Acta Ecologica Sinica 29 (2009) 171–175 Contents lists available at ScienceDirect Acta Ecologica Sinica journal homepage: www.elsevier.com/locate/chnaes Impacts of recreational trampling on sub-alpine vegetation and soils in Northwest Yunnan, China Yang Mingyu a,b,*, Luc Hens c, Ou Xiaokun a, Robert De Wulf b a Institute of Ecology and Geobotany, Yunnan University, Kunming 650091, China Department of Forest and Water Management, University Gent, Coupure 653, B9000 Gent, Belgium c Department of Human Ecology, Vrije Universiteit Brussel, Laarbeeklaan 103, Belgium b a r t i c l e i n f o Keywords: Recreational trampling Experimental approach Sub-alpine environment Protected areas a b s t r a c t Controlled trampling experiments were undertaken to assess impacts of recreation in a sub-alpine environment in an upper-Mekong mountainous protected area in China. Hiking and recreational horse-riding were applied at different trampling intensity to two typical, widespread vegetation types (Carex grassland and low Rhododendron shrubland) and trampling effects were assessed to study vegetation resistance and soil compaction. The results indicate: (1) low shrub vegetation is highly vulnerable to trampling damage while the graminoids-dominated grassland is more resistant; (2) dry soil with low organic matter, which often is found in the shrubland, is more susceptible to compaction than wet soil and (3) horses cause substantially more damage than hikers at equivalent trampling levels. These data are useful to develop a visitor’s management strategy that allows to minimize the impact from recreation on the vegetation. Ó 2009 Ecological Society of China. Published by Elsevier B.V. All rights reserved. 1. Introduction Nature reserves, heritage sites, national parks, and wilderness areas, all aim at safeguarding the natural and cultural resources [1]. Human impacts, including encroaching development, pollution, introduction of exotic species, and recreation or tourism, increasingly threaten these natural assets. In particular, recreation and tourism activities offer a perennial and growing management challenge [2–4]. However, there is a need to combine the protection of natural resources with opportunities offered by contemporary tourism. Given the increasing access, trampling becomes a problem as it represents the major disturbance affecting vegetation and surface profile in the protected areas. Therefore, it becomes a topic of considerable importance for ecological impact study. Previous studies quantified the impacts of trampling by field surveys [5]. According to Sun and Walsh [6], this approach allows for fast results, but is limited in its ability to accurately quantify the relationship between the seriousness of the impact and the usage intensity. Experimental measurement on trampling offers a complementary method to the field survey. It allows to develop a more precise relationship between ecosystem response and levels of recreation. * Corresponding author. Address: Institute of Ecology and Geobotany, Yunnan University, Kunming 650091, China. Tel.: +86 13577111010 (mobile); fax: +86 0871 5165581. E-mail addresses: mingyuyang@gmail.com (M.Y. Yang), human.ecology@vub. ac.be (L. Hens), xkou@ynu.edu.cn (X.K. Ou), Robert.DeWulf@UGent.be (R.D. Wulf). Wagar [7] was the first to propose an experimental approach to assess the tolerance of vegetation to human use. Since then, many vegetation types worldwide have been examined. These include, for example, heathlands [8,9], dune and coastal systems [10–12], tropical and sub-tropical vegetation [13–15], alpine and sub-alpine areas [16–19], and arctic ecosystems [20]. Despite all these studies, our understanding of the impacts of recreational trampling on vegetation is still limited as most studies focus on obvious changes and short time frames. Therefore, data on long term and low pressures are lacking and few studies produce results that lead to broader generalizations [21,22]. This study investigates the effects of recreational trampling on a sub-alpine environment in an upper-Mekong mountainous area in China where nature-based tourism prevails [23–25]. Controlled levels of trampling were applied to two sub-alpine plant communities and their initial responses were measured to study the