Benchmarking Polar WRF in Antarctica

The Advanced polar-optimized Weather Research and Forecasting Model (Polar WRF) is being developed at The Ohio State University's Byrd Polar Research Center for high latitude simulations. Accurate simulations can provide both successful weather forecasts for the U.S. Antarctic Program via the Antarctic Mesoscale Prediction System and a means to study high latitude climate processes in Antarctica. This study details ongoing model verification efforts over Antarctica.

The overarching objectives are to identify optimal configurations for Polar WRF simulations over the Antarctic and to provide performance evaluation for a full annual cycle using detailed observations. The ERA-40 Reanalysis, RAMP digital elevation high horizontal resolution (200 m) topography, USGS vegetation, and the Reynolds SST data sets are used to specify initial, lateral and lower boundary conditions for the present simulations. An initial configuration using the Noah Land surface model, RRTM longwave, the MRF PBL scheme, Single Moment 5-class microphysics and Grell convection is applied. The model is integrated in forecast mode, allowing re-initialization every 48 hours with 27-vertical levels and 120x120 grid points in the x and y direction at 60-km resolution. Surface data for the detailed diagnosis were obtained from archives of the University of Wisconsin's Automatic Weather Station (AWS) program at the Antarctic Meteorological Research Center (AMRC), and the National Climatic Data Center (NCDC) while upper air data was obtained from the British Antarctic Survey (BAS), the University of Wyoming, and the Baseline Surface Radiation Network (BSRN).

Preliminary results show a warm bias in 2-m air temperature in both the austral winter and summer months of July and January 1993. We have identified three potential sources of this warm bias; ERA-40 initializes the model soil temperatures with a warm bias, the constant surface albedo of 0.8 may be too low for the interior and high mixing ratios in the middle troposphere are leading to higher cloud fractions and hence heating from long-wave radiation. Using the WRF Single Moment 5-Class scheme results in better simulations of the air temperature compared to the Morrison scheme. We initially use 1993 to facilitate comparisons of the Polar WRF performance with those from Polar MM5.