Improvements to the WRF Seasonal Hindcasts over South Africa by Bias Correcting the Driving SINTEX-F2v CGCM Fields

Abstract

Abstract In an attempt to improve the forecast skill of the austral summer precipitation over South Africa, an ensemble of 1-month-lead seasonal hindcasts generated by the Scale Interaction Experiment–Frontier Research Center for Global Change (SINTEX-F2v) coupled global circulation model is downscaled using the Weather Research and Forecasting (WRF) Model. The WRF Model with two-way interacting domains at horizontal resolutions of 27 and 9 km is used in the study. Evaluation of the deterministic skill score using the anomaly correlation coefficients shows that SINTEX-F2v has significant skill in precipitation forecasts confined to western regions of South Africa. Dynamical downscaling of SINTEX-F2v forecasts using the WRF Model is found to further improve the skill scores over South Africa. However, larger improvements in the skill scores are achieved when the WRF Model is forced by a form of bias-corrected SINTEX-F2v forecasts. The systematic biases in the original fields of the SITNEX-F2v forecasts are removed by superimposing the SINTEX-F2v 6-hourly anomalies over the ERA-Interim 6-hourly climatological fields. The WRF Model forced by the bias-corrected SINTEX-F2v shows significant skill in the forecast anomalies of precipitation over most parts of South Africa. Interestingly, the WRF Model runs with the bias correction did not help to improve the SINTEX-F2v forecast of 2-m air temperatures. Perhaps this is because of the large biases in the precipitation forecast by the WRF Model driven by the bias-corrected SINTEX-F2v. These results are important for potentially improving seasonal forecasts over South Africa.