基于加权马尔可夫模型的标准化降水指数干旱预测研究

基于不同时间尺度标准化降水指数的干旱监测结果,以规范化的各阶白相关系数为权重,采用加权马尔可夫链方法对未来干旱状态进行预测和分析.以关中平原和渭北旱塬36个气泉站39年逐月降水量为分析数据.系统地分析了该方法在不同时间尺度(从1个月到1年)上的预测能力及存在的问题.结果表明:对所选的5个时间尺度该方法都有一定的预测能力,并且随着时间尺度的增加,预测正确率也相应提高.同时,该方法对无旱的预测比较准确.对干旱的发生也有一定预测能力,可以作为早期干旱预警的参考.但是,该方法对干旱状态突变的预测能力较弱;随着干旱程度的加重其预测能力也逐渐降低.

Prediction of drought occurrence based on the Standardized Precipitation Index and the Markov Chain Model with weights

Values of Standardized Precipitation Index(SPI) are calculated from monthly precipitation data collected from 36 weather stations in Guanzhong Plain and Weibei tablelands.The Markov chain model with weights is applied to predict SPI drought intensity by using standardized self-coefficients as weights.The prediction temporal scales of the model are set to 1 month,3 months,6 months,9 months and 12 months.The results show that this Markov chain model has ability to forecast SPI drought intensity at the five temporal scales.The longer the temporal scale,the better the predication.However,the forecasting ability is weak when there is a sharp change or increasing of drought intensity.