1 Introduction

Alternaria and Cladosporium are highly prevalent and their spores represent a major component of airborne biological particles. Their spores are a common trigger of allergy and asthma symptoms and plant diseases (Guarneri et al., 2008; Weryszko-Chmielewska et al., 2018). Several studies have shown that weather is an important factor in determining airborne spore concentrations (Lacey, 1996). Among these, temperature, wind and precipitation play crucial roles in spore production, aerosolization, transport and deposition (Rodríguez-Rajo et al., 2005). Other factors besides weather are less known, such as biotic factors, for example the effect of animal activity on airborne spore concentrations. It has long been recognized that the growth of Alternaria and Cladosporium on plant leaves follows upon a heavy infestation of the trees, cereals and other plants by different species of plant-sucking insects (Hemiptera) (Friend, 1965; Hughes, 1976) including aphids.

Aphids are a predominantly northern temperate group and more than 75% of the species are known from the Palearctic region (Holman, 2009). Prior to the introduction of molecular markers, much was known about aphid population structure and movement, especially through the pioneering studies at Rothamsted that evolved into the Europe-wide suction trap network (Taylor, 1951; Allison & Pike, 1988, Macaulay et al., 1988). Of the present-day subfamilies of Aphididae, one, in particular, was successful at exploiting the rapid expansion of numbers and diversity of herbaceous flowering plants in the Tertiary period. This subfamily, the Aphidinae, with 2750 + extant species, is predominantly a northern temperate group, with life cycles closely tied to temperate seasonality and the phenology of temperate plants (van Emden & Harrington, 2007).

Hemipteran insects, especially aphids (Aphidoidea), have sucking and piercing mouthparts capable of piercing the foliage or some other part of the host plant and then extracting large volumes of plant sap. Phloem sap offers a near continuous supply of small organic compounds, principally sugars and amino acids, that require the minimum of digestive processing, and consequently, the assimilation efficiency of aphids is exceptionally high. Thanks to a special adjustment of the digestive tract with the filter ventricle, they can quickly get rid of unnecessary liquids contained in plant sap. In species of Hemiptera which have filter ventricle there is a twice- and three times more intense excretion of honeydew drops. This sugary material covers the surfaces of the host plants, drips down around them onto the surrounding objects. Honeydew then serves as nutrient for the development of fungi (Auclair, 1963; Flessa et al., 2012; Magyar et al., 2016). Certain fungi, especially sooty moulds, saprophytes and some leaf pathogens utilize honeydew while colonizing leaves by attachment of spores and their germination, mycelial growth and spore production. Among the fungal genera feeding on honeydew, Alternaria and Cladosporium are common in the temperate region (Flessa et al., 2012). Despite having an apparently strong impact on fungi, the effect of the population dynamics of aphids on the airborne levels of fungal spores has never been studied. Therefore, in this study, we test the hypothesis that there should be a positive relationship between the occurrence of aphids and airborne fungal spore levels. In addition, we select aphid species and weather patterns that may serve as probable indicators in forecasting the spore concentration of airborne Alternaria and Cladosporium.

2 Materials and methods

2.1 Sampling locations

Global maps of airborne spore (Buters et al., 2018) and aphid (Harrington et al., 2012) monitoring stations were reviewed and compared, to select nearby (< 60 km) monitoring stations with parallel spore and aphid data. Based on these criteria, three study areas were selected: Kecskemét/Szolnok (Hungary, 1996–1997), Leicester/Wellesbourne (UK, 2006–2017) and Poznań/Winna Góra (Poland, 2000–2013). For the sake of clarity and simplicity of the message, the monitoring stations will be referred to as: Szolnok (Hungary), Leicester (UK), and Poznań (Poland). Because of the short monitoring period, the modelling analysis has not been performed on the Szolnok data; however, it was included in the manuscript to show the monitoring data of the third, a climatically different sampling area. Distance between spore and aphid monitoring stations are shown in Fig. 1.

