In October 2012, Hurricane/Superstorm Sandy killed dozens of people and caused more than $60 billion in damage, making it one of the Northeast’s most destructive storms in decades and the second-costliest hurricane in U.S. history. Computer forecast models did not reach strong agreement on Sandy’s unusual track until several days before landfall, which limited the amount of time available to protect lives and property.
One factor hindering the skill of weather prediction models is that most track the atmosphere at points separated by 10 km (6 miles) or more. This is too coarse to directly simulate the showers and thunderstorms that help drive hurricane behavior. Increased computer power now supports some models with resolution as fine as 3 km (2 mi). Experiments with even finer-scale models (less than half a mile) by the National Center for Atmospheric Research and partners are now helping shed light on the costs and benefits of improved resolution. Such research, once it is evaluated and deemed useful, can be transferred to national and global weather prediction centers to help improve model skill and forecast quality.
One of the most detailed simulations of a weather event ever conducted has been achieved for Hurricane Sandy. This project utilized ARW, the NCAR-based advanced research version of the multiagency WRF model (Weather Research and Forecasting). ARW simulated a large region surrounding Sandy, from eastern North America to the western North Atlantic, at a horizontal resolution of 500 meters (1,640 feet), with 150 vertical levels. The atmosphere’s behavior was tracked every second over a four-day period at more than 4 billion points. The results capture key aspects of Sandy in detail, including the intensification of winds just before landfall and the bands of heavy snowfall produced by Sandy across the Appalachians.
The ARW modeling was done by researchers at NCAR, Cray Research, and the National Center for Supercomputing Applications (NCSA) at the University of Illinois at Urbana-Champaign. It involved a Cray XE6 supercomputer named Blue Waters, located at NCSA, as well as the NCAR-based VAPOR platform (Visualization and Analysis Platform for Ocean, Atmosphere, and Solar Researchers).
Primary: National Science Foundation
Seeking additional funding for expansion of ultra-fine-scale computing as part of a proposed forecasting and warning testbed project for the 2013 Atlantic hurricane season.
National Center for Supercomputing Applications, University of Illinois at Urbana-Champaign
U.S. Office of Naval Research