Abstract: Based on 83 meteorological stations observations from 2000 to 2016 over the Loess Plateau, the standardized precipitation evapotranspiration index (SPEI) was calculated. The rotational empirical orthogonal function (REOF) was used to analyze the spatial and temporal distribution characteristics of drought over the Loess Plateau. Moreover, the normal difference vegetation index (NDVI) was also used to assess the impacts of drought on vegetation activity under different land use types and elevations. The results are as follows: (1) The Loess Plateau can be divided into four sub-regions: the southeast area, the northeast area, the southwest area and the northwest area. Except for the northwest area, the remaining sub-areas showed significant humidification. (2) The NDVI values from April to October during growing season increased significantly, and the NDVI positively correlated with SPEI, and the correlation coefficient from June to September was significant. (3) The land use types in the study area were divided into four categories: cultivated land, grass land, forest land and unused land. During growing season, the percentage of area of positive correlation between NDVI and SPEI was the largest in grass land (62.54%), and it was second in cultivated land (52.97%) and the samllest in forest land. (4) The percentage of area of positive correlation between NDVI and SPEI increased firstly and then decreased with rise of elevation. The values reached the maximum when the elevation ranged from 1500 to 2000 meters.

Key words:  Loess Plateau, vegetation change, meteorological drought, standardized precipitation evapotranspiration index (SPEI), normal difference vegetation index (NDVI)

摘要： 基于黄土高原83个地面气象站2000—2016年观测数据，计算标准化降水蒸散发指数（SPEI，standardized precipitation evapotranspiration index），分析黄土高原干旱时空变化特征，并结合对应时段的归一化植被指数（NDVI, normal difference vegetation index），研究不同土地利用类型及不同海拔高度植被对气象干旱的响应。结果表明：（1）通过旋转经验正交函数（REOF，rotational empirical orthogonal function）可将黄土高原划分为东南区、东北区、西南区和西北区；除西北区外均呈明显湿润化特征。（2）生长季各月黄土高原NDVI均呈显著增大趋势，NDVI均与SPEI呈正相关关系，其中两者在6—9月相关性最强（P<0.05）。（3）对于研究区内4种土地利用类型（耕地、草灌、森林和未利用土地），生长季内NDVI与SPEI呈显著正相关的面积百分比在草灌类型最大（62.54%），耕地类型次之（52.97%），而森林类型最小（25.62%）。（4）6—9月NDVI与SPEI呈显著正相关的面积百分比随海拔的升高先增加后减小，最大值出现在海拔1500~2000 m。

关键词: 关键词：黄土高原, 植被变化, 气象干旱, SPEI, NDVI