基于马尔可夫链的降水异常等级预测

Prediction of precipitation abnormal grade using Markov Chain

基于东江流域1951~2000年的降水序列,计算分析东江流域不同分析尺度的标准化降水指标序列,并依据标准化降水指标值对降水异常程度进行等级划分。运用马尔可夫链分析研究区降水异常等级状态间的转移概率,以及不同程度降水异常的平均持续时间等特性,并进一步根据系统统计特性进行未来1~3个月降水预测的决策风险进行评估。研究结果可为水资源管理,尤其是干旱期水资源配给提供一定的帮助。

SPI series of precipitation in the Dongjiang River Basin from 1951 to 2000 is calculated under different time-scales and precipitation abnormal grade is classified accordingly.Transition probability matrix between the abnormal grades and average durations of different precipitation grades are obtained by Markov Chain analysis.Furthermore,the reliability of precipitation prediction in future 3 months based on statistical precipitation series is evaluated.The results can provide support for water resources management,especially water resources allocation in dry periods.