Measurement of above- and belowground plant pathogens
For foliar fungal diseases, we recorded foliar fungal disease severity following the methods provided in Liu et al. (2017). In brief, we visually recorded disease severity (i.e. % leaf area covered by fungal lesion;Vi ) from five leaves randomly selected from five individuals for each plant species in each plot. We recorded all available leaves for species with less than 25 leaves. We then calibrated our records by comparing diseased leaves to reference images of known disease severity. We identified the foliar fungal pathogens using an Olympus CX33 light microscope (Shinjuku, Japan) following identification manuals, including the Fungal Identification Manual (Wei, 1979), Plant Disease Diagnosis (Lu, 1997), and also previous studies in this area (Zhang, 2009; Liu et al., 2019; Liu, Lu, et al., 2020). We defined community pathogen load (PL ) following Mitchell et al. 2002 as:
\begin{equation} PL=\frac{\sum_{i=1}^{S}{b_{i}V_{i}}}{\sum_{i=1}^{S}b_{i}}\nonumber \\ \end{equation}
where S was the total number of plant species in a certain plot, and bi was the aboveground biomass of plant species i . We then defined a ‘disease proneness index’ (hereafter ‘Pi ’) for each species as the average severity index (Vi ) across 30 plots of plant species i . We then calculated a ‘community proneness index’ (hereafter ‘Proneness ’) for each plot by calculating a plant aboveground biomass-weighted average of the Pifor each plot (Liu et al., 2017):
\begin{equation} Proneness=\frac{\sum_{i=1}^{S}{b_{i}P_{i}}}{\sum_{i=1}^{S}b_{i}}\nonumber \\ \end{equation}
where Proneness was the expected community pathogen load based on constituent host plant species, which was measured independent of the actual disease in a given plot (Liu et al., 2017). Specifically, each species in each plot was assigned a value of disease proneness based on averaging its disease severity (Vi ) in this plot and weighted by its aboveground biomass. Despite a similar mathematical formula for PL andProneness , these values represent two distinct characters of plant community (the actual amount of disease in a community and the amount that would be expected based on the identity of species present and their relative abundances alone). PL and Proneness are not always correlated with one another (e.g. Liu et al., 2019), and are often used together in disease ecology studies (e.g. Mitchell et al., 2002; Johnson et al., 2013; Liu et al. 2017; Liu et al., 2019). We log-transformed community pathogen load (PL ) and disease proneness index (Proneness ) to achieve normality of residuals in the following analysis.
For soil fungal pathogens, we defined fungal taxa as putative plant fungal pathogens when they include any pathogenic species which were reported to induce any plant disease symptoms (e.g. canker, rot, leaf spot, blight, rust and mildew) (Liang et al., 2016), as determined by references to published data (Tedersoo et al., 2014), paper inISI Web of Science (if any paper reported their pathogenicity) and the FUNGuild algorithm (Nguyen et al., 2015). Indeed, even if some genera with mixed feeding strategies (e.g. parasitic, mutualistic and saprophytic) were presumed as pathogens based on the above method in this study, they still represent pathogen potential. For instance, genera belong to Dothideomycetes (e.g. Alternaria ,Epicoccum , Fusicladium ), Leotiomycetes (e.g. Coma ,Erysiphe , Scytalidium ), Sordariomycetes (e.g.Fusarium , Neonectria , Valsa ) and other classes were presumed as plant pathogenic. All plant pathogenic genera identified in the field study are listed in Table S2.1. We then calculated the accumulated OTU number of soil fungal pathogens (sfpOTUs ; OTU richness of soil fungal pathogens), and also the relative abundance of soil fungal pathogens (sfpRA ; copy number of soil fungal pathogens divided by the total number of copy number of soil fungus) for each sample. We log-transformed soil fungal pathogen relative abundance (sfpRA ) to achieve normality of residuals in the following analysis.
Limitations and caveats in methodology
This study includes three complementary measurements of plant pathogens: the relative abundance of pathogens causing foliar disease, the relative abundance of soil pathogens (sfpRA ), and the richness of soil pathogens (sfpOTUs ). We measured damaged on the leaves and fungal pathogen communities in the soil. Although these two approaches do not provide identical assessments of pathogen richness and disease outcomes in foliar and soil- compartments, we believe that this approach is justified, as it reflects the most commonly used approach in these two respective fields of research, and represents comprehensive characteristics of both above- and belowground fungal plant pathogen communities, and thereby provides a more comprehensive understanding of how pathogen communities respond to changing environmental conditions. Furthermore, although the measurements are not identical, they are often positively correlated with one another, and this correlation often transcends study systems (Rottstock et al., 2014; Liu et al., 2016; Halliday et al., 2017; Halliday et al., 2020b). Therefore, we feel confident that the distinct measurements in our study can provide insight into the biogeographic pattern of pathogens across elevation gradients.
In this study, we measured foliar fungal diseases and soil pathogens as representatives of above- and belowground plant pathogens respectively to bring together research from two different fields that tend to study pathogens in different ways, which provides complementary information using complementary measurement approaches. However, we applied these measurements with caveats that visual assessment for foliar fungal diseases does not include all pathogens and can result in an incomplete assessment of pathogen diversity, while sequencing-based assessment for soil fungal pathogens is not directly related to any particular disease outcome. Overall, these measurements still reflect the most commonly used approach in these two respective fields of research, and thereby provide a more comprehensive understanding of how pathogen communities respond to changing environmental conditions.
Root diseases resulted by root-borne pathogens can also affect host mortality, growth and productivity, and further influence ecological succession and biogeochemistry process, thereby regulating ecosystem functioning (e.g. Hansen & Goheen, 2000; Healey et al., 2016). Future studies could combine surveys of foliar and root fungal diseases by both visual measurements and sequencing to more comprehensively explore the responses of above- and belowground plant pathogens to abiotic and biotic factors and their impacts on ecosystem functioning.
The results of the field survey might be sensitive to limitations of the empirical approach. For soil fungal sequencing, fungal ITS1 region may suffer from certain taxonomic biases, like high proportion of mismatches and biased amplification of certain fungal taxa (e.g. basidiomycetes; Tedersoo & Lindahl, 2016). However, fungal ITS1 region indeed possess some advantages which outperformance to other regions, for example, it is easily discriminate fungal taxa from plants and provides wider richness and taxonomic coverage (Mbareche et al., 2020). Future studies could overcome these challenges by incorporating multiple sequencing regions, and using more advanced methods (e.g. exact sequence variants (ESVs), which generate a greater resolution than OTU-based methods; Mbareche et al., 2020). In fact, quantitative PCR is a good choice to calculate the abundance of pathogens (Tellenbach et al., 2010), although the small datasets prevent us from further analyses regarding the absolute abundance of pathogens. In fact, unlike the foliar fungal disease, the relative abundance of soil pathogen profiles cannot indicate the absolute abundance of pathogens. Hence, we can only conclude that elevation had no associated with the relative abundance of soil pathogens. Our empirical results also stem from a single location in a single year, and thus our results might be sensitive to local environmental conditions that are characteristic of the particular year of sampling. Although our empirical results largely agree with the results of the meta-analysis, this does not negate the limitations of the empirical study (i.e., relatively small sample size, single gradient, single year). Future studies conducted over multiple elevational gradients and multiple years remain the gold standard for empirical field surveys along elevational gradients. Large-scale and long-term studies of biotic and abiotic drivers of disease across environmental contexts remain a pressing need if ecologists want identify the underlying effects of temporal dynamics and spatial heterogeneity in the community ecology of infectious disease.