Beschreibung:
<jats:title>Abstract</jats:title><jats:sec><jats:title>Aim</jats:title><jats:p>To assess richness patterns of ferns and lycophytes in relation to climatic factors in Eastern Asia, analysing eight elevational gradients, situated along a latitudinal gradient.</jats:p></jats:sec><jats:sec><jats:title>Location</jats:title><jats:p>East and Southeast Asia (4° S to 43.3° N).</jats:p></jats:sec><jats:sec><jats:title>Methods</jats:title><jats:p>Ferns and lycophytes were recorded in 480 plots of 400 m<jats:sup>2</jats:sup> along eight elevational gradients. We related macroclimatic factors to local fern species richness using generalized linear models. We tested the predictive power of macroclimatic variables by a leave‐one‐(gradient‐) out cross‐validation (<jats:styled-content style="fixed-case">LOOCV</jats:styled-content>) and tested the predictive power of each model using Spearman's rank correlation, and predicted the respective models spatially. Model residuals were related to two available local (plot scale) factors, terrain inclination and terrain heterogeneity using simple linear regression.</jats:p></jats:sec><jats:sec><jats:title>Results</jats:title><jats:p>Local fern species richness declined linearly towards higher latitudes, and simultaneously elevational richness peaks shifted towards lowlands with increasing distance from the equator, transitioning from unimodal to linear patterns. Temperature was the macroclimatic factor with the highest predictive power for fern species richness. In combination with other variables, especially cloud cover, macroclimatic factors could explain up to about 60% of species richness distributions.</jats:p></jats:sec><jats:sec><jats:title>Main conclusions</jats:title><jats:p>Macroclimatic factors, such as temperature and cloud cover, are able to predict a substantial proportion of the variation in local fern species richness in East and Southeast Asia. Using predictive models based on <jats:styled-content style="fixed-case">LOOCV</jats:styled-content> substantially increases the accuracy of species richness predictions as compared to simply searching for correlates along all gradients.</jats:p></jats:sec>