CEE professor Arif Masud has been awarded 1.5 M node hours by the National Science Foundation (NSF) to develop tools for coupling different types of models vital to big-picture climate modeling. With this NSF provided computational resource, Masud and his team hope to improve methods for hurricane prediction.
Current Earth system models incorporate a variety of different phenomena at a variety of different scales. To ensure the highest possible accuracy from such complex systems, most models use a multiscale procedure called “nesting” in which a telescopic hierarchy is implemented to coordinate how different models interact at their respective scales.
Despite the success of these models in producing detailed outputs, they lack sound mathematical footings to merge global and local models that account for different variables and different resolutions. To address this issue, Masud and his students have developed a new computational approach called the Variational Multiscale Discontinuous Galerkin (VMDG) method. Unlike nesting, VMDG provides a unified framework capable of integrating mathematical models that differ in established variables and resolutions.
With the 1.5M node hours provided by NSF -Division of Mathematical Sciences in conjunction with the National Center for Supercomputing Applications (NCSA), Masud and his team will continue to develop VMDG’s ability to couple disparate models, with a focus on improving hurricane prediction.
Hurricanes form locally but are strongly influenced by global atmospheric dynamics. Changes to global convective trends have shown to impact localized heating patterns, increasing the intensity and frequency of high turbulent weather events like hurricanes. Thus, the tools in development in Masud’s lab become vitally important for the future of using local and global information to predict extreme climate events.
This ongoing research is funded by Sandia National Laboratories.