Dawson, who works in the Institute for Computational Engineer and Sciences at the University of Texas at Austin, says his research focuses primarily on modeling and simulation of processes in the coastal ocean.

He says government funding is allowing scientists to do predictive simulation of storm surges due to hurricanes and tropical storms.

"By 'predictive simulation' I am referring to the development of computer models which can be used in real time to forecast storm surge as hurricanes approach land, to study the impacts of historical hurricanes and attempt to reproduce actual measurements which were taken during the storm, and to study future scenarios," Dawson says.

He notes that the revolution in modeling of storm surge is occurring because of government funding of the computational infrastructure available through the National Science Foundation TeraGrid and basic research funding in computational science and engineering.

"We are now able to do high-resolution storm-surge predictions within the time frame required by emergency managers," Dawson says. "This would have been impossible five years ago."

At the same time, he voices concern that federal funding for coastal ocean modeling research has been piecemeal across different agencies and focused more on short-term projects rather than long-term priorities.

"I would welcome any effort to promote longer-term, focused, sustained funding of research in this area," he says.

He says the National Oceanographic and Atmospheric Administration (NOAA) is now re-evaluating the computer models it uses to determine which ones should be used in the future for storm-surge forecasting.

He says he believes that future forecast models "should be performed at the highest fidelity possible given the computational resources available and the uncertainties inherent in any hurricane forecast.

"We must attempt to quantify these uncertainties where possible," Dawson says.

"It is also important that we work closely with emergency-management personnel to understand the type of information that is needed and to develop ways in which risk can best be conveyed to the public," he adds.