resistance of vegetation and soil. Data and analysis show important for several reasons. First, these types of vegetation are abundant in the region and feature high scenic value [26]. Tourists undoubtedly wall over and camp on these plants, that offer comfortable hiking and horse-riding activities. Second, alpine and sub-alpine areas are more subject to trampling damage than most other environments [16,27]. Management objectives commonly emphasize the protection of such vegetation. It is however, unclear which management option (dispersal or containment of tourists) should be applied. Also the limits of sustainable use are unknown. More experimental evidence is needed to establish a management that targets sustainable use 1872-2032/$ - see front matter Ó 2009 Ecological Society of China. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.chnaes.2009.07.005 172 Y. Mingyu et al. / Acta Ecologica Sinica 29 (2009) 171–175 by tourists of these (sub-) alpine ecosystems. This study generates data that contribute to develop scientifically sound management practices. The aims of this study are to determine:  vegetation damage to plant communities resulting from different levels of trampling;  soil exposure and compaction that result from trampling;  the extent and nature of damage trampling caused by walkers as compared to that caused by horses. 2. Study area The study was conducted in the Laojun mountain area, China (N 26°380 –27°150 ; E 99°360 –100°100 ; 2100–4200 m a.s.l) which is part of the Hengduan range, at the eastern end of the Himalayas. This is a scenic mountainous area that is listed as a world heritage site by UNESCO in 2003. Two sub-alpine study sites, each with a vegetation type that is characteristic for sub-alpine ecosystems, were selected. The annual precipitation averages 600 mm with 160 days of snow cover. The temperature ranges from 3.2 °C in January to 15 °C in July. The Carex grassland grows on poorly drained, acidic soils along the rivers in the flat bottom of the valley, at an elevation of 3214 m a.s.l. Carex spp., Juncus sp., Sanguisorba filiformis are abundant in these plant communities. The soils are covered by a thick (typically > 1 cm) organic layer. Another type, the low Rhododendron shrubland, is located on well drained, stony slope soils at 3244 m a.s.l. The community is dominated by Rhododendron hippophaeoides and Rhododendron racemosum that 30–50 cm tall. Under the rhododendron shrub, the most abundant associated plant species are Polygonum sphaerostachyum, Potentilla fulgens, and Kobresia sp. Only a thin organic layer of less than 1 cm covers the soil under these rhododendrons. 3. Methods 3.1. Experiments 3.1.1. Trampling The standard experimental trampling protocol recommended by Cole and Bayfield [21] is slightly modified for this study. Minor modifications concern the trampling plot layout and the impact indicators used. Selected plots for each type of vegetation were established in areas not showing any evidence of recent disturbance and of pre-existing tracks. The homogeneity of the vegetation was the main criterion to select the experimental areas. Four replicate plots (type A), each comprising five treatments lanes, were established to study the effect of hikers trampling on the two plant communities. Each lane was 0.5 m wide and 2.5 m long and separated by a 0.5 m wide buffer. The lane size allowed walkers to maintain their natural gait when walking within the lane. Treatments were randomly assigned to lanes. The experiment took place in May 2006. All trampling was completed within one day, eliminating variation due to changes in weather, considered in the method described by Cole [17]. The control lane received no trampling. Other lanes received 25 (low), 75 (moderate), 200 (high1) and 500 passes (high 2), respectively. A pass was a one way walk conducted by volunteers (mean weight 62 ± 5.4 kg, mean pack weight 5 ± 2.1 kg) wearing lug sole boots and adopting a natural walling style. Walkers started trampling 1m before the start of each marked lane to ensure that the impact resulted from a natural gait. Four replicate plots (type B) were established to measure trampling differences between humans and horses. Two controlled treatments were applied both to hikers and horses: 30 passes (low) and 100 passes (high). To make sure that the horse can reach a normal stride within a lane, experimental lanes expanded to 1 m wide and 5 m long. Horses with freshly clipped toenails and weights of about 200 ± 15 kg were used for trampling. Horses were led (not ridden) by lead ropes attached to a halter. Other experimental designs were the same as the plots for the hikers alone. 