, A language and environment for statistical computing. Vienna (Austria). Roach D. & Wulf R.D., Maternal Effects In Plants. Annual Review of Ecology and Systematics. Saad L., Parmentier I., Colinet G., Malaisse F., Faucon M.- P., Meerts P. & Mahy G., Investigating the Vegetation-Soil Relationships on the Copper-Cobalt Rock Outcrops of Katanga (D. R. Congo), an Essential Step in a Biodiversity Conservation Plan. Restoration Ecology 20: Sauer D. & Burroughs R., Disinfection of seed surfaces with sodium hypochlorite. Phytopathology 76: Schemske D.W., Husband B.C., Ruckelshaus M.H., Goodwillie C., Parker I.M. & Bishop J.G., Evaluating Approaches to the Conservation of Rare and Endangered Plants. Ecological Society of America 75: Street R.A., Kulkarni M.G., Stirk W.A., Southway C. & Van Staden J., Toxicity of metal elements on germination and seedling growth of widely used medicinal plants belonging to Hyacinthaceae. Bulletin of environmental contamination and toxicology 79: Tadros T.T.M., Evidence of the presence of an edapho-biotic factor in the problem of serpentine tolerance. Ecology 38: UNEP, Annual Report. Whiting S.N., Reeves R.D. & Baker A.J.M., Conserving biodiversity: Mining, metallophytes and land reclamation. Mining Environmental Management 10:

158

159 Axe 1 -Article 6 Soumis - Specialist or refuge model for endemic metallophytes living in south-eastern D.R. Congo: implications for ex situ conservation BOISSON Sylvain, LANGE Bastien, GARIN Olivier, TSHOMBA WETSHY Axel, SÉLECK Maxime, LE STRADIC Soizig, NGOY SHUTCHA Mylor, VERBRUGGEN Nathalie, MAHY Grégory & FAUCON Michel-Pierre Abstract Metalliferous ecosystems, located in south-eastern Democratic Republic of Congo, host more than 550 metallophytes on isolated copper hills, including 57 taxa recognised as copper endemics. Mid-term conservation strategy of these species needs to establish structures and actions, in collaboration with the mining industry, to reproduce copper endemics. Considering the limits of the regeneration methods, the characterisation of the relationship between copper endemics and the copper in soil is essential to implement concrete actions. This study aims at determining the causes of edaphic endemism of four copper endemics to discuss the required edaphic conditions in ex situ conservation programs. The ex situ response to copper and the in situ Cu edaphic niches of taxa were characterised and discussed. The in situ distributions along this copper gradient (i.e., in situ copper edaphic niche) were calculated by means of a generalised additive model. The ex situ response of three populations of the four taxa to copper was tested, starting from the germination of seeds to the growth of seedlings in artificial copper-enriched soils. The results highlighted three patterns of ex situ responses to copper, while the populations had little effect on emergence and growth. Qualitative comparison with the in situ copper edaphic niche highlighted that Crotalaria cobalticola presented the status specialist model. In contrast, Diplolophium marthozianum and Gladiolus ledoctei corresponded to refuge model. The last species, Triumfetta welwitschii var. rogersii, did not show a clear pattern. 159

160 Introduction Metallophytes are plant species and taxa that can colonise metal-rich soils. These very restrictive habitats are generally of a small size and are geographically isolated from each other, within a landscape matrix presenting regular background metal concentrations (Antonovics et al. 1971, Baker 1987). The toxicity effect of metal and the insularity of areas are known to be major source of microevolution processes, leading to original communities, including rare, endemic and tolerant taxa (Kruckeberg & Kruckeberg 1990, Macnair & Gardner 1998, Van Rossum et al. 2004, Escarré et al. 2013). Edaphic endemism of taxa could be studied using genetic point of view by studying biotic depletion, allopatric speciation or genetic differentiation of individuals occurring in populations (Kruckeberg & Rabinowitz 1985, Macnair & Gardner 1998, Rajakaruna 2004) or from ecological point of view (Gankin & Major 1964, Meyer 1986). Three assumptions could explain the presence of edaphic endemic taxa in metalliferous soils. First assumption is that the edaphic endemics species are outcompeted from normal adjacent soils by dominant vegetation and take refuge in metalliferous soils, where interspecific competition is weaker. This is described as endemism refuge model and could be observed on serpentine ecosystems (Gankin & Major 1964, Anacker 2014). Second, edaphic endemic taxa could be specialised to metalliferous soils because they are characterised by a high physiological need in metal. They are more competitive on metalliferous soils (Meyer 1986, Peng et al. 2012, Faucon, Chipeng, et al. 2012). These taxa are called endemic specialists. Finally, edaphic endemic taxa could be restricted to metalliferous soils because the presence of heavy metals in soil relax the pathogen pressure (Boyd 2007, Fones et al. 2010). Ecosystems established on the natural copper (Cu) soils of south-eastern D. R. Congo support unique metallophyte communities (Malaisse et al. 1999). More than 150 Cu hills, presenting a large Cu gradient, occur in the so-called Katangan Copperbelt (Duvigneaud & Denaeyer-De Smet 1963, Cailteux et al. 2005). While concentrations less than 100 mg kg -1 occur in the non-metalliferous soils of the region, typical hills show Cu variations along a topographical gradient, with available Cu concentrations of up to 10,000 mg kg -1 at the top and a variation of 100 to 3,500 mg kg -1 on the slopes (Duvigneaud & Denaeyer-De Smet 1963, Leteinturier 2002). While the excess of Cu in soils is toxic for most plant species, more than 550 taxa grow on the Cu hills, of which 57 are considered to be Cu endemics. Among them, 33 have been recognised as absolute metallophytes (i.e., occurring strictly on the Cu- 160

161 enriched soils, and thus called strict Cu endemics) and 24 have been recognised as facultative metallophytes (i.e., broad copper endemics, with more than 75 percent of known populations occurring on copper-cobalt (Cu-Co) rich soils) (Leteinturier 2002, Faucon et al. 2010, Champluvier 2011). Heterogeneity of heavy metal concentrations within and among sites could lead to divergent selection in the population and thus contribute to ecotypic differentiation (Ilunga wa Ilunga et al. 2013). Studies on some of these metallophytes have highlighted a large intraspecific variability due to the heterogeneous, intraspecific genetic structure among populations (Peng et al. 2012, Faucon, Chipeng, et al. 2012) and presenting different tolerances to heavy metals (Faucon, Chipeng, et al. 2012). In the Katangan Copperbelt, the recent revival of mining activities has led to the destruction and degradation of these unique ecosystems, which host a number of species of a high conservation value (Brooks & Malaisse 1990, Leteinturier et al. 1999, Saad et al. 2012). To date, approximately 80 percent of Cu endemics have been identified as threatened (i.e., critically endangered, endangered or vulnerable) according to IUCN criteria (Faucon et al. 2010), and represent a conservation priority. From a biodiversity conservation point of view, in situ conservation appears to be the best strategy. However, this would require the preservation of a portion of copper endemic populations by establishing protected areas (Saad et al. 2012). This option, however, is economically unrealistic, as the demand for Cu and Co rises worldwide, with their exploitation providing a significant contribution to the economy of the D. R. Congo. The absence of a list of copper endemics in environmental protection laws and regulations in the D. R. Congo, further hampers the implementation of conservation programmes at a national scale. In this context, ex situ conservation strategies (i.e., a botanical garden or seedbank) of endemics should be promoted during the exploitation phase of Cu hills, in order to provide plant material for post-mining restoration programmes (Faucon et al. 2011, 2015, Faucon, Tshilong, et al. 2012, Saad et al. 2012). Improved understanding of the factors that influence species distribution, and the causes of endemism, would contribute significantly to the success of regeneration of Cu endemics, in ex situ conservation strategies. These factors, which potentially influence regeneration, advocate for a knowledge on species ecology and their response to Cu (Bizoux et al. 2004, Whiting et al. 2004, Boisson, Ortmans, et al. 2016). To date, most studies focusing on Cu endemics have been performed in axenic conditions (Chipeng et al. 2010, Faucon, Chipeng, et al. 2012, Godefroid et al. 2013, Boisson, 161

162 Le Stradic, et al. in press, Boisson, Ortmans, et al. in press). Results from laboratory tests are not systematically transferable to a botanical garden context because of the broad differences of metal availability in substrates, as well as the climatic conditions. In this context, this study aims at identifying the causes of edaphic endemism of four copper endemics (Crotalaria cobalticola, Diplolophium marthozianum, Gladiolus ledoctei and Triumfetta welwitschii var. rogersii) by characterising their ex situ response to copper during one vegetation season and determining their in situ Cu edaphic niches in their natural habitat. Considering that the conservation actions are urgent, we further discuss the guidelines for the ex situ conservation programs in relationship with their edaphic endemism status. Methods Study Site The study was performed in the south-eastern D. R. Congo (in the Katangan Copperbelt) where Cu hills were formed during the mineralisation of the most resistant siliceous rock of the Roan Series (Duvigneaud & Denaeyer-De Smet 1963, François 1973, Cailteux et al. 2005). These hills are scattered in the dominant vegetation, the Miombo woodland, with altitude in the area ranging from 1200 to 1400 m. Two major plant communities have been identified along the Cu gradient: (1) the steppe, on soils highly enriched in Cu (from 3,500 mg kg -1 to 10,000 mg kg -1 of available Cu in soils, on the top of a Cu hill); and (2) the steppic savannah, on the slopes and bases of the hills (up to mg kg -1 of available Cu in soil) (Ilunga wa Ilunga et al., 2013; Séleck, Bizoux, et al., 2013). This region has a subtropical climate, with a rainy season extending from November to the end of March, and a dry season from May to September. October and April are transition months between the two seasons. Studied copper endemics We focused the analysis on four Cu endemic taxa from the Katanga Copperbelt as representative of the diversity of distribution along the Cu gradient. Crotalaria cobalticola P.A.Duvign. & Plancke (Fabaceae) is an annual. It is a strict Cu endemic with a proposed endangered (EN) IUCN status (Faucon et al., 2010). This erect herb, which is silky-pubescent, with numerous branched stems and leaves composed of three linear-filiform leaflets, occurs in the steppe of the Cu hills. Flowers appear in April/May and fruits usually mature between May and June. Diplolophium marthozianum P.A.Duvign. (Apiaceae) is a perennial broad Cu 162

163 endemic, occurring in the highest Cu enriched soils of the steppic savannah of the Cu hills. Its proposed IUCN status is endangered (EN) (Faucon et al., 2010). This herb has basal leaves, which are largely petiolated, with a limb wider than it is long. Flowering peaks from February to April. Mature seeds are dispersed throughout the dry season, from June to August. Gladiolus ledoctei P.A.Duvign. & Van Bockstal (Iridaceae) is a perennial that has been considered to be a broad Cu endemic. Its habitat principally include the steppe and, to a lesser extent, the steppic savannah of the Cu hills. Its proposed IUCN status is endangered (EN, (Faucon et al., 2010). This glabrous and erect herb grows from a corm, bearing up to five to seven caulinar leaves. Zygomorphic flowers appear in the second half of the rainy season, from late January to March/April. Seeds ripen between April and May. Triumfetta welwitschii var. rogersii (N.E.Br.) Brummitt & Seyani (Malvaceae) is a perennial and strict Cu endemic. This taxon is found on the soils of the steppic savannah having the highest Cu concentrations. Its proposed IUCN status is data deficient (DD), for lack of information. This suffrutex, with a deep, woody rootstock (i.e., xylopod), produces annual shoots. Leaves are linear-lanceolate, lamina linear, acute and apiculate at the apex. Flowers appear right after fires, at the end of the dry season, in September/October. Fruits mature shortly after and can co-occur with flowers. In situ modelling of Cu edaphic niches The in situ copper edaphic niches were modelled using the species presence-absence database from Séleck et al. (2013). This database contains phytosociological data of 172 square-metre plots (1 m 2 ) surveyed along systematic grids on three copper hills between the towns of Tenke (10 36'36" S, 26 07'12 E ) and Fungurume (10 37'12"S, '12"E). Data from each vegetation plot was linked with soil data corresponding to the available Cu concentration in soil (i.e., EDTA-extractable in the study), sampled at a depth of 15 cm. Edaphic niches along the natural Cu gradient (i.e., Cu edaphic niches) of each selected taxon were then modelled with generalised additive models (GAM), using a binomial likelihood (Hastie et al., 1986). The resulting species response curves represent the probability of occurrence of the taxon along the Cu gradient. For each taxon, the best fit model among three degrees of smoothing (3, 4 or 5) was selected, based on the Aikaike Information Criterion (AIC) (Aikaike, 1987). The optimum Cu concentration of each taxon was then estimated by identifying the highest probability of occurrence. The borders of the Cu edaphic niche widths (or amplitudes) were determined by the lower and upper limits of 80 percent of the area 163

164 under the species response curves (Gégout et al., 1998) followed by the calculation of the Cu edaphic niche widths by the difference between the upper and the lower limits. Response to Cu in ex situ conditions PLANT MATERIAL AND SOIL PREPARATION The experiment was performed in the experimental botanical garden of the University of Lubumbashi (Faculty of Agronomy, 11 36'36"S, '48" E). Climatic conditions of the experimental garden closely approximate seeds origins. For each taxon, seed samples conserved at the University of Lubumbashi (UNILU) were used. Each sample represented a seed lot of one population collected on one copper hill and were conserved in ultra-drying conditions (i.e., relative humidity [R.H.] < 5%, at room temperature), except for the seeds of T. welwitschii var. rogersii, which were conserved at ambient R.H., at room temperature for only two months (Table 1). For each taxon, three seed lots from three distinct copper hills were chosen, representing three distinct populations. Three types of substrates were created from a homogeneous arable soil (ferralsol) at the University of Lubumbashi: (1) a control substrate, with 0 mg kg -1 soil Cu-addition; (2) an intermediately Cu-enriched soil, with 100 mg kg -1 soil Cu added; and, (3) a highly Cu-enriched soil, with 1,000 mg kg -1 soil Cu added. Soil enrichment with Cu was achieved through the addition of Cu(II) sulfate pentahydrate (CuSO 4.5H 2 O) in powder form. A quantity of 0.02 g of compost per kilogram of soil was mixed with sieved soil (2 mm) to reproduce the organic matter (OM) content usually found in the copper hill soils (C org : 30, Séleck et al. 2013). Soils were then homogenised after the Cu sulfate pentahydrate and compost addition. Plastic nursery bags were filled with 1.5 kg (c. 20 cm height) of the different substrates. Bags were perforated to facilitate water drainage and set in the garden for one week, prior to sowing. 164

165 Table 1. Detailing the collection sites, geographical coordinates, conservation methods and duration, and identity (ID) for each population of the studied taxa. Taxon Collecting site Coordinate (utm) Conservation method C. cobalticola Fungurume V Ultradrying Room T C Kabwelunono [1] Ultradrying Room T C Kavifwafwaulu [4] Ultradrying Room T C D. marthozianum Fungurume V Ultradrying Room T C Kakalalwe [1] Ultradrying Room T C Kavifwafwaulu [4] Ultradrying Room T C G. ledoctei Fungurume V Ultradrying Room T C Shimbidi Ultradrying Room T C Kavifwafwaulu Ultradrying Room T C T. welwitschii var. rogersii Fungurume VIII [1] Ambient R.H. Room T C Kabwelunono [1] Ambient R.H. Room T C Kavifwafwaulu [4] Ambient R.H. Room T C Conservation duration ID 28 months A 28 months B 28 months C 16 months A 24 months B 27 months C 41 months A 41 months B 28 months C 2 months A 2 months B 1 months C 165

166 EXPERIMENTAL DESIGN AND MEASURES For each combination of species and population, five seeds were sown per bag, between 26 November and 3 December 2013, at the beginning of the rainy season. A prior test showed that the weight of seedlots was not significantly different within taxa. Ten repetitions (i.e., ten bags) for each population of each species were cultivated in the three substrates (control, 100 and 1,000 mg kg -1 Cu-enriched soils). Bags were then placed in a shade house to keep the seedlings in favourable conditions. The experiment took place from the end of November 2013 to the end of April After the 15 February (i.e., when all species had emerged), the tallest individual was kept in each bag and the others were removed, in order to limit competition. During the experiment, all bags were watered every two days, and were moved randomly in the shade house every week. Emergence (i.e., an individual with cotyledons) was recorded weekly from December 2013 to February In February, the most developed individual was kept per bag. At the end of the rainy season 2014 (i.e., last week of April), growth of individuals was assessed through height measurement (i.e., the distance between soil surface and the apical point of the plant). In addition, for T. welwitschii var. rogersii and D. marthozianum, the number of leaves was counted, and for C. cobalticola, the number of main stems was counted. As C. cobalticola is an annual, the number of flowers by individual was also counted (mid-april), and the length of the root system was measured at the end of the experiment (first week of May). SOIL ANALYSES At the end of the experiment, soils were randomly collected from 12 bags representing each modality of species Cu concentration, in order to have four replicates of soil treatment. Soil samples were air-dried and sieved to 2 mm prior to the analyses. The ph was measured in Potassium chloride (KCl) 1N with a glass electrode in a 2:5 soil:solution ratio, after a 2h equilibration time. Total carbon (C) content was measured after oxidising to Carbon dioxide (CO 2) by heating the soil to at least 900 C, in a flow of oxygen-containing gas. The amount of CO 2 released was then measured by the Gas Chromatography (GC) method (Margesin et al., 2005). Concentrations of Cu extractable from the soil were determined using Ammonium acetate EDTA (CH 3 COONH 4 EDTA) 1N (Lakanen et al., 1971).The soil:solution ratio was 1:5 and the ph of the extraction solution buffered at The supernatant was filtered through a 166

