Impact of data assimilation on Chukchi/Beaufort Seas mesoscale modeling

Offshore development in the Chukchi/Beaufort Seas requires an improved understanding of the surface wind field, a crucial parameter for assessing and predicting dispersal and movement of oil spills. A study has thus been established to investigate the mesoscale features of the surface wind field throughout this region. In this study, we focus on investigating the impact of data assimilation on the simulation of the Chukchi/Beaufort Seas mesoscale surface wind field with the Weather Research and Forecasting model (WRF) and its variational data assimilation system (WRF-Var). In the application of WRF-Var, model background error (BE) is a key element required for the assimilation procedure. We have conducted an investigation into how varying the model background error affects the results of assimilating in-situ surface observations. The background error covariances we tested include CV3 (a built-in global BE file in WRF-Var) and several domain-dependent CV5 BE files. The latter were calculated using 12-hourly forecast results and 24-hourly forecast over 2-month and 1-year periods, followed by the calculation of separate BE at 00, 06, 12, and 18 UTC, in order to capture the diurnal variation of the model background error. The results show that the customized domain-dependent BE produces much better assimilation results than the global CV3 BE, and the CV5 BE calculated over a 1-year period additionally helps to improve the assimilation. When the diurnal variation of background error was included in the CV5 BE, a further improvement was achieved.