3.1.2. Trampling response variables A subplot (50  50 cm in type A and 100  100 cm in type B) was positioned in the center of each treatment lane. Each subplot consisted of 25 quadrates which were established to facilitate the measurements. Indices were recorded in each of quadrates in each treatment lane immediately after trampling. Measurements consisted of: (1) visual assessment of the canopy coverage of each vascular plant species (only green plants were taken into account, while mosses and lichens were not part of the evaluation); (2) determination of the vegetation canopy height; (3) visual assessment of the cover of organic soil litter (soil organic material and plant litter); and (4) soil compaction was estimated using a pocket soil penetrometer. 3.2. Data analysis To quantify trampling disturbance of vegetation and soil, the data were calculated as a percentage or a proportion of the pretreatment share. For example, the relative cover (RC) was calculated as: RC ¼ cover on trampled lane  100 : initial cover on control lane in the same plot Other values were calculated in a similar wary. The RC is 100% in the absence of any change caused by trampling. Deviations from 100% provide an estimate of the trampling effects. Differences caused by trampling intensity or user type after trampling were tested using a t-test. All statistical tests throughout this study were performed using SPSS 14.0. Table 1 Initial frequency (%) and mean percent cover (%) of species in Carex and Rhododendron plant communities. Species Plant community Grassland Frequency Shrubland Cover Shrubs Rhododendron hippophaeroides Rhododendron racemosum Graminoids Carex spp. Juncus sp. Poa sp. Kobresia sp. Forbs Sanguisorba filiformis Oenanthe sinense Primula poissonii Plantago major Iris bulleyana Polygonum sphaerostachyum Potentilla fulgens Epilobium palustre Ranunculus yunnanensis Primula yunnanensis 45 30 20 20 10 20 10 10 40 30 30 30 15 10 10 10 10 10 5 5 Frequency Cover 25 15 25 20 10 5 10 5 30 5 5 3 2 10 3 1 1 1 Y. Mingyu et al. / Acta Ecologica Sinica 29 (2009) 171–175 173 Before experiments started, the pre-trampling species abundance on experimental plots at both sites were assessed (Table 1). Results of hikers trampling in the two sites under the four treatments are shown in Figs. 1 and 2. The null hypothesis that treatments had no effect on the vegetation and the soil was tested by a one-sample t-test (a = 0.05) that allows to compare the trampling results with the control lanes. Results of horse and hiker comparison in the two sites under two treatments are presented in Figs. 3 and 4. The null hypothesis is that type of use has no different effect on the vegetation and the soil. This hypothesis is tested using a paired-samples t-test (a = 0.05). vegetation cover (Fig. 1a). On the Rhododendron low shrubland, reductions of vegetation are most noticeable after 75 and 200 passes. The highest level of trampling results in 29% of the original cover (Fig. 1a). Immediately after trampling, the relative height of the plants on the Carex grassland is significantly affected by the trampling, except for the low pass level (P = 0.11). Moderate and high passes differ significantly in comparison to the control lanes in particular. The vegetation is significantly lower after 500 passes (P < 0.001). In the Rhododendron shrubland, a significant reduction of the height of the vegetation occurs already at 25 passes (P = 0.013). The reduction in the height is most evident in the lanes trampled by 200 and 500 passes (P < 0.001) (Fig. 1b). 