摘要 : The purpose of this study was to conduct a yield prediction from 2001 to 2050 on a national level by developing the Gentan probability-based model that combined yields of 47 prefectures. First, the Gentan probabilities were estima... 展开 The purpose of this study was to conduct a yield prediction from 2001 to 2050 on a national level by developing the Gentan probability-based model that combined yields of 47 prefectures. First, the Gentan probabilities were estimated for the 22 prefectures that had reliable harvesting data by the maximum likelihood method. Second, these prefectures were clustered into 3 groups based on the estimated Gentan probability patterns, and then the synthesized Gentan probabilities were estimated for each cluster from the aggregated stand areas and harvested areas of prefectures which belonged to the same cluster. Third, the other 25 prefectures without reliable harvesting data were classified into the preceding 3 clusters by discriminant analysis with socioeconomic explanatory variables. Fourth, harvesting volume were calculated for each prefecture based oh the estimated Gentan probabilities: The former 22 prefectures employed the Gentan probabilities derived from their own harvesting data while the latter25 prefectures employed the synthesized Gentan probabilities corresponding to their cluster number classified. Finally, a yield prediction was conducted from 2001 to 2050 based on the 5 scenarios that increased the means and standard deviations of Gentan probabilities in 47 prefectures by -10%, -5%, 0%, 5% and 10% every 5 years during the prediction periods. The results showed that harvesting areas decreased as the means and standard deviations got raised, and that the scenario of 0% increase brought about moderate result. To actualize this scenario, a standard rotation age in the Regional Management Plan should be raised 5-7% every 5 years for next 50 years. 收起