Investigating Lateral Boundary Forcing of Weather Research and Forecasting (WRF) Model Forecasts for Artillery Mission Support

Abstract

Abstract : Accuracy of global numerical weather prediction (NWP) models is highly dependent on initial atmospheric conditions and to a lesser extent on physical parameterizations. Generally, model error grows as a result of poor initial conditions. Limited area models (LAMs) have an additional source of error/variability from their lateral boundaries, forced from global models or other LAMs. Lateral boundary forcing can be a substantial source of error/variability for LAMs due to interpolation of coarse atmospheric data down to fine grids and high-resolution topography. Domain size has been identified as a way to control the spatial spin-up and internal variability of LAMs caused by lateral boundary conditions (LBCs). Previous research has found that larger LAM domains provide additional freedom for the model to create its own solution, whereas smaller domains are dominated by the forcing from the lateral boundaries. U.S. Army artillery uses LAM output to correct trajectories from standard atmosphere conditions to those of the current atmosphere. Sensitivity testing to the Weather Research and Forecasting model s LBCs over Europe is performed herein. Initial results from limited cases suggest modest difference between large and small domain experiments, except that the magnitude of error is a function of domain size.