A mesoscale real-time four-dimensional data assimilation (RT-FDDA) and short-term forecasting system has been developed at NCAR/RAP for the U.S. Army Test and Evaluation Command (ATEC) at the White Sand Mission Range (WSMR). Built on the Penn State/NCAR mesoscale model (MM5), the system continuously assimilates all available observations from various sources and provides updated 3-dimensional analyses and short-term forecasts every 3 hours. The data assimilation engine is based on the Newtonian Relaxation (nudging) method. The nudging method was mostly tested on synoptic scale, but few studies has been done with fine resolution grid (grid increment of 1-3km) that can resolve fine scale terrain and very local circulations. At the fine scale, choosing the appropriate nudging parameters, such as (1) the influence radius of observations, (2) nudging coefficients and (3) weight functions, may have critical effect on the quality as well as the correctness of the data assimilation result.

A series of sensitivity simulations are conducted in this study by gradually changing the three nudging parameters. A clear-sky case with lifecycles of local thermal driven circulations is chosen for the tests. Parallel RT-FDDA runs are performed using different nudging parameters. Besides simply varying the magnitude of the constant parameters, terrain-adjusted and flow-dependent nudging weights are developed. In each model run, sixteen analysis/forecast cycles are conducted for the 48 h period. To objectively assess the impact of the nudging parameter, verification statistics are calculated for the analysis/forecasts against a subset of the rawinsonde and surface observations that is withheld from the data assimilation.