Fig. 1
figure 1

Distance between aphid (triangle) and spore (diamond) monitoring stations. Climate data of spore monitoring stations are from Climate-Data.org. Red line: average temperature per month. Blue bars: total rain. Maps were modified from Wikipedia and Google Maps

Full size image

2.2 Aphid sampling

Daily sum of aphid data was collected with a Rothamsted type suction trap in Szolnok, Wellesbourne and Winna-Góra for the same periods as spores (Burkard Manufacturing, Rickmansworth, UK, and Rothamsted Insect Survey, Rothamsted, UK, Maculay et al., 1988). Air was drawn in at 3000 m3 air /hour, at a height of 12.2 m. The diameter of the collecting tube was 20 cm. The collecting jar was changed daily at 7.00 a.m. during the experiment. The collected specimens were preserved in 70% ethanol. Each of the examined specimens was mounted on slides, and they were identified according to diagnostical keys (Blackman & Eastop, 1994, 2000; Remadiére & Remadiére, 1997; Taylor, 1984). A total of 36 aphid taxa were monitored at the studied areas (Table 1) Among these taxa, those having long datasets and occurring at more than one monitoring station were selected for further statistical analysis; moreover, to represent the genus Aphis, A. craccivora and A. fabae forms were also selected for further analysis from Poznań and Leicester, respectively (N = 9).

Table 1 Aphid taxa monitored in Leicester (UK), Poznań (Poland) and Szolnok (Hungary). Selected taxa are highlighted in bold
Full size table

2.3 Spore sampling

Daily average concentrations of Alternaria and Cladosporium spore data were collected in Kecskemét, Leicester and Poznań for 2 (1996–1997), 12 (2006–2017) and 14 (2000–2013) years, respectively, using volumetric spore traps of the Hirst design (Hirst, 1952, Burkard Manufacturing, Rickmansworth, UK) situated at roof level (at a height of 20, 12 and 33 m a.g.l. in Kecskemét, Leicester and Poznań, respectively) (Fig. 1). Air is sucked into the trap at a rate of 10 l/min through a 2 × 14 mm orifice. Behind the orifice the air flows over a rotating drum that moves past the inlet at 2 mm/h and is covered with an adhesive coated, transparent plastic tape. Particles in the air impact on the tape to give a time related sample (Emberlin, 2000). Following its removal from the trap, the tape is divided into segments corresponding to 24-h periods (48 mm in length). Each segment is mounted between a glass slide and cover slip using a mixture that contains gelatine, glycerine, phenol, distilled water and basic fuchsine (Kecskemét and Poznań) or lactophenol cotton blue (Leicester). The samples are then examined by light microscopy, and spores were identified at × 400 magnification. Spores were counted following the two methods described by Mandrioli et al. (1998), i.e. 1 longitudinal transect or 12 vertical transects of the slide. Daily average (00:00–24:00) spore counts were converted into concentrations and expressed as spores/m3 air (Galán et al., 2017). Because of their high number in the Hungarian samples, spores were enumerated in two microscope field of sight (with a diameter of 0.46 mm) using the 12 vertical transects method. The sum of the spore counts of the 24 areas of each slide was multiplied with a factor (× 32) to produce daily average concentrations.

2.4 Data analysis

Weather data used for the computations have been kindly provided by the Hungarian Meteorological Service, the Leicester City Council Air Quality group, and the Institute of Meteorology and Water Management-National Research Institute (Poznań-Ławica meteorological station).

We classified the monthly sum of aphid individual numbers (\({N}_{i}, i\) is for the taxa) and the monthly sum of spore concentrations (\({S}_{A}\), \({S}_{C}\) of Alternaria and Cladosporium, respectively) that are notably different from each other and name them as “low”, “medium” and “high”. Weather data were classified according to the monthly mean temperature (cold, mild and hot) and monthly sum of precipitation (dry and wet). These classifications were preformed based on graphical representations to find relevant “cut points” in the distributions of the monthly sum of aphid individual numbers and the monthly sum of spore concentrations as well as the weather data. We did this first and careful approach, with the aim of using the classification results to find well-recognizable patterns between classes of occurrences. Other methods of classification were also considered in preliminary analysis (see Supplementary Materials for details). In what follows, we provide more details about the way of classification we performed.

2.4.1 Determination of weather classes

For each month at each site a lower threshold was determined based on the monthly precipitation (mm) and lower and upper thresholds determined for monthly mean temperature (°C). Below the lower thresholds, temperature classes were classified as “cold” and precipitation classes—as “dry”. Above the lower threshold, precipitation classes were called “wet”. Above the upper threshold, temperature classes were called “hot”. Between the two thresholds, temperature classes were called “mild”. The thresholds are given in Supplementary Table 2.

