滇西南地区NDVI变化及其对不同时间尺度干湿变化的响应

Changes in NDVI and Its Response to Changes in Dry and Wet Weather at Different Time Scales in the Southwestern Yunnan,China

利用2001—2013年MODIS/NDVI归一化植被指数时间序列,采用变异系数、一元线性趋势变化等多项指标,结合SPEI指数和DEM分级数据提取不同高程地区的NDVI变化信息,分析滇西南地区不同海拔梯度NDVI时空变化规律及对不同时间尺度干湿变化的响应,结果表明:(1)研究区不同海拔梯度四季和年NDVI波动变化较小,NDVI变化趋势随海拔梯度变化存在一定的差异性,但总体上升趋势明显,四季和年NDVI改善区域占研究区46.9%~74.0%。(2)各海拔梯度NDVI变化百分率多集中在-10%~10%范围;以1500 m为界,变化百分率超过(小于)10%(-10%)的区域在高于(低于)1500 m的3个海拔梯度上。(3)各海拔梯度NDVI与2~8个月时间尺度SPEI指数相关性最为密切,表明研究区NDVI对降水及蒸散发导致的水分盈亏的响应具有滞后效应和累积效应;1—5月及春、冬季NDVI对SPEI指数的响应较强,7—9月和夏季NDVI与SPEI指数呈负相关。(4)2个典型干旱年中,NDVI负距平像元主要分布在1000~1750 m的3个海拔梯度上,受干旱影响较小的植被集中于NDVI≥0.6的高植被区。

Based on the MODIS/NDVI normalized vegetation index time series from 2001 to 2013, a variety of indexes including variation coefficient and one-way linear trend change were used to extract the NDVI change information in different elevation areas based on the SPEI index and DEM classification data, and analyzed the spatial and temporal variation of NDVI at altitudes in the southwestern Yunnan region and its response to changes in dry and wet weather at different time scales. The results showed that ： （ 1 ） The fluctuations of seasonal and annual NDVI at different altitudes in the study area were small, and there was a certain difference in the change trend of NDVI with elevation gradients, but the overall upward trend was obvious. Four seasons and annual NDVI improvement areas accounted for 46.9% -74.0% of the study area. （2）The percentage change of NDVI for each elevation gradient was mostly concentrated in the - 10 to 10% range; with 1500 m as threshold, the area where the percentage of change exceeded （less than） 10% （ -10% ） focused on three altitude gradients with higher than （less than） 1500 m. （3）NDVI at each elevation was most closely related to the 2-8 month time scale SPEI index, indicating that the NDVI in the study area had a hysteresis effect and a cumulative effect on the response to precipitation and evapotranspiration caused by precipitation; From January to May, and in spring as well as winter, the NDVI response to SPEI was stronger and the NDVI was negatively correlated with SPEI index in July-September and summer. （4） In the two typical drought years, NDVI negative anomaly pixels were mainly distributed in three altitude gradients from 1000 to 1750 m. Vegetation that was less affected by drought concentrated on NDVI ≥10.6 Vegetation area.