Evaluation and Bias Correction of Marine Surface Winds from CMIP5 GCMs for Wave Climate Modelling in the Western North Pacific

Abstract

The western North Pacific (WNP) is commonly indicated as region with high occurrence of high waves due to typhoons for decades. The increasing trend of typhoon frequency and intensity can be observed through General Circulation Models (GCMs) output from CMIP5. CMIP5 consist of a set of ensemble data including the marine surface winds for future wave climate modelling. In this study, validation of surface winds from CMIP5 GCMs is conducted by inspecting their quality and uncertainty over the study area, the western North Pacific (WNP). The monthly surface winds of ERA-Interim reanalysis data from 1979 to 2008 is taken into account as reference data, and monthly surface winds of 59 ensembles in CMIP5 GCMs over the same period are investigated as modelled data. First, the time series of reference and modelled grid-point marine surface winds are compared to find out their correlations. Then, bias correction methods using a variance correction, quantile mapping based on distribution correction, and quantile mapping based on Weibull correction are applied to modelled winds. The time series of surface winds at grids are also investigated with climate indices. Several points depict good agreements between reference and averaged ensemble data located on sea region, but relations between seasonal indices are still unclear. The proposed bias correction method in this study exhibits good improvement of the modelled marine surface winds for the study period in the WNP. In further study, the patterns and correlations with indices in addition to the bias correction method will be used to improve the future marine surface winds from CMIP5 GCMs for future wave climate modeling.