As a result of the classification with the thresholds given in Supplementary Table 2, we got classes with the following properties:

  1. 1.

    The means, medians, first quartiles and third quartiles of the wet groups of precipitation data were at least 3.2; 2.7; 3.1; 2.7 times higher, respectively, than those in the dry groups in every month in Leicester and at least 2.3; 2.1; 2.3; 1.6 times higher in Poznań, respectively. Note that Poznań weather is notably more wet than that in Leicester, so twice as much rain in Poznań would be the equivalent of 2.5–3 times more rain in Leicester.

  2. 2.

    The means, medians, first quartiles and third quartiles of the cold groups of temperature data were on average 1.3 °C higher than those in the mild groups that were 2 °C lower than the hot groups on average in every month in Leicester. In Poznań, these were 1.2 °C; 1.8 °C mean differences, respectively. Note that 1.0 °C difference in monthly mean temperature can make difference, as it means 30 °C extra temperature sum in a month.

  3. 3.

    To check our previous classification, we ran hierarchical (Ward’s method) and k-means clustering (with the number of clusters 2 for precipitation and 3 for temperature data) and compared the results of these two methods with our classification outcomes using Fisher’s exact tests. We found that there was very strong agreement in the clusters of all the three outcomes of classification (all Fisher’s exact tests were highly significant with p < 0.001) and where there were some rare differences between the cluster memberships of our grouping and one of the clustering methods (hierarchical or k-means), there the cluster memberships of the hierarchical and k-means clustering were also slightly different.

After these careful checks, we decided to accept our classification as they are able to show coarse relationships as patterns between weather, spore concentration and aphid numbers, which was our primary aim.

2.4.2 Determination of spore concentration classes

Depending on months and sites, lower and upper technical thresholds were fixed, regarding the monthly sum of spore concentration of Alternaria and Cladosporium. First, for each month i (i runs from March to August), concentration values of different years were ordered from the smallest to the highest. Below the lower thresholds, concentration classes of the relevant year were set as “low” while above the upper threshold, classes were set as “high”. Between the two thresholds, class “low” is changed to “medium” only if the ratio \({S}_{C\left(j+1\right)}/{S}_{C\left(j\right)}\) or \({S}_{A\left(j+1\right)}/{S}_{A\left(j\right)}\) (j is for the rank of the concentration value) is higher than 1.2. (It means that a concentration in a year is said to be “medium” if it is between the lower and upper threshold and it is at least 20% higher than in the year classified as having “low” class of the highest concentration). The thresholds are given in Supplementary Table 3.

With this 20% “jump” in the values, we could separate classes between which the difference was notable. The “jumps” higher than 20% are referring the slopes of a hidden change in spore numbers. This way we could express that there is no pronounced difference if we measure, let us say, 300 or 310 Alternaria spores in July in Leicester, but we wanted to differentiate years with 300 spores from another with 400 or even with 500 spores (Supplementary Table 3, row “July”).

Again, we checked our classification running hierarchical (Ward’s method) and k-means clustering (with the number of clusters 3) and compared the results of these two methods with our classification outcomes using Fisher’s exact tests. We detected again that there was very good agreement in the clusters of all the three outcomes of classification (all Fisher’s exact tests were highly significant with p < 0.001) and where there were some rare differences, there the hierarchical and k-means clustering results did not agree, either.

2.4.3 Determination of aphid number classes

Depending on months and sites, again lower and upper thresholds were determined, regarding the monthly sums of aphid numbers of 9 taxa (Acyrthosiphon pisum, Aphis fabae, Aphis craccivora, Drepanosiphum platanoidis, Euceraphis punctipennis, Metopolophium dirhodum, Myzus cerasi, Pemphigus spp., Sitobion avenae), all were monitored at both sites Leicester and Poznań. Note that A. fabae and A. craccivora were observed in Leicester and in Poznań, respectively. Again, for each month i (i runs from April to July), numbers of individuals collected in different years were ordered from the lowest to the highest. Below the lower thresholds, aphid number classes of the relevant year were set as “low” while above the upper threshold, classes were set as “high”. Between the two thresholds, class “low” is changed to “medium” only if the ratio \({N}_{k\left(j+1\right)}/{N}_{k\left(j\right)}\) (j is for the rank of aphid number, k is for the species) is higher than 1.5 or 1.2 in case of Leicester and Poznań, respectively (except for Euceraphis punctipennis; in this case we used the constant 1.5 at both sites).

