Although the Atlantic Ocean has been relatively quiet this season, events such as Hurricane Rita remind us that forecasting chaotic storms is an imperfect science, but there are ways we can make it better. And that starts with how we predict storm surge.
At this time in 2005, the U.S. was beginning to recover from Hurricane Katrina while Hurricane Rita was forming off the coast of Africa. Rita became one of the most intense hurricanes ever to form in the Atlantic.
Initial forecasts showed the storm making landfall near Galveston, which caused panic and led to mass evacuations in the Houston-Galveston metropolitan area.
The initial forecasts turned out to be wrong; the hurricane took a northward turn. It made landfall near the Texas-Louisiana border on Sept. 24, 2005, with a much greater storm surge than had been forecast, causing widespread damage and flooding.
Never miss a local story.
A hurricane forecast consists of two main components: a forecast of the hurricane wind field, usually reported as the category and track of the hurricane, and a forecast of the resulting storm surge.
Both forecasts are made using sophisticated computer models.
Although wind and storm surge are both important components of a hurricane, storm surge is more deadly and causes most of the property damage.
It’s still difficult for these models to accurately capture a storm’s intensity, but they have shown significant improvement during the past decade, especially at predicting the storm’s track.
Storm surge modeling, however, is in a less advanced stage. There have been improvements, but the National Hurricane Center still relies primarily on one model, called the SLOSH model, which was developed decades ago.
Recently, other models have demonstrated significant predictive capability and are undergoing adoption by the National Hurricane Center to enhance storm surge predictions.
More of an emphasis ought to be put on storm surge. The National Hurricane Center does its best to predict, but what is really needed is more rapid adoption of new but well-studied simulators into operational use.
During Hurricane Ike in 2008, storm surge reached as far as 10 miles inland east of Houston, up to about 15 feet in some places. If this hurricane had struck the Texas coast a little farther to the west, that 15 feet of water would have been in the city of Houston and its southern suburbs.
Understanding the timing, potential depth and the extent of storm surge is essential for evacuating people from low-lying areas and for deploying emergency management personnel.
However, a storm surge forecast is only as good as the wind forecast, because surge is wind-driven.
There is a move toward combining the wind and surge forecasting capabilities into one overall modeling framework, with the advantage that surge could be calculated simultaneously with the wind field.
This is something that should be adopted fully because it would immediately give forecasters more information about potential surge effects.
There is also a move toward developing improved risk assessment. For example, providing a “cone of uncertainty” for storm surge, in the same way that one is developed for the hurricane track, would provide information to the public in a way that is both informative and understandable.
This is a good step and one that should be implemented now.
These advances require interdisciplinary teams of researchers to collaborate and must be fostered by sustained financial research support over several years.
Moving forward, lawmakers should recognize that investments in predictive simulation technology will lead to big benefits to society in terms of better preparedness.
Researchers should continue to try new things in developing new hurricane prediction methods and work closely with forecasters to more rapidly, yet prudently, adopt this technology.
We can, and should, be doing more when it comes to storm surge prediction.
Clint Dawson is a professor of aerospace engineering and engineering mechanics in the Cockrell School of Engineering at the University of Texas at Austin.