Spatial and temporal analysis of drought using entropy-based standardized precipitation index: a case study in Poyang Lake basin, China

Drought is a frequent and worldwide disaster causing huge losses in agriculture and damages in natural ecosystems every year. Precise assessment and prediction of droughts are important for regional water resources planning and management. An alternative distribution, based on the entropy theory, was used to derive a unified probability distribution function (PDF) for different cumulative precipitation series to calculate the Standardized Precipitation Index (SPI). Thirteen meteorological stations located within the Poyang Lake basin with daily precipitation records from 1958 to 2011 were selected for spatial and temporal analysis of basin-scale droughts. The entropy-based distribution is proved to be flexible enough for modeling aggregated precipitation at different time scales by the Kolmogorov-Smirnov (K-S) test. Most severely and extremely dry months were recorded in spring and winter in the Poyang Lake basin over the study period. Negative trends of the short-term entropy-based SPIs and corresponding aggregated numbers of rainy days in spring and autumn are detected based on the Mann-Kendall test, which implies an upgrade in difficulties to mitigate the agricultural droughts in the Poyang Lake basin. Once droughts occurred, regions with less frequent drought would face severer drought degree. The lower Ganjiang River, lower and middle Fuhe River, as well as the Xinjiang River are identified to be the most vulnerable regions with highest drought intensities. Droughts could occur at any periods and move from region to region in the Poyang Lake basin; thus, well preparation for potential droughts is needed.