Nielsen-Gammon, J. W., R. T. McNider, W. M. Angevine, A. B. White, and K. Knupp, 2007: Mesoscale model performance with assimilation of wind profiler data: Sensitivity to assimilation parameters and network configuration. J. Geophys. Res., 112, D09119.
Abstract
The performance of the nudging scheme in the MM5 numerical model is evaluated for a 4-day simulation of an ozone episode during the TexAQS-2000 field study. Quality-controlled horizontal wind data from five boundary layer profilers spaced 30–50 km apart are used as input data for the nudging as well as for quantification of model errors. Observations from selected profilers are withheld to provide independent data to evaluate the performance of the model. A series of experiments are conducted to estimate the overall model performance as well as the dependence of the model performance on nudging parameters and on profiler network configuration. Within the profiler network, nudging generally reduces 4-day average biases in the wind components to less than 0.3 m s–1 and yields rms errors averaging 1.75 m s–1 to 2.08 m s–1, with the best model performance at about 500–1500 m above ground level. Rms errors without data assimilation are about twice as large. Model errors, with and without assimilation, are generally largest at night. Wind fields that are more accurate in a statistical sense can be obtained by simpler analysis schemes, but the consistent dynamical evolution of all meteorological variables, such as the pattern of temperatures associated with corrected sea breeze location and intensity, are lost. Tests with a reduced profiler network showed that a single profiler in an appropriate upstream location can produce model performance equivalent to that from a four-profiler network surrounding the validation station.