How much is too much? Toward prediction of factors limiting the biotic condition of highly modified fluvial systems of the western US

The goals of our study were: (1) to describe influences of streamflow patterns on bethnic macroinvertebrate assemblages (BMIs) and (2) to document response patterns of BMIs to landscape-level surrogates of catchment modification and hydrologic infrastructure (e.g.,  number of canals) across multiple geographic scales. Continuous discharge records from 114 sites in 11 Western States were used to calculate streamflow parameters.   These parameters were selected to represent magnitude, duration, frequency, timing, and variability at three levels of flow (high, low, and central tendency).  GIS surrogates of landscape alteration and hydrologic alteration were derived for 332 sites. Quantitative macroinvertebrate samples from each of the sites were used to calculate multivariate response variables as well as metrics of macroinvertebrate community structure   including abundance, richness, diversity, evenness, functional feeding groups, and relative abundances of specific taxa.  A nonparametric screening procedure and quantile regression were used to identify relationships between physical variables and macroinvertebrate assemblages. Streamflow variability was associated with the largest number of macroinvertebrate metrics and associations with daily streamflow variability were inverse to associations with monthly variability. Variability in flows was associated with macroinvertebrate abundance, richness, and diversity.  The magnitude of high flows was associated with relative abundance and richness of EPT taxa and non-insect taxa, while the frequency of high flows was associated with dominant taxa, diversity, and evenness. Richness of tolerant organisms and EPT richness were the most responsive metrics, showing statistically significant patterns for 21 and 24 interactions, respectively, of 28 possible interactions on the geographic scale.  EPT richness, Shannon diversity index, and functional feeding group diversity showed significant patterns of decline with increasing catchment modification.   These results serve as the foundation for our future work of developing models to predict biological response to catchment and hydrologic modifications in Pacific Northwest and California.