The “jump” parameters 1.2 and 1.5 were used to express a remarkable great amount of change in the aphid numbers that was observed at the sites, with the aim to create three main groups the within-variances of which remained still low compared to the between-variances. Our classification results were later checked and confirmed by classification methods described above. The thresholds are given in Supplementary Table 4.

Weather, spore and aphid classes are shown in a heat map (Fig. 2). Relations between aphid numbers and spore concentrations before August were investigated, since after these month, deciduous plants lose all their leaves and herbaceous plants are harvested in the temperate region. No more than one month of lag effect between aphid and fungal counts were considered, based on the relative short time needed for spore development of these fungi. Nevertheless, longer lags may also be considered, as a possible prolonged effect; these are shown in the Supplement. As the sample sizes in the datasets from Hungary are relatively small, they were not analysed statistically.

Fig. 2
figure 2

Heat map showing classes of weather, aphid and spore classes. Their description is found in Data analysis chapter. Colour codes (right bottom corner): C: cold; M mild; H: hot; W: wet; D: dry weather; 1: low; 2: medium; 3: high aphid/spore concentration; 0: zero; N: no data available. Abbreviations: temp.: temperature, Alt.: Alternaria spp., Cla.: Cladosporium spp.; for the abbreviations of aphid names (Ac. pis…Si. ave) see Table 1

Full size image

Crosstabulation was calculated and Fisher’s exact tests were run to select significant relationships between weather and aphid number classes, aphid number and spore concentration classes, and weather and spore concentration classes. The adjusted standardized residuals in cells above 1.96 or below − 1.96 (Agresti, 2002) that indicates that the number of cases in those cells is significantly higher or lower, respectively, than it would be expected if null hypothesis were true, with a significance level of 0.05 were also considered in order to find out the over- and underrepresented cells in the crosstabs. We aware that the selected significant results do not necessarily refer to causal relationships, so our conclusions are formulated in a cautious way.

To find out the importance order of the variables that can affect the spore concentration rates, hundreds of Random Forest models were run with optimized parameters of the algorithm “the number of trees”: ntree = 500 and “the number of random subsets of the explaining variables”: mtry = 2. For the most successful models, the variable importance was calculated based on the overall mean decrease in accuracy (Kuhn, 2020). Finally, the most frequent order was chosen to report. Note that these methods provide no information on whether a variable is important as a main effect or as part of an interaction. Model development was conducted in R 4.1.3 (R Core team, 2021) with packages “caret” (Kuhn, 2020).

3 Results

3.1 Associations of aphid numbers and spore concentrations

Various weather classifications (i.e. hot vs. cold, dry vs. wet) were evenly distributed in the studied areas (Fig. 2). Weather in April and May had a significant effect on aphid numbers in both areas. Significant associations were also found between spore concentrations and meteorological factors. Cold weather was associated with low spore concentrations in most cases. (The detailed results of direct associations of weather and aphid numbers, as well as weather and spore concentrations are shown on flowcharts in the Supplement, Figs. S1-S8.)

The main focus of this study, namely the effects of aphid numbers on spore concentrations, showed significant associations (Fisher’s test p < 0.05 with significant adjusted standardized residuals of absolute values above 1.96). The results are summarized on flowcharts (Figs. 3, 4, 5 and 6). In Poznań, effects were usually observed one month earlier than in Leicester.

Fig. 3
figure 3

Flowchart showing significant positive associations between weather, aphid and spore concentrations in Leicester. Abbreviations: Alt.: Alternaria spp., Cla.: Cladosporium spp.; for the abbreviations of aphid names see Table 1. Read the flowchart like this: “When the weather in month 5 (i.e. May) was mild and dry, then in month 6 (June) Aphis fabae had significantly more frequently low concentrations. Then in month 6 (June) Cladosporium had significantly more frequently low spore concentrations”

Full size image
Fig. 4
figure 4

Flowchart showing significant negative associations between weather, aphid and spore concentrations in Leicester. Abbreviations: Alt.: Alternaria spp., Cla.: Cladosporium spp.; for the abbreviations of aphid names see Table 1. Read the flowchart like this: “When the weather in month 3 (i.e. March) was hot and wet, then in month 7 (July) Drepanosiphum platanoidis never had low concentrations Then in month 7 (July) Alternaria never had high spore concentrations”

