Data Assimilation Experiments of an Antarctic Cyclone Using WRF 4DVAR

In this research, we try to examine the performance of the WRF 4 dimensional variational (4D-Var) data assimilation system in the analysis and prediction of an Antarctic cyclone that penetrate the western Antarctic Ice Sheet (WAIS) from 03 to 06 October 2007. The cyclone underwent an extensive development with the wind strength near the Ross Ice Shelf increases from < 10 knots on October 3 at 1200 UTC to more than 50 knots on October 6 at 0900UTC. Strong winds affect the safety of aircraft taking part in operations of the United States Antarctic Program (USAP). Because of scarcity of data in Antarctica and the modest performance of models in transition seasons (fall and spring) this case is selected to demonstrate the impact of data assimilation technique in Antarctica. Various observational data for this case has been collected, including conventional observations, automatic weather station (AWS) data in Antarctica, AMSU-A(B) satellite radiance, retrieved MODIS satellite winds, QuikScat winds over the Southern Oceans around the Antarctica, COSMIC radial occultation data. We carried out three sets of experiments, i.e. CTRL experiments that using the GFS analysis as initial conditions, WRF 3DVAR experiments that assimilated one-time observational data, and WRF 4DVAR experiments that assimilate observations in a 6-hour window. We compare the performance of these experiments and evaluate the 4DVAR technique on the case study. The WRF 4DVAR shows its advantage over its 3DVAR component in Antarctic applications.