Improving skill precipitation forecast of the European Centre for Medium-Range Weather Forecasts (ECMWF) Integrated Forecasting System (IFS) model by using the quantile mapping method for Central Vietnam

Abstract

This study applies the Quantile Mapping (QM) post processing method to improve the precipitation forecast skill of the deterministic forecasts produced by the European Centre for Medium Range Weather Forecasts (ECMWF) Integrated Forecasting System (IFS). The study area focuses on Central Vietnam, using observational rainfall data from 41 meteorological stations and ECMWF-IFS forecast data over a 10 year period (2013-2022). The QM method is applied to forecasts with lead times of up to 10 days. The results show that QM significantly enhances the forecasting skill of the IFS model across all lead times, with varying degrees of improvement depending on forecast lead time and rainfall thresholds. However, the impact of QM on heavy rainfall events is limited, with skill improvements being statistically insignificant for lead times up to 72 hours in summer and 144 hours in winter.