167 S&S 595 folded filter and analysed using a flame atomic absorption spectrometer (Varian 220). DATA TREATMENT The percentage of emergence per bag was compared using a two-way ANOVA (population Cu concentration). The growth (i.e., height, number of leaves or main stems and length of the root system) and the number of flowers per individual (one per bag) were also compared by a two-way ANOVA (population Cu concentration). Prior to analysis, normality and homoscedasticity of the data were checked, with log transformations were required for the number of leaves of T. welwitschii var. rogersii and for number of flowers of C. cobalticola. Means were structured via the Tukey HSD test, with an error rate of 5 percent. Soil compounds were compared with a Kruskal-Wallis test followed by a Bonferroni mean adjustment comparison. Spearman correlation tests were performed between height and number of leaves (or stems), to identify the relationship between these responses. All analyses were performed using the R statistical software (R Development Core Team, 2010). Results In situ Cu edaphic niches The four Cu endemic species present distinct, while overlapping, in situ Cu edaphic niches (Figure 1). Crotalaria cobalticola had its Cu optimum at the highest concentrations (10,136 mg kg -1 Cu), with a truncated response curve, and a lower limit of its Cu edaphic niche, at 4547 mg kg -1 Cu (Figure 1). This taxon could be considered the most tolerant Cu endemic of our study. Gladiolus ledoctei occurred in a range of 1,525 to 9,499 mg kg -1 Cu, with a Cu optimum at 7,468 mg kg -1 Cu (Figure 1). Its response curve was skewed, with a longer tail toward the lowest Cu concentrations. Diplolophium marthozianum presented a Cu edaphic niche optimum at 3,243 mg kg -1 Cu, with a symmetrical niche width (5,387 mg kg -1 Cu, Figure 1). Triumfetta welwitschii var. rogersii had its in situ Cu edaphic niche optimum at 1,838 mg kg -1 Cu. Its Cu niche width extended from 201 mg kg -1 Cu up to 5,194 mg kg -1 Cu, and had a skewed response with a tail toward the highest Cu concentrations (Figure 1). 167

168 Figure 1. Response of the taxa (C. cobalticola, D. marthozianum, G. ledoctei and T. welwitschii var. rogersii) according to copper concentrations (ex situ, mg kg -1 ) added in soils and correspondent copper edaphic niches along natural available copper gradient (in situ, mg kg -1 ). Data of response were collected five months after the beginning of the experiment. The number of main stems was represented for C. cobalticola and the number of leaves was for D. marthozianum, G. ledoctei, T. welwitschii var. rogersii (Mean ± SE). Means sharing the same letters are not significantly different, according to the Tukey HSD test. The copper edaphic niches were modelled by generalised additive models (GAM). Dots are the niche optima of taxa determined by the gradient value of the highest probability of occurrence. Thick and straight lines are the niche widths (or amplitudes) calculated by the difference between the borders of area of 80 percent under the species response curves. 168

169 Response to Cu in ex situ conditions As expected, soil analyses revealed that the EDTA-extractable Cu concentrations were significantly higher in the 1,000 mg kg -1 Cu-enriched soils than the two other soils (control and 100 mg kg -1 Cu-enriched soils), (X 2 = 8.346; p-value < 0.05, Table 2). The ph values were inversely proportional to the Cu concentrations added in soils with a significantly lower ph in 1,000 mg kg -1 Cu-enriched soils than in the control and the 100 mg kg -1 (X 2 = 7.731; p-value < 0.05, Table 2). The means stayed between 5.0 ± 0.3. Surprisingly, a nearly significant variation of the total C was highlighted, with the highest percentage found in the 1000 Cuenriched soils (X 2 = 6.039; p-value < 0.05, Table 2). Table 2. Physico-chemical conditions (Mean ± SD) including the total organic matter (Corg, %), the EDTA-extractable copper (Cu, mg kg -1 ) and the ph KCl of the three substrates of the test (n=4). Control 100 mg kg -1 1,000 mg kg -1 Kruskal-Wallis test Edaphic content X 2 p-value C org ( ) 27.2 ± 3.5 b 30.5 ± 1.5 ab 33.6 ± 2.2 a Cu (mg kg -1 ) 96.7 ± 12.4 b ± 38.1 b ± a ph KCl 5.2 ± 0.1 b 5.1 ± 0.1 b 4.7 ± 0.0 a Global means of the percentages of emergence differed according to taxa. Gladiolus ledoctei and C. cobalticola presented the highest percentage of emergence, respectively 49.6 ± 3.0% and 44.4 ± 4.8% (Mean ± SE), in comparison to T. welwitschii var. rogersii and D. marthozianum, which had percentages of emergence of 13.1 ± 2.1% and 11.7 ± 1.5%, respectively (Table 3). From this, it is apparent that the percentage of emergence of the four taxa was not affected by Cu concentration (Table 3). In contrast, the percentage of emergence of the population of T. welwitschii var. rogersii from Fungurume VIII (24.7 ± 4.5%) was significantly higher than the population from Kavifwafwaulu [4] (10.7 ± 3.4%) and Kabwelunono [1] (4.0 ± 1.8%, F = 9.323, p-value < 0.001). For two Cu endemics, C. cobalticola and T. welwitschii var. rogersii, no interaction between the population and Cu concentrations were identified for the percentage of emergence (C. cobalticola: F = 0.334; p-value = 0.855; T. welwitschii: F = 0.608, p-value = 0.657). Interactions between both factors affected the percentage of emergence for G. ledoctei (F = 3.665, p-value < 0.01) and D. marthozianum (F = 2.500, p-value = 0.049), 169

170 but no significant difference between the combination of the population and the Cu concentrations, were evident in the Tukey HSD test. No significant interaction existed between the populations and the Cu concentration on the growth for the four tested species. Effects of each factor were highlighted independently. Considering that the height and the number of main stems of individuals of C. cobalticola were correlated and presented the same pattern of variation (Spearman coefficient correlation = 0.80, p-value < 0.001), the number of stems was only detailed in the results. The number of mains stems was not significantly different between the populations, but was significantly different between Cu concentrations of soils. The control soils presented individuals with significantly fewer stems (9.8 ± 2.1 stems) than the 100 mg kg -1 Cuenriched soils (17.8 ± 1.8 stems) and the 1,000 mg kg -1 Cu-enriched soils (21.5 ± 2.0 stems, Figure 1; Table 3). The same pattern of response to Cu was also observed for the length of the root system, with significantly longer roots observed in individuals growing on 1,000 mg kg -1 Cu-enriched soils (5.5 ± 0.6 cm) and on 100 mg kg -1 Cu-enriched soils (5.3 ± 0.8 cm), than on the control soils (2.6 ± 0.7 cm, Table 3). While the number of flowers per individual did not significantly vary between Cu concentrations (Table 3), the percentage of flowering individuals was higher on 100 and 1,000 mg kg -1 Cu-enriched soils (respectively 36% and 44%) than on the control soils (25%). No effect of the population on the growth traits was demonstrated (Table 3). The individuals of D. marthozianum presented a distinctly different pattern in relation to Cu concentrations (Figure 1; Table 3). Considering that the number of leaves was significantly correlated with the height (r = 0.73, p-value < 0.001), the number of leaves was chosen for the comparisons. The number of individuals leaves was greater on the control soils (4.6 ± 0.6 leaves) and in the 100 mg kg -1 Cu-enriched soils (4.9 ± 0.6 leaves), than on the 1,000 mg kg -1 Cu-enriched soils (2.8 ± 0.4 leaves, Figure 1; Table 3). Additionally, an effect of population was observed, in which the number of leaves of individuals from the Kakalalwe population [1] (4.9 ± 0.5 leaves) was significantly higher than the mean number of leaves of individuals from the Fungurume V (2.9 ± 0.3 leaves) and Kavifwafwaulu [4] (3.3 ± 0.3 leaves, Table 3) populations. For G. ledoctei, the height of individuals and the number of the individuals leaves were not significantly different, either between populations (Table 3) or Cu concentrations (Figure 1; Table 3). 170

171 Since a significant positive correlation was observed between the number of leaves and the height of T. welwitschii var. rogersii (r = 0.56, p-value < 0.001), only the results for the number of leaves according to the Cu concentrations were presented. The number of leaves for plants that were grown on the control soils (16.3 ± 1.6 leaves) and on 100 mg kg -1 Cuenriched soils (15.1 ± 1.6 leaves) were significantly greater than the number of leaves for plants sown on 1,000 mg kg -1 Cu-enriched soils (8.3 ± 1.4 leaves, Figure 1; Table 3). However, populations did not affect the number of leaves (Table 3). Table 3. Percentage of emergence and growth trait variations, according to the Cu concentrations added to the soil and the populations for the four studied taxa C. cobalticola, D. marthozianum, G. ledoctei and T. welwitschii var. rogersii. Legend: Letters A, B, C used for the population refer to the Table 1. Traits Percentage of emergence (%) Cu concentrations added in soil Populations Control 100 mg kg -1 1,000 mg kg -1 F p-value A B C F p- value C. cobalticola 42.7 ± ± ± 4.1 Number of main stems 9.8 ± 2.1 b 17.8 ± 1.8 a 21.5 ± 2.0 a Root system length (cm) 2.6 ± 0.7 b 5.3 ± 0.8 ab 5.5 ± 0.6 a Number of flowers by individuals Percentage of emergence (%)* 2.7 ± ± 2.2 D. marthozianum 13.3 ± ± ± ± 2.4 Number of leaves 4.6 ± 0.6 a 4.4 ± 0.3 a 2.6 ± 0.4 b Percentage of emergence (%)* G. ledoctei 52.7 ± 5.4 Height (cm) 9.4 ± 0.6 Number of leaves 1.3 ± 0.1 Percentage of emergence (%) 46.0 ± ± ± ± ± ± 0.1 T. welwitschii var. rogersii 12.0 ± ± ± 2.8 Number of leaves 16.3 ± 1.6 a 15.1 ± 1.6 a 8.3 ± 1.4 b ± ± ± ± ± ± ± ± ± ± ± 0.3 b ± 0.5 a ± ± ± ± ± ± ± 4.5 a ± 3.4 b < ± 1.2 *Significant interaction, Tukey HSD test was performed when the p-value was lower than ± ± ± ± ± ± <0.001 ± 0.3 b 54.7 ± ± ± <0.001 ± 1.8 c 13.8 ±

172 Discussion In ex situ conditions, the Cu concentrations in the soil did not influence seedling emergence for any of the four endemic species. Such results imply that the seed coat could play a major role as a barrier between the embryo and the direct environment, as demonstrated in a study on non-metallophytes (Di Salvatore et al., 2008). In addition, no difference in terms of emergence rate among populations was evident for three species, C. cobalticola, G. ledoctei and D. marthozianum. This contrasts with other studies performed on the European metallophyte populations of Noccea caerulescens (syn. Thlaspi caerulescens) (Roosens et al., 2003) or Crepidorhopalon tenuis and Haumanisatrum katangense in D.R. Congo (Faucon et al. 2012, Peng et al. 2012), where significant differences in response to Cu were found among populations from distinctly metalliferous sites. Considering that the intraspecific variations correspond to a combination of environmentally-induced and genetic differences between populations (Mengoni et al., 2000; Van Rossum et al., 2004; Yang et al., 2005), the non-significance of our results could be due to the experimental design and the fact that we have a limited range of environmental conditions (i.e., no variation in rainfall, drainage, light, insolation, etc.) at the botanical garden. Only the seedling emergence of T. welwitschii var. rogersii was significantly affected by the population s origin. This suggests that variation in emergence rates exists between seed lots and that this might have been inherited from the maternal populations. Roach Wulf (1987) showed that environmental conditions during seed maturation of maternal population are likely to have significant effects on germination behaviour of the next generation (i.e., maternal effect). Small scale variations of edaphic conditions in metalliferous sites have been identified as being responsible for the physiological and evolutionary process variation among metallophytes (Bizoux et al., 2004; Ilunga wa Ilunga et al., 2013). Variations in emergence rates might also be due to the ex situ conservation method of each seed lot. The storage time and the method used for ex situ seed conservation could affect the viability and the emergence percentage for desiccation-tolerant (i.e., orthodox) or desiccation-sensitive (i.e., recalcitrant) species. The ex situ conservation strategy in the seedbank should be examined, to identify the best storage method for each species (Godefroid et al., 2013). Furthermore, the large variation found between the percentage of emergence of studied taxa suggests that the ex situ conservation action should adjust the number of seeds collected, depending on the taxon. Species occurring in the Cu communities are characterised by distinct life forms 172

173 that potentially influence the percentage of emergence (Duvigneaud et al., 1963a; Leteinturier, 2002; Godefroid et al., 2013). Three different growth patterns in terms of ex situ Cu response have been highlighted. The first pattern, observed for C. cobalticola, can be defined as a favourable response of growth to increasing Cu concentrations in soil. This species showed the largest individuals, five months after beginning the experiment on the highest Cu-enriched soils (1,000 mg kg -1 ), suggesting that this species requires high Cu concentrations for optimal development (Baker, 1987). Enhanced growth of some metallophytes in the presence of growing metal concentrations was already found on H. katangense and C. perennis (Chipeng et al., 2010; Faucon, Chipeng, et al., 2012), both occurring in the steppes of the Cu hills, as well as, other hyperaccumulators occurring on serpentine soils (Millie Burrell et al., 2012; Ghasemi et al., 2014). However, the enhanced vegetative growth of the individuals did not affect the number of flowers of C. cobalticola. The second growth pattern that could be recognised was an increase of growth in response to decreasing Cu concentrations. This was the case with D. marthozianum and T. welwitschii var. rogersii, presenting most developed individuals in the lowest Cu concentrations ( 100 mg kg -1 ) and in the control. The third pattern, shown by G. ledoctei, corresponds to an equivalent emergence and growth response whatever the Cu concentrations in soil. In general, for almost all the species (except for G. ledoctei), increasing Cu sensitivity (i.e., reduction of the Cu tolerance) is observed between the stages of germination and early growth. The patterns of C. cobalticola, D. marthozianum and T. welwitschii var. rogersii correspond to an ontogenetic niche shift explained by a reduction, or a change, of the ecological niche during the growing stages (i.e., Cu tolerance in this study) over time (Parish et al., 1985; Eriksson, 2002; Miriti, 2006). The growth response to Cu, in ex situ conditions, is in accordance with in situ edaphic niches (modelled for Cu in soil) for the four Cu endemics. Crotalaria cobalticola presented the most individuals in the highest Cu-enriched soils and had its Cu edaphic niche optimum at 10,139 mg kg -1 Cu. Conversely, D. marthozianum and T. welwitschii var. rogersii shared the same pattern, with better growth in the lowest Cu-enriched soils, since they had intermediate Cu niche widths between 0 and 6,000 mg kg -1 Cu. Finally, G. ledoctei did not present a significant 173

174 growth response to variation of Cu in ex-situ conditions. This species had the broadest Cu edaphic niche, with an optimum between D. marthozianum and C. cobalticola. Based on the comparison between the Cu edaphic niches and the growth response to Cu of the Cu endemics, it may be assumed that the biotic interactions, such as competition, could limit the distribution of Cu endemics occurring mostly in the steppe. This is the case for the taxon D. marthozianum, which is usually found in the steppe, while in ex situ conditions, this species grew better in the lowest Cu-enriched soils. The asymmetric Cu edaphic niche and the equivalent growth response across all Cu concentrations, suggest that G. ledoctei also follows this pattern. These two species correspond best to the refuge endemism model, which states that edaphic, endemic species are rejected from the dominant vegetation to Cuenriched soils, where interspecific competition is hypothesised to be lower (Gankin et al., 1964). C. cobalticola, on the other hand, performed best in the highest Cu-enriched soils, fitting a specialist endemism model (i.e., to Cu), as was already observed by Boisson, Le Stradic, et al. (in press). The tolerance to the highest Cu concentrations usually comes at a physiological cost (Macnair, 1993; Macnair et al., 2000; Chipeng et al., 2010; Faucon, Chipeng, et al., 2012), implying that individuals have lower performance on non-metal enriched soils. Although the range of Cu concentrations tested between soils ( 100 to 1,000 mg kg -1 Cu) was relatively large, the addition of Cu sulphate and organic matter might have significantly affected the availability of other edaphic compounds, such as the nutrients calcium (Ca) and phosphorus (P) and P (Córdova et al., 2011). The complex interactions between soil compounds and heavy metals contribute indirectly to floristic variation and heterogeneity of plant communities in Cu outcrops (Saad et al., 2012; Séleck, Bizoux, et al., 2013; Lange et al., 2014; Pourret, Lange, Houben, et al., 2015). The presence of iron-oxides (FeOx) and Mnoxides (MnOx) in soils could be decreasing the toxicity of Cu by decreasing its availability (Kabala et al., 2001; Pourret, Lange, Houben, et al., 2015). The response of tolerant plant species can be studied in vitro by varying the edaphic compounds and Cu concentration in agar, and excluding all other biotic and abiotic factors. 174

175 Conclusion The present study demonstrated that ex situ conservation of endemic metallophytes should take the variety of patterns found between taxa in Cu communities, such as endemics, into account. It also illustrates that knowledge of their in situ habitat, gained by studying the edaphic niches along a metal gradient, is just as important as the taxa s responses under artificial conditions. In the Katangan Copperbelt, the habitats created by mining activities host new edaphic conditions and which need to be assessed in the field, in order to expand the ex situ conservation of Cu endemics at a large scale. Furthermore, this study explored the patterns among four Cu endemics representing less than 10 percent of the identified strict and broad Cu endemics. Further studies should focus on other Cu endemics, and on other determinant factors, such as biotic interactions, in order to improve ex situ conservation strategies. Acknowledgments The present work has been made possible thanks to the financial intervention of the Fonds de la Recherche dans l Industrie et dans l Agriculture (FRIA) of the FNRS (Fonds National de la Recherche Scientifique), Belgium. FNRS. The Coopération Universitaire au Développement (CUD) is acknowledged for the funding support for the Faculty of Agronomy s seedbank at the University of Lubumbashi (PIC REMEDLU) and the fellowship of Olivier Garin who performed his Master s thesis in D.R. Congo. The authors also thank Tenke Fungurume Mining sarl (TFM, Freeport McMoran Copper & Gold) for financial and logistic support in the field. References Aikaike H (1987) Factor analysis and AIC. Psychometrika 52: Anacker BL (2014) The nature of serpentine endemism. Am J Bot 101: doi: /ajb Antonovics J, Bradshaw AD, Turner RG (1971) Heavy metal tolerance in plants. Adv Ecol Res 7:1 85. Baker AJM (1987) Metal Tolerance. New Phytol 106: doi: /j tb04685.x Bizoux J, Brevers F, Meerts P, et al (2004) Ecology and conservation of Belgian populations of Viola calaminaria, a metallophyte with a restricted geographic distribution. Belgian J Bot 137: Boisson S, Le Stradic S, Commans M, et al (2016a) Copper tolerance of three Crotalaria species from southeastern D. R. Congo at the early development stage. Biotechnol Agron Soc Environ 20: Boisson S, Ortmans W, Maréchal J, et al (2016b) No copper required for germination of an endangered endemic species from the Katangan Copperbelt (Katanga, DR Congo): Diplolophium marthozianum. Boyd RS (2007) The defense hypothesis of elemental hyperaccumulation: status, challenges and new directions. Plant Soil 293: doi: /s Brooks RR, Malaisse F (1990) Metal-enriched sites in south central africa. In: Shaw J (ed) Heavy Metal Tolerance in Plants: Evolutionary Aspects. CRC Press, Inc, New York, pp Cailteux JLH, Kampunzu AB, Lerouge C, et al (2005) Genesis of sediment-hosted stratiform copper cobalt deposits, central African Copperbelt. J African Earth Sci 42: doi: /j.jafrearsci

