Real-time total water forecasts for the Washington, DC Metropolitan region: Integrating coastal, riverine and urban flood forecasting at a range of spatio-temporal scales

Abstract:

Recent hurricane seasons highlighted the importance of integrating urban, riverine, coastal and ocean models for total water prediction in large metropolitan areas located along estuaries. This presentation will demonstrate that coupling the current state-of-the-art technology for flood forecasting across different environmental domains can significantly improve the current capacity of real-time flood prediction. A case study incorporating multi-scale and multi-temporal physical process for total water predictions, including large scale oceanic process, off-shore and near shore waves, estuarine and coastal processes, riverine flows, urban runoff, and watershed hydrology systems will be presented focusing on the Washington, DC region. Like many US coastal areas, this metropolitan region is vulnerable to multi-flood hazards, subjected to high flood levels from inland urban rainfall-runoff processes, high stream flows, as well as tide and surge driven coastal inundation from the Chesapeake Bay. The flood forecasting multi-model framework (ADCIRC, SWAN, NAM, XBeach, WRF-Hydro, NWM and HEC-RAS 2D) developed here provides several parameters at a range of spatial and temporal scales providing information directly to the Washington/Baltimore National Weather Service Office. Results demonstrate that areas subject to compound flooding are a result of a dynamic and complex interaction between storm surge signals, riverine flow, local winds and urban runoff. The proposed system improved the current ability to predict total water levels in real-time for short-range predictions (3.5 days) by adding multiple ensemble boundary conditions from a range of models. Results also demonstrate that continental scale hydrological models can be locally implemented improving local-scale hydrological predictions at short-range temporal scale. We will also show an attempt to expand real-time flood forecasts at sub-seasonal timescale. While the National Capital Region is especially vulnerable to these multi-flood hazards, several major cities along the coast are equally or more susceptible. We expect that this approach can demonstrate viable alternatives to the current forecast systems and provide guidance for the development of the new generation of water forecast frameworks.