The effects of land use on stream nitrate concentrations: From the catchment scale to the regional scale

This work furthers the understanding of processes occurring in catchments that affect stream nitrate concentrations using two different approaches: a temporally intensive case study of three headwater catchments with varying land use (through storm event monitoring) and a spatially intensive study on the regional scale (through statistical modeling) of 1st-4 th order catchments. At the catchment scale, stream nitrate concentrations during three storm events were monitored in three catchments with different land uses (forested, agricultural, residential) to determine how land use affects nitrate "patterns" during storm events. Overall, results of storm event nitrate concentrations suggest that varying nitrate inputs have a large affect on nitrate dynamics. While within-storm nitrate concentration response patterns in the residential catchment were the same as the patterns in the reference forested catchment (a "concentration" pattern throughout the year), a "dilution" pattern was observed in the fall and winter and a "concentration" pattern was observed in the spring in the agricultural catchment. At the regional scale, a statistical model was developed using land use and either topographic index (TI) or hydrologic landscape regions (HLRs) to predict stream nitrate concentrations during lowflow. Including TI and HLRs (in the form of primary hydrologic flowpaths) significantly improved chloride predictions, but did not improve nitrate predictions. Results of the linear regressions imply that the hydrologic setting of the catchments are adequately represented (from chloride, which is tightly linked to hydrology), and nitrate is more strongly affected by processes such as denitrification and plant uptake during lowflow. Agricultural effects were seen both on the smaller catchment scale and the regional scale. Different patterns were observed in the agricultural catchment during storm events, and chloride was elevated in the Willamette Valley where agricultural activity is concentrated. The temporal pattern of nitrate during storm events was found to be largely controlled by the spatial organization of land cover, whereas the spatial pattern of land cover did not control stream nitrate concentrations sufficiently to improve predictions of nitrate during lowflow. Future work should determine whether or not the spatial pattern of land cover, TI, and HLRs improves nitrate predictions during storm events.