176 Champluvier D (2011) New and overlooked Acanthaceae taxa from D.R.Congo, Rwanda and Burundi: (1) the genus Barleria. Plant Ecol Evol 144: doi: /plecevo Chipeng FK, Hermans C, Colinet G, et al (2010) Copper tolerance in the cuprophyte Haumaniastrum katangense (S. Moore) P.A. Duvign. & Plancke. Plant Soil 328: doi: /s z Córdova S, Neaman A, González I, et al (2011) The effect of lime and compost amendments on the potential for the revegetation of metal-polluted, acidic soils. Geoderma 166: doi: /j.geoderma Di Salvatore M, Carafa A, Carratù G (2008) Assessment of heavy metals phytotoxicity using seed germination and root elongation tests: a comparison of two growth substrates. Chemosphere 73: doi: /j.chemosphere Duvigneaud P, Denaeyer-De Smet S (1963) Etudes sur la végétation du Katanga et de ses sols métallifères. Communication n 7 Cuivre et végétation au Katanga. Bull Soc Roy Bot Belgique 96: Eriksson O (2002) Ontogenetic niche shifts and their implications for recruitment in three clonal Vaccinium shrubs: Vaccinium myrtillus, Vaccinium vitis-idaea, and Vaccinium oxycoccos. Can J Bot 80: doi: /B Escarré J, Lefèbvre C, Frérot H, et al (2013) Metal concentration and metal mass of metallicolous, non metallicolous and serpentine Noccaea caerulescens populations, cultivated in different growth media. Plant Soil 370: doi: /s z Faucon M, Le Stradic S, Boisson S, et al (2015) Implication of plant-soil relationships for conservation and restoration of copper-cobalt ecosystems. Plant Soil. doi: /s Faucon M-P, Chipeng F, Verbruggen N, et al (2012a) Copper tolerance and accumulation in two cuprophytes of South Central Africa: Crepidorhopalon perennis and C. tenuis (Linderniaceae). Environ Exp Bot 84: doi: /j.envexpbot Faucon M-P, Meersseman A, Shutcha MN, et al (2010) Copper endemism in the Congolese flora: a database of copper affinity and conservational value of cuprophytes. Plant Ecol Evol 143:5 18. doi: /plecevo Faucon M-P, Parmentier I, Colinet G, et al (2011) May Rare Metallophytes Benefit from Disturbed Soils Following Mining Activity? The Case of the Crepidorhopalon tenuis in Katanga (D. R. Congo). Restor Ecol 19: doi: /j X x Faucon M-P, Tshilong BM, Van Rossum F, et al (2012b) Ecology and Hybridization Potential of Two Sympatric Metallophytes, the Narrow Endemic Crepidorhopalon perennis ( Linderniaceae ) and its More Widespread Congener. Biotropica 44: Fones H, Davis C a R, Rico A, et al (2010) Metal hyperaccumulation armors plants against disease. PLoS Pathog 6:1 13. doi: /journal.ppat François A (1973) L extrémité occidentale de l'arc Cuprifère Shabien. Etude géologique-département de géologie de la Gécamines, Likasi (République du Zaire) Gankin R, Major J (1964) Arctostaphylos myrtifolia, its biology and relationship to the problem of endemism. Ecology 45: Gégout J-C, Pierrat J-C (1998) L autécologie des espèces végétales : Une approche par régression non paramétrique. Ecologie 29: Ghasemi R, Chavoshi ZZ, Boyd RS, Rajakaruna N (2014) A preliminary study of the role of nickel in enhancing fl owering of the nickel hyperaccumulating plant Alyssum inflatum Nyár. (Brassicaceae). South African J Bot 92: doi: /j.sajb Godefroid S, Van de Vyver A, Massengo Kalenga W, et al (2013) Germination capacity and seed storage behaviour of threatened metallophytes from the Katanga copper belt (DR Congo): implications for ex situ conservation. Plant Ecol Evol 146: doi: /plecevo Hastie T, Tibshirani R (1986) Generalized additive models. Stat Sci 1: Ilunga wa Ilunga E, Séleck M, Colinet G, et al (2013) Small-scale diversity of plant communities and distribution of species niches on a copper rock outcrop in Upper Katanga, DR Congo. Plant Ecol Evol 146: doi: /plecevo Kabala C, Singh BR (2001) Fractionation and mobility of copper, lead, and zinc in soil profiles in the vicinity of a copper smelter. J Environ Qual 30: doi: /jeq x Kruckeberg a R, Rabinowitz D (1985) Biological Aspects of Endemism in Higher Plants. Annu Rev Ecol Syst 16: doi: /annurev.es Kruckeberg AR, Kruckeberg A (1990) Endemic metallophytes: their taxonomic, genetic and evolutionary attributes. In: Shaw J (ed) Heavy Metal Tolerance in Plants: Evolutionary Aspects. CRC Press Inc, New York, pp Lakanen E, Erviö R (1971) A comparison of eight extractants for the determination of plant available micronutrients in soil. Acta Agral Fenn 123: Lange B, Faucon M-P, Meerts P, et al (2014) Prediction of the edaphic factors influence upon the copper and cobalt accumulation in two metallophytes using copper and cobalt speciation in soils. Plant Soil 379: doi: /s y Leteinturier B (2002) Evaluation du potential phytocénotique des gisements cupriferes d Afrique centro-australe en vue de la phytoremédiation de sites pollués par l'activité. Leteinturier B, Baker AJM, Malaisse F (1999) Early stages of natural revegetation of metalliferous mine workings in South Central Africa: a preliminary survey. Biotechnol Agron Soc Environ 3: Macnair M, Gardner M (1998) The Evolution of Edaphic Endemics. In: Howard D, Berlocher S (eds) Endless 176

177 Forms. Species and Speciation. Oxford University Press, New York, pp Macnair M, Tilstone G, Smith S (2000) The Genetics of Metal Tolerance and Accumulation in Higher Plants. In: Terry N (ed) Phytoremediation of contaminated soil and water. CRC Press Inc, p 408 Macnair MR (1993) The genetics of metal tolerance in vascular plants. New Phytol 124: doi: /j tb03846.x Malaisse F, Baker AJM, Ruelle S (1999) Diversity of plant communities and leaf heavy metal content at Luiswishi copper/cobalt mineralization, Upper Katanga, Dem. Rep. Congo. Biotechnol Agron Soc Environ 3: Margesin R, Schinner F (2005) Manual for Soil Analysis - Monitoring and Assessing Soil Bioremediation: Monitoring and Assessing Soil Bioremediation. Springer Mengoni A, Gonnelli C, Galardi F, et al (2000) Genetic diversity and heavy metal tolerance in populations of Silene paradoxa L. (Caryophyllaceae): a random amplified polymorphic DNA analysis. Mol Ecol 9: Meyer S (1986) The ecology of gypsophile endemism in the eastern Mojave Desert. Ecology 67: Millie Burrell A, Hawkins AK, Pepper AE (2012) Genetic analyses of nickel tolerance in a North American serpentine endemic plant, Caulanthus amplexicaulis var. barbarae (Brassicaceae). Am J Bot 99: doi: /ajb Miriti MN (2006) Ontogenetic shift from facilitation to competition in a desert shrub. J Ecol 94: doi: /j x Parish J, Bazzaz F (1985) Ontogenetic niche shifts in old-field annuals. Ecology 66: Peng H, Wang-Müller Q, Witt T, et al (2012) Differences in copper accumulation and copper stress between eight populations of Haumaniastrum katangense. Environ Exp Bot 79: doi: /j.envexpbot Pourret O, Lange B, Houben D, et al (2015) Modeling of cobalt and copper speciation in metalliferous soils from Katanga (Democratic Republic of Congo). J Geochemical Explor 149: doi: /j.gexplo R Development Core Team (2010) A language and environment for statistical computing. Vienna (Austria) Rajakaruna N (2004) The edaphic factor in the origin of species. Int Geol Rev 46: Roach D, Wulf RD (1987) Maternal Effects In Plants. Annu. Rev. Ecol. Syst. 18: Roosens N, Verbruggen N, Meerts P, et al (2003) Natural variation in cadmium tolerance and its relationship to metal hyperaccumulation for seven populations of Thlaspi caerulescens from western Europe. Plant Cell Environ 26: Saad L, Parmentier I, Colinet G, et al (2012) Investigating the Vegetation-Soil Relationships on the Copper-Cobalt Rock Outcrops of Katanga (D. R. Congo), an Essential Step in a Biodiversity Conservation Plan. Restor Ecol 20: doi: /j X x Séleck M, Bizoux J-P, Colinet G, et al (2013) Chemical soil factors influencing plant assemblages along coppercobalt gradients: implications for conservation and restoration. Plant Soil 373: doi: /s Van Rossum F, Bonnin I, Fenart S, et al (2004) Spatial genetic structure within a metallicolous population of Arabidopsis halleri, a clonal, self-incompatible and heavy-metal-tolerant species. Mol Ecol 13: doi: /j X x Whiting SN, Reeves RD, Richards D, et al (2004) Research Priorities for Conservation of Metallophyte Biodiversity and their Potential for Restoration and Site Remediation. Restor Ecol 12: Yang XE, Jin XF, Feng Y, Islam E (2005) Molecular mechanisms and genetic basis of heavy metal tolerance/hyperaccumulation in plants. J Integr Plant Biol 47: doi: /j x 177

178

179 Axe 2 Conservation et phytostabilisaton

180

181 Axe 2 -Article 7 Publié - Concilier la phytostabilisation des sols pollués avec la conservation de la flore cupro-cobalticole dans la région de Lubumbashi (R.D. Congo) : une stratégie nouvelle pour valoriser les écosystèmes extrêmes? BOISSON SYLVAIN, COLLIGNON JULIEN, LANGUNU SERGE, LEBRUN JULIE, NGOY SHUTCHA MYLOR & MAHY GRÉGORY Résumé Parallèlement au développement économique, les villes tropicales s étendent depuis la zone urbaine créant une partie distincte, la zone périurbaine. Au Katanga (République Démocratique du Congo), les activités minières ont des impacts non négligeables sur la santé des populations humaines et l envi- ronnement. Cependant, les espèces végétales vivant naturellement sur les gisements cupro-cobaltifères présentent une haute valeur conservatoire et un intérêt potentiel en réhabilitation des sites pollués en éléments traces métalliques. Cette étude vise à évaluer la possibilité de concilier la conservation de deux espèces métallicoles (Crotalaria cobalticola et Anisopappus davyi) avec la phytostabilisation des sites pollués. L établissement des deux espèces a donc été évalué de la graine à la plantule dans une zone périurbaine polluée et un site témoin phytostabilisé par la graminée Microchloa altera. Le taux de levée, la survie et la croissance ont été analysés en fonction de la présence du couvert végétal et du type d amendement (calcaire, matière organique ou aucun). Les résultats soulignent l importance et la nécessité de la couverture végétale dans le succès d établissement des plantules des deux espèces. 181

182 Introduction L Homme et ses activités industrielles transforment drastiquement le paysage et les écosystèmes de la Terre (Vitousek, 1997 ; Comín, 2010 ; Bogaert et al., 2014). Tandis que l urbanisation, l intensification de l agriculture et l industrialisation ont profondément marqué le paysage de l hémisphère nord au 20e siècle, ces changements se produisent maintenant avec le développement de la plupart des pays du Sud. La croissance démographique et économique d une ville dépend de l urbanisation et de l industrialisation en perpétuelle interaction (Quigley, 2009). En s étendant par la périphérie, une ville voit sa zone rurale colonisée par des éléments urbains formant une nouvelle zone aux caractéristiques propres appelée «périurbaine» (André et al., 2014). Cette zone, souvent moins organisée et moins densément peuplée que la zone urbaine, présente des surfaces agricoles et industrielles plus importantes (Ravetz et al., 2013). En Afrique centrale, la ville de Lubumbashi est située au centre de l Arc cuprifère dans la partie katangaise (République Démocratique du Congo). Cette région contient des gisements de cuivre (Cu) et de cobalt (Co) parmi les plus importants au monde associés avec d autres minerais (Cailteux et al., 2005 ; Kampunzu et al., 2009). Ces ressources naturelles ont été exploitées de manière artisanale pendant plusieurs siècles, mais la période coloniale a fortement bouleversé ces activités en industrialisant les processus de production (UMHK, 1956). Notamment, plusieurs sites d exploitation et de transformation ont été créés dans la région au 20e siècle, dont celui de la GECAMINES (Générale des Carrières et des Mines) qui a développé la ville de Lubumbashi (Prasad, 1989). Pendant plusieurs dizaines d années, l extraction chimique des métaux a conduit à la pollution des sols et des eaux environnants par la retombée de particules métallifères à l ouest de la ville (cône de pollution) due aux vents dominants (UMHK, 1956 ; Shutcha et al., 2010). L urbanisation a amené des habitants à s installer dans ces zones polluées situées en périphérie de la zone urbaine où la végétation est fortement fragmentée à cause de la présence d éléments traces métalliques dans l air, l eau et le sol (Manda et al., 2010 ; Shutcha et al., 2010 ; Vranken et al., 2013). Aujourd hui, les habitants et les animaux de ces zones présentent dans leurs tissus des teneurs en cobalt et en cuivre anormalement élevées comparativement aux standards internatio- naux (Banza et al., 2009). Les voies d exposition sont principalement l eau, les particules dans le sol et la consommation de fruits et de légumes contenant des hautes 182

183 teneurs en métaux (Shutcha et al., 2010 ; Mpundu Mubemba Mulambi et al., 2013 ; Cheyns et al., 2014). Afin de remédier aux problèmes d exposition aux métaux à Lubumbashi, une étude récente a proposé d utiliser des végétaux locaux afin de diminuer la dispersion des éléments traces métalliques dans le sol, l eau et l air (Shutcha et al., 2010). Cette technique, la phytostabilisation, présente l avantage d être peu coûteuse, facile d entretien et favorable pour le paysage, contrairement aux méthodes chimiques et physiques (Berti et al., 2000). Elle demande cepen- dant l utilisation d espèces tolérantes à l excès des concentrations en métaux présents dans le sol et ayant, entre autres, une bonne couverture du sol. Sur les affleurements métallifères du Katanga, une flore de plus de 550 espèces végétales, appelées métallophytes, est adaptée à des conditions environnementales extrêmes (Duvigneaud et al., 1963 ; Leteinturier, 2002). Dans cet écosystème, les teneurs disponibles en cuivre et en cobalt du sol dépassent respectivement mg/kg et mg/kg, soit des teneurs environ 100 fois plus élevées que les sols de la région (Leteinturier, 2002). Parmi ces espèces, une cinquantaine est présente dans plus de 75 % des cas sur des sites présentant des sols naturelle- ment enrichis en métaux. Elles sont qualifiées d endémiques (Faucon et al., 2010) et certaines présentent un intérêt scientifique (Chipeng et al., 2009 ; Faucon et al., 2012), par exemple les (hyper)accumulatrices de métaux ( (Leteinturier, 2002 ; Faucon et al., 2009 ; Lange et al., 2014). Néanmoins, l intensification des activités minières d extraction aura, jusqu à la fin de l exploitation, des impacts irrémédiables sur cette végétation unique au monde. Aujourd hui, plus de 65 % des espèces endémiques sont en danger critique d extinction selon les critères de l UICN et 9 % se sont déjà éteintes (Faucon et al., 2010). Dans un effort de conservation de ces espèces, plusieurs études ont mis en évidence les relations entre les communautés végétales, les espèces et les éléments du sol (Saad et al., 2012 ; Ilunga wa Ilunga et al., 2013 ; Séleck et al., 2013). Cependant, aucun test n a été effectué sur l usage des espèces endémiques d intérêt scientifique en stabilisation et en réhabilitation de sols pollués dans un contexte urbanisé. Ce type d approche permettrait d établir une dynamique de développement durable dans les zones périurbaines et polluées d une ville tropicale issue d un contexte minier. 183

184 Dans ce cadre, la présente étude vise à tester une nouvelle approche qui consiste à combiner la phytostabilisation des sols pollués et la conservation d espèces cupro-cobalticoles endémiques et d intérêt scientifique dans une zone périurbaine et polluée de la ville de Lubumbashi (Katanga). Deux espèces d intérêt ont été semées en 2013 dans des parcelles mises en place en 2009, présentant deux types de traitement (couverture végétale; amendement), et situées dans deux sites de la ville de Lubumbashi : un site contrôle et un site localisé dans le cône de pol- lution. Leur germination et leur développement ont été suivis durant une saison de végétation afin d évaluer leur installation dans les sites. Matériel et Méthodes Espèces étudiées Deux espèces métallicoles d intérêt conservatoire des sites cuprifères du Katanga ont été sélectionnées. Crotalaria cobalticola P.A. Duvign. & Plancke (Fabaceae) est une annuelle endémique stricte des sites cuprifères du Katanga dont le statut UICN est «en danger» (EN). Elle a été identifiée comme une hyperaccumulatrice de cobalt avec un maximum d accumulation d environ mg/kg de matière sèche dans les tissus (Duvigneaud et al., 1959 ; Brooks et al., 1987) (Copperflora : Anisopappus davyi S.Moore (Asteraceae) est une pseudo-pérenne identifiée comme une hyperaccumulatrice de cuivre et de cobalt (Brooks et al., 1987) (Copperflora : php?id_e=13). Pour chacune des espèces, trois populations de graines ont été collectées entre 2010 et 2011 dans l Arc cuprifère Katangais entre Tenke et Fungurume puis stockées à la banque de graines de la Faculté des Sciences Agronomiques de l Université de Lubumbashi (Tableau 1). Les populations de graines ont été conservées à une humidité relative de 2-5 % et à température ambiante sauf pour une population d A. davyi qui a été conservée à humidité ambiante et à 5 C. Tableau 1. Caractéristiques des accessions utilisées dans les mélanges de graines des deux espèces. Espèce Accession Site de récolte Date de récolte Date de mise en conservation Mode de conservation HI 100 bis Kabwelunono 15/05/ /06/2010 R Anisopappus davyi HI 166 bis Fungurume IV 11/10/ /10/2010 U HI 298 bis Kakalalwe 01/07/ /08/2011 U HI 105 bis Kavifwafwaulu 19/05/ /06/2010 U Crotalaria cobalticola HI 107 bis Goma 3 20/05/ /06/2010 U HI 237 bis Kakalalwe 12/05/ /07/2011 U R = Réfrigération (T C = 5 C, H.R. ambiante); U = Ultraséchage (T C = 25 C, H.R. < 5 %). Les dates sont exprimées en jj/mm/aaaa 184