4.1. Effects of trampling on vegetation 4.2. Effects of trampling on soil In general, trampling destroys the vegetation and the effect is close dependent: the more trampling, the more serious the effect. This applies to both vegetation types: the Carex grassland and the Rhododendron shrubland. The vegetation cover significantly is reduced in all trampled lanes (P < 0.05) as compared to the control lanes (mean for grassland = 80%; mean for shrubland = 65%). Low to moderate levels of trampling disturbance (25–75 passes) have less effect on the Carex grassland than higher levels of trampling (200–500 passes). Five hundred passes result in a most significant decrease of the vegetation cover and leave only 39% of the original The organic cover in the grassland decreases slightly as a result of 25 passes (P = 0.093). However, the difference with the control lane is not statistically significant. Statistically, significant reduction of the organic matter in the soil is observed after 200 passes (P = 0.002). In the lanes trampled by 500 passes, only 46% of the original organic litter remains. A similar trend is observed in the shrubland. After 500 passes only 60% relative organic soil cover remains (Fig. 2a). Soil compaction is observed on all trampled lanes (P < 0.05) in the shrubland (Fig. 2b). After 25 passes the mean penetration resis- 4. Results Fig. 1. The relationships between vegetation cover (a), vegetation height (b) and trampling intensity in two plant communities. Values are means ± 1SE. Fig. 2. The relationships between soil organic litter cover (a), penetration resistance (b) and trampling intensity in two plant communities. Values are means ± 1SE. 174 Y. Mingyu et al. / Acta Ecologica Sinica 29 (2009) 171–175 Fig. 3. Comparison between hiker and horse trampling on vegetation cover (a) and vegetation height (b) in two plant communities. Values are means ± 1SE. Fig. 4. Comparison between hiker and horse trampling on soil organic litter (a) and soil penetration resistance (b) in two plant communities. Values are means ± 1SE. tance is 1.21 kg/cm2. After 500 passes, the value increases to 2.91 kg/cm2 which is almost four times the initial situation. On the grassland, significant a increase of the soil compaction is observed only after 200 and 500 passes (mean for 200 passes = 2.44 kg/cm2; mean for 500 passes = 2.88 kg/cm2). 4.3. Comparison between hikers and horses In both vegetation types, horses cause more vegetation loss than hikers at the two trampling intensities (P30 < 0.001; P100 = 0.04) (Fig. 3a). The vegetation height however is not significantly affected after 100 house-passes as compared to the situation after 100 passes by hikers (P = 0.107) (Fig. 3b). The relative soil organic liter is affected in a similar wary by low (30 passes) horse and hiker trampling (P = 0.107). More intensive trampling (100 passes) in contrast results in statistically significant difference (P = 0.008) (Fig. 4a). Soil compaction is significantly less on lanes that are trampled by hikers as compared to lanes trampled by horses at both intensities (P30 = 0.013; P100 < 0.001) (Fig. 4b). 5. Discussion and conclusion The results clearly show that trampling causes severe vegetation cover loss in two sites. The two plant communities show approximately a 60% cover loss when trampled with 500 passes. The Carex grassland is slightly less affected than the Rhododendron shrubland. The relative height of the Carex grassland is largely unaffected by low and moderate trampling. The effect is more profound at higher trampling levels: 68% height loss after 500 passes is assessed. In the low Rhododendron shrub, the vegetation height is significantly reduced to 58% of the original level after 75 passes and to 42% of its original high after 500 passes. This result indicates that Carex grassland is more resistant to trampling than the low Rhododendron shrubland. Modeling work by previous studies [14,17] suggest that vegetation resistance to trampling is largely a function of plant stature, erectness and whether the plants are graminoids (grasses and grass-like plants), forbs (herbaceous plant other than graminoids) or shrubs. In this study, the Carex grassland is abundant with matted graminoids species which seem to show more adaptations and resistance feasibility to trampling than other life forms. Shrubs and erect forbs in the Rhododendron shrubland are easily flattened even after low human use and could reach levels of disturbance exceeding the ability to recover by intensive tramping. The results indicate that two soils types are more resistant to low to moderate impacts. Because