基于MODIS-EVI和CI的洞庭湖流域植被指数对气象干旱的响应

基于洞庭湖流域2000~2012年98个气象站点的综合气象干旱指数（CI）和MODIS增强型植被指数（EVI）数据，对研究时段内生长季干旱强度和EVI时空变化特征，典型春、夏、秋季干旱年月和湿润年月干旱强度和EVI的空间分布及相关性进行分析，旨在研究EVI指数能否反映大范围气象干旱事件，探讨该指数作为流域干湿状况度量标准的潜力。结果表明：（1）在年际变化上，EVI在2001、2005、2009年出现低谷值，与其对应年份的干旱强度出现高峰值；在季节变化上，干旱强度的高值段主要集中在夏秋季，低值段主要集中在冬春季；夏季EVI值最大，冬季EVI值最小；（2）去除年际变化和季节性影响的生长季干旱强度、EVI和森林覆盖率年际波动变化明显，并且干旱强度高峰值的年份对应EVI和森林覆盖率出现低值的年份；（3）典型干旱年月干旱强度高值区的空间分布与EVI距平低值区的空间分布存在高度一致性，其中夏、秋季干旱年月的干旱强度和EVI的相关系数分别通过了0.01、0.1的显著性水平检验；（4）除部分地区外，典型春夏秋季湿润年月的干旱强度空间分布与EVI距平的空间分布对应较差，两者的相关系数未通过显著性检验。上述结果表明，在典型干旱年月，流域植被的生长状况能够响应大范围的区域气象干旱情况。

Responses of Vegetation Index to Meteorological Drought in Dongting Lake Basin Based on MODIS-EVI and CI

Based on the comprehensive meteorological drought index (CI) calculated from 98 meteorological stations and the MODIS enhanced vegetation index (EVI) data from 2000 to 2012 in the Dongting Lake Basin, we analyzed spatial and temporal variations of drought intensity and EVI during growing seasons. We also analyzed spatial distributions of drought intensity and EVI in spring, summer and autumn of typical dry and wet years as well as the correlation between drought intensity and EVI. We aimed to answer the question whether variations in EVI can reveal large-scale meteorological drought events, and to explore the potential of EVI as an indicator of dry and wet conditions in this basin. The results showed that (1) the lowest value of EVI occurred in 2001, 2005 and 2009, while the highest value of drought intensity also appeared in these three years. For drought intensity, high values were mainly distributed in summer and autumn, while low values were present in winter and spring. For EVI, the maximum and minimum values were present in summer and winter, respectively. (2) The drought intensity, EVI and forest coverage that have been corrected for inter-annual and seasonal effects during the growing season showed significant inter-annual fluctuations. The year with high drought intensity was consistent with that with low forest coverage. (3) The spatial distribution of high drought intensity matched well with that of low EVI in typical drought years. The correlation coefficients between drought intensity and EVI in summer and autumn of dry years were statistically significant (P<0.01 and 0.1 for summer and autumn, respectively). (4) The spatial distribution of drought intensity and EVI anomalies did not match in spring, summer and autumn in wet years. No significant correlations were found between drought intensity and EVI anomalies during these time periods. Our results suggested that growth of vegetation can well reflect large-scale meteorological drought in the Dongting Lake Basin.