基于气象旱涝指数的旱涝急转事件识别方法

基于长江流域212个气象站点1961-2017年的日降水资料,借助标准化加权平均降水指数（SWAP）,结合多门槛游程理论,提出一种识别旱涝急转事件的新方法。方法应用于旱涝急转事件高发的长江流域,分别从典型站点旱涝事件分析、区域典型旱涝急转事件分析、旱涝急转事件时空分布规律分析等角度,探讨了长江流域1961-2017年旱涝急转事件规律。结论显示：①SWAP指数对于旱涝事件具有良好的识别能力。②聚类方法可聚合相似旱涝急转事件,2011年长江中下游旱涝急转事件中干旱事件占主导地位,持续时间远长于洪涝事件。③ 长江流域旱涝急转事件呈现明显的区域规律：上游发生频率较低,中下游偏高;此外,长江流域多数分区近期旱涝急转事件发生频率呈现上升趋势。研究结果表明,基于SWAP指数并结合多门槛游程理论的方法能够比较准确地识别旱涝急转事件,可进一步应用于旱涝急转事件的预测及评估中。

A method to identify the drought-flood transition based on the meteorological drought index

A new method was proposed to identify drought-flood transition events by combining a drought-flood index (Standard Weighted Average Precipitation, SWAP) with the multi-threshold theory. This method was tested in the Yangtze River Basin using daily precipitation data from 212 stations for the 1961-2017 period. With this method, the meteorological drought and rainstorm flood in the representative station were identified, and representative regional drought-flood transition events and spatiotemporal patterns of drought-flood transition were analyzed. The results show that: SWAP is an effective index to identify the meteorological drought and rainstorm flood. K-means clustering can classify similar drought-flood transition events into one cluster. The drought event plays a dominate role in drought-flood transition events for the middle and lower reaches of the Yangtze River Basin in 2011, and the drought lasts for a much longer duration than the flood during the drought-flood transition event. Drought-flood transition events show an obvious regional pattern for the Yangtze River Basin with low frequency for the upper reaches and high frequency for the middle and lower reaches. In addition, the drought-flood transition frequency presents an increasing trend recently for most parts of the Yangtze River Basin. Overall, the results imply that the proposed method combining meteorological drought index with multi-threshold theory is capable of identifying drought-flood transition events, and can be further used for predicting drought-flood transition events.