Shifting taxonomic and functional community composition of rivers under land use change

Abstract

Land use intensification has led to conspicuous changes in plant and animal communities across the world. Shifts in trait‐based functional composition have recently been hypothesized to manifest at lower levels of environmental change when compared to species‐based taxonomic composition; however, little is known about commonality in these responses across taxonomic groups and geographic regions. We investigated this hypothesis by testing for taxonomic and geographic similarities in the composition of riverine fish and insect communities across gradients of land use in major hydrologic regions of the conterminous United States. We analyzed an extensive dataset representing 556 species and 33 functional trait modalities from 8,023 fish communities, and 1,434 taxa and 50 trait modalities from 5,197 aquatic insect communities. Our results demonstrate abrupt threshold changes in both taxonomic and functional community composition due to land use conversion. Functional composition consistently demonstrated lower threshold responses compared to taxonomic composition for both fish (urban p = 0.069; agriculture p = 0.029) and insects (urban p = 0.095; agriculture p = 0.043) communities according to gradient forest models. We found significantly lower thresholds for urban versus agricultural land use for fish (taxonomic and functional p < 0.001) and insects (taxonomic p = 0.001; functional p = 0.033). We further revealed that threshold responses of functional composition were more geographically consistent than those of taxonomic composition to both urban and agricultural land use change. Traits contributing the most to overall functional composition change differed along urban and agricultural land gradients, and conformed to predicted ecological mechanisms underpinning community change. This study points to reliable early‐warning thresholds that accurately forecast compositional shifts in riverine communities to land use conversion, and highlight the importance of considering trait‐based indicators of community change to inform large scale land use management strategies and policies.

This article is protected by copyright. All rights reserved.