摘要
针对干旱预警预测十分困难的问题,基于湖北省16个气象站点1960~2009年的逐月降水资料,计算了各站点的标准化降水指数（SPI）,并利用马尔科夫链模型预测了湖北省未来干旱的发生概率,得出了各个干旱等级的发生概率、预计持续时间、初次转移到无旱状态的时间、重现时间以及预见期为1~3个月的短期概率。与实际统计频率对比发现,马尔科夫链模型能很好地估算各个干旱等级的发生概率,且大多数站点未来干旱等级预测概率的误差都控制在20%以内,说明该模型对于未来干旱情况具有一定的预测能力,可以为早期干旱预警提供依据。
In order to study drought prediction which is a very difficult task among drought management,monthly standardized precipitation index（SPI）of 16 meteorological stations are calculated based on the monthly precipitation data during 1960-2009 in Hubei Province.Furthermore,Markov chain model is used to estimate the probability of different drought severity classes,the expected time in each class of severity,the expected time for the SPI to change from a particular class to none-drought class,the recurrence time to a particular drought class and the short-term prediction of 1-3months.Compared with actual statistical frequency,the prediction results indicate that Markov chain model can simulate the drought probability in a high precision.Moreover,the error of short-term prediction results of most stations is limited to 20% and the model has predictive ability for drought situation in the future.So,this model can provide a basis for early drought warning.
出处
《水电能源科学》
北大核心
2015年第4期6-9,51,共5页
Water Resources and Power
基金
国家自然科学基金项目(51279140
51339004
51279139)
中央高校基本科研业务费专项资金项目(2042014kf1012
2042014kf0033)
关键词
干旱
标准化降水指数
马尔科夫链模型
干旱短期预测
drought
standardized precipitation index
Markov chain model
short-term drought prediction