185 Sites d étude L étude a été effectuée à Lubumbashi (altitude 1500 m; température : C; précipitations annuelles : 1300 mm, saison des pluies de novembre à avril). L expérimentation a été mise en place dans des zones expérimentales de deux sites aménagés entre 2006 et 2009 (Shutcha et al., 2010; Shutcha et al., 2013). Le dispositif expérimental au sein des deux sites est complètement aléatoire. Le premier site est une zone habitée située à 2 km à l ouest de l usine d extraction de la GECAMINES (11 40 S, E). La zone a été caractérisée comme périurbaine selon la clé élaborée par André et al. (2014). Les teneurs en cuivre dans le sol dépassent mg Cu/kg de sol par endroit dues aux retombées atmosphériques métalliques pendant plusieurs dizaines d années (Shutcha et al., 2010 ; Narendrula et al., 2012). En 2009, 50 parcelles de 1 m2 ont reçu trois modalités d amendements calcaires et organiques croisées (Or0 : pas d amendement orga- nique, Or1 : 4,5 kg compost/m 2, Or2 : 22,5 kg compost/m 2, Ca0 : pas d amendement calcaire, Ca1 : 0,25 kg calcaire/m 2, Ca2 : 0,5 kg calcaire/m 2, Ca3 : 1 kg calcaire/m 2, T : sans amendement) (Figure 1). Cinq nouvelles parcelles non couvertes et sans amendement (T) ont été délimitées dans la continuité des anciennes parcelles. Le second site est le champ expérimental (CE) de la Faculté des Sciences Agronomiques de l Université de Lubumbashi (11 27 S, E). Au total, 32 parcelles de 1 m2 y ont été délimitées et présentent des modalités d amendement similaires au site GECAMINES (Or2 : 22,5 kg compost/m2, Ca3 : 1 kg calcaire/m2, T : sans amendement). Les sols ont été enrichis en cuivre en 2006 à l aide de sulfate de cuivre hydraté (CuSO4.5H2O) afin d obtenir 24 parcelles avec des teneurs d environ mg Cu/kg de sol et 8 parcelles avec des teneurs d environ 500 mg Cu/kg de sol (Figure 1). Au sein des deux sites, 16 éclats de souches/m² de l espèce végétale Microchloa altera (Rendle) Stapf (Copperflora : ont été installés en 2008 avec succès dans chaque parcelle (Figure 10.1) (Shutcha et al., 2010). M. altera est une graminée (Poaceae) pérenne cespiteuse formant des touffes denses d environ 50 cm de hauteur. En février 2013, chaque parcelle était couverte à plus de 90 % par M. altera. 185

186 Figure 1. Dispositif expérimental au site GECAMINES (haut) et au champ expérimental de la Faculté des Sciences Agronomiques de l Université de Lubumbashi (bas). Légende : Or0 : pas d amendement organique, Or1 : 4,5 kg compost/m 2, Or2 : 22,5 kg compost/m 2, Ca0 : pas d amendement calcaire, Ca1 : 0,25 kg calcaire/ m 2, Ca2 : 0,5 kg calcaire/m 2, Ca3 : 1 kg calcaire/m 2, T : sans amendement. Les cases grises correspondent aux parcelles (1 m 2 ) où Crotalaria cobalticola a été semée et les cases blanches correspondent aux parcelles (1 m 2 ) ensemencées avec l espèce Anisopappus davyi. Mise en place et suivi Pour chaque espèce, les trois accessions ont été mélangées de manière égale avant d être semées. Les espèces ont été semées à raison de 30 graines par parcelle le 26/02/2013 suivant une grille de 6 lignes et 5 colonnes. Dans toutes les parcelles sans amendement (T) du site GECAMINES, 15 graines de chaque espèce ont été semées afin de conserver une densité constante dans toutes les parcelles ainsi que le même nombre de répétitions (GECAMINES : n=5 ; CE : n=2). À partir du mois d avril, les parcelles ont été arrosées tous les deux jours avec 75 cl d eau. Les températures et les précipitations journalières ont été mesurées au niveau de l aéroport international de Luano à Lubumbashi (11 35 S, E). 186

187 Mesures et suivi La levée (apparition des cotylédons) et la survie (plantule morte) ont été observées tous les 4 jours pendant 5 semaines puis tous les 7 jours. La hauteur (distance entre le collet et l apex) et le nombre de feuilles de chaque plantule ont été mesurés une fois par semaine à partir du 01/04/2013 jusqu au 08/05/2013 (fin d expérimentation). Les taux de germination (apparition de la radicelle 2 mm) des mélanges de populations utilisées ont été vérifiés en conditions contrôlées à l Unité Biodiversité et Paysage de Gembloux Agro-Bio Tech en mai 2013 (boites de Petri, 25 C, 12 h de photopériode). Traitement des données Les proportions ont été transformées afin de respecter l homoscédasticité, le cas échéant. La quantité de données était insuffisante pour tester la normalité. Les variables de levée et de survie des espèces ont été calculées à l échelle de la parcelle avant les comparaisons. Ces variables ont été comparées via une analyse de la variance à deux critères de classification selon la couverture et l amendement de sol (ANOVA 2). Étant donné le nombre hétérogène de plantules par parcelle, les hauteurs ont été modélisées avec un modèle linéaire mixte avec comme facteurs fixes la couverture et l amendement de sol et, comme facteur aléatoire de groupement, la parcelle. Les moyennes ont été structurées via le test de Tukey ou de la différence franchement significative (HSD) avec un taux d erreur de 5 %. Toutes les analyses ont été effectuées à l aide du logiciel R ( Resultats Taux de levée En conditions contrôlées, les taux de germination des mélanges utilisés étaient de 40,0 ± 17,6 % pour C. cobalticola et de 54,4 ± 6,9 % pour A. davyi. Au champ expérimental, aucun effet de l interaction entre la couverture et les amendements n a été mis en évidence sur le taux de levée de C. cobalticola et de A. davyi. Dans les parcelles avec une couverture végétale, les taux de levée de A. davyi et de C. cobalticola valaient respectivement 14,2 ± 5,6 % et 42,5 ± 23,0 %. Les taux de levée dans les parcelles sans couverture végétale étaient significativement inférieurs aux valeurs dans les parcelles avec 187

188 couverture. Ils vont de 6,2 ± 4,9 % (F = 9,75 ; P < 0,05) pour A. davyi et 25,8 ± 11,6 % (F = 8,23 ; P < 0,05) pour C. cobalticola (Tableau 2). Les modalités d amendement seules n ont pas influencé signifi- cativement la levée de A. davyi contrairement à C. cobalticola qui présentait des taux de levée supérieurs dans les parcelles enrichies à 1000 mg/kg et amendées au compost (50,0 ± 4,7 %), à la chaux (37,5 ± 20,0 %) et non amendées (30,0 ± 4,7 %) par rapport aux parcelles avec 500 mg/kg de cuivre et non amendées (19,2 ± 3,5 %) (F = 4,91 ; P < 0,05) (Figure 2). Au site GECAMINES, les taux de levée dans les parcelles non amendées (T) sans couverture végétale étaient nuls pour les deux espèces. Dans les parcelles non amendées (T) couvertes, les taux de levée étaient de 41,3 ± 22,8 % pour A. davyi et 45,3 ± 22,3 % pour C. cobalticola. Entre les amendements, les taux de levée des deux espèces n étaient pas signi- ficativement différents et valaient en moyenne 50,5 ± 24,4 % pour A. davyi et 38,0 ± 13,2 % pour C. cobalticola. Figure 2. Effet des amendements sur la levée (moyenne ± écart type, %) de Crotalaria cobalticola (F = 4,91 ; P < 0,05) au champ expérimental (CE). 500 et ppm sont respectivement les parcelles avec 500 et 1000 mg Cu/kg. Les moyennes ne partageant pas les mêmes lettres sont significativement différentes (P < 0,05). Légende : T : sans amendement; Ca3 : 1 kg calcaire/m2 ; Or2 : 22,5 kg compost/m2. 188

189 Hauteur (cm) Levée (%) Croissance et survie Au champ expérimental, aucune interaction entre les deux types de traitement (couverture végétale amendement) n a été mise en évidence. Les hauteurs des plantules de C. cobalticola mesurées après 2 mois étaient significativement supérieures (41,7 ± 17,6 mm) dans les sols contenant mg Cu/kg non amendés (T) par rapport aux sols contenant 500 mg Cu/kg non amendés (T) (27,0 ± 11,3 mm). Le taux de survie des plantules de A. davyi était nul un mois après le semis même si sous couverture végétale il est resté supérieur plus longtemps. Pour C. cobalticola, les taux de survie dans toutes les parcelles sont restés supérieurs à 50 % jusqu au 21/03/2013. En fin d expérimentation, on constate un taux de survie plus élevé dans les parcelles couvertes sans être différentes significativement. Au site GECAMINES, l hétérogénéité dans les hauteurs n a pas permis d établir des tests statistiques de comparaison. Pour C. cobalticola, aucune différence significative des hauteurs moyennes entre les différents amendements n a été constatée. Les plantules mesuraient en moyenne 27,9 ± 6,4 mm. La hauteur de C. cobalticola était significativement corrélée avec le nombre de feuilles (R = 0,88 ; P < 0,001). En fin d expérience, les taux moyens de survie de A. davyi et de C. cobalticola étaient respectivement de 3,6 ± 4,4 % et 7,5 ± 8,4 %. Tableau 2. Taux de levée (%) par parcelle (30 graines) et hauteur après deux mois (mm) au champ expérimental de la Faculté des Sciences Agronomiques (Université de Lubumbashi, République Démocratique du Congo), et dans les sols sans amendement (T) du site GECAMINES. Les taux de germination ont été également mesurés en boite de Petri. Les valeurs P indiquent si les deux valeurs précédentes sont significativement différentes (P < 0,05). Légende : ND = pas de données; - = non analysé. Champ expérimental GECAMINES (T) Boite de Petri Couvert Sol nu P Couvert Sol nu P 42,5 ± 23,0 25,8 ± 0,021 45,3 ± 22,3 0,0 ± 0,0-40,0 ± 17,6 C. cobalticola 11,6 A. davyi 14,2 ± 5,6 6,2 ± 4,9 0,014 41,3 ± 22,8 0,0 ± 0,0-54,4 ± 6,9 C. cobalticola 29,14 ± 4,5 37,0 ± 0,001 31,7 ± 9,3 ND - - A. davyi ND ND - ND ND

190 Discussion Cette étude a démontré la nécessité d un couvert végétal dans l installation des deux espèces métallicoles d intérêt dans la région de Lubumbashi. La zone périurbaine polluée (GECAMINES) a été un site propice à la mise en place de nouvelles zones couvertes afin de permettre à la population d être moins exposée aux éléments traces métalliques et de conserver une partie de la biodiversité végétale de la région. Même si cette expérimentation a concerné la partie écologique du problème, cette démarche nécessite également de tenir compte des acteurs sociopolitiques des villes concernées. Leur implication dans ces méthodes peu coûteuses permettrait d établir un plan d aménagement pour cette zone périurbaine, identifiée initialement comme non constructible (Bruneau et al., 1990). La levée était supérieure sous un couvert pour les deux espèces, sans être systématiquement influencée par le type d amendement. En effet, il est difficile de distinguer l effet de la concentration en cuivre et l effet de l amendement sur la levée au champ expérimental. La couverture végétale de l espèce M. altera a probablement permis d améliorer les propriétés physico-chimiques du sol et de sa surface. Cette interaction favorable est définie comme la facilitation (Callaway, 2007 ; Bonanomi et al., 2011). Les touffes de graminées créent et alimentent la litière qui favorise la porosité du sol, l infiltration de l eau et la quantité de nutriments disponibles (Janeau et al., 1999 ; Parraga-Aguado et al., 2013). Le couvert végétal réduit également la température au niveau du sol et l évaporation de l eau offrant des conditions microclimatiques différentes d un sol nu (Maestre et al., 2001). Cet effet a été particulièrement marqué au site GECAMINES qui présentait de vastes zones de sols dénudés extrêmement sensibles à l érosion éolienne et exposées à la chaleur. On constate néanmoins qu en conditions contrôlées, les taux de germination des deux espèces sont faibles comparativement à une étude précédente (Godefroid et al., 2013). La technique de conservation des espèces utilisées dans cette étude est l utraséchage (T C ambiante, H.R. < 5 %), ce qui expliquerait les différences avec l étude menée par Godefroid et al. (2013) sur les graines conservées à -18 C et une H.R. < 5 %. Aussi, les milieux métallifères sont des environnements propices aux processus de spéciation intense qui peuvent conduire à la formation de plusieurs populations soumises à une adaptation locale (Leimu et al., 2008 ; Fones et al., 2010 ; Faucon et al., 2012). Malgré la toxicité du Cu, la germination de certaines espèces cuprophytes semble également être favorisée par la présence de Cu dans le substrat (Chipeng et al., 2009). Dans les parcelles, la levée de C. cobalticola au champ 190

191 expérimental est proche du taux de germination en boites de Petri contrairement à l espèce A. davyi. Cette différence n ayant pas été observée entre le site GECAMINES et les boites de Petri, le mélange des trois populations de A. davyi a probablement été moins homogène au champ expérimental. Au contraire, la présence d un couvert végétal a contraint la croissance des espèces d intérêt dans leurs premiers stades même si elle est restée généralement faible pendant la période du test. Une plus faible croissance pourrait être la conséquence d une entrée en compétition pour la lumière, l eau et les nutriments des espèces testées avec M. altera. Cette tendance de facilitation de l émergence et de la survie opposée à un ralentissement de la croissance sous un couvert végétal a déjà été citée dans plusieurs études (Callaway et al., 1996 ; Suding et al., 1999 ; Walker et al., 1999 ; Foster, 2002). Les amendements ont significativement influencé la hauteur au champ expérimental avec des individus plus développés dans les parcelles les plus enrichies en cuivre et non amendées pour C. cobalticola. Des études antérieures ont effectivement mis en évidence l action favorable des métaux à des concentrations précises sur la croissance chez certains métallophytes (Keller et al., 2004 ; Street et al., 2007). Au site GECAMINES, aucun effet des amendements sur la croissance des deux espèces n a été observé. Córdova et al. (2011) ont également démontré une productivité identique des plantes entre des sols normaux, amendés à la chaux ou au compost. À Lubumbashi, Shutcha et al. (2010) avaient également observé l absence de l influence de la chaux et du compost sur les performances de M. altera. Les taux de survie en fin d expérimentation sont très faibles (< 10 %) en raison du passage vers la saison sèche (mai). La période de transition (avril) a été caractérisée par une diminution de 23 % de l humidité relative de l air et par un arrêt progressif des précipitations. Ces conditions climatiques augmentent l évapotranspiration des jeunes plantules qui sont incapables de résister à ce type de stress (Callaway, 2007). L étude de Shutcha et al. (2010) avait démontré qu un amendement calcaire ou organique était favorable à la survie de plants transplantés. Cet effet semble avoir été inexistant au site GECAMINES durant le stade d installation des plantules. Une prochaine étude devrait débuter au début de la saison des pluies (novembre) et assurer un suivi continu durant toute la saison des pluies afin de reproduire la phénologie complète de ces espèces. 191

192 Conclusion Les couverts phytostabilisants sont des environnements propices à l établissement d espèces à conserver, quelles que soient les conditions d amendements. Ils présentent un réel intérêt dans les grandes villes des pays tropicaux où le développement est intense afin d apporter des zones herbacées de haute biodiversité. L association de la phytostabilisation et de la conservation peut être un succès, à condition de connaître toutes les variables influençant le développement des espèces utilisées. La conservation des métallophytes d intérêt doit passer par ce type de stratégie afin de valoriser les espèces conservées en banque de graines et afin de réhabiliter les sites fortement dégradés. Ces approches recréant de nouvelles communautés végétales demandent encore d être étudiées afin de maximiser les chances d établissement en conditions naturelles et d élargir le nombre d espèces utilisées. Les zones périurbaines polluées de Lubumbashi représenteront sans doute à elles seules les vestiges vivants de la végétation cupro-cobalticole du Katanga. Remerciements Les auteurs remercient les habitants du site GECAMINES de Lubumbashi qui ont veillé au maintien de la parcelle. Cette étude a été soutenue financièrement par les Fonds d Aide à la Mobilité Étudiante (FAME) Fédération Wallonie-Bruxelles qui a permis à Julien Collignon de voyager en République Démocratique du Congo et les Fonds pour la Formation à la Recherche dans l Industrie et dans l Agriculture (FRIA) Fonds de la recherche scientifique (FRS-FNRS) qui financent les recherches doctorales de Sylvain Boisson. Références André M. et al., Vers une synthèse de laconception et une définition des zones dans le gradient urbain-rural. Biotechnol. Agron. Soc. Environ., 18(1), Banza C.L.N. et al., High human exposure to cobalt and other metals in Katanga, a mining area of the Democratic Republic of the Congo. Environ. Res., 109(6), Berti W.R. & Cunningham S.D., Phytostabilization of metals. In: Raskin I., Ensley B.D. (eds). Phytoremediation of toxic metals: Using plants to cleanup the environment. New York: John Wiley & Sons, Inc., Bogaert J., Vranken I. & André M., Biocultural landscapes. In: Hong S.K., Bogaert J., Min Q. (eds). Biocultural landscapes: Diversity, functions and values. Dordrecht, The Netherlands: Springer, Bonanomi G., Incerti G. & Mazzoleni S., Assessing occurrence, specificity, and mechanisms of plant facilitation in terrestrial ecosystems. Plant Ecol., 212(11), Brooks R.R., Naidu S.M., Malaisse F. & Lee J., The elemental content of metallophytes from the copper/cobalt deposits of Central Africa. Bull. Soc. R. Bot. Belg., 119(2), Bruneau J.C. & Pain M., Atlas de Lubumbashi. Nanterre, France : Université Paris X, 133. Cailteux J.L.H., Kampunzu A.B., Lerouge C., Kaputo A.K. & Milesi J.P., Genesis of sediment-hosted stratiform copper cobalt deposits, central African Copperbelt. J. Afr.Earth Sci., 42(1-5), Callaway R.M., Positive interactions and interdependence in plant communities.dordrecht, The Netherlands: Springer. 192