Full size image
Fig. 5
figure 5

Flowchart showing significant positive associations between weather, aphid and spore concentrations in Poznań. Abbreviations: Alt.: Alternaria spp., Cla.: Cladosporium spp.; for the abbreviations of aphid names see Table 1. Read the flowchart like this: “When the weather in month 4 (i.e. April) was wet, then in month 4 (April) Drepanosiphum platanoidis had significantly more frequently low concentrations. Then in month 4 (April) Alternaria had significantly more frequently low spore concentrations. Spore concentrations were so even if month 4 (April) was cold”

Full size image
Fig. 6
figure 6

Flowchart showing significant negative associations between weather, aphid and spore concentrations in Poznań. Abbreviations: Alt.: Alternaria spp., Cla.: Cladosporium spp.; for the abbreviations of aphid names see Table 1. Read the flowchart like this: “When the weather in month 4 (i.e. April) was hot, then in month 5 (May) Aphis craccivora never had low concentrations. Then in month 5 (May) Cladosporium never had high spore concentrations”

Full size image

A probable direct effect of weather on the spore concentration was considered by the observed significant relationship of weather variables and spore concentrations. In what follows, we list the results with significant associations based on Fisher’s exact tests and adjusted standardized residuals (p < 0.05).

Acyrthosiphon pisum counts were associated with spore numbers in Leicester. Cold weather in April and June were followed by low aphid counts and low Alternaria spore counts in June.

Aphis craccivora showed a similar pattern. Cold weather in April was observed with low aphid numbers in May, which were followed by low Cladosporium spore concentrations in the air in June. Conversely, after hot weather in April the numbers of A. craccivora in May was higher, and Cladosporium spore concentrations increased in May. In this scenario, a probable direct effect of weather on the spore concentration should also be considered.

Aphis fabae populations were associated negatively with mild and dry conditions in May. Due to this their numbers were low in June, and there were low Cladosporium spore levels. When July was rainy, this aphid never had low populations in July, and Alternaria spore concentrations were never low in August. A hot May was followed by a moderate number of this aphid species in June and July. Moderate numbers of Cladosporium spores were measured in June, and Alternaria spore levels dropped in August.

Drepanosiphum platanoidis showed a positive association with hot and wet weather in March, and hot weather in July in Leicester. Proliferation of aphid populations in July coincided with high Alternaria spore levels in July. In Poznań, wet weather in April was associated with low aphid and Alternaria spore numbers in the same month.

Euceraphis punctipennis’ effects on spore numbers seem to be suppressed by the direct influences of meteorological factors in Leicester. In Poznań, however, it showed stronger, more direct effects. Here, aphid numbers were increased by hot and wet weather in May; the resulting high aphid counts in June led to high Alternaria spore concentrations in June and July. Hot and dry weather had only a moderately positive effect, while cold and wet weather acted negatively, by decreasing aphid and Alternaria spore numbers in June and July, respectively.

Metopolophium dirhodum was affected negatively by cold weather in Leicester in April, with a result of low aphid and Alternaria spore counts in June. The results seem to show the effects of aphids on spore counts, as the weather showed no direct effect on spore levels.

Myzus cerasi seemed to have only a moderate effect on spore counts, mostly by increasing population numbers in hot May and June in Leicester. This species had no detectable effect on airborne spore concentrations in Poznań.

Sitobion avenae had no detectable effect on airborne spores in Leicester or Poznań.

3.2 The relative importance of the aphids on the spore concentration

To estimate the relative importance of aphid species on the airborne spore concentrations, random forest analysis was performed where meteorological variables were also included. The analysis showed the importance of meteorological factors at the first place in most cases. However, aphid species also had important contribution in some periods (Fig. 7).

Fig. 7
figure 7

Relative importance of aphids and meteorological factors on Alternaria and Cladosporium spore concentrations in Leicester and Poznań. T: temperature classes, R: precipitation classes. 3–7: months

Full size image

In April, temperature was the prevalent factor having a major effect on the counts of Alternaria and Cladosporium spores (both in 100%) in Poznań. The effect of April number of D. platanoidis was considerable too, being at the second place among the factors with 44% for Cladosporium and 61% for Alternaria.

