1Aim: General trends in spatial patterns of macroscopic organisms diversity can be reasonably 2 well predicted from correlative models, using for instance topo-climatic variables for plants and 3 animals allowing inference over large scales. By contrast, soil microorganisms diversity is gener-4 ally considered as mostly driven by edaphic variables and, therefore, difficult to extrapolate on a 5 large spatial scale based on predictive models. Here, we compared the power of topo-climatic vs. 6 edaphic variables for predicting the diversity of various soil protist groups at the regional scale. 7 Location: Swiss western Alps. 8 Taxa: Full protist community and nine clades belonging to three functional groups: parasites 9 (Apicomplexa, Oomycota, Phytomyxea), phagotrophs (Sarcomonadea, Tubulinea, Spirotrichea) 10 and phototrophs (Chlorophyta, Trebouxiophyceae, Bacillariophyta).11Methods: We extracted soil environmental DNA from 178 sites along a wide range of elevations 12 with a random-stratified sampling design. We defined protist Operational Taxonomic Units as-13 semblages by metabarcoding of the V4 region of the ribosomal RNA small sub-unit gene. We 14 1 assessed and modelled the diversity (Shannon index) patterns of all selected groups as a function 15 of topo-climatic and edaphic variables using Generalized Additive Models.
16Results:The respective significance of topo-climatic and edaphic variables varied among taxo-17 nomic and -to a certain extent -functional groups: while many variables explained significantly 18 the diversity of phototrophs this was less the case for parasites. Generally, topo-climatic vari-19 ables had a better predictive power than edaphic variables, yet predictive power varied among 20 taxonomic and functional groups.
21Main conclusions:Topo-climatic variables are, on average, better predictors of protist diversity at 22 the landscape scale than edaphic variables, which opens the way to wide-scale sampling designs 23 avoiding costly and time-consuming laboratory protocols. However, predictors of diversity differ 24 considerably among taxonomic and functional groups; such relationships may be due to direct 25 and/or indirect, e.g. biotic influences. Future prospects include using such spatial models to 26 predict hotspots of diversity or pathogens outbreaks. 27