193 Callaway R.M., DeLucia E., Moore D., Nowak R. & Schlesinger W., Competition and facilitation: contrasting effects of Artemisia tridentata on desert vs. montane pines. Ecology, 77(7), Cheyns K. et al., Pathways of human exposure to cobalt in Katanga, a mining area of the D.R. Congo. Sci. Total Environ., 490, Chipeng F.K. et al., Copper tolerance in the cuprophyte Haumaniastrum katangense(s. Moore) P.A. Duvign. & Plancke. Plant Soil, 328(1-2), Comín F.A., Ecological restoration: a global challenge. Cambridge, UK: Cambridge University Press, 291 p. Córdova S., Neaman A., González I., Ginocchio R. & Fine P., The effect of lime and compost amendments on the potential for the revegetation of metal-polluted, acidic soils. Geoderma, 166(1), Duvigneaud P. & Denaeyer-De Smet S., Cuivre et végétation au Katanga [Copper and vegetation in Katanga]. Travaux du Centre scientifique et médical de l Université Libre de Bruxelles en Afrique centrale. Bull. Soc. R. Bot. Belg., 96, Duvigneaud P. & Timperman J., Études sur le genre Crotalaria. Bull. Soc. R. Bot. Belg., 91(2), Faucon M.P. et al., Soil influence on Cu and Co uptake and plant size in the cuprophytes Crepidorhopalon perennis and C. tenuis (Scrophulariaceae) in SC Africa. Plant Soil, 317(1-2), Faucon M.P. et al., Copper endemism in the Congolese flora: a database of copper affinity and conservational value of cuprophytes. Plant Ecol. Evol., 143(1), Faucon M.P. et al., Copper tolerance and accumulation in two cuprophytes of South Central Africa: Crepidorhopalon perennis and C. tenuis (Linderniaceae). Environ. Exp. Bot., 84, Fones H. et al., Metal hyperaccumulation armors plants against disease. PLoS Pathog., 6(9), e Foster B., Competition, facilitation, and the distribution of Schizachyrium scopariumalong a topographic-productivity gradient. Ecoscience, 3, Godefroid S. et al., Germination capacity and seed storage behaviour of threatened metallophytes from the Katanga copper belt (D.R. Congo): implications for ex situ conservation. Plant Ecol. Evol., 146(2), Ilunga wa Ilunga E., Séleck M., Colinet G., Meerts P. & Mahy G., Small-scale diversity of plant communities and distribution of species niches on a copper rock outcrop in Upper Katanga, D.R. Congo. Plant Ecol. Evol., 146(2), Janeau J.L., Mauchamp A. & Tarin G., The soil surface characteristics of vegetation stripes in Northern Mexico and their influences on the system hydrodynamics: an experimental approach. CATENA, 37(1-2), Kampunzu A.B., Cailteux J.L.H., Kamona A.F., Intiomale M.M. & Melcher F., Sediment-hosted Zn Pb Cu deposits in the Central African Copperbelt. Ore Geol. Rev. 35(3-4), Keller C. & Hammer D., Metal availability and soil toxicity after repeated croppings of Thlaspi caerulescens in metal contaminated soils. Environ. Pollut., 131(2), Lange B. et al., Prediction of the edaphic factors influence upon the copper and cobalt accumulation in two metallophytes using copper and cobalt speciation in soils. Plant Soil, 379(1-2), Leimu R. & Fischer M., A meta-analysis of local adaptation in plants. PLoS One, 3(12), e4010. Leteinturier B., Évaluation du potential phytocénotique des gisements cupriferes d Afrique centro-australe en vue de la phytoremédiation de sites pollués par l activité. Thèse de doctorat : Faculté des Sciences Agronomiques de Gembloux (Belgique). Maestre F., Bautista S., Cortina J. & Bellot J., Potential for using facilitation by grasses to establish shrubs on a semiarid degraded steppe. Ecol. Appl., 11(6), Manda B., Colinet G. & André L., Évaluation de la contamination de la chaîne trophique par les éléments traces (Cu, Co, Zn, Pb, Cd, U, V et As) dans le bassin de la Lufira supérieure (Katanga). Tropicultura, 28(4), Mpundu Mubemba Mulambi M. et al., Évaluation des teneurs en éléments traces métalliques dans les légumes feuilles vendus dans les différents marchés de la zone minière de Lubumbashi. J. Appl. Biosci., 66, Narendrula R., Nkongolo K. & Beckett P., Comparative soil metal analyses in Sudbury (Ontario, Canada) and Lubumbashi (Katanga, D.R. Congo). Bull. Environ. Contam. Toxicol., 88(2), Parraga-Aguado I. et al., The importance of edaphic niches and pioneer plant species succession for the phytomanagement of mine tailings. Environ. Pollut, 176, Prasad M.S., Production of copper and cobalt at Gecamines, Zaire. Miner. Eng. 2(4), Quigley J.M., Urbanization agglomeration and economic development. In: Spence M., Annez P.C. & Buckley R.M. (eds). Urbanization and growth, Washington, DC.: World Bank E-Library. Commission on Growth and Development, Ravetz J., Fertner C. & Nielsen T.S., The dynamics of peri-urbanization. In: NilssonK. et al. (eds). Peri-urban futures: Scenarios and models for land use change in Europe. Heidelberg, Germany: Springer, Saad L. et al., Investigating the vegetation-soil relationships on the copper-cobalt rock outcrops of Katanga (D.R. Congo), an essential step in a biodiversity conservation plan. Restor. Ecol., 20(3), Séleck M. et al., Chemical soil factors influencing plant assemblages along copper-cobalt gradients: 193

194 implications for conservation and restoration. Plant Soil, 373(1/2), Shutcha M., Meerts P., Kazadi S., Faucon M.P. & Visser M., Low caryopsis production of the toothbrush grass (Microchloa altera) from Katanga (D.R. Congo) could limit the revegetation of trace metal contaminated lands by seeding. Restor. Ecol., 31(3), Shutcha M.N. et al., Phytostabilisation of coppercontaminated soil in Katanga: an experiment with three native grasses and two amendments. Int. J. Phytoremediation, 12(6), Street R., Kulkarni M.G., Stirk W., Southway C. & Van Staden J., Toxicity of metal elements on germination and seedling growth of widely used medicinal plants belonging to Hyacinthaceae. Bull. Environ. Contam. Toxicol., 79(4), Suding K.N. & Goldberg D.E., Variation in the effects of vegetation and litter on recruitment across productivity gradients. J. Ecol. 87(3), UMHK, Union Minière du Haut Katanga , Bruxelles : Éd. Cuypers L., 283 p. Vitousek P.M., Human domination of Earth s ecosystems. Science, 277(5325), Vranken I. et al., The spatial footprint of the non-ferrous mining industry in Lubumbashi. Tropicultura, 31(1), Walker L. & Powell E., Regeneration of the Mauna Kea silversword Argyroxiphium sandwicense (Asteraceae) in Hawaii. Biol. Conserv., 89(1),

195 Axe 2 -Article 8 En preparation - Using phytostabilisation as a way to conserve threatened endemic species from the Southeastern D.R. Congo BOISSON Sylvain, LE STRADIC Soizig, SÉLECK Maxime, COLLIGNON Julien, GARIN Olivier, NGOY SHUTCHA Mylor, MAHY Grégory Abstract Outcrops of the southeastern D.R. Congo are recognized as some of the largest copper-cobalt bodies of the world. They support unique vegetation with near than 600 metallophytes including rare and endemic species. Since 1900s, the exploitation of the copper outcrops in D.R. Congo has considerably increased, impacting environment and public health. Phytostabilisation of polluted areas represents a suitable technology to decrease the bioavailability of heavy metals in highly polluted soils. This technic has been successfully implemented in the vicinity of Lubumbashi with the grass Microchloa altera. The long term maintenance requires a good understanding of the interspecific relationship like competition and facilitation. This study aims to test the success of establishment of four species from the Katangan Copperbelt by also assessing the potential role of M. altera as nurse plant. Two annual species and two perennial species were sown in experimental design with soil amendments crossed with vegetation cover. The emergence, the growth and the survival were followed during the vegetation season and the resprout was measured for perennial species. M. altera had a distinct effect on the emergence and the survival of the annual and perennial species but affected negatively the growth of individuals for both species. 195

196 Introduction Copper outcrops are widespread all over the world (Prasad, 1989; Shallari et al., 1998; Gonnelli et al., 2001; Tembo et al., 2006; Ke et al., 2007; Lorestani et al., 2011; Saad et al., 2012; Battogtokh et al., 2013). Southeastern D. R. Congo hosts some of the largest ore bodies of copper (Cu) and cobalt (Co) in the world that are scattered over more than 300 km long and 50 km large, named the Katangan Copperbelt (Duvigneaud et al., 1963a; Cailteux et al., 2005). These outcrops were formed before the Quaternary and the mineralized zone of the Roan Series were made up during Mines Series (François, 1973). Sediments covered partially Cu-Co outcrops that lead to the formation copper hills having soils with increasing depths from the summit to the bottom of the outcrop. These soils are naturally rich in copper and cobalt with highest metal concentrations at the hill summit (Duvigneaud et al., 1963a). Those copper hills support an unique biodiversity with near than 600 metallophytes occurring in different plant communities determined by edaphic factors like Cu, Co and ph (Leteinturier, 2002; Saad et al., 2012; Séleck, Bizoux, et al., 2013). Two plant communities have been identified according to the copper and the cobalt concentrations: the steppic savannah, on the bases and the slopes of the hills (less than 3500 mg kg -1 of available Cu in soil and 50 mg kg -1 of available Co), and, the steppe, on soils highly enriched in Cu and Co (from mg kg -1 to 10,000 mg kg -1 of available Cu in soils and from 50 mg kg -1 to 1,000 mg kg -1 of available Co in soils) and located on the top of a Cu hill (Séleck et al., 2013). Among all recorded taxa, 56 have been identified as copper endemic including 44 as threatened of extinction (i.e., VU, EN, CR) considering IUCN criterions (Faucon et al., 2010). In situ and ex situ conservation strategies of species have already been implemented with a mining company but they focused on small to medium scale conservation areas (Semereab et al., 2009; Le Stradic et al., 2015). Mining constitutes one of several sources of the spread of heavy metal in soils (Li et al., 2001; Järup, 2003; Blacksmith Institute, 2007; Narendrula et al., 2012; Sharma et al., 2014; Sherameti et al., 2015). Southeastern D. R. Congo, the mining industry was initiated in early 1900s in the city of Lubumbashi (previously Elisabethville). Copper and Cobalt were produced in several hydrometallurgical plants and smelters, which also produced zinc, arsenic, and cadmium as by-products (Prasad, 1989). These activities increased the concentration of metal input in environment, generating unfavourable physico-chemical soil 196

197 conditions, limiting vegetation establishment and conducts to bare grounds and/or water pollution that put health of human populations at risk (Banza et al., 2009; Manda et al., 2010; Shutcha et al., 2010; Cheyns et al., 2014). Phytostabilisation of polluted areas represents a suitable technology to decrease the bioavailability of heavy metals in highly polluted soils (Berti et al., 2000). While this method allows to reduce significantly human or animal exposure to heavy metals (Zhang et al., 2010; Kacprzak et al., 2014; Shutcha et al., 2015), large scale experiments often lead to the establishment of monospecific communities. Actually the criterions of the species selection (i.e., high biomass production, dense root systems or large quantity of propagules) restrict greatly the choice to few species (Rizzi et al., 2004; Mench et al., 2006; O Dell et al., 2006; Mendez et al., 2008; Shutcha et al., 2010; Parra et al., 2014; Boisson et al., 2015). Facilitation is a positive interaction assembling ecological communities and is recognized as an important factor conditioning the success of the phytoremediation strategies (Frérot et al., 2006; Brooker et al., 2008; Parraga-Aguado et al., 2013; Wang et al., 2014). It occurs when a plant species (i.e., nurse plant) improve the establishment of other species by changing the micro-climatic surface conditions and the nutrient content (Brooker et al., 2008; Soliveres et al., 2010; Bonanomi et al., 2011; Beltrán et al., 2012). Identification of nurse plants is important to improve the germination, the growth and the survival of different plant species in heavy metal polluted environments (Frérot et al., 2006; Padilla et al., 2006; Parraga-Aguado et al., 2013; Wang et al., 2014). While conservation of threatened endemic or rare species failed, their integration in project as phytostabilisation became a new challenge officially recognized by the Global Strategy for Plant Conservation (Secretariat of the Convention on Biological Diversity, 2014). Considering the growing awareness about the biodiversity in the mining practices, the threatened metallophyte from the southeastern D.R.C. need to be associated with rehabilitation or phytostabilisation strategies (Whiting et al., 2004; International Council on Mining and Minerals (ICMM), 2006; Baker et al., 2010; Faucon et al., 2010, 2011). Some tolerant grasses have been reported as candidates for phytostabilisation programs (Shutcha et al., 2010; Boisson et al., 2015) and, Microchloa altera has been successfully tested in situ in the vicinity of Lubumbashi (Shutcha et al., 2015). This species can constitute the basis to start 197

198 experiments in order to create a new dynamic of species colonization in phytostabilized areas. The present work aimed to test the success of establishment of four species from the Katangan Copperbelt with high conservation value, Anisopappus davyi, Crotalaria cobalticola, Crotalaria peschiana, Triumfetta welwitschii var. rogersii, in polluted areas in the vicinity of Lubumbashi. Also, the potential role of Microchloa altera as nurse plant was evaluated by comparing the emergence and the growth of the studied species between covered and noncovered parcels. Methods Plant material and seed collection The four studied taxa occur in southeastern D. R. Congo, particularly, in the Katangan Copperbelt (Leteinturier, 2002; Faucon et al., 2010). Anisopappus davyi S.Moore (Asteraceae) is a pseudo-annual species occurring in natural steppes with copper-rich soils and in disturbed soils. It is not an endemic species but it has been recognized as a hyperaccumulator and an indicator of copper in soils (Faucon et al., 2010). The flowering period occurs between January and April and the fruiting time ranges between April and August. The species grows slowly and has a branched root system of 10 to 15 cm depth at adult stage. Crotalaria cobalticola P.A.Duvign. & Plancke (Fabaceae) is an annual species. The species is present in Cu-Co outcrop steppe. This strict endemic species was proposed as endangered (EN) by Faucon (2010) according to the IUCN classification. Flowers appear in April-May and fruits mature shortly after, between May and June. It has a branched root system of 10 cm depth at adult stage. Crotalaria peschiana P.A. Duvign. & Timp. (Fabaceae) is a perennial species. It is also a strict endemic species, proposed as critically endangered (CR), which occurs in the steppic savannah of Cu-Co outcrop (i.e. the lower part of the outcrop) in the Katangan Copperbelt (Duvigneaud et al., 1959, Faucon et al., 2010). Main flowering occurs during the dry season in August-September, and seeds are produced shortly after. Second flowering (few-flowered) can occur at the end of the growing season (i.e., April- May). Root system is a xylopodia (i.e., tuberous roots). Triumfetta welwitschii var. rogersii (N.E.Br.) Brummitt & Seyani (Malvaceae) is a perennial suffrutex being identified as a strict endemic species of the Katangan Copperbelt but it has been classified as data deficient (proposed IUCN status: DD). This taxon is found in both the steppe and the steppic savannah 198

199 of Cu-Co outcrops. Flowers appear after the fires, usually at the end of the dry season, in September-October. Fruits mature shortly after and co-occur with the flowers. It produces a deep and woody tap-root. Seeds of each taxon were taken from three seed samples from the seedbank of the University of Lubumbashi (D.R. Congo) excepted for C. peschiana for which only one population was selected because the species has been reported in too few sites to allow collection of numerous seeds (Faucon et al., 2010). During seed collection, a minimum of 50 mature and well-developed individuals was targeted to obtain numerous mature seeds and conserved in distinct conditions (Table 1), as the four species present different conservation purposes (i.e. long term conservation for A. davyi and C. cobalticola and short term conservation for C. peschiana and T. welwitschii). A. davyi and C. cobalticola were stored at ultradrying (i.e. silicagel) and 25 C while C. peschiana and T. welwitschii were stored at 5 C to use their seeds in propagation programs. Seedlots of each species were constituted before experiment by pooling equitably the seeds from different seed samples. Table 1. General description of studied species. H: hyperaccumulator, SE: strict endemic Conser vation status Proposed IUCN status. Life form A. davyi H - Pseudoannual Root system type Fruiting time Collecting sites Number of individu al Branched April-August Fungurume IV Kabwelunono Kakalalwe 1 C. cobalticola SE CR Annual Branched May-June Goma 3 C. peschiana SE CR Perennial August- September T. welwitschii SE DD Perennial Taproot September- October Kakalalwe Kavifwafwaulu Kazinyanga 75 Fungurume VIII Kabwelunono Kavifwafwaulu

200 Study site The study was performed in Lubumbashi located in Southeastern Democratic Republic of Congo at the extreme south of the Katangan Copperbelt in altitude around m. This area has a subtropical climate with a rainy season extending from November to the end of March and a dry season from April to October. This study was based on the experiment of phytostabilisation performed by Shutcha et al. (2010). The work was performed in semi-controlled conditions in the experimental garden of the University of Lubumbashi (11 27 S, E). The area contained plots artificially enriched by copper in the experimental field of the University of Lubumbashi since 2008 (Shutcha et al., 2010). There are 24 1m 2 plots of 1m 2 having soils enriched with copper sulphate (around 1,000 mg kg -1 ) with three different modalities of amendment (i.e., with lime, with organic matter or unamended) crossed with two modalities of vegetation cover (with or without vegetation cover). In the plots with amendments, 1 kg of limestone or 22.5 kg of organic matter was mixed with the first 15 cm-soil in In order to limit the border effect, all plots with the same modality of vegetation cover were spatially grouped. The vegetation cover consisted of transplanted individuals of Microchloa atlera (Rendle) Stapf (Poaceae) in It is a caespitose perennial grass forming compact tufts. This species is frequent on metal disturbed sites in southeastern D.R.Congo where it is identified as the first coloniser of disturbed soils by mining activities and can form monospecific stands (Duvigneaud et al., 1963a; Shutcha et al., 2010). Experiment in controlled conditions Two distinct experimentations were performed (Figure 1). For both annual (Year 1, Figure 1) and perennial (Year 2, Figure 1) species, 30 seeds of one species were sown by 1 m 2 plot at one centimetre depth according to a grid of six lines and five columns. A border of minimum 10 cm was considered on each side. Soils were slightly ploughed at the location of seeding in order to create favourable germination conditions and to avoid the seed runoff. Two replications by species were performed. The first year (Year 1), annual species, A. davyi and C. cobalticola were sown in February The percentage of emergence and the survival were measured once a week until the last week of April The height of each individual and the number of leaves were 200

