马尔可夫链在太湖流域梅雨期预报中的应用

Application of Markov chain to forecasting plum rain period of Taihu Lake Basin

依据1954～1991年太湖流域的入、出梅日期以及梅雨集中期，应用马尔可夫链预报了太湖流域1992年和1993年的梅雨期，并利用其平稳分布分析了梅雨期系统在各状态下的概率．应用痞果表明：在太湖流域气候条件稳定的情况下，历史数据越多越精确，预测也就越可靠；以马氏链预测梅雨状态的方法是一种概率预报方法，需通过大量的预测检验才能显示出该方法的准确性和可靠性．

According to the data of beginning and ending of the plum rain period and the main periods of plum rains from 1954 to 1991, the Markov chain was applied to forecasting the plum rain period of 1992 and 1993 for the Taihu Lake Basin, and the probability of the plum rain period in different states was analyzed based on its stability distribution. The analysis shows that, with long and accurate historical data series, the present model can give satisfactory plum rain period forcasting for the Taihu Lake Basin under the stable climate condition. However, as a probability forecasting model, the accuracy and reliability of the present model should be verified through a great amount of data forecasting and testing.