of relatively thick organic layer, loss of litter is less pronounced on the grassland soil than on the shrubland soil at low to moderate impacts. However, on both soils, once the surface roots are destroyed by higher trampling intensity, the bare ground increases significantly: only 37% organic litter on grassland soil and 43% organic litter on shrubland soil remain after Y. Mingyu et al. / Acta Ecologica Sinica 29 (2009) 171–175 500 passes. Such removal or reduction of the litter and humic layers initiates a cycle with profound implications for the ‘‘health” of soils as Manning [28] indicated. The susceptibility to compaction of the two soils types increase when organic matter is lost. However, the shrubland is particularly vulnerable to trampling. All trampled lanes show statistically significant more compaction than the control lanes. In contrast, compaction of the grassland soil occurs only after 200 passes. Hammitt and Cole [29] argue that soils with wide range of particle sizes, a low organic content, and that are wet most of time are prone to compaction when trampled. The control values of penetration resistance for grassland and shrubland soils are 1.03 kg/cm2 and 0.73 kg/cm2, respectively. These values partially reflect the difference in texture and moisture status of soil at the time of the experiment. Yet, because local farmers use the grassland study site for decades as a communal free grazing area during summer time, livestock trampling also affects the soil compaction during grazing. This might explain why the control of the compaction value on the grassland is slightly higher than the one on the shrubland and why dry shrubland soils are more susceptible to compaction. Trampling by horse causes substantially more damage to vegetation and surface profile in both vegetation types than impacts of hikers despite the fact that two vegetation types are different in trampling responses. Difference between horses and hikers are more serious on the less resistant vegetation types and at the lower trampling intensity. Conversely, difference between horses and hikers are likely to be less pronounced in more resistant vegetation types or at higher trampling intensities. 6. Management implications In many protected areas of the upper-Mekong mountainous area, management plans call for protection and conservation of the natural heritage. Current management strategies do not limit access in remote or sensitive areas. The increasing number of tourists in the area results in more trampling and consequently this presents an increasing management challenge. Site-specific information on the response of plant communities to recreation and disturbance by tourists is necessary to underpin management decisions. This work provides a basis to develop prescriptive visitor management strategies for these protected areas in sub-alpine environments. Although applied trampling experiments to plant communities and soils do not exactly mimic disturbance from actual recreational use in the study sites, it does provide an effective means for examining the responses to recreational disturbance while controlling or evaluating the influence of extraneous variables. The study results suggest that rotation of tracks or dispersal of visitors could be a useful strategy to maintain trampling levels below damage thresholds, so that sites do not deteriorate. Where use levels cannot be held substantially below thresholds or sensitive to trampling, damages, such as devegetation, compaction, loss of organic matter, are inevitable on all trampled sites. Secure concentration of visitors on a minimal number of sites is a more desirable management option. Meanwhile, managers need to be aware that users differ in their potential to cause impact. The profound difference among non-motorized users appears to be between horses and hikers in the study sites. Managers can utilize this information in several ways: they can zone sites to different View publication stats 175 types of users or to confine the more damaging user types to certain preferably more resistant areas. References [1] IUCN, Guidelines for Protected Area Management Categories, Gland, Switzerland and Cambridge, UK, 1994. [2] C. Hunter, H. Green, Tourism and the Environment: A Sustainable