201 measured at the end of the growing season between the 15 th and the 18 th April. The second year (Year 2), perennial species, C. peschiana and T. welwitschii, were sown in November 2013 and the emergence and survival were followed from February 2014 to March The height and the number of leaves were measured between the 15 th and the 18 th April 2014 (Figure 1). For perennial species only, we measured the resprouting and the growth (number of leaves and the height) at the end of the growing season in April 2015 (Year 3). February May November February May November February May Year Year 1 1 Year 2 Year 3 End of the study Sowing of annual species Sowing of perennial species Vegetative measure on annual species Vegetative measures on perennial species Resprouting and vegetative measures on perennial species Period of measures Growing season Transition and dry season Figure 1. Planning of the study performed between 2013 and 2015 in the experimental controlled conditions in the experimental garden of the Faculty of Agronomy (University of Lubumbashi, DRC) Experiment in in situ conditions The four tested species were also sown in a polluted area located in the vicinity of Lubumbashi in order to simulate a practical case. The area was contaminated by the deposition of metalliferous fallout during forty years from the copper smelter (Gécamines) located at 2 km west (Prasad, 1989; Shutcha et al., 2015). This site has been populated despite the elevated concentrations of copper, cobalt and aluminium in soil (Faucon, Chipeng, et al., 2012; Narendrula et al., 2012; Shutcha et al., 2015). Phytostabilisation test has been implemented in 2008 with four modalities of limestone amendments (control, 0.25, 0.5 or 1 kg m -2 ) crossed with three modalities of organic matter (control, 4.5 or 22.5 kg m -2 ) (Shutcha et al., 2015). Each 1m 2 plots received 16 transplanted individuals of M. altera in order to establish a vegetation cover. Before our experimentation, M. altera successfully colonized all the parcels representing a total surface of 72 m 2. In 2014, the Cu and Ca concentration, the ph and the organic matter proportion (C ) were not different between the initial treatments: 2,826 ± 2,230 mg kg -1 Cu, ± 53.7 mg 100g -1 Ca, 13.7 ±

202 of organic carbon and a ph KCl of 6.5 ± 0.7. The four species were sown using the equivalent method than experiment in controlled condition, using a grid of 30 seeds. In total, 720 seeds of each species were sown in 60 plots order to assess their establishment. The emergence, the survival at the end of the first growing season and the resprouting were recorded after one dry season (March 2015). Considering that the edaphic conditions were homogenous, no statistical analyses were performed. Soil analysis Samples were taken from two composite sub-sample at the first 15 cm soil in Prior to physico-chemical analyses, soil samples were air-dried and sieved to 2 mm. First, the ph was measured in 1N KCl with a glass electrode in a 2:5 soil:solution ratio after 2h equilibration time. Then, total organic and nitrogen content were measured in soil after dry combustion. Carbon and nitrogen were oxidized to CO 2, NO x and N 2 by heating the soil to at least 900 C in a flow of oxygen-containing gas. The amount of CO 2 released was then measured by Gas Chromatography (GC) method the nitrogen content was measured by means of thermal conductivity detection (Margesin et al., 2005). Finally, soil extractable concentrations of Cu and Mn were measured using 0.5N CH3COONH 4 EDTA (Lakanen et al., 1971). The soil/solution ratio was fixed to 1:5. For the extraction solution, the ph was buffered at 4.65 (Kucak et al., 1998; Faucon et al., 2009; Saad et al., 2012). The supernatant was filtered through a S&S 595 folded filter and analysed using a flame atomic absorption spectrometer (Varian 220), following the norm NF X Data treatment Kruskal-Wallis rank sum tests followed by Bonferroni tests were performed to compare the nutrient content and metal concentrations. The percentage of emergence was calculated as the ratio between the maximum of emergence (for the annual species) or the maximum number of individuals during February (for the perennial species) and the total sown seeds. The percentage of survival was the ratio between the number of seedlings occurring at the end of the growing season (April 2013 for annual species and April 2014 for perennial species) and the maximum number of emergences. These two parameters were calculated by species and by plot. They were transformed into arcsin of square formula prior to compare them in a two-way ANOVA (vegetation cover x amendments), when normality and homoscedasticity were not verified. Height and the number of leaves were compared using 202

203 linear mixed effect model (LME) with the 1m 2 plot as grouping factor. Then, p-value was determined with a likelihood test, which consist in model comparison. When data dit not present heteroscedasticity and normality, they were log or square-root transformed. For the two perennial species, the percentage of respouting individuals was calculated by the ratio between the number of resprouts and the total number of individuals at the end the growing season For this analysis, treatments were combined before to perform LME. Means were structured by Tukey's range test (HSD) with an error rate of 5%. Analyses were carried out in R software version (R Development Core Team, 2010). Results Experiment in controlled conditions: Soils comparison As expected, amendments did not induce differences in available copper concentrations (1,034 ± 248 mg kg -1 Cu, X 2 =2.66, p-value = 0.26, Table 2) but lead to distinct calcium concentrations in soil (Table 2). The soil enriched with lime had higher Ca concentrations (34.2 ± 10.4 mg 100g -1 Ca, X 2 =12.7, p-value < 0.01) than soils enriched with the organic matter (12.6 ± 4.9 mg 100g -1 Ca) and unamended soils (17.2 ± 11.4 mg 100g -1 Ca). The ph was also higher in soil enriched with lime (4.6 ± 0.2, X 2 = 9.06, p-value < 0.05, Table 2). Finally, the organic matter and the total nitrogen content did not vary among soils. According to these edaphic element contents, two kinds of soils can be distinguished: the soil with lime and the soil with organic matter or unamended. However, in order to conform to the test of Shutcha et al. (2010, 2015), the three categories were kept in the following results. Table 2. Edaphic content in copper (Cu, mg kg-1), ph, calcium (Ca, mg 100g -1 ), organic matter (Corg, %) and nitrogen (Ntot, %). Means were compared using a Kruskal-Wallis test. Lime O.M. UA X 2 p-value Cu (mg kg -1 ) 1040 ± 146 a 939 ± 297 a 1122 ± 272 a ph KCl 4.6 ± 0.2 a 4.2 ± 0.1 b 4.3 ± 0.2 b 9.06 < 0.05 Ca (mg 100g -1 ) 34.2 ± 10.4 a 12.6 ± 4.9 b 17.2 ± 11.4 b 12.7 < 0.01 C org ( ) 13.8 ± ± ± N tot (%) ± ± ±

204 Experiment in controlled conditions: Emergence, survival and growth of annual species (Year 1) Mean percentages of emergence of these annual species during the first test (Year 1) were 14.2 ± 5.6 % for A. davyi and 34.2 ± 18.5 % for C. coblaticola. However, the presence of a vegetation cover of M. altera increased greatly the number of emergences for both species in (Table 3). Percentages of emergence of A. davyi in presence of vegetation cover were 16.1 ± 4.9 % contrary to 5.0 ± 2.8 % without cover (F= 19.0, p-value < 0.01) and for C. cobalticola, the percentage of emergence was 49.4 ± 19.6 % under vegetation which was significantly greater than the percentage without cover (28.8 ± 11.9 %, F= 7.09, p-value < 0.05, Table 3). No significant effect of the interaction or the amendments on the percentage of emergence was observed for the two studied species. At the end of the growing season of the Year 1, all individuals of A. davyi died whereas C. cobalticola had higher percentage of survival in soil amended with lime (53.8 ± 24.9 %) than other soils (F = 4.71, p-value = 0.06, Table 4). Due to the low percentage of emergence and the absence of seedling survival in the test, the growth of the species A. davyi was not compared to the other species. For C. cobalticola the vegetation cover had an impact on height and number of leaves of seedlings (F height = 4.59, p- value < 0.05; F leaves = 11.8, p-value < 0.01). The seedlings were higher and had more leaves in plot without cover (2.9 ± 1.1 cm; 3.6 ± 1.0 leaves) than in plot with a vegetation cover (2.5 ± 0.7 cm; 2.4 ± 0.8 leaves). Table 3. Percentage of emergence (%) of the four studied species among the amendment and the vegetation cover treatment. Data were analysed by a two-way ANOVA followed by a Tukey HSD range test. A. davyi 10.8 ± 8.8 C. cobalticola C. peschiana T. welwitschii Lime O.M. UA F 37.5 ± ± ± ± ± ± ± 3.2 Interactions were not significant 10.8 ± ± ± ± p- value With vegetation cover Without cover ± 4.9 a 5.0 ± 2.8 b ± 19.6 a 28.8 ± b ± 8.3 b 26.7 ± a ± ± F p- value < 0.01 <

205 Experiment in controlled conditions: Emergence, survival and growth of perennial species (Year 2) C. peschiana and T. welwitschii, the two perennial species, recorded very low percentages of emergence the second test (Year 2) were 20.5 ± 5.7 % for C. peschiana and 7.5 ± 1.8 % for T. welwitschii. However, the presence of a vegetation cover affected marginally the percentage of emergence of C. peschiana (Table 3), which was slightly lower in presence of vegetation cover (11.1 ± 8.3 %) than without cover (26.7 ± 10.5 %, F= 4.89, p-value = 0.07). Amendment or interaction between amendment and plant cover did not impact the number of emergence of these taxa. At the end of the growing season of Year 2, C. peschiana and T. welwitschii presented high survival percentage 82.1 ± 23.2 % and 92.7 ± 25.1 % respectively (Table 4). A nearly significant effect of the vegetation cover was observed for C. peschiana (F = 3.86, p-value = 0.09), with a greater percentage of survival in plot without cover (91.7 ± 20.4 %, Table 4). For C. peschiana, interaction between the vegetation cover and the amendment affected the seedling height (F = 3.66, p-value < 0.05). Analyses revealed that the height of the individuals of C. peschiana were significantly greater in unamended plots without vegetation cover (4.4 ± 2.0 cm) than the plot amended with lime (2.3 ± 1.2 cm) and organic matter (1.7 ± 1.1 cm) without vegetation cover. However, the vegetation cover had a significant effect on the number of leaves (L ratio = 5.47, p-value < 0.05), with a higher value 10 ± 7 leaves in plots without vegetation cover than in plots with vegetation cover (7 ± 3 leaves). Finally, height of individuals of T. welwitschii was significantly higher in plots without vegetation cover than in plots with vegetation cover (F = 5.63, p-value < 0.05). Table 4. Percentage of survival (%) of the four studied species among the amendment and the vegetation cover treatment. Data were analysed by a two-way ANOVA followed by a Tukey HSD range test. Lime O.M. UA F p- value With vegetat ion cover Without cover A. davyi C. cobalticola 53.8 ± 24.2 ± 14.6 ± ± 24.9 a 9.9 ab 13.8 b 20.0 C. peschiana 87.5 ± 25.0 T. welwitschii 95.8 ± 8.3 Interactions were not significant 73.8 ± ± ± ± ± ± ± ± ± 40.0 F p- value

206 Experiment in controlled conditions: Resprout and growth after the dry season (Year 3) For C. peschiana, the soils with lime presented greater percentage of resprout (respectively 85.7 ± 28.6 % and 53.3 ± 21.0 %) than the plots with organic matter (8.3 ± 16.7 %) and unamended (12.9 ± 17.7 %, F = 15.5, p-value < 0.01). The percentage of resproutings of T. welwitschii was also affected by the amendments (F = 5.79, p-value < 0.05) with higher values in the unamended plots (87.5 ± 25.0 %) compared to the soil with lime (10.0 ± 20.0 %) and organic matter (25.0 ± 50.0 %). No significant differences of the percentage of resprouting were highlighted between the vegetation cover for both species. Crotalaria peschiana was the only species to present individual growth variations according to soil and vegetation cover. Their height and their number of leaves were significantly influenced by the combination of the amendments and the vegetation cover treatments (height: L ratio = 9.99, p-value < 0.05, number of leaves: L ratio = 19.1, p-value < 0.001). The unamended plots without cover had more developed individuals (19.5 ± 6.2 cm, 14 ± 7 leaves) than the amended soil with lime without cover (7.5 ± 5.1 cm, 7 ± 3 leaves). In situ experiment In in situ conditions, in February 2013 (Year 1), percentage of emergence of C. cobalticola was 36.8 % and A. davyi was 37.2 % that are respectively lower and higher than the results found in the experiment in controlled conditions. At the end of the growing season of Year 1, the percentage of survival of A. davyi was 7.5 % and 13.6 % for C. cobalticola. Compared with the experiment in controlled conditions, the percentage of survival of A. davyi was higher while the percentage of survival of C. cobalticola was more than twice less. For perennial species, the percentage of emergence in February 2014 (Year 2) was 11.7 % for C. peschiana and 11.5 % for T. welwitschii, similarly than the experiment in controlled conditions. The percentage of survival of C. peschiana was 21.4 % and 64.3 % for T. welwitschii at the end of the first growing season, both being lower than the experiment in controlled conditions. In March 2015, the percentage of resprouting was 11.1 % (2 individuals) for C. peschiana and 68.5 % for T. welwitschii (37 individuals). 206

207 Discussion Results highlighted that combining the phytostabilisation of the polluted and the conservation of endemics is complex but feasible in a tropical context. Germination is the first step of the vegetation establishment and it is a crucial step in phytostabilisation. The results highlighted that the vegetation cover of M. altera had a greater positive effect on the emergence of the annual species, in contrast to the perennial species. Percentage of germination of A. davyi and C. cobalticola are approximately twice and more under the vegetation cover. M. altera acted as a nurse plant facilitating the establishment of annual species (Brooker et al., 2008; Frérot et al., 2006). This cover change the above-ground microclimatic conditions reducing probably the water run-off (Levine, 2013), buffering the extreme temperature (R.M. Callaway and Callaway, 2007) and increase the shading (Bader et al., 2007). This facilitation mechanism is the most reported mechanism in tropical ecosystems followed by the increase of soil fertility and associational refuge (Bonanomi et al., 2011). However, the higher germination percentages of these species in Petri dishes (data not shown) suggest that these conditions were not still optimal. M. altera creates an important litter that could create heterogeneous conditions of shading being greater than the natural communities of the copper hills (Callaway and Lawrence, 1997; Donath et al., 2007; Ilunga wa Ilunga et al., 2015). The pattern was almost opposite for the perennial species suggesting that the life forms could involve the pattern. Our observations do not support the results of the review of the terrestrial ecosystems indicating that the assistance was mainly provided to individuals of the same growth form (Bonanomi et al., 2011). Both annual and perennial species presented lower growth under the vegetation cover of M. altera signifying that competition appeared to be higher than facilitation. Considering that M. altera is a cespitose species forming dense tuft of 60 cm high, the species could enter in competition for light and resources when first leave appears. In natural communities, A. davyi and C. crotalaria are found mostly in the steppe characterized by low and open vegetation (Ilunga wa Ilunga et al., 2015). Even as the annual species tend to have a greater growth rate than the perennial (Garnier, 1992), the balance between competition and facilitation depends on several factors such as stress and resource gradients (Brooker et al., 2008). The survival of species was relatively low and inexistent for A. davyi. Furthermore, the cover of M. altera did not affect the percentage of survival of taxa, excepted for C. peschiana that was higher than other treatment. However, no flower were measured on the annual species 207

208 suggesting that seeds should be established just before the rainy season or that seedling should be transplanted in the field in order to complete the life cycle. The translocation is an efficient technics but need more labor than seed dispersion (Mench et al., 2006). During the Year 3, t some individuals of C. peschiana in non-covered plots produced flowers (data not shown) in the field in controlled conditions. The soil amendment slightly affected the survival of the annual species C. cobalticola with better results in soils amended with lime. The resprout and the growth of the perennial species C. peschiana after one dry season (Year 3) were also positively affected the calcium concentrations. Calcium have a significant effect on the plant physiology (Jones and Lunt, 1967; Rengel, 1992) and on heavy metal availability (Parra et al., 2014; Remon et al., 2005). Cordova (2011) showed that the combination between the lime and the organic matter increase the plant productivity ion metalliferous soils. Even if calcium concentrations and ph still presented difference with original amendments (Shutcha et al., 2010)., the organic matter content has been stabilized in all treatment at an average 14. However, T. welwitschii showed better value of resprouting in unamended soils. This could be explained by the difference of tolerance to copper of both taxa. T. welwitschii occurs in highest copper concentrations than C. peschiana. Furthermore, the percentage of survival and resprout of C. peschiana in the in situ experiment (max mg kg -1 Cu) were lower than the experiment in controlled conditions (max mg kg -1 Cu). T. welwitschii had similar mean values between both sites. While M. altera presents advantages for phytostabilisation of polluted soils because it forms dense cover (Shutcha et al., 2015, 2010), strong evidence exists about the positive effect of species association in polluted soils (Frérot et al., 2006; Padilla and Pugnaire, 2006). These interactions should consider plant having distinct species traits (Ilunga wa Ilunga et al., 2015). Mine bare soils are generally low in nutrients and organic matter (Bradshaw and Chadwick, 1980). Plants have the ability to change the soil properties in order to improves the performance of conspecifics and enhances the probability of that the species monopolizing its local habitat, this process is called positive plant-soil feedback (Van der Putten et al., 2013). Action of adding species that are able to increase nutrient content are essential to preserve long term vegetation on these soils (Gan et al., 2013; Whiting et al., 2004). Primary successional stages are often characterized by the symbiosis between plants and nitrogen-fixing bacteria, that are considered as positive plant soil feedbacks (Van 208