In May, no significant factors were detected.

In June, the situation was similar in Leicester to those observed formerly in Poznań, temperature and D. platanoidis being at the first and second place, respectively. Significant effects were found with April temperature for Alternaria spore concentrations and April and May temperature for Cladosporium spore concentrations. The effect of D. platanoidis was stronger in Leicester than in Poznań, as Alternaria spore concentrations were affected here by this aphid by 80% and Cladosporium spore concentrations – by 79%. May and June numbers of this aphid species had a significant effect on June concentrations of Alternaria and Cladosporium spores, respectively.

In July, the effect of aphids was dominant in both cities. In Poznań, the importance of June numbers of E. punctipennis weighted 100% for Alternaria spores (meantime, June temperature was responsible for 39%). In Leicester, June numbers of A. pisum was the main factor (100%) but the contribution of June temperature and July precipitation were also high (92% and 87%, respectively) on Alternaria spores. Some effect of June number of M. cerasi (64%, 4th place) was also observed on this fungus. For Cladosporium, apparently, aphids were not a relevant factor in this month.

In August, the contribution of July number of D. platanoidis was high on Cladosporium spores (80%) in Leicester, but placed only as the 2nd one on the factors’ rank. Here, July precipitation had the first place (100%). The effect of aphids on Alternaria spores was negligible in August. Some aphid species seem to have a considerable effect on Alternaria and Cladosporium spore concentrations from mid-spring to mid-summer.

To summarize, D. platanoidis showed major effect on spore concentrations, especially when this aphid’s numbers were high in April, May, June and July in Leicester, and in April in Poznań. The contribution of A. pisum, E. punctipennis and M. cerasi (in descending order of importance) was considerable too (above 60%) in July. Other aphid species showed no effects on airborne levels of monitored fungal spores in Leicester or Poznań. This does not mean that there is no other aphid that has pronounced effect, just that our dataset was not appropriate to show it.

4 Discussion

In this study, the effects of aphids on spore seasons of Alternaria and Cladosporium were analysed over an 18-year period (2000–2017). Weather strongly affected both aphid populations and fungal spore concentrations in the air. In most cases, a direct effect of weather on fungal spore levels was found. Direct effect of aphids on atmospheric spore levels was identified only in two cases: (1) A. pisum levels had a direct effect on the Alternaria spore concentrations in Leicester in cool Junes. If June was cold, the A. pisum concentrations were significantly lower as were the Alternaria spore concentrations. (2) June E. punctipennis high levels had a direct effect on the Alternaria spore concentration in Poznań in July. When the weather in May was hot and wet, then in June E. punctipennis levels were significantly higher, and in July Alternaria spore levels were significantly higher. At other times, D. platanoidis, E. punctipennis, A. pisum and M. cerasi had a considerable, but possibly indirect (or secondary) effect on airborne spore concentrations.

D. platanoidis often forms abundant colonies on Acer spp., wherever they are grown in Europe North America, North Africa, Central Asia, Australia and New Zealand. It is an important source of honeydew (Crane & Walker, 1985). D. platanoidis and E. punctipennis are regarded to be the main honeydew-producers (Dixon, 1963; Prabucki, 1972). E. punctipennis feeds predominantly on downy birch (Betula pubescens) and exudes copious amounts of honeydew which coats the leaves on which it feeds. A. pisum is feeding on about 20 genera in the family Fabaceae, having a worldwide distribution in temperate climates. It is a major pest of pea (Pisum) and alfalfa (Medicago). In A. pisum, honeydew droplet volume, frequency and rate of excretion were proportional to the susceptibility of the host pea variety on which the aphids were feeding (Auclair, 1959). Therefore, the selection of less susceptible plants was proposed. This action may be important not only for a higher yield of pea crops, but also for reducing honeydew and fungal proliferation. The colonies of M. cerasi aphid are often very large and frequently thickly covered in honeydew. This aphid migrates from cherry (Prunus cerasus, P. avium) as the primary host to eyebrights (Euphrasia spp.), bedstraws (Galium spp.) and speedwell (Veronica spp.) as the secondary hosts. Cinara pectinatae was not present in our samples, but it is regarded to be the main honeydew producer in Poland (Rybak-Chmielewska et al., 2013) mainly in mountainous areas of the Southern regions wherever it is foraging on its host plant, silver fir (Abies alba) (Durak et al., 2011; Szelegiewicz, 1968). It is speculated that fungal concentrations in the air may be high near silver fir forests attacked by aphids, but this theory needs further studies.

