Assessment of candidate distributions for SPI/SPEI and sensitivity of drought to climatic variables in China

作者：Huaijun Wang, Yaning Chen, Yingping Pan, Zhongsheng Chen, Zhiguo Ren · 发表于：International Journal of Climatology · 年份：2019 · DOI：10.1002/joc.6081 · 被引用次数：110 · 研究领域：Hydrology and Drought Analysis、Climate variability and models、Plant Water Relations and Carbon Dynamics

Drought occurs more frequently in the context of climate change and threatens water security worldwide. An appropriate fitting distribution is crucial for accurately identifying drought using drought indices. Here, seven two‐parameter distributions (Gamma, Gumbel, Logistic, Log–Logistic, Log‐normal, Normal, and Weibull) and four distributions (general logistic, generalized extreme value, Normal, and Pearson Type III) are applied to calculate the standardized precipitation index (SPI) and standardized precipitation evapotranspiration index (SPEI), respectively, to choose the most appropriate fit for the 541 stations in China. The results show that, in most cases, Gamma and general logistic were the best distributions for SPI and SPEI, respectively. Nevertheless, we noted that Pearson Type III and generalized extreme value also fit the SPEI series well, indicating the importance of distribution fitting assessment for various regions. The annual and seasonal drought evolutions across China were clarified, with drought decreasing significantly in western Northwest China (annually and each season), northern North China (spring and winter), the Tibetan Plateau (spring), and the lower reaches of the Yangtze River (winter) and increasing mainly in South China (spring) and western South China (summer and autumn). The intensity, duration and severity of light, moderate, and severe drought were also detected; the results suggest that drought in China is mainly concentrated around the ch...