The influence of spatial heterogeneity, hydrologic variability, and scale on nitrogen and water export in urban residential ecosystems

Urban ecosystems are a major source of non-point source (NPS) nitrogen (N) pollution to aquatic and coastal ecosystems. It is well established that N export from urban watersheds is both greater than that from undeveloped watersheds and approaches the level of exports from agriculture. In addition, high variation in N export among urban watersheds has been found. Understanding this variation is an important question for urban landscape ecology and for improving management of NPS pollution. Two factors that may contribute to this variability are: 1) the heterogeneous distribution of sources and sinks of N among different elements of the landscape, and 2) hydrologic variability induced by changes in runoff volumes and flow paths as a result of urbanization.;Our current understanding of the relationship between landscape heterogeneity and variation in N export is mixed. The studies that establish urban watersheds as N sources relative to natural watersheds are typically conducted at coarse spatial scales and describe landscape heterogeneity in terms of land use. Land use refers to the human activity taking place in these landscapes and may not be related to the ecological processes governing N and water dynamics. Land cover, on the other hand, refers to the physical elements that make up the landscape. Consequently, land cover can vary widely within a given land use. Therefore, variation in N export from residential watersheds may be better assessed using land cover to describe landscape heterogeneity and this dictates analysis at finer scales.;The overarching goal of this dissertation was to determine the relationships among N export, land cover heterogeneity, and hydrologic variability in urban residential watersheds. We used a fine scale spatial approach and quantified variation in land cover to determine what, if any, elements of land cover act as sources and sinks for N in the landscape, and how this land cover interacts with hydrology to explain variability in urban N exports.;We measured N export in 2 urban ecosystems: Baltimore, MD and Sacramento, CA, using a land cover classification to quantify the amount of building, pavement, herbaceous vegetation, woody vegetation, and bare soil in each watershed. The land cover classification was developed in Baltimore as part of the Baltimore Ecosystem Study-LTER. We applied it to Sacramento where we divided hydrologic and spatial variability along two axes---at the watershed scale during storm events and at the sub-watershed scale during both dry and wet season baseflow. We also mapped waste water and storm drainage infrastructure that may affect N export dynamics in each system.;We found that, in Baltimore, septic infrastructure was tightly linked to high proportions of herbaceous land cover and high N exports. Because septic infrastructure and herbaceous vegetation occurred together, it was not possible to determine whether either of them was independently related to high N levels. However, the presence of septic systems may potentially override any land cover signal. In Sacramento the effect of infrastructure on N export was also significant and, although expected to coincide with high proportions of impervious cover, was not statistically related to land cover. During base flow conditions, water yield was highly correlated with N yield. Water yield was also correlated to high proportions of building cover, establishing an indirect relationship between building cover and N yield. The correlation between building cover and higher water yield, especially during dry summer months, indicated a water subsidy from residential landscape irrigation that increased N export---fundamentally altering the hydrology of the system by creating perennial flows in what would ordinarily be intermittent streams. Nitrogen export was significantly correlated with the amount of runoff generated during storm events and the density of drainage infrastructure in each watershed. Therefore, land cover heterogeneity did not directly predict N export during storms or base flows.;Our research demonstrates how septic system infrastructure that contains major N sources and drainage infrastructure that modifies hydrology affect landscape controls on N dynamics in urban systems. These findings highlight the complex, multiple dimensional changes to ecosystem processes incurred by patterns of urbanization. Increasing the retention of water on urban landscapes, restoring biological sinks such as sediment and vegetation, and buffering the input of storm water may be the most immediately effective approaches for reducing urban NPS nitrogen pollution in urban ecosystems.