Some statisticaltransformation methods for bias correction of daily precipitation from meteorological models to the station scale – a case study in Binh Dinh province

Abstract

     Global Climate Model and Regional Climate Model are widely used to simulate regional climate despite large errors of models. Some bias correction techniques are applied to get more accuracy results. This research studies some common bias correction methods for daily precipitation to the station scale, a case study in Binh Dinh province. Eight methods of three Transformation groups: distribution derived transformations, parametric transformations, nonparametric transformations are selected for review. Mean absolute error index estimated from Cross-validation technique is used for ranking the methods. The nonparametric transformations is ranked as the best method in reducing biases of both of precipitation intensities and wet days. Distribution derived transformations is less effective. Large errors are occurred in the stations which have extreme precipitation due to the limitations of extrapolation of these techniques.