Projection of future precipitation change over South Korea by regional climate models and bias correction methods

Abstract

In this study, the effects of bias correction methods on the daily precipitation in South Korea produced by five regional climate models (RCMs), and the impact of bias-corrected precipitation on projected climate change is investigated. We use four bias correction methods (Linear Scaling (LS) Power Transformation (PT), Quantile Mapping for the entire period (QME), and Quantile Mapping for each month (QMM)). We perform pre-processing corrections for the dry period in advance of QME and QMM for the wet period. All bias correction methods improve both long-term temporal and spatial averaged precipitation in the present-day period. However, the second peak of the annual precipitation cycle in the QME method is underestimated. LS shows poor correction skills for the intensity and frequency of extreme precipitation. The pre-processing for the dry period also helps to correct the intensity and frequency of daily precipitation. The corrected precipitation characteristics could vary depending upon the bias correction method. Thus, bias correction methods must be carefully chosen according to the statistical features such as the mean and extreme values that should be corrected. For future analysis, PT and QMM are only applied to improve daily precipitation. RCMs simulate the increase in precipitation mainly in the southern regions of Korea. RCMs also show that the second precipitation peak of the annual cycle is significantly strengthened. The intensity of extreme precipitation is increased significantly in the projection of the two scenarios. Bias correction can contribute to the improvement of precipitation variability retaining the characteristics of raw RCM data.