Evaluation of surface winds over the Korean Peninsula and its surrounding seas

Abstract

Surface winds over ocean can influence not only on turbulent heat fluxes but on coastal erosion mainly through changes in ocean wave. To project a future change in surface wind, a global climate model has been widely used but for detailed regional information, dynamical downscaling with a regional model has been used. The purpose of this study is to evaluate the surface winds over the Korean Peninsula and its surrounding seas simulated by CORDEX-EA (Coordinated Regional Climate Downscaling Experiment for East Asia) multi-model ensemble. While in many studies the evaluation of climate models has been conducted based on reanalysis data, we first verified the ERA-interim and NCEP/DOE reanalyses based on an in-situ observational data and then evaluated the regional models in comparison with CMIP5 global models for 1989-2005. Although surface wind at the boundary of CORDEX domain was similar between these two reanalyses, the wind over the Korean Peninsula was different between them and was overestimated by more than two times compared to the in-situ data. The regional models simulated a relatively stronger wind over the east coast of South Korea during summer and over the east and westboundaries of the Japan Basin during winter, whereas the CMIP5 global models were incapable of simulating such a wind pattern varying with season over the coastal regions of the East Sea. Thus we suggest that a regional modelling is necessaegional information, dynamical downscaling with a regional model has been used. The purpose of this study is to evaluate the surface winds over the Korean Peninsula and its surrounding seas simulated by CORDEX-EA (Coordinated Regional Climate Downscaling Experiment for East Asia) multi-model ensemble. While in many studies the evaluation of climate models has been conducted based on reanalysis data, we first verified the ERA-interim and NCEP/DOE reanalyses based on an in-situ observational data and then evaluated the regional models in comparison with CMIP5 global models for 1989-2005. Although surface wind at the boundary of CORDEX domain was similar between these two reanalyses, the wind over the Korean Peninsula was different between them and was overestimated by more than two times compared to the in-situ data. The regional models simulated a relatively stronger wind over the east coast of South Korea during summer and over the east and westboundaries of the Japan Basin during winter, whereas the CMIP5 global models were incapable of simulating such a wind pattern varying with season over the coastal regions of the East Sea. Thus we suggest that a regional modelling is necessa