209 der Putten et al., 2013). The Fabaceae C. cobalticola and C. peschiana could present nodules and thus would have the ability to fix nitrogen from the air as the N nutrition for plants. C. cobalticola tolerate highest copper concentrations than C. peschiana (Boisson et al., 2016). While these association should be difficult in soils polluted by cadmium because of nitrogenfixation is inhibited (Furini, 2012), copper seems to not reduce the nodule biomass in a copper tolerant Fabaceae (Gan et al., 2013). The recreation of vegetation succession in a bare soil would create a heterogeneous environment, having small-scale spatial variability of soil parameters that represents opportunities for successful recruitment and establishment of tolerant and non-tolerant plant species (Mench et al., 2010). This process should assure the succession of vegetation at long term that represents the main purposes of the phytostabilisation. References Bader, M.Y., Van Geloof, I., Rietkerk, M., High solar radiation hinders tree regeneration above the alpine treeline in northern Ecuador. Plant Ecol. 191, doi: /s Baker, A.J.M., Ernst, W.H.O., Van der Ent, A., Malaisse, F., Ginocchio, R., Metallophytes: the unique biological resource, its ecology and conservational status in Europe, central Africa and Latin America, in: Batty, L., Hallberg, K. (Eds.), Ecology of Industrial Pollution. pp Banza, C.L.N., Nawrot, T.S., Haufroid, V., Decrée, S., De Putter, T., Smolders, E., Kabyla, B.I., Luboya, O.N., Ilunga, A.N., Mutombo, A.M., Nemery, B., High human exposure to cobalt and other metals in Katanga, a mining area of the Democratic Republic of Congo. Environ. Res. 109, doi: /j.envres Battogtokh, B., Lee, J.M., Woo, N., Contamination of water and soil by the Erdenet copper molybdenum mine in Mongolia. Environ. Earth Sci. 71, doi: /s y Beltrán, E., Valiente-Banuet, A., Verdú, M., Trait divergence and indirect interactions allow facilitation of congeneric species. Ann. Bot. 110, doi: /aob/mcs089 Berti, W.R., Cunningham, S.D., Phytostabilization of metals, in: Raskin, I., Ensley, B.D. (Eds.), Phytoremediation of Toxic Metals: Using Plants to Cleanup the Environment. John Wiley & Sons, Inc., New York, pp Blacksmith Institute, The World s Worst Polluted Places: The Top Ten of The Dirty Thirty. New York. Boisson, S., Le Stradic, S., Collignon, J., Séleck, M., Malaisse, F., Ngoy Shutcha, M., Faucon, M.-P., Mahy, G., Potential of copper-tolerant grasses to implement phytostabilisation strategies on polluted soils in South D. R. Congo. Environ. Sci. Pollut. Res. doi: /s Boisson, S., Le Stradic, S., Commans, M., Dumont, A., Leclerc, N., Thomas, C., Mahy, G., Copper tolerance of three Crotalaria species from southeastern D. R. Congo at the early development stage. Biotechnol. Agron. Soc. Environ. 20, Bonanomi, G., Incerti, G., Mazzoleni, S., Assessing occurrence, specificity, and mechanisms of plant facilitation in terrestrial ecosystems. Plant Ecol. 212, doi: /s Bradshaw, A., Chadwick, M., The restoration of land: the ecology and reclamation of derelict and degraded land. University of California Press. Brooker, R.W., Maestre, F.T., Callaway, R.M., Lortie, C.L., Cavieres, L. a., Kunstler, G., Liancourt, P., Tielbörger, K., Travis, J.M.J., Anthelme, F., Armas, C., Coll, L., Corcket, E., Delzon, S., Forey, E., Kikvidze, Z., Olofsson, J., Pugnaire, F., Quiroz, C.L., Saccone, P., Schiffers, K., Seifan, M., Touzard, B., Michalet, R., Facilitation in plant communities: the past, the present, and the future. J. Ecol. 96, doi: /j x Cailteux, J.L.H., Kampunzu, A.B., Lerouge, C., Kaputo, A.K., Milesi, J.P., Genesis of sediment-hosted stratiform copper cobalt deposits, central African Copperbelt. J. African Earth Sci. 42, doi: /j.jafrearsci Callaway, R.M., Lawrence, W.R., Competition and Facilitation : a Synthetic Approach To Interactions in Plant Communities. Ecology 78, Cheyns, K., Banza Lubaba Nkulu, C., Ngombe, L.K., Asosa, J.N., Haufroid, V., De Putter, T., Nawrot, T., Kimpanga, C.M., Numbi, O.L., Ilunga, B.K., Nemery, B., Smolders, E., Pathways of human exposure to cobalt in Katanga, a 209

210 mining area of the DR Congo. Sci. Total Environ. 490, doi: /j.scitotenv Córdova, S., Neaman, A., González, I., Ginocchio, R., Fine, P., The effect of lime and compost amendments on the potential for the revegetation of metal-polluted, acidic soils. Geoderma 166, doi: /j.geoderma Donath, T.W., Bissels, S., Hölzel, N., Otte, A., Large scale application of diaspore transfer with plant material in restoration practice - Impact of seed and microsite limitation. Biol. Conserv. 138, doi: /j.biocon Duvigneaud, P., Denaeyer-De Smet, S., Etudes sur la végétation du Katanga et de ses sols métallifères. Communication n 7 Cuivre et végétation au Katanga. Bull. Soc. Roy. Bot. Belgique, Travaux du Centre scientifique et médical de l Université libre de Bruxelles en Afrique centrale 96, Duvigneaud, P., Timperman, J., Etudes sur la végétation du Katanga et de ses sols métallifères. Communication n 3 Etudes sur le genre Crotalaria. Bull. Soc. Roy. Bot. Belgique 91, Faucon, M.-P., Chipeng, F., Verbruggen, N., Mahy, G., Colinet, G., Shutcha, M., Pourret, O., Meerts, P., Copper tolerance and accumulation in two cuprophytes of South Central Africa: Crepidorhopalon perennis and C. tenuis (Linderniaceae). Environ. Exp. Bot. 84, doi: /j.envexpbot Faucon, M.-P., Colinet, G., Mahy, G., Ngongo Luhembwe, M., Verbruggen, N., Meerts, P., Soil influence on Cu and Co uptake and plant size in the cuprophytes Crepidorhopalon perennis and C. tenuis (Scrophulariaceae) in SC Africa. Plant Soil 317, doi: /s Faucon, M.-P., Meersseman, A., Shutcha, M.N., Mahy, G., Luhembwe, M.N., Malaisse, F., Meerts, P., Copper endemism in the Congolese flora: a database of copper affinity and conservational value of cuprophytes. Plant Ecol. Evol. 143, doi: /plecevo Faucon, M.-P., Parmentier, I., Colinet, G., Mahy, G., Ngongo Luhembwe, M., Meerts, P., May Rare Metallophytes Benefit from Disturbed Soils Following Mining Activity? The Case of the Crepidorhopalon tenuis in Katanga (D. R. Congo). Restor. Ecol. 19, doi: /j x x François, A., L extrémité occidentale de l'arc Cuprifère Shabien. Etude géologique-département de géologie de la Gécamines, Likasi (République du Zaire). Frérot, H., Lefèbvre, C., Gruber, W., Collin, C., Santos, a. Dos, Escarré, J., Specific Interactions between Local Metallicolous Plants Improve the Phytostabilization of Mine Soils. Plant Soil 282, doi: /s Furini, A. (Ed.), Plants and Heavy Metals, SpringerBriefs in Molecular Science. Springer Netherlands, Dordrecht. doi: / Gan, J., Xiong, Z.-T., Li, J.-P., Chen, D.-Q., Differential response to copper stress in the reproductive resources and allocation of metallophyte Kummerowia stipulacea. Ecotoxicol. Environ. Saf. 89, doi: /j.ecoenv Garnier, E., Growth analysis of congeneric annual and perennial grass species. J. Ecol. 80, doi: / Gonnelli, C., Galardi, F., Gabbrielli, R., Nickel and copper tolerance and toxicity in three Tuscan populations of Silene paradoxa. Physiol. Plant. 113, doi: /j x Ilunga wa Ilunga, E., Mahy, G., Piqueray, J., Séleck, M., Shutcha, M.N., Meerts, P., Faucon, M.-P., Plant functional traits as a promising tool for the ecological restoration of degraded tropical metal-rich habitats and revegetation of metal-rich bare soils: A case study in copper vegetation of Katanga, DRC. Ecol. Eng. 82, doi: /j.ecoleng International Council on Mining and Minerals (ICMM), Good Practice Guidance for Mining and Biodiversity. ICMM, London. Järup, L., Hazards of heavy metal contamination. Br. Med. Bull. 68, doi: /bmb/ldg032 Jones, R., Lunt, O., The function of calcium in plants. Bot. Rev Kacprzak, M., Grobelak, A., Grosser, A., Prasad, M.N. V., Efficacy of Biosolids in Assisted Phytostabilization of Metalliferous Acidic Sandy Soils with Five Grass Species. Int. J. Phytoremediation 16, doi: / Ke, W., Xiong, Z., Chen, S., Chen, J., Effects of copper and mineral nutrition on growth, copper accumulation and mineral element uptake in two Rumex japonicus populations from a copper mine and an uncontaminated field sites. Environ. Exp. Bot. 59, doi: /j.envexpbot Kucak, A., Blanuša, M., Comparison of two extraction procedures for determination of trace metals in soil by atomic absorption spectrometry. Arh. Hig. Rada Toksikol. 49, Lakanen, E., Erviö, R., A comparison of eight extractants for the determination of plant available micronutrients in soil. Acta Agral. Fenn. 123, Le Stradic, S., Séleck, M., Lebrun, J., Boisson, S., Handjila Minengo, G., Faucon, M., Enk, T., Mahy, G., Comparison of translocation methods to conserve metallophyte communities in the Katangan Copperbelt, DRC. Environ. Sci. Pollut. Res. doi: /s Leteinturier, B., Evaluation du potential phytocénotique des gisements cupriferes d Afrique centro-australe en vue de la phytoremédiation de sites pollués par l'activité. Fac. des Sci. Agron. Gembloux. Levine, J.M., Complex Interactions in a Streamside Plant Community. Ecology 81, Li, X., Poon, C.S., Liu, P.S., Heavy metal contamination of urban soils and street dusts in Hong Kong. Appl. Geochemistry 16, doi: /s (01)

211 Lorestani, B., Cheraghi, M., Yousefi, N., Phytoremediation potential of native plants growing on a heavy metals contaminated soil of copper mine in Iran. World Acad Sci Eng Techno 5, Manda, B., Colinet, G., André, L., Evaluation de la contamination de la chaîne trophique par les éléments traces (Cu, Co, Zn, Pb, Cd, U, V et As) dans le bassin de la Lufira supérieure (Katanga/. Tropicultura Mench, M., Lepp, N., Bert, V., Schwitzguébel, J.P., Gawronski, S.W., Schröder, P., Vangronsveld, J., Successes and limitations of phytotechnologies at field scale: Outcomes, assessment and outlook from COST Action 859. J. Soils Sediments 10, doi: /s x Mench, M., Vangronsveld, J., Lepp, N., Bleeker, P., Ruttens, A., Geebelen, W., Phytostabilisation of metal-contaminated sites, in: Morel, J.-L., Echevarria, G., Goncharova, N. (Eds.), Phytoremediation of Metal- Contaminated Soils. Springer, Trest, pp Mendez, M.O., Maier, R.M., Phytostabilization of mine tailings in arid and semiarid environments--an emerging remediation technology. Environ. Health Perspect. 116, doi: /ehp Narendrula, R., Nkongolo, K., Beckett, P., Comparative Soil Metal Analyses in Sudbury (Ontario, Canada) and Lubumbashi (Katanga, DR-Congo). Bull. Environ. Contam. Toxicol. 88, O Dell, R.E., Claassen, V.P., Relative performance of native and exotic grass species in response to amendment of drastically disturbed serpentine substrates. J. Appl. Ecol. 43, doi: /j x Padilla, F.M., Pugnaire, F.I., The role of nurse plants in the restoration of degraded environments. Front. Ecol. Environ. doi: / (2006)004[0196:tronpi]2.0.co;2 Parra, A., Zornoza, R., Conesa, E., Gómez-López, M.D., Faz, A., Seedling emergence, growth and trace elements tolerance and accumulation by Lamiaceae species in a mine soil. Chemosphere 113, doi: /j.chemosphere Parraga-Aguado, I., Gonzalez-Alcaraz, M.N., Alvarez-Rogel, J., Jimenez-Carceles, F.J., Conesa, H.M., The importance of edaphic niches and pioneer plant species succession for the phytomanagement of mine tailings. Environ. Pollut. 176, doi: Prasad, M.S., Production of copper and cobalt at Gecamines, Zaire. Miner. Eng. 2, doi: / (89) R Development Core Team, A language and environment for statistical computing. Vienna (Austria). R.M. Callaway, Callaway, R., Positive interactions and interdependence in plant communities. Remon, E., Bouchardon, J.L., Cornier, B., Guy, B., Leclerc, J.C., Faure, O., Soil characteristics, heavy metal availability and vegetation recovery at a former metallurgical landfill: Implications in risk assessment and site restoration. Environ. Pollut. 137, doi: /j.envpol Rengel, Z., Role of calcium in aluminium toxicity. New Phytol Rizzi, L., Petruzzelli, G., Poggio, G., Guidi, G.V., Soil physical changes and plant availability of Zn and Pb in a treatability test of phytostabilization. Chemosphere 57, doi: /j.chemosphere Saad, L., Parmentier, I., Colinet, G., Malaisse, F., Faucon, M.-P., Meerts, P., Mahy, G., Investigating the Vegetation-Soil Relationships on the Copper-Cobalt Rock Outcrops of Katanga (D. R. Congo), an Essential Step in a Biodiversity Conservation Plan. Restor. Ecol. 20, doi: /j x x Secretariat of the Convention on Biological Diversity, Global Biodiversity Outlook 4. Montréal. doi: /aje/kwq338 Séleck, M., Bizoux, J.-P., Colinet, G., Faucon, M.-P., Guillaume, A., Meerts, P., Piqueray, J., Mahy, G., Chemical soil factors influencing plant assemblages along copper-cobalt gradients: implications for conservation and restoration. Plant Soil 373, doi: /s Semereab, E., Rensonnet, A., Lebrun, J., Handjila, G., Saad, L., Parmentier, I., Senterre, B., Guillaume, A., Weiskopf, T., Malaisse, F., Mahy, G., Biodiversity Action Plan Tenke Fungurume Mining Gembloux. Shallari, S., Schwartz, C., Hasko, A., Morel, J.L., Heavy metals in soils and plants of serpentine and industrial sites of Albania. Sci. Total Environ. 209, Sharma, K., Basta, N.T., Grewal, P.S., Soil heavy metal contamination in residential neighborhoods in post-industrial cities and its potential human exposure risk. Urban Ecosyst. 18, doi: /s Sherameti, I., Varma, A., Heavy Metal Contamination of Soils: Monitoring and Remediation, Soil Biology. Springer International Publishing. Shutcha, M.N., Faucon, M.-P., Kamengwa Kissi, C., Colinet, G., Mahy, G., Ngongo Luhembwe, M., Visser, M., Meerts, P., Three years of phytostabilisation experiment of bare acidic soil extremely contaminated by copper smelting using plant biodiversity of metal-rich soils in tropical Africa (Katanga, DR Congo). Ecol. Eng. 82, doi: /j.ecoleng Shutcha, M.N., Mubemba, M.M., Faucon, M.-P., Luhembwe, M.N., Visser, M., Colinet, G., Meerts, P., Phytostabilisation of copper-contaminated soil in Katanga: an experiment with three native grasses and two amendments. Int. J. Phytoremediation 12, doi: / Soliveres, S., DeSoto, L., Maestre, F.T., Olano, J.M., Spatio-temporal heterogeneity in abiotic factors modulate multiple ontogenetic shifts between competition and facilitation. Perspect. Plant Ecol. Evol. Syst. 12, doi: /j.ppees

212 Tembo, B.D., Sichilongo, K., Cernak, J., Distribution of copper, lead, cadmium and zinc concentrations in soils around Kabwe town in Zambia. Elsevier, Chemosph. 63, doi: /j.chemosphere Van der Putten, W.H., Bardgett, R.D., Bever, J.D., Bezemer, T.M., Casper, B.B., Fukami, T., Kardol, P., Klironomos, J.N., Kulmatiski, A., Schweitzer, J.A., Suding, K.N., Van de Voorde, T.F.J., Wardle, D.A., Plant-soil feedbacks: The past, the present and future challenges. J. Ecol. 101, doi: / Wang, J., Ge, Y., Chen, T., Bai, Y., Qian, B.Y., Zhang, C.B., Facilitation drives the positive effects of plant richness on trace metal removal in a biodiversity experiment. PLoS One 9, e doi: /journal.pone Whiting, S.N., Reeves, R.D., Richards, D., Johnson, M.S., Cooke, J.A., Malaisse, F., Johns, R., Mcintyre, T., Purvis, O.W., Salt, D.E., Schat, H., Zhao, F.J., Baker, A.J.M., Research Priorities for Conservation of Metallophyte Biodiversity and their Potential for Restoration and Site Remediation. Restor. Ecol. 12, Zhang, X., Xia, H., Li, Z., Zhuang, P., Gao, B., Potential of four forage grasses in remediation of Cd and Zn contaminated soils. Bioresour. Technol. 101, doi: /j.biortech

213 Axe 2 -Article 9 Publié - Potential of copper tolerant grasses to implement phytostabilisation strategies on polluted soils in South D.R.Congo BOISSON Sylvain, LE STRADIC Soizig, COLLIGNON Julien, SÉLECK Maxime, MALAISSE François, NGOY SHUTCHA Mylor, FAUCON Michel-Pierre, & MAHY Grégory Abstract Phytostabilisation (i.e. using plants to immobilise contaminants) represents a well-known technology to hamper heavy metal spread across landscapes. Southeastern D.R. Congo, Microchloa altera, a tolerant grass from the copper hills, was recently identified as a candidate species to stabilise copper in the soil. More than 50 grasses compose this flora, which may be studied to implement phytostabilisation strategies. However, little is known about their phenology, tolerance, reproductive strategy or demography. The present study aims to characterize the Poaceae that may be used in phytostabilisation purposes based on the following criteria: their ecological distribution, seed production at two times, abundance, soil coverage and the germination percentage of their seeds.we selected seven perennial Poaceae that occur on the copper hills. Their ecological distributions (i.e. species response curves) have been modelled along copper or cobalt gradients with generalised additive models using logic link based on 172 presence-absence samples on three sites. For other variables, a total of 69 quadrats (1 m 2 ) were randomly placed across three sites and habitats. For each species, we compared the number of inflorescence-bearing stems (IBS) by plot, the percentage of cover, the number of seeds by IBS and the estimated number of seeds by plot between sites and habitat. Three species (Andropogon schirensis, Eragrostis racemosa and Loudetia simplex) were very interesting for phytostabilisation programs. They produced a large quantity of seeds and had the highest percentage of cover. However, A. schirensis and L. simplex presented significant variations in the number of seeds and the percentage of cover according to site. 213