The amount of honeydew produced depends on several factors, i.e. aphid species and stage, host plant and its susceptibility, water stress and climate (Auclair, 1958, Kunel, 1997, Fischer et al., 2005, Morley & Lewis, 2014). In addition to honeydew, colonization and abundance of phylloplane fungi are also influenced by other factors, such as water availability, temperature and UV radiation (Breeze & Dix, 1981; Newsham et al., 1997; Zak, 2002). The effect of aphid honeydew on fungi has been investigated in several studies. In a field experiment, the aphids (Sitobion avenae) and the honeydew they excreted significantly increased the quantity of Cladosporium on wheat during one week, at early July, one week after the peak aphid population. Cladosporium was significantly increased by the presence of aphid colonies in a glasshouse experiment too (Ajayi & Dewar, 1983). In another study (Fokkema et al., 1983) the colonization of flag leaves of wheat by Cladosporium cladosporioides was highly stimulated by honeydew when it was added to the fungal inoculum (in a mixed population of fungi). After a growth period of 6 days, the population densities of fungi showed a tenfold increase in the presence of honeydew. Although the positive effect of the aphids on fungal populations on leaves is well-documented, their effect on airborne fungal levels is more complex and can depend on several factors. (i) We measured aphid population by air sampling. There are three major peaks of aphid aerial populations (McManus, 1979). The first is in May–June, when they migrate from winter to summer hosts; the second wave occurs in July, when the winged forms redistribute; the third is in September and October, when most aphids return to their winter hosts (McManus, 1979). Apparently, the most honeydew is produced around the second migration wave. In the first wave, flying forms are not producing too much honeydew unlike non-winged forms. (ii) Insects besides aphids also have sucking mouthparts (like cicadas, leafhoppers and scale insects) and can produce large amounts of honeydew. (iii) The intensity of honeydew production varies greatly from year to year according to the annual population size of honeydew-producing aphids and other insects. Honeydew production may depend on weather and other environmental factors, like the activity of aphid-associated ants. It has been shown repeatedly that the presence of a nest of ants (especially the red wood ant, Formica polyctena) increases the honeydew production as they feed on honeydew and in return they protect honeydew-producing insects from predators. In a 9-year set of records average honey yields were 43% higher in a forest area with ants than in a comparable area without them (Wellenstein, 1983). The above-mentioned factors (i–iii) are strongly limiting our observations, despite this, some effects on airborne spores of Alternaria and Cladosporium could be shown in some aphid species.

On cereal crops, the most common species of sooty moulds are Alternaria and Cladosporium, but the latter usually predominates (Ajayi & Dewar, 1983). On trees, the most common sooty mould species is a mixture of Cladosporium and Aureobasidium pullulans covering the living leaves of a great variety of hosts, especially Tilia, Salix, Ulmus (and Celtis according to the author’s observation) in Europe and North America (Hughes, 1976; Snieškienė et al., 2016). Alternaria alternata and Cladosporium oxysporum were found along with 45 other saprophytic phylloplane fungi (e.g. Capnodium, Curvularia and Torula spp.) as a member of the sooty mould biomass, colonizing citrus leaves attacked by citrus black fly, citrus whitefly and citrus psylla (Srivastava & Thakre, 2000). Invasive hemipterans like Metcalfa pruinosa (Wilson & Lucchi, 2007) and Coccus pseudomagnoliarum (Fetykó et al., 2013) are devastating pests on urban green areas, their honeydew and consequently sooty moulds are covering not only street trees, but also stone and metal objects, e.g. fences, pavement or even, cars (Friend, 1965; Hughes, 1976). In a study performed by Levetin and Dorsey (2006), the types of fungi found on leaf surfaces were compared with those in the atmosphere. Cladosporium and Alternaria spp. were common on the leaves, and they had a major contribution to the air spora as well. These authors found a significant correlation between leaf surface fungi and airborne fungi. In Poznań, the highest concentrations of Cladosporium and Alternaria spores are often observed in July (Grewling et al., 2019; Kasprzyk et al., 2016). Presumably, aphids may be a “triggering factor” for spore season start in Poland, but more data are needed. Alternaria and Cladosporium are saprotrophic fungi, which feed on honeydew but also on dead plant biomass. Being primary colonizers of the plant surface, these fungi have the advantage of gaining access to readily available organic compounds in freshly fallen leaves. The amount of litter is increasing due to plant senescence in the temperate regions in late summer and autumn, when this large quantity of plant biomass provides food for these fungi. The airborne concentration of Alternaria and Cladosporium spores is still high in these months, but the contribution of honeydew to the fungal growth may be less than that of the decaying plant biomass. Large amounts of Alternaria and Cladosporium spores are released during the harvest of infested crops which could also contribute to high levels of spores in late period of the season (Olsen et al., 2019a, 2019b; Skjøth et al., 2012; Weikl et al., 2015). Therefore, it is plausible that litterfall and harvest may mask the effect of honeydew on the airborne spores of Alternaria and Cladosporium in autumn.

