Trait co‐occurrence networks reveal changes in community functional structure from local to landscape scales

Abstract The functional composition of communities reflects species performance and its interplay with local and regional drivers. Yet trait‐based analyses often overlook regional heterogeneity (geographic and environmental differences among communities), despite its central role for community assembly. This limits our ability to assess functional change across broad spatial extents or long time periods, which is essential to understand the functional consequences of global change. To overcome this limitation, here, we expand the network analysis of traits (NATs) to assess community functional structure along spatial and environmental gradients by shifting the focus from species–trait associations to trait–trait co‐occurrences. We sequentially aggregated communities by (1) spatial proximity, (2) environmental similarity or (3) random selection to quantify how spatial and environmental gradients determined the density of links between trait co‐occurrences (i.e. trait associations) within networks (measured as edge density) and the position of individual traits within those networks (measured as trait strength and betweenness). We developed this approach using long‐term datasets of aquatic benthic invertebrate communities from Sweden: 56 lakes (spanning 21 years) and 39 streams (spanning 12 years) and compared their patterns with those obtained using in silico metacommunities. We found clear scale dependence on functional structure and contrasting regional drivers across ecosystems. While streams were only structured by geographic distance, the influence of environmental filtering increased over time in lakes, potentially signalling the effects of climate change on their functional structure. These results demonstrate that the relative importance of spatial and environmental drivers of functional structure can differ among freshwater ecosystems and change over time, highlighting the need to consider both regional heterogeneity and temporal dynamics when assessing functional metacommunity structure. Read the free Plain Language Summary for this article on the Journal blog.