Marsha Berger Research Group

The group develops adaptive mesh and solution-refinement techniques for computational fluid dynamics to capture multiscale flow features. Researchers design and analyze adaptive numerical methods that reduce discretization error while concentrating computational effort where needed. The team implements parallel scientific computing frameworks and domain-decomposition strategies to scale adaptive solvers on high-performance systems. They study algorithmic mechanisms that combine adaptivity with parallelization to accelerate convergence and enable large-scale CFD simulations.