The Weather Research and Forecasting (WRF) model is a U.S. community numerical weather prediction (NWP) tool, collaboratively developed from contributions by numerous research institutions and government agencies. WRF is the primary numerical weather prediction model used by NARAC to generate meteorological fields when the coverage provided by weather observations and/or the resolution of publicly available forecast data are insufficient to capture the complex flow patterns necessary for accurate plume predictions.
NARAC is investigating and developing operational meteorological data assimilation methods such as observational nudging and 4-dimensional data assimilation (FDDA). These methods incorporate observational data to improve WRF model initialization and the accuracy of predicted meteorological fields. NARAC also is researching the use of multi-physics model ensembles to evaluate uncertainty in WRF predictions.
To improve simulations of turbulent mixing in the planetary boundary layer, LLNL has developed advanced turbulence models that are currently available in the public version of the WRF modeling tool. In collaboration with UC Berkeley, LLNL is leading the ongoing development and implementation of an immersed-boundary method in WRF to improve simulated wind flow when steep terrain or buildings are present in a numerical domain.