Baltimore Ecosystem Study Institute of Ecosystem Studies

2014 BES Annual Meeting Presentation and Poster Abstracts

Can stream restoration mitigate flashy hydrologic response in urban watersheds?
Miller, Andrew
Co-Authors: Garth A. Lindner (first author), Andrew J. Miller (second author, presenting)

Abstract: In this study we worked in collaboration with the Baltimore County Department of Environmental Protection and Sustainability and with the contractors responsible for designing a major stream restoration project in the DR5 tributary of Dead Run. The DR5 watershed (1.63 km2) has 45% impervious cover and typically exhibits an extremely rapid and peaked hydrologic response to storm events leading to formation of deeply incised gullies in the headwaters. The restoration project, planned for 2015, will cover more than 1400 m of stream channel encompassing almost the entire drainage network upstream of the gage. Three supplementary stream gages were installed in the headwaters to calibrate a 2-dimensional hydrodynamic model for flood-wave simulation. Lidar and field-surveyed channel topography data were combined to develop a DEM (digital elevation model) describing existing conditions. A second DEM was developed based on the proposed restoration design, and two additional DEMs representing extreme modification of the channel and riparian zone were developed in order to assess whether any restoration scenario was capable of attenuating the flood wave along the restoration reach. Short-duration storm events with flood-peak recurrence intervals of 1.1 to 3.2 yr were routed through each of the restoration scenarios and resulting hydrographs at the DR5 gage were compared with the actual hydrographs. The contractors’ restoration design and the first “extreme” design resulted in very modest (<10%) reductions and delays of ~5 minutes in the hydrograph peak, and the only significant attenuation observed was in connection with a scenario involving large ponds with dead storage occupying the floodplain. Routing of these same urban flood hydrographs through a 2-km-long modeling domain representing the broad riparian zone of Horsehead Branch resulted in up to 60% attenuation and delay of the flood peak by 30-70 minutes. The results of these analyses suggest that restoration in headwater basins is unlikely to accomplish significant mitigation of the flood wave, whereas wide, low-gradient floodplains in larger watersheds have greater capacity to store and attenuate flashy urban hydrographs.