Climate is an important driver for stream diatom distributions

Abstract

AbstractAimThe species distributions of macroorganisms are widely studied, yet microbial distributions at species level remain poorly resolved. We explored the relative contributions of climatic and local environmental factors in explaining the distribution patterns of unicellular diatoms.LocationA geographical gradient of c. 1200 km in Finland (60°–70° N).MethodsWe modelled the distributions of 157 diatom taxa sampled in 227 stream sites using climatic and local environmental predictors with four different modelling techniques: generalized linear models, generalized additive models, boosted regression trees and Random Forest. We used models with three separate sets of predictors: environment only, climate only and the full set of predictors. The model performances were evaluated using the area under the curve of a receiver operating characteristic plot and the true skill statistic values by a four‐fold cross‐validation approach.ResultsWe found that the predictive performance of the full models was highest, indicating the importance of both the local environment and large‐scale climatic factors in diatom distribution patterns. However, climate‐only models outcompeted the environment‐only models in predicting diatom distributions. The explanatory variables had varying importance across species and growing degree days and precipitation had the highest relative importance in the full models. We also found that the predictability of the distributions varied greatly among species, but the differences among families were typically small.Main conclusionsOur results suggest that at a broad geographical scale climate‐related factors are important determinants of diatom distributions and may be stronger drivers than local variables. The inclusion of both climatic and local environmental factors in species distribution modelling facilitates the understanding of the joint effects of these drivers on microorganisms in future conditions. From an applied perspective, our study demonstrated that species distribution models serve as an important tool in explaining and predicting microbial distributions.