Relationship? Routledge, London, 1995. [3] D. Newsome, S.A. Moore, R.K. Dowling, Natural Area Tourism: Ecology, Impacts and Management, Channel View Publications, Clevedon, 2002. [4] G. Worboys, M. Lockwood, T.D. Lacy, Protected Area Management: Principles and Practice, Oxford University Press, London, 2005. [5] F.B. Goldsmith, Ecological effects of visitors in the countryside, in: A. Warren, F.B. Goldsmith (Eds.), Conservation in Practice, John Wiley & Sons, London, 1974. [6] D. Sun, D. Walsh, Review of studies on environmental impacts of recreation and tourism in Australia, Journal of Environmental Management 53 (4) (1998) 323–338. [7] J.A. Wagar, The Carrying Capacity of Wildland for Recreation. Forest Science Monographs 7, Society of American Foresters, Washington, 1964. [8] N.G. Bayfield, Recovery of four montane heath communities on Cairngorm, Scotland, from disturbance by trampling, Biological Conservation 15 (3) (1977) 165–179. [9] S. Gallet, F. Roze, Long-term effects of trampling on Atlantic Heathland in Brittany (France): resilience and tolerance in relation to season and meteorological conditions, Biological Conservation 103 (3) (2002) 267–275. [10] T. Hylgaard, M.J. Liddle, The effect of human trampling on a sand dune ecosystem dominated by Empetrum nigrum, Journal of Applied Ecology 18 (2) (1981) 559–569. [11] U.V. Andesen, Resistance of Danish coastal vegetation types to human trampling, Biological Conservation 71 (3) (1995) 23–230. [12] S. Lemauvie, F. Roze, Response of three plant communities to trampling in a sand dune system in Brittany, Environmental Management 31 (2) (2003) 227– 235. [13] D. Sun, Trampling resistance, recovery and growth rate of eight plant species, Agriculture, Ecosystem, and Environment 38 (4) (1992) 237–265. [14] D. Sun, M.J. Liddle, Plant morphological characteristics and resistance to simulated trampling, Environmental Management 17 (4) (1993) 511–521. [15] L.M. Talbot, S.M. Turton, A.W. Graham, Trampling resistance of tropical rainforest soils and vegetation in the wet tropics of north east Australia, Journal of Environmental Management 69 (1) (2003) 63–69. [16] J. Whinam, E.J. Cannell, J.B. Kirkpatrick, et al., Studies on the potential impact of recreational horse riding on some alpine environments of the Central Plateau, Tasmania, Journal of Environmental Management 40 (2) (1994) 103– 117. [17] D.N. Cole, Experimental trampling of vegetation II: predictors of resistance and resilience, Journal of Applied Ecology 32 (1) (1995) 215–224. [18] J. Whinam, M. Comfort, The impacts of commercial horse riding on sub-alpine environments at Cradle Mountain, Tasmania, Australia, Journal of Environmental Management 47 (1) (1996) 61–70. [19] E. Thurston, R.J. Reader, Impacts of experimentally applied mountain biking and hiking on vegetation and soil of a deciduous forest, Environmental Management 27 (3) (2001) 397–409. [20] C.A. Monz, The response of two arctic tundra plant communities to human trampling disturbance, Journal of Environmental Management 64 (2) (2002) 207–217. [21] D.N. Cole, N.G. Bayfield, Recreational trampling of vegetation: standard experimental procedures, Biological Conservation 63 (3) (1993) 209–215. [22] M. Liddle, Recreation Ecology, Chapman and Hall, London, 1997. [23] Y.G. Peng, X.C. Wang, A study on tourism resources and its exploitation of Northwest Yunnan, Journal of Economic Geography 19 (3) (1999) 111–114 (in Chinese). [24] G.H. Yang, Research on ecotourism development in Northwest Yunnan, Journal of Inquiry into Economic Issues 9 (2000) 104–107 (in Chinese). [25] G.P. Nyaupane, D.B. Morais, L. Dowler, The role of community involvement and number/type of visitors on tourism impacts: a controlled comparison of Annapurna, Nepal and Northwest Yunnan, China, Tourism Management 27 (6) (2006) 1373–1385. [26] Editing Board of Yunnan Vegetation, Yunnan Vegetation, Science Press, Beijing, 1987 (in Chinese). [27] K.L. Bell, L.C. Bliss, Alpine disturbance studies: Olympic National Park USA, Biological Conservation 5 (1) (1973) 25–32. [28] R.E. Manning, Impacts of recreation on riparian soils and vegetation, Water Resource Bulletin 15 (1) (1979) 30–43. [29] W.E. Hammitt, D.N. Cole, Wildland Recreation: Ecology and Management, second ed., John Wiley & Sons, New York, 1998.