Nitrogen hotspots along a longitudinal transect in residential streams: The role of infrastructure and land cover at multiple scales
Urban ecosystems are a major source of non-point source (NPS) nitrogen (N) pollution to aquatic and coastal ecosystems. Urban watersheds are also highly variable in N export. Land use/land cover heterogeneity is commonly assumed to predict how elements of the landscape contribute to this variability. Despite widespread application of this approach, land use/land cover based models often do not explain significant variation in N export among urban watersheds. Recently published studies suggest that watershed-scale landscape controls over N export may be superseded by the effects of land cover adjacent to streams, by infrastructure, or by stream modification and channelization. The goal of our research was to determine whether variation in stream N concentration in urbanized streams was best explained by proximity to 1) specific land cover elements at multiple scales, 2) storm drains, or 3) modified sections of stream channels. We measured longitudinal variability in N concentration for nine residential watersheds of Sacramento, CA. We assessed the proportion of pavement, building, and herbaceous and woody vegetation to determine the effect of urban land cover on N concentration at various spatial scales. These scales included catchment, riparian, and adjacent buffer zones (at widths of 10, 25, 50, and 100m) relative to sampling locations. We also analyzed the spatial relationships among N concentration, drainage outfalls, and channel or riparian condition.
We found that hotpots of N concentration coincided with the location of storm drain outfalls and buried stream headwaters. Land cover was not generally related to these hotspots. However, woody and herbaceous land cover in the riparian zone had a weak but significant relationship to changes in N concentration. Nitrogen hotspots were strongly correlated with the location of storm drain outfalls. Hotspots also occurred where portions of the natural stream channel were replaced with impervious concrete or steel culverts. There was evidence for significant N uptake along unmodified reaches of the stream channels. Our research highlights how modification of hydrologic flow paths by infrastructure, and in-stream processes by channel modification, supersede landscape controls, such as land cover, for determining variation in N concentration in urban streams.