214 Introduction The long-term persistence of heavy metal elements in the environment has serious negative consequences, such as soil and water contamination, ecosystem damage and public health issues (Shaw 1990; Hadjiliadis 1997; Lorestani et al. 2011). Harsh physico-chemical conditions of areas polluted by heavy metals, such as low water retention capacity (Ernst 1996), extreme ph, high heavy metal concentrations, low organic matter and nutrient concentrations (Bradshaw and Chadwick 1980), high electrical conductivity (Conesa et al. 2006)or low cation exchange capacity, strongly inhibit vegetation establishment, leading to bare ground of several hectares, which in turn results in a high exposure risk to human health by erosion or food-chain contamination (Li et al. 2006; Harmanescu et al. 2011;Clary 2014). Because of these problems linked to the heavy metal toxicity of polluted sites, phytoremediation strategies have evolved over the years as valuable options to rehabilitate polluted areas. Among them, phytostabilisation has been recognised as an adaptable and cost-effective strategy restoring a functional ecosystem in the highest heavy metal-polluted areas (Dubourguier et al. 2001). Phytostabilisation is employed to treat heavy metal-polluted soils involving the establishment of a persistent plant cover that prevents pollution spread by erosion, water percolation, leaching and wind dispersal of toxic dust (Berti and Cunningham 2000; Wong 2003). Phytostabilisation can also diminish the metal bioavailability by their precipitation to less soluble compounds (Wong 2003). The selection of plant species is crucial; they should have characteristics to grow under specific site conditions (i.e. high heavy metal concentrations in soil, pollution, climate, etc.) and to meet the objectives of (phyto)remediation (i.e. to decrease water and soil pollution; Freitas et al. 2004; Ali et al. 2013;Suchkova etal. 2014). They should also be native to the region that needs to be depolluted (Remon et al. 2005), grow quickly, have dense root and shoot systems preventing heavy metal dispersion by water and/or wind erosion (Berti and Cunningham 2000), have a large quantity of available propagules (Henson et al. 2013) and should preferentially disperse by seeds to allow the implementation of phytostabilisation on a large scale (Mench et al. 2006; Bert et al. 2008). The reclamation methods for phytostabilisation require elevated seed production, which usually results in a more patchy vegetation cover than using transplants (Mendez and Maier 2008; Mench et al. 2010; Shutcha et al. 2010). Finally, species have to be able to pro- mote soil development process by a long-term 214

215 succession in the polluted areas (Mendez and Maier 2008). Some grass species have already been identified as successful species in the phytostabilisation of heavy metal-polluted areas in Europe by semi-field and field experiments (Rizzi et al. 2004; O Dell and Claassen 2006; Mench et al. 2010; Parraga-Aguado et al. 2014). The Katangan Copperbelt in the southeast of Katanga (Democratic Republic of Congo) comprises one of the most important metallogenic areas of the world, especially in regards to copper and cobalt, and therefore, provides interesting economic opportunities formining activities (Duvigneaud and Denaeyer-De Smet 1963; Brooks and Malaisse 1985). One of the main sources of heavy metal pollution is the mining and smelting of metalliferous ores. In Katanga, humans impacted metalliferous ecosystems (e.g. copper hills), initially with little artisanal ore extraction in small areas, which also limited the pollution. However, the strong revival of the min- ing industry in the last decades radically increasedmining and smelting activities, leading to increased release of large amounts of SO 2 and metalliferous dust into the atmosphere. Heavy metals have accumulated mainly in the soil surface horizons in the surrounding area preventing vegetation establishment, leading to bare soils and putting local human populations at risk (Banza et al. 2009; Shutcha et al. 2010; Vranken et al. 2013; Cheyns et al. 2014). Greater than 550 species occur on the copper hills, including mainly graminoid, forb and small shrub species, as well as several (hyper) accumulators (Leteinturier 2002; Faucon et al. 2009; Lange et al. 2014). This area has been recognized as a hotspot for metallophytes, with more than 10 % of endemic species comprising this vegetation (Wild and Bradshaw 1977; Brooks and Malaisse 1985; Faucon et al. 2010). Among the 550 reported species, Poaceae constitute a family having dominant and structuring species of herbaceous plant communities with a high plant cover. About 50 grass species were identified on metalenriched soils on the copper hills ( This native metallophyte diversity represents a great source of potential candidate species for phytostabilisation (Leteinturier and Malaisse 1999; Shutcha et al. 2010). One study showed that metal-polluted soils in Lubumbashi (capital of Katanga province, D.R.C.) can be phytostabilised by the local toothbrush grass, Microchloa altera (Shutcha et al. 2010). This perennial species only inhabits metal-enriched soils in southeast Katanga (Duvigneaud and Denaeyer-De Smet 1963; Faucon et al. 2010). Excepted for M. altera, little is known about the phenology, reproductive strategy, copper tolerance, demography or ecological distribution of other 215

216 dominant grass species, which currently limits their use in phytostabilisation strategies. The aim of the present work is to identify new native Poaceae species that can serve as candidates for phytostabilisation in mining sites in the same climatic region. We selected seven dominant native Poaceae species occurring on copper hills and we evaluated their ecological distribution along the copper and cobalt gradients (i.e. realized niches), capacity to provide soil cover, seed production, and germination capacities in order to assess if these species can be used in phytostabilisation projects. Methods Study area Our study was conducted in the Tenke Fungurume area, in the center of the Katangan Copperbelt, between the cities of Lubumbashi and Kolwezi (Katanga province, D. R. Congo). The region(ca.1,000 km 2 ) contains more than 40 copper hills distributed along two parallel low ridges between the cities of Tenke (10.61 S, E) and Fungurume (10.62 S, E) at an altitude of approximately 1,300 m (Fig. 1, Table 1). The area presents a subtropical climate with a rainy season (from November to the end of March) and a dry season (from April to September). Mean annual temperature is about 20 C. Temperatures range from 15 to 17 C at the beginning of the dry season to C in September October.During the dry season, temperatures can reach 5 C at night (Campbell 1996). Fig. 1. Location of the sampled sites in the Katangan Copperbelt between Tenke and Fungurume (Democratic Republic of Congo). 216

217 The dominant natural vegetation of the landscape is comprised of an open deciduous forest (Miombo) dominated by trees of the Caesalpinioideae subfamily (Campbell 1996). Copper hills are very characteristic and visible in the land- scape because they are covered by herbaceous vegetation, in contrast to the surrounding Miombo forest, as heavy metal concentrations in the soil prevent tree development. On the copper hills, copper and cobalt concentrations vary continuously along the topographic gradient (Fig. 2). The kriging method was based on 83 plots corresponding to a part of data of Séleck et al. (2013). Default options of the Geostatistical Wizard available from of the Geostatistical Analyst Extension on ArcGis 10.0 were used. In doing so, values of available copper and cobalt concentrations were log transformed (ESRI 2010). Plant species were found in two main plant communities according to the heavy metal concentration and ph. The steppe, characterized by shallow soils with lower vegetation, is located on the top of the hills where copper concentrations are high (3,500 10,000 mg Cu kg soil -1 ). The steppic savannah, located on the slope and bottom of the hills where soils are deeper, presenting the lowest copper concentrations (0 3,500 mg Cu kg -1 ; Saad et al. 2012; Ilunga wa Ilunga et al. 2013; Séleck etal. 2013). Table 1. Geographic data (UTM zone 35S) and description of the sampled sites. Sites Coordinates (utm) Altitude minmax (m) Fungurume V Kazinyanga Kavifwafwaulu I Goma II Surface (ha) Description Natural Cu/Co outcrop; steep slopes ranging from 0 (bottom) to 30 (top); dominated by steppic savannah in the bottom Natural Cu/Co outcrop; steppic savannah for the 2/3 on medium slope and steppes on the 1/3 top Natural Cu/Co outcrop; steppic savannah for the 2/3 in a gentle slopes and steppes on the 1/3 top Natural Cu/Co outcrop; steppic savannah for the 2/3 and steppes on the 1/3 top with an escarpment of the rocky outcrops 217

218 Fig. 2. Kriging map of Fungurume V showing the distribution of available copper or cobalt concentrations in soils of the copper hills. Studied species Seven Poaceae species were selected according to their frequent occurrence in copper hill communities: Andropogon schirensis Hochst.ex A.Rich., Eragrostis racemosa (Thunb) Steud, Hyparrhenia diplandra (Hack.) Stapf, Loudetia simplex (Nees) C. E. Hubb., Monocymbium ceresiiforme (Nees) Stapf, Tristachya bequaertii De Wild., and Trachypogon spicatus (L. f.) Kuntze (Table 2, Fig. 3). They are the dominant Poaceae species in the copper plant communities (Table 2), usually having broad geographic distribution throughout the African continent (west tropical, west central, east tropical, east central, etc., Table 2), which means that their use in phytostabilisation will be potentially possible in these areas. Plant determinations were realized using several floras (Bamps ( ); Board of Trustees Kew Royal Botanic Gardens ( ) and Flora of Tropical East Africa (Kew Royal Botanic Gardens )) and completed with more recently published taxonomic 218

219 literature for particular genera and species (Leteinturier 2002). Species descriptions came from Clayton et al. (2006) and Van Oudtshoorn (2012). Fig. 2. Pictures of the seven selected Poaceae species: a) Loudetia simplex, b) Tristachya bequaertii, c) Eragrostis racemosa, d) Monocymbium ceresiiforme, e) Trachypogon spicatus, f) Andropogon schirensis, g) Hyparrhenia diplandra. 219

220 Table 2. Main characteristics of the seven selected Poaceae species. Species description and geographic range are given according to Clayton et al. (2006) and Van Oudtshoorn (2012). Underground systems and root system architecture are given according to Willem (2011), seed size and fruiting period are field observation data. Morphological and physiological characteristics Species Species description Underground system (cm) Andropogon schirensis Hochst. ex A.Rich. Red- brown grass relatively densely tufted. Unbranched culms with two terminal racemes arranged in «V» with spikelets by pairs, each containing a fertile spikelet without pedicel and a sterile spikelet pedicelate. Root system Architecture Seeds size (mm) Fruiting period Fascicle January to April Geographic range Africa: west tropical, west-central tropical, northeast tropical, east tropical, southern tropical, and south. Eragrostis racemosa (Thunb) Steud Hyparrhenia diplandra (Hack.) Stapf Short dense tufted grass with sturdy and erected culms. Most of leaves are basilar and panicles contain green-olive spikelets. Dense tufts with robust and erected straws. Inflorescence is composed of racemes by pairs surmounted by a red-brown spathe. Each spikelets contains pairs of spikelets with one fertile sessile and one sterile pedicelate Fascicle < 2 February to March Fascicle April to May Africa: west-central tropical, northeast tropical, east tropical, southern tropical, south, and western Indian ocean. Africa: west tropical, west-central tropical, northeast tropical, east tropical, southern tropical, and western Indian ocean. Asiatemperate. Loudetia simplex (Nees) C.E. Hubb. Hard tufted grass, unbranched culms. Inflorescence is an open to contracted lightbrown panicle with long awns. Inflorescences Fascicle May to April Africa: west tropical, west-central tropical, northeast tropical, east tropical, southern tropical, south, andwestern Indian ocean. Monocymbium ceresiiforme (Nees) Stapf Trachypogon spicatus (L.f.) Kuntze Inflorescences with short racemes, partially or entirely covered with orange-brown spathes. Spikelets are by pairs: a fertile (sessile) and a sterile (pedicelate). Dense tufts with erected and unbranched stems. Nodes are characterized by a white hairy ring. Inflorescence is a racem with spikelets pairs overcomed by a long white hairy barb. Fertile spikelets have a larger pedicel than sterile ones (sometimes sessile) Fascicle March to May Fascicle March to May Africa: west tropical, west-central tropical, northeast tropical, east tropical, southern tropical, south, and western Indian ocean. Africa: west tropical, west-central tropical, northeast tropical, east tropical, southern tropical, and south. Tristachya bequaertii De Wild. Dense tufts with many basal leaves. Panicles with groups of three spikelets (often fertile) with a long barb. The spikelets external hull is black and densely hairy Fascicle 6-15 March Africa: west-central tropical, east tropical, and southern tropical. Ecological distribution assessment A vegetation survey was performed in March 2008 and 2009, at the period of peak vegetation. Three natural copper hills were chosen: Fungurume V, Kavifwafwaulu I and Kazinyanga (Table 1). Data were from the transect study of Séleck et al. (2013). On each site, six to eight parallel transects were established from the bottom to the top of the hill. On each transect, plots were placed at every change of 5 m in altitude to cover the geological and edaphic variability in order to obtain a systematic sampling, which led to the establishment of a total of 172 1m 2 vegetation sample plots according to a grid on the three sites (Fungurume V, 83 plots; Kavifwafwaulu I, 37 plots; Kazinyanga, 52 plots). In each plot, the presence or absence of each Poaceae was recorded and four soil subsamples were collected at the four plot corners at a depth of 15 cm. All soil samples were air-dried and sieved to 2mm. Bioavailability of Cu andcowere determined after extraction with 50 ml of CH 3 COONH 4 + EDTA (0.5N, ph 4.65) for 30 min (soil to solution ratio of 1:5; Kucak and Blanuša 1998; Faucon et al. 2009; Saad et al. 2012). The supernatant was filtered through an S&S 595 folded filter and analysed using a flame atomic absorption spectrometer (VARIAN 220), following the norm NF 220

221 X The limits of quantification of the VARIAN 220 were 10 mg kg -1 for copper and 0.25 mg kg -1 for cobalt. The laboratory is a member of the REQUASUD network that organizes quality and repeatability tests to assure the validity of applied methods and measures. Ecological distribution (i.e., response curves) of each selected Poaceae species along the Cu and Co gradients were obtained with generalized additive models (GAM) using a binomial likelihood (Hastie et al., 1986) to model the probability of occurrence ( ) at each metal concentrations ( ): ( ( )), where ( ) is the logit function, ( ) is smooth functions (non-parametric) and k is the number of degree of smoothing. For each species, the best fitted model among three degrees of smoothing k (3, 4 or 5) was selected based on the Akaike Information Criterion; AIC (Aikaike, 1987). The selected models for A. schirensis, E. racemosa, M. ceresiiforme and T. spicatus had three degrees of smoothing, four for L. simplex and T. bequaertii and five for H. diplandra. The optimum value of each species along the two gradients was estimated by the projected value of the highest probability of occurrence and plotted on the x-axis. The borders of niche width (or amplitude) were determined by the projection of the lower and upper limits of 80% of the area under the curve and also plotted on the x-axis (Gégout et al., 1998). Niche width was calculated by the difference between the upper and the lower limits. Species characteristics: seed production and inflorescence abundance In March 2013, the Kavifwafwaulu I site had already been exploited. Goma II, a hill close to the city of Tenke, was chosen in place of Kavifwafwaulu I for the analysis (Table 1). For each species, permanent 1 m 2 plots were randomly placed on each site within the population in order to obtain a minimum of eight plots by species. When several selected species occurred in the same plots, all these species were taken into account. A total of 68 permanent plots were established in the three sites. Each plot was delimited with sticks and the habitat was recorded (steppic savannah or steppe). For each Poaceae species present in the plot, the percentage of cover was noted, total inflorescence-bearing stems (IBS) were counted and three IBS were randomly collected. For each collected IBS, seeds were counted at the laboratory using a microscope. A first sampling was obtained at the end of March 2013 (t1) 221

222 and a second sampling was acquired in April 2013 (t2). Seed production by area (seeds per square meter) was calculated by multiplying the number of IBS by the number of square meters (IBS per square meter), while the mean of the seed number was calculated by multiplying the number of IBS by the plot (seeds per IBS). To analyse the reproductive (i.e. seeds per IBS and seeds per square meter) and the vegetative traits (i.e. inflorescence abundance and coverage), data were first analysed using a crossed two-way analysis of variance considering the site (Fungurume V, Goma II or Kazinyanga) as a random factor and the habitat (steppe versus steppic savannah) as a fixed factor. In the absence of interactions, data were analysedwith a one-way ANOVA considering site or habitat as a single factor. In order to take into account normality and homoscedasticity, data were transformed; the percentage of cover was transformed using a square root conversion and the other data were transformed using a logarithm conversion [log (x+1)]. Means were structured by Tukey s HSD test with an error of 5%. Analyses were carried out in R version (R Development Core Team 2010). Germination tests and seed mass Germination tests were performed for A. schirensis, H. diplandra, M. ceresiiforme and T. spicatus, species for which the number of harvested fresh seeds were sufficient. For A. schirensis only, the large number of seeds collected in both steppic savannahs and steppes allowed germination tests to be performed for each habitat. The well-developed seeds were selected and placed into Petri dishes (five replicates of 30 seeds) with filter paper to germinate. Petri dishes were placed into a germination chamber under a 12h photoperiod and maintained at temperatures of 25 C and 15 C for diurnal and nocturnal temperatures, respectively. For the species A. schirensis, seeds from steppic savannahs and steppes were separately weighed (10 repetitions of 10 seeds) using a balance with a precision of 0.01 mg. The transformed values of the germination percentages by the square root of arcsine were compared between sites using an analysis of variance with one criterion of classification (one-way ANOVA) in order to account for normality and heteroscedasticity. For A. schirensis, germination percentages were analysed by a two-way ANOVA (site habitat). Seedlot weights of A. schirensis were compared according to habitat using a one-way ANOVA. Means were structured by Tukey s range test (HSD) with an error rate of de 5 %. The ANOVAs were based on the measure of the F statistic to determine the p-value that was considered as 222

223 significant when it was lower than Analyses were all were carried out in R version (R Development Core Team 2010). Results Ecological distribution along copper and cobalt gradients Tristachya bequaertii (n=48 occurences) and Trachypogon spicatus (n=64) had their copper and cobalt optima equal to the lower limits (29 mg Cu.kg -1 ; 2 mg Co.kg -1 ), these two species occur at the bottom of the hill in the steppic savannahs. Loudetia simplex (n=131) had also its copper and cobalt optimum equal to the lower limits but its niche widths were broader than the two first species. It was found in the steppic savannahs and in the steppes. Hyparrhenia diplandra (n=49) was present in the steppic savannah but it had its optimum in high available copper concentrations (1,534 mg Cu.kg -1 ) and in the highest cobalt concentrations (265 mg Co.kg -1 ). Monocymbium ceresiiforme (n=71) occurred in the steppe and in the steppic savannah. It had the broadest niche width (optimum: 5,487 mg Cu.kg -1 ) among the seven selected Poaceae. Its cobalt optimum was at 138 mg Co.kg -1. Finally, Andropogon schirensis (n=131) and Eragrostis racemosa (n=29) were the most copper tolerant species having their copper optima at 10,136 mg Cu.kg -1. Cobalt optima were at 120 mg Co.kg -1 for A. schirensis and at 2 mg Co.kg -1 for E. racemosa. However, A. schirensis had larger copper and cobalt niche widths and occurred in the two habitats (Fig. 3). Fig. 3. Realized niches of the seven Poaceae species modeled with the generalized additive model (GAM) along the available Cu and Co gradients. Dots are the niche optima of taxa determined by the gradient value of the highest probability of occurrence. Grey straight lines were the niche widths calculated by the difference between the limits of area of 80 % under GAM-curve. 223