ベイズ的アプローチによる日本域における地球温暖化に伴う地表気温上昇量の予測

摘要

A Bayesian approach is applied to project surface air temperature (SAT) changes over Japan using available simulations from the Intergovernmental Panel on Climate Change fourth assessment report. The Bayesian model averaging (BMA) can produce projections constrained by observations and weighted probability density functions (PDFs). We project annual mean and seasonal mean (winter and summer) SAT changes using the BMA and the arithmetic ensemble mean (AEM) and investigate the difference between the two methods. Additionally, we investigate the sensitivity of training period to the PDFs produced by BMA and the influences of long-term and short-term variations on them. These results indicate that the annual mean SAT change projected by BMA is significantly different from that projected by AEM and that BMA is sensitive to training periods and short-term variations.