Regional impacts of urbanization on stream channel geometry: A case study in semiarid southern California

Abstract

Urbanization often increases storm runoff, peak discharges and rates of stream channel erosion. Coastal California has experienced rapid urbanization over the past several decades and has the potential for stream channel degradation. Several counties in California have implemented Hydromodification Management Plans (HMPs) to protect channels from erosion, but few studies have quantified the impact of urbanization on channel geometry in diverse geological settings at the county scale. A synoptic survey of field sites (N=56) by the California Environmental Data Exchange Network (CEDEN) and additional field surveys (N=24) were used to develop regional hydraulic geometry curves relating bankfull cross-sectional area (Axs), width (w), mean depth (d), and discharge (Qbf) to watershed area (Aw) in San Diego County. Regional curves were compared for urban and reference sites and to other regional curves developed for southern California. Multiple regression models were used to identify dominant watershed and channel controls on geometry, including Aw, percent impervious cover (I%), mean annual precipitation, underlying geology, longitudinal slope, hydrologic soil group, and channel particle size. For the reference streams, regional curves were statistically significant for w and Axs (p<0.05). The regional curves for urban channels (I%>20%) had significantly larger w, d, Axs, and Qbf for a given watershed size. A majority (68%) of the urban channels and 78% of the small urban channels (Aw<10km2) were enlarged. Enlargement of channels in small watersheds disrupted the correlation between Aw and bankfull dimensions, and I% was the only significant predictor of channel geometry in urban watersheds. Channel response differed by channel substrate: sand-bedded channels incised and experienced extreme enlargement of up to 115× the Axs of reference sites, while gravel-bedded channels widened and showed less enlargement (<7× reference Axs). Diverse channel responses to urbanization were observed at the county scales with significant scatter about the regional curves and regression equations. Urban channels have a high probability of enlargement, with greater risk for sand-bedded streams, but the magnitude of enlargement was not predictable from watershed or channel characteristics.