Prediction of Rainfall with High-Resolution QPF Model Using Public-Resource Distributed Computing

In this paper, an attempt has been made to predict and validate rainfall over the Korean Peninsula using a high resolution QPF model during a very strong precipitation event. The prediction of the intensity of precipitation depends on the vertical velocity and has been incorporated into cross-platform-based package to use the cost effective Public-Resource Distributed Computing System of Korea. The prediction has been made for the rainfall associated with the Typhoon Nabi (Jolina) during the period of 1200 UTC 5 September to 1200 UTC 7 September 2005. The maximum intensities of the predicted 3 hourly rainfall accumulations and the spatial distributions compare well with the observed ones. The root mean square error (RMSE) of the QPF model is less compared to that of the Regional Data Assimilation Prediction System (RDAPS) and the correlation coefficient (R-2) of the QPF model with the realized rainfall is higher than that of the RDAPS. This high resolution QPF model appears to be useful for a near-real time forecast of precipitation.