The Ecological Memory of Fish Assemblages in Tropical Agroecosystems With Different History of Landscape Changes

Abstract

ABSTRACT Environmental changes can have persistent effects on the local and regional dynamics of multispecies assemblages. However, the extent of historical legacies on current diversity patterns may depend on the trajectory and intensity of the environmental change. Here, we investigate how persistent decadal land use and land cover (LULC) changes are related to the taxonomic and functional structure of present‐day fish assemblages in regions with distinct histories of land use transformation. We selected streams in Brazilian river basins where the LULC change has started more than 150 years ago (state of São Paulo), in the 1970s (state of Rondônia), and more intensively in the 2000s (state of Mato Grosso). We used a dataset comprising 219 streams, 240 species, 13 instream variables and 13 variables describing past LULC changes. We used varying‐coefficient models (VCM) to test how instream habitat and past land use are related to taxonomic and functional fish richness and rarity. Although diversity metrics were slightly better explained by instream variables, we found stronger landscape legacy effects on fish diversity in RO and MT, where LULC changes have occurred more recently. We did not detect landscape legacy effects or strong biodiversity–environment relationships in the region that experienced deforestation earliest (SP), probably because fish assemblages are now dominated by generalist species that are more resistant to further environmental change. Our results indicate that landscape legacies may leave more detectable signals in stream communities in regions where the first major LULC changes happened not too long ago. Thus, the influence of past processes on current fish assemblage diversity and structure may decline with increasing time since the initial LULC change. Our study shows that past LULC changes can shape present‐day biodiversity patterns, and that differences in species pool and landscape change histories can lead to different assemblage–environment relationships. Incorporating historical land use variables, especially in recent deforested regions, can therefore improve our understanding of current patterns of biodiversity. Our findings highlight the importance of preserving and restoring riparian corridors to prevent species loss locally and regionally, because, even in long‐altered regions, streams with riparian cover may still support rare species. Overall, integrating land use history and regional context is crucial for effective conservation and biodiversity monitoring in ecosystems, especially those ones under the first major LULC changes.