Predicting the Next Storm Surge Flood

Sea Technology, Aug 2007 by Stamey, Barry, Wang, Harry, Koterba, Michael

Rapid Prototype Development of a Regional Capability to Address A National Problem

On August 29, 2005, hurricane Katrina became the worst natural disaster in the recent history of the United Stales and was indelibly etched in the memories of its citizens. The costliest and one of the deadliest hurricanes ever, Katrina caused unprecedented devastation. Ocean storm surge, combined with elevated flood water levels in the Mississippi River, led to unprecedented water-level rise in the canal system, which, aided by local winds, eventually topped and then breached the levees.

Katrina underscores a critical consideration in forecast modeling-although atmospheric models predict storm development, movement and intensity, the hydrodynamic modeling capability to predict and visualize flooding remains limited. This is especially relevant when the combined effects of wind-driven ocean storm surges, tides and downriver discharge lead to rapid, intense flooding.

Earlier in the summer of 2005, perhapsserendipitously, the National Oceanic and Atmospheric Administration's (NOAA) National Weather Service's (NWS) Sterling, Virginia, Weather Forecast Office (WFO) and Noblis Inc. (formerly Mitretek Systems Inc.) discussed the challenge of predicting Potomac River storm surge and flooding inundation in the metropolitan Washington, D.C., region. Within this area, cities such as Alexandria, Virginia, have experienced flooding related to storm surge. tides and river discharge from tropical storms and nor'easters transiting the Chesapeake Bay.

In 2003, hurricane Isabel dumped several inches of rain in the upper Potomac River watershed, and downstream flooding along the river was expected. What was not anticipated was the approximately six to eight-tool storm surge that moved some 120 miles upstream from the Chesapeake Bay. The surge caused significant flooding in Old Town Alexandria and neighboring communities, damaged thousands of homes and businesses and caused losses in the tens of millions of dollars throughout the Chesapeake Bay. It was followed about 48 hours later by downstream flooding that further paralyzed residents for several days.

NWS forecasters currently use a general hydrodynamic model developed in the 1980s to predict tidal gauge water-level heights. Emergency managers (EMs) use this information to estimate what areas will flood, but they rely heavily on local history to extrapolate gauge flood heights to the shoreline and inland. Although this approach enables some advanced decision-making on where to deploy (and not deploy) emergency responders and crews, the inundation extent, timing and depth are often quite uncertain. EMs and first responders are responsible for preparing communities for disasters and then reacting to protect citizens and their property. Critical resources and time are expended deploying spotters and awaiting word for information as the storm unfolds. EMs have indicated the urgent need for better local in formation about expected surge and inundation.

Emerging hydrodynamic models can help reduce uncertainty by projecting inundation over land rather than just the water height at the shoreline. Models can incorporate the combined effects of ocean storm surge, riverine discharge, tides and winds to provide visualized forecasts of flooding and assist in better response planning.

To explore this potential, Noblis; the Virginia Institute of Marine Science (VIMS); and NWS Sterling and Wakefield, Virginia, WFOs and the Chesapeake Bay Observing System (CBOS) through both the NOAA Chesapeake Bay Office and the U.S. Geological Survey (USGS) agreed in January 2006 to an ad hoc, self-funded collaboration. The objectives were to rapidly develop an initial prototype system that would integrale high-resolution atmospheric and hydrodynamic stomi surge models; evaluate the prototype's ability to predict land inundation in the Washington, D.C., area; provide flooding results to EMs using emerging visualization technologies and evaluate improved methods; and document what would be needed for an operational NWS prototype.

The initial prototype uses advanced modeling and visualization techniques to depict expected inundation at a spatial resolution of less than a city block (about 50 meters) and a vertical resolution of about 30 centimeters in a time-step display of one hour or less.

EMs' response to this prototype has been positive and supportive. It is the foundation for the Chesapeake Inundation Prediction System (CIPS) and is a potential tool for NWS WFOs to better support EMs, first in the Chesapeake Bay region and then for replicating similar approaches in other coastal and Great Lakes regions. CIPS is a CBOS initiative within the Mid-Atlantic Coastal Ocean Observing Regional Association (MACOORA) of the Integrated Ocean Observing System (IOOS) in coordination with the Chesapeake Research Consortium (CRC).

Rapid Prototype

High-resolution atmospheric and hydrodynamic models are initialized and validated with existing observational network data. Model-determined changes in water-level elevation are directly simulated across high-resolution elevation landscape light detection and ranging (LIDAR) data. Emerging geographic information systems (GIS) visualize and animate the expected storm surge and inundation.