云南省植被NDVI时间变化特征及其对干旱的响应

基于云南省74个气象站点的1997—2012年逐日降水资料和逐旬SPOT-NDVI值,利用标准化降水蒸散指数(SPEI)多尺度分析了云南省干旱时间和强度演变与NDVI时间动态特征及其相关性分析,进而探讨气候变化对植被的影响。结果表明,1999—2013年云南省年平均NDVI值和年最大NDVI值均呈现波浪式的发展趋势,其趋势线斜率分别为0.0017和0.0011;NDVI年内各月变化情况大体上相同;不同季节NDVI的年际变化特征呈现出显著差异。1997—2012年不同时间尺度SPEI均体现出干旱化加剧的趋势,并随SPEI的时间尺度增大而增大;3个月尺度的SPEI值(SPEI3)结果表明,各月的变化呈现先增大后减小的趋势;SPEI3反映出多年季节水平的干旱强度为:冬季秋季春季夏季。总体上,云南省的年均NDVI与SPEI的相关性极弱,年最大NDVI与SPEI呈正相关;多年月均NDVI与不同尺度SPEI的相关性较强且存在滞后性;不同季节NDVI与SPEI的相关性及滞后性有较大差异,其中冬季NDVI、秋季NDVI与其当年当季SPEI的负相关性较强。

Temporal dynamics of vegetation NDVI and its response to drought conditions in Yunnan Province

The characteristics of drought at different time scales between 1997 and 2012 in Yunnan Province were studied using a new drought index (standardized precipitation evapotranspiration index, SPEI) calculated based on daily precipitation data from 74 weather stations. In addition, we used a time series data set (1999-2013) of the normalized difference vegetation index (NDVI) and time series analysis to assess temporal dynamics of the NDVI, and we explored the relationships between NDVI and SPEI. Correlations between NDVI and multiscale SPEI were examined, including annual, monthly, and seasonal long-term dynamics. Annual average and annual maximum NDVI in Yunnan Province showed an upward trending wave shape with slopes of 0.0017 and 0.0011, respectively. Monthly NDVI showed the same pattern, and the wave crest in each year was concentrated in July through September. Interannual variability in the NDVI showed significant seasonal differences, with higher NDVI in summer and autumn compared to spring and winter. A relatively consistent and dynamic trend was found in the time scales of mean annual SPEI. Drought intensity increased in small increments from 1997 to 2012, and the trend was more obvious at larger time scales. The 3-month SPEI value (SPEI3) initially increased and then decreased, and its wave crest was concentrated in May through July. The seasonal SPEI3 value reflected the level of drought strength as follows: winter > autumn > spring > summer. We found that the correlation between NDVI and SPEI at different time scales was complex and variable. The relationship between annual average NDVI and annual SPEI was weak, while the annual maximum NDVI and annual SPEI were positively correlated, especially for SPEI3, SPEI6, and SPEI12. No significant correlation was observed between monthly NDVI and SPEI; in contrast, years with average monthly NDVI showed a strong response to SPEI at different time scales. The coefficients between years for average monthly NDVI and SPEI were strongly affected by the SPEI time scale, with a 1- and 2-month lag. The NDVI in autumn and winter showed a strong negative correlation with SPEI for the same season, and the autumn NDVI was slightly correlated with SPEI in summer of the same year. The lag effect was also found in other seasons to different extents.