Beta diversity of stream fish communities along anthropogenic environmental gradients at multiple spatial scales.

Despite the importance of assessing beta diversity to understand the effects of human modifications on biological communities, there are almost no studies that properly addressed how beta diversity varies along anthropogenic gradients. We developed an algorithm to calculate beta diversity among a set of sites included in a moving window along any given environmental gradient. This allowed us to assess beta diversity among sites with similar conditions in terms of human modifications (e.g., land use or instream degradation). We investigated beta diversity using stream fish community data and indicators of human modification quantified at four spatial scales (whole catchment, riparian, local, and instream). Variation in beta diversity was dependent on the scale of human modifications (catchment, riparian, local, instream, and all four scales combined) and on the type of diversity considered (taxonomic or functional). We also found evidence for non-linear responses of both taxonomic and functional beta diversity to human-induced environmental alterations. Therefore, the response of beta diversity was more complex than expected, as it depended on the scale used to quantify human impact and exhibited opposite responses depending on the location along the environmental impact gradient and on whether the response was taxonomic or functional diversity. Anthropogenic modifications can introduce unexpected variability among stream communities, which means that low beta diversity may not necessarily indicate a degraded environmental condition and high beta diversity may not always indicate a reference environmental condition. This has implications for how we should consider beta diversity in environmental assessments.

A comparative analysis reveals weak relationships between ecological factors and beta diversity of stream insect metacommunities at two spatial levels.

The hypotheses that beta diversity should increase with decreasing latitude and increase with spatial extent of a region have rarely been tested based on a comparative analysis of multiple datasets, and no such study has focused on stream insects. We first assessed how well variability in beta diversity of stream insect metacommunities is predicted by insect group, latitude, spatial extent, altitudinal range, and dataset properties across multiple drainage basins throughout the world. Second, we assessed the relative roles of environmental and spatial factors in driving variation in assemblage composition within each drainage basin. Our analyses were based on a dataset of 95 stream insect metacommunities from 31 drainage basins distributed around the world. We used dissimilarity-based indices to quantify beta diversity for each metacommunity and, subsequently, regressed beta diversity on insect group, latitude, spatial extent, altitudinal range, and dataset properties (e.g., number of sites and percentage of presences). Within each metacommunity, we used a combination of spatial eigenfunction analyses and partial redundancy analysis to partition variation in assemblage structure into environmental, shared, spatial, and unexplained fractions. We found that dataset properties were more important predictors of beta diversity than ecological and geographical factors across multiple drainage basins. In the within-basin analyses, environmental and spatial variables were generally poor predictors of variation in assemblage composition. Our results revealed deviation from general biodiversity patterns because beta diversity did not show the expected decreasing trend with latitude. Our results also call for reconsideration of just how predictable stream assemblages are along ecological gradients, with implications for environmental assessment and conservation decisions. Our findings may also be applicable to other dynamic systems where predictability is low.