Honeydew and sooty moulds are mostly regarded as a cosmetic problem on the plants, as it is unsightly. However, as a major source of airborne spores of allergenic Alternaria and Cladosporium, they deserve more attention. Informing the public about Alternaria and Cladosporium spore concentrations is a preventive tool to fight against allergic diseases. High levels of these fungi are predicted by forecast models, which are usually based on meteorological factors. Biotic factors, such as availability of food for fungi (in this case, aphid-produced honeydew), are rarely considered. Meteorological factors can only partially explain the variation of airborne concentration of fungi in most studies (in 1.0–49.7%, Hjelmroos, 1993, Grinn-Gofroń et al., 2016, Ščevková & Kováč, 2019). We propose that the unexplained part of these models may be the effect of biotic factors, such as aphid activity, as their contribution was significant in our analysis (Fig. 7). The predictive value of aphids (at least in 4 species) on spore concentration shows that aphid monitoring data (like weather data) should be used in the preparation of Alternaria and Cladosporium forecast in Leicester and Poznań, in spring and summer. Unfortunately, aphid monitoring data are rarely available near spore monitoring stations. In Hungary, aphid samples were available only for two years, and only two of the taxa effective in Leicester and Poznań (A. pisum and M. cerasi) were counted. High June A. pisum numbers were followed by high July Alternaria spore concentration in one year; however, this observation cannot be accepted without further monitoring and analysis. Therefore, we propose the continuation of the operation of suction- traps to monitor aphids in Hungary, as well as in other countries where fungal forecast  models are developed.

With an accelerating climate change and growing global population, monitoring of crop diseases is urgently required (Mothapo et al., 2022). Diseases and pests that can cause distinctive plant symptoms, like sooty moulds (including Alternaria and Cladosporium) are good candidates for remote sensing (Yang, 2022). With the advances in remote sensing technologies, it seems feasible in the near future to detect sooty moulds on crops with remote sensing and integrate the data into forecast models of airborne spore concentrations.

It is important to mention that the species composition of aphids is different in different parts of the world. Some aphid species that are absent or rare in areas investigated in this study may have high number and honeydew production in other countries. The contribution of insect data to the spore numbers merits further studies in areas where honeydew production is especially high, like in Greece, Turkey and other Mediterranean areas; in New Zealand, in certain parts of North America and in the famous Black Forest in Germany (Kunkel, 1997, Crane et al. 1985).

5 Conclusion

It is already known that some aphid species, due to their feeding habit and excretion of honeydew, have a major effect on fungal proliferation. Setting out from this, by detecting significant relationships, we aimed to select the most relevant species probably having this effect. However, statistical methods that can detect relationships are not appropriate for proving causal effects in an exact way. The reason of it is that significant association can be the result of direct or indirect causality (in a specified direction) or the outcome of confounding when both the related variables are depending on a third explanatory variable. Nevertheless, based on our results, the aphid species having significant association with the considered weather variables as well as Alternaria and Cladosporium spore concentration deserve special attention in the future. Therefore, regular observations of these species are highly recommended since reliable aphid data can be useful in forecasting of airborne concentrations of allergenic spores of Alternaria and Cladosporium. For a precise proof of the potential causal relation, not only observations but also further well-designed experiments are necessary.