A physically based model of the 14 km 2 Dead Run watershed in Baltimore County, MD was created to test the impacts of detention basin storage and soil storage on the hydrologic response of a small urban watershed during flood events. The Dead Run model was created using the Gridded Surface Subsurface Hydrologic Analysis (GSSHA) algorithms and validated using U.S. Geological Survey stream gaging observations for the Dead Run watershed and 5 subbasins over the largest 21 warm season flood events during [2008][2009][2010][2011][2012]. Removal of the model detention basins resulted in a median peak discharge increase of 11% and a detention efficiency of 0.5, which was defined as the percent decrease in peak discharge divided by percent detention controlled area. Detention efficiencies generally decreased with increasing basin size. We tested the efficiency of detention basin networks by focusing on the ''drainage network order,'' akin to the stream order but including storm drains, streams, and culverts. The detention efficiency increased dramatically between first-order detention and second-order detention but was similar for second and third-order detention scenarios. Removal of the soil compacted layer, a common feature in urban soils, resulted in a 7% decrease in flood peak discharges. This decrease was statistically similar to the flood peak decrease caused by existing detention. Current soil storage within the Dead Run watershed decreased flood peak discharges by a median of 60%. Numerical experiment results suggested that detention basin storage and increased soil storage have the potential to substantially decrease flood peak discharges.

Urban development has been observed to lead to variable magnitudes of change for stormflow volume and directions of baseflow change across cities. This work examines temporal streamflow trends across the flow duration curve in 53 watersheds during periods of peak urban development, which ranged from 1939 to 2016. We used U.S. Geological Survey streamgage records combined with pre-development and urbanization characteristics to identify 20 years for analysis in each urbanizing watershed. Each urbanizing gage was paired with a nearby reference gage representing climatic trends over the same time period. Results indicated that urbanization, as measured by housing density, did not homogeneously alter the flow duration curve. Urbanization led to widely variable trends in low flow, where half of the urbanizing gages had increasing flow at the 10th non-exceedance percentile, and the other half had declining low flow. High flows generally increased in streams as the area urbanized. The largest increases in high flows were in streams in semi-arid and arid areas. The largest urban flow changes had transformations in wastewater infrastructure, water supply infrastructure, and flood control facilities. Isolating flow changes due to urbanization from those of reference sites will serve to better identify and manage synergistic effects of urban development and climate change on flooding and water availability. Plain Language Summary We analyzed water flows in 53 U.S. streams that drain areas with housing density increases of at least 40% over 20 years. Streams had periods of low and high flows within that time frame. Low flows went up in about half the streams and went down in the other half. The largest decreases in low flows were seen where septic systems were converted to municipal sanitary sewer systems. High flows generally increased in streams as the area urbanized. The largest increases in high flows were in streams in semi-arid and arid areas. Using historical records of changes in streamflow can help predict future changes in flow with further urbanization and separate the effects of urban development on streamflow from those of reference conditions.

An inconsistency is noted in the transformation to modify random normal variates to random skewed variates used in the Thomas and Fiering hydrologic generation model.

No abstract

No abstract