Spatial modelling of temporal dynamics in stream fish communities under anthropogenic change

Abstract

Aim: Understanding temporal changes in aquatic communities is essential to address the freshwater biodiversity crisis. In particular, it is important to understand the patterns and drivers of spatial variation in local community dynamics, generalizing temporal trends from discrete locations to entire landscapes that are the main focus of management. Here, we present a framework for producing spatially continuous views of community dynamics, focusing on stream fish affected by hydropower development. Location: River Sabor, NE Portugal. Methods: We sampled stream fish at thirty sites between 2012 and 2019. Community trajectory analysis was used to quantify the directionality and velocity of community change, and the geometric resemblance of community trajectories between sites. Geostatistical models for stream networks were used to relate metrics describing community dynamics to environmental variables, while controlling for Euclidean and hydrologic spatial dependencies, and to map spatial variation in community dynamics across the watershed. Results: Trajectories in multivariate space underlined strong temporal dynamics, with local communities deviating and returning to previous states, but without evidence for directional changes. Accordingly, directionality values were low and not consistently affected by environmental variables. The velocity of community change varied markedly across the watershed and it was strongly affected by stream order and elevation, with faster changes observed in lowland streams draining into hydroelectric reservoirs and with a high proportion of exotic species. Pairwise distances between community trajectories were strongly related to hydrologic and environmental distances between sites. Main conclusions: Local stream fish communities were in a loose equilibrium across the watershed, but they fluctuated at a faster rate closer to a hydroelectric reservoir. Integrating community trajectory analysis and geostatistical modelling provides a relatively simple framework to understand how, where and why temporal community dynamics vary across dendritic stream networks and to visualize spatial patterns of community change over time in relation to anthropogenic impacts