Demonstrating the bias-correction impact on regional climate model (RegCM) over the Democratic People's Republic of Korea: Implication for temperature and precipitation

This study tests the nonlinear bias-correction method of regional climate model (RCM) outputs using observations from meteorological stations in the Democratic People's Republic of Korea (DPRK). The correction technique applies to both monthly precipitation and temperature simulated by the RCM, RegCM4.7 for the period 1991-2011. The correction parameters of both monthly precipitation and temperature are investigated for different periods within the observational time to assess the stability of this correction method over time. The results show that this nonlinear bias-correction approach improves significantly the accuracy of the RCM mean monthly precipitation and temperature as well as the values of probability distribution over the study area. The bias-corrected forecast has significant improvement for both monthly precipitation and temperature from about 18.9 mm and 0.7 degrees C of mean absolute error (MAE), and about 16 mm and 0.3 degrees C of the root mean square error over the uncorrected ones, respectively. The correlation coefficient, skill score and the MAE skill are improved significantly over the study region. The successful performance of the bias-correction technique used in this study during the calibration and validation periods will be helped to capture more accurate current climate conditions and to predict future climate change over the DPRK.

Automated summary

This study tests the nonlinear bias-correction method of RCM outputs and finds that it significantly improves the accuracy and probability distribution of the model. The bias-corrected forecast shows significant improvement in precipitation and temperature, and also improves the correlation coefficient, skill score, and MAE skill.