A Study on Superensemble Simulation of East Asia Summer Climate Projection

In this study, the superensemble simulation method is applied to combine the predicted results of various global circulation models: MPIM, NCAR, CCCma, CCSRNIES, GFDL and HCCPR. Climate factors considered are the 2m temperature, precipitation, sea level pressure and wind flow. Due to various domain and grid of each model output, all data (including the global reanalysis data ERA40) are interpolated to an identical domain (East Asia). Then, linear combination of all model outputs are done by designating a weighting coefficient for each model output by either the equal weight, which is the ensemble mean, or the inverse of square error summation (ISES) method. The training period for ISES is from 1980 to 1999, while the verification period will be from 2000 to 2005. Then the super-ensemble projections for the 2040s and 2090s decades are obtained and studied. Since we are mainly concerned with the summer climate projection, the variation of the intensity of the Pacific High and the expansion or contraction of its domain will affect significantly the strength of the southwesterly flow from South China Sea into Mainland China and therefore the intensification of temperature increase and rainfall enhancement over most of East Asia.