Proportions of imperviousness, drainage density, detention ponds and their distribution relative to the catchment outlet, have been independently linked with stormwater runoff and discharge in urban hydrologic systems. Limited research has examined the combined effects of urban land cover and drainage network structure on stormwater discharge. This study used HERCULES (High Ecological Resolution Classification for Urban Landscapes and Environmental Systems), a patch-based landscape classification approach, to investigate the effects of landscape structural heterogeneity on watershed stormwater discharge across a rural-urban gradient in Baltimore, MD. Patches were classified within five subbasins according to the percent cover of several land cover and drainage network characteristics that influence the generation of stormwater runoff and discharge, and the distribution of patches was examined across subbasins. Hydro-NEXRAD precipitation and USGS stream discharge records were analyzed for the subbasins during a four-month period in summer 2007 to evaluate differences in maximum rainfall rates, peak discharge, and runoff ratios. Linear regression models were developed to test for relationships between the proportion and spatial distribution of land cover and drainage network structure variables, rainfall intensity, and peak discharge.
Results/Conclusions
Results showed that more heavily urbanized subbasins had greater frequencies of patches with high proportions of impervious surfaces, and surface channels overlapping roads and draining to storm sewers compared to less urbanized ones. A changepoint was observed in rainfall rate at 0.81 mm/h below which no apparent relationship existed with discharge. Above that threshold, significant differences in discharge and runoff ratios were found between less and more developed subbasins, which were accentuated by events of higher rainfall intensity. Linear regressions identified significant relations between maximum peak discharge and a) the combination of high impervious cover with a high proportion of stream channels overlapping roads, and b) high impervious cover together with patches situated in the upstream portion of the subbasin. None of the land cover or drainage network parameters were meaningful predictors of minimum or average peak discharge for the subbasins and precipitation events considered. The findings demonstrate the utility of combining land cover and drainage network structure together to illuminate interactions between precipitation and heterogeneous components of the urban landscape that govern hydrologic responses at the watershed scale.