Land cover-nitrogen interactions in small watersheds: patterns of variation within a semi-arid urban system

Background/Question/Methods We examined the effect of fine-scale variation in land cover on stream N export from 26 residential sub-watersheds in metropolitan Sacramento. Linking landscape structure and ecological processes in urban areas is important for building scientific understanding and for improving management. A key process in urban areas is non-point source (NPS) loading of N, which adversely affects aquatic environments. Previous NPS studies primarily compared urban to rural areas, often weakly explaining variability in stream N. Additionally, high variability in N export has been documented within urban systems. There is a need to begin developing ecological models specific to urban landscapes. To address this, we classified land cover for 26 sub-watersheds using HERCULES (High Ecological Resolution Classification for Urban Landscapes and Environmental Systems). HERCULES classifies high-resolution aerial photographs into 5 land cover elements: buildings, pavement, herbaceous and woody vegetation, and bare soil. All these sub-watersheds are residential, but they differ substantially in land cover, as identified by HERCULES. Focusing on land cover may better explain variability in N export.  Streams were sampled for N concentration and discharge during both wet and dry seasons. Correlation analysis and linear mixed models described the relationships between land cover elements and forms of stream N in sub-watersheds.

Results/Conclusions

The proportion of pavement in the sub-watersheds, a commonly used indicator of urban intensity, did not strongly correlate with increased levels of N in streams. Instead, high proportion of vegetation cover was associated with greater stream N content. The use of fertilizers or decomposition in vegetated areas is a likely source of this N. Pavement cover also did not correlate with differences in baseflow, while the proportion of vegetation cover in a sub-watershed was positively correlated with discharge. In semi arid urban environments, lack of dry season precipitation coupled with high use of lawn irrigation could account for observed increases in baseflow seen in highly vegetated areas. Lawn irrigation in the dry season created perennial flows in what would ordinarily be ephemeral streams. Pavement likely plays a role in modifying stream nutrient dynamics by increasing flow during storms. However, areas of high vegetation appear to be a primary source for N in urban stream systems. Such results indicate that residential lawns should be a high priority for intervention as an NPS source in urban systems. Management activities in residential areas may be best directed toward achieving reductions of residential fertilizer use, not solely the minimization of pavement cover.