Open AccessFlash Flood Forecasting for Small Urban Watersheds in the Baltimore Metropolitan Region
Author(s) -
Julie Rose N. Javier,
James A. Smith,
Katherine L. Meierdiercks,
Mary Lynn Baeck,
Andrew J. Miller
Publication year - 2007
Publication title -
weather and forecasting
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.393
H-Index - 106
eISSN - 1520-0434
pISSN - 0882-8156
DOI - 10.1175/2007waf2006036.1
Subject(s) - flash flood , flood forecasting , environmental science , hydrology (agriculture) , flood myth , hydrograph , hydrological modelling , watershed , 100 year flood , weather radar , meteorology , storm , rain gauge , radar , surface runoff , precipitation , climatology , geology , geography , telecommunications , ecology , geotechnical engineering , archaeology , machine learning , computer science , biology
The utility of distributed hydrologic models in combination with high-resolution Weather Surveillance Radar-1988 Doppler (WSR-88D) rainfall estimates for flash flood forecasting in urban drainage basins is examined through model simulations of 10 flood events in the 14.3 km2 Dead Run watershed of Baltimore County, Maryland. The hydrologic model consists of a simple infiltration model and a geomorphological instantaneous unit hydrograph–based representation of hillslope and channel response. Analyses are based on high-resolution radar rainfall estimates from the Sterling, Virginia, WSR-88D and observations from a nested network of 6 stream gauges in the Dead Run watershed and a network of 17 rain gauge stations in Dead Run. For the three largest flood peaks in Dead Run, including the record flood on 7 July 2004, hydrologic model forecasts do not capture the pronounced attenuation of flood peaks. Hydraulic controls imposed by valley bottom constrictions associated with bridges and bridge abutments are a dominant element of the extreme flood response of small urban watersheds. Model analyses suggest that a major limitation on the accuracy of flash flood forecasting in urban watersheds is imposed by storm water management infrastructure. Model analyses also suggest that there is potential for improving model forecasts through the utilization of information on initial soil moisture storage. Errors in the rainfall field, especially those linked to bias correction, are the largest source of uncertainty in quantitative flash flood forecasting. Bias correction of radar rainfall estimates is an important element of flash flood forecasting systems.