Effects of Land Use on Fish Assemblages of San Francisco Bay Streams: a Spatial Analysis Approach

There are approximately 70 small streams that flow directly into the San Francisco Bay (Bay), many of which are surrounded by highly urbanized landscapes. Urban development often leads to altered hydrographs and stream channel morphology, as well as high concentrations of contaminants. These physical changes can have ecological consequences. Previous research has shown that urbanization can result in overall decreases in diversity and shifts from assemblages dominated by natives to ones dominated by nonnative and tolerant species. Broad scale spatial analyses utilizing GIS watershed tools can be useful in elucidating correlations between land use and stream fish communities. Using a large dataset on the distribution and abundance of stream fishes collected at 275 sites throughout the Bay region from 1993-1999, we take a spatial analysis approach to examine the relationship between land use and fish assemblages in Bay streams. Specifically, we (1) characterize the fish assemblage data into multiple metrics (e.g., percent native) and (2) use GIS tools to delineate individual watershed basins, from which we (3) characterize the composition (e.g., percent impervious surface) and configuration (e.g., road adjacency to streams) of land cover for individual watershed landscapes at multiple scales. Finally, we (4) use a multivariate approach to analyze the relationships between land use and fish assemblage metrics.

The Bay region encompasses a broad diversity of land use types, ranging from ranchland to densely urbanized cities. This region also contains a diverse fish fauna; 39 species were collected at varying abundances and distributions, of which 18 (46%) were native and 21 (54%) were nonnative. Preliminary results in the largest watershed draining to the Bay (Alameda Creek) reveal that the number and the percentage of nonnative individuals are positively associated with sub-watersheds characterized by high percentages of impervious surfaces. In larger sub-watersheds, there were stronger associations between land use composition and fish metrics when local scale watershed buffers were used. Multivariate analyses also indicated that the addition of local stream habitat variables (e.g., conductivity) to the land use model better predicted the presence of nonnative fishes. Ongoing work is focused on understanding whether these same patterns emerge when considering the entire Bay region. Our preliminary results build on earlier research that shows the influence of land development on native fishes. We suggest that managers consider both the composition of altered landscapes (e.g., percent impervious surfaces) and their configuration within the landscape to better conserve native fish communities in urbanized regions.