PS 60-27
A decade later: Consequences of land use on stream physical habitat and fish assemblages
Stream ecosystems are highly associated with the adjacent terrestrial environment, so that the conversion of natural landscapes in agroecosystems is one of the most meaningful threats to their biodiversity. In this context, identifying temporal changes in these environments and communities are essential to conservation and restoration actions. In this study, we aimed to identify the main changes in both physical habitat structure and fish fauna in streams from agricultural landscapes before and after ten years. In 2003 (T0), we sampled 44 stream reaches (75 m long) that were resampled in 2013 (T10) using standardized methodology to collect environmental variables (through visual estimates) and fish (electrofishing gear). To select the variables that explain the most the environment variability between the two periods we used a Principal Component Analysis. To investigate environment and fish fauna homogenization we conducted a PERMDISP Analysis. To compare overall fish assemblage structure (abundance and biomass) we used a Similarity Analysis (ANOSIM). These analyses were conducted in PRIMER 6.0. In addition, to identify which species presented significantly changed abundance between T0 and T10, we conducted an Indicator Species Analysis (IndVal) in the PCOrd 5.0.
Results/Conclusions
Depth, width, current, and habitat volume were significant lower in T10 than in T0. The Physical Habitat Index (PHI), % of grasses in stream banks, % of runs and riffles, % of wood debris, % of shrubs, width, depth, current, and habitat volume explained 55.9% of stream reach variations between T0 and T10. Stream reaches in T0 were in poorer conditions than ten years ago: narrower, shallower, and with pasture grasses invading stream channel. Furthermore, this set of variables exhibited low multivariate dispersion in T10 (F=17.38, P=0.001), suggesting physical habitat homogenization probably as a consequence of increased siltation. We registered 53 species in T0 and 47 in T10. Fish abundance and biomass decreased 21.5 % and 13 %, respectively, in T10 (abundance: T0=10,039, T10=7,880; biomass: T0=25.5, T10=22.2). Notwithstanding, there were no significant differences between periods regarding overall fish assemblage structure (R=0.025, P=0.03) and multivariate dispersion (F=1.10, P=0.3). IndVal Analysis, on the other hand, revealed that the abundance of several species changed significantly between T0 and T10. Hence, our results evidenced a rapid habitat and fish fauna modification in a relatively short-time period, probably related to the intense land use in this region.