多种气象干旱指数在新疆干旱评价中的应用对比研究

Evaluation and comparability of four meteorological drought indices during drought monitoring in Xinjiang

选取新疆50个气象测站1960-2010年间逐日降水、气温资料,采用模糊C均值聚类方法,并结合新疆地形地貌及水汽来源等地理特征,将全疆分为北疆、南疆、天山北坡、天山南坡及天山东段5个区域,对比分析标准化降水指数(SPI)、标准化降水蒸散发指数(SPEI)、自适应帕默尔指数(scPDSI)及有效干旱指数(EDI)的优缺点及其在各分区干湿监测过程中的适用性。SPI、SPEI、EDI对降水、气温响应较快,且干湿趋势变化所呈总体特征一致,EDI较SPI及SPEI对降水量变化更加敏感,SPI较EDI对"干"的评定较重,对"湿"的判定较轻,SPEI综合考虑气温及降水对"干"、"湿"的共同作用,能够监测到SPI及EDI无法监测到的一些特大旱情,而scPDSI对干湿事件的判断结果与其他三种指数相差较大,甚至会出现完全相反的判断结果,且scPDSI对旱情严重程度的判断偏轻,对干湿变化响应较慢,不适用于新疆干湿监测。SPI与SPEI相关性较好,SPI与EDI相关性明显好于SPEI与EDI的相关性,而scPDSI与其他三种干旱指数的相关性均较弱。另外,从干旱发生频率、干旱站次比及干旱强度三个角度进一步分析不同等级干旱事件的发生规律,整体上,干旱发生具有明显的区域性和季节性特征,不论是年或者四季干旱,干旱影响范围均呈缩减趋势。

Drought is perhaps the most complex natural hazard. Drought monitoring is an essential component of drought risk management. It is usually performed with the use of various drought indices. A number of drought indices have been introduced and applied in the drought monitoring in different regions to date all over the world. Xinjiang was divided into five sub-regions with Fuzzy Clustering Method along with considering its geographic and geomorphic conditions and different vapor sources： northern Xinjiang,southern Xinjiang,northern slope of Tianshan Mountain,southern slope of Tianshan Mountain,and the eastern part of Tianshan Mountain. Comparability analyses and correlation analyses have been carried out between the Standardized Precipitation Index（ SPI）,the Standardized Precipitation Evapotranspiration Index（ SPEI）,the Effective Drought Index（ EDI）,and the Self-calibrating PDSI（ sc PDSI） to investigate advantages and disadvantages of four drought indices and to pick out the most applicable drought index for the five sub-regions in Xinjiang based on daily precipitation and temperature data from 50 meteorological stations in Xinjiang. SPI,SPEI,and EDI respond quickly to precipitation and temperature. The tendency and changing characteristics were in consistent with each other. Drought classes and wet classes defined with SPI are higher and lower than with EDI,respectively. Since SPEI can take into account of combined action of precipitation and temperature in the evaluation of drought and wet,some severe droughts that cannot be detected with SPI and sc PDSI can be monitored with SPEI. The correlation between SPI and EDI is better than it between SPEI and EDI. The difference of variation tendency between SPI,SPEI,EDI and sc PDSI is very large and the relationship between sc PDSI and other drought indices is worst. Therefore,sc PDSI is not applicable to drought monitoring in Xinjiang. In addition,drought frequency,drought severity,drought stations proportion of different classes of annual and seasonal drought have been analyzed with SPEI12. No matter annual or seasonal droughts,there are obvious regional and seasonal characteristics. And the influence area has been reduced to an extent.