为研究黄土高原地区退耕还林（草）后，植被覆盖变化及其对水热条件的响应，利用1999-2013年SPOT VGT NDVI 1km/10d分辨率数据，采用最大合成法、一元线性回归法和偏相关分析法，系统分析了黄土高原地区NDVI（归一化植被指数）的时空分布及变化趋势，及其与气候因子的关系。结果表明：黄土高原1999-2013年年最大NDVI的平均值为0.31，NDVI较高的区域位于黄土高原南部，而西北部植被覆盖度较低；自1999年开始，黄土高原地区NDVI呈极显著（P < 0.01）增加趋势，年最大NDVI的变化斜率为0.0099；不同季节（春、夏、秋、冬）和生长季的植被状况均呈现良性发展趋势；1998-2013年间，黄土高原地区气候呈现不显著的“冷湿化”特征；NDVI年际（及生长季和季节）变化与降雨和温度的相关性不显著，而在月时间尺度上，呈显著的相关性，并且月NDVI与当月降雨量的相关性要强于与当月温度的相关性；植被生长对温度的响应存在一个月的滞后期，而对降雨的响应无滞后效应。
The Chinese Loess Plateau is one of the best-known areas in the world. It is located in the middle reaches of the Yellow River basin, and experiences the heaviest soil erosion in the world. In 1999, the Chinese government initiated the Grain-for-Green Program (GGP) for ecological restoration and soil erosion control. The main objective of the GGP was to rehabilitate forests and grasslands in an effort to control soil erosion. The project suggested converting all croplands with slopes of greater than 15° to grassland or forest. By the end of 2003, 79000 km2 of cropland had been returned to forest or grassland. Thus, it is necessary to assess the trend in land cover change following the implementation of the GGP in the Loess Plateau. The objectives of this study were 1) to investigate the spatio-temporal patterns of vegetation cover in Loess Plateau based on the 1999-2013, 1 km/10 d resolution SPOT VEGETATION Normalized Difference Vegetation Index (NDVI) data and 2) to investigate their responses to climatic factors. The NDVI data were provided by the Flemish Institute for Technological Research (VITO). The datasets can be downloaded from http://cdc.cma.gov.cn. The climatic factors were precipitation and temperature in this study. The precipitation and temperature data were downloaded from the China Meteorological Data Sharing Service System (http://www.escience.gov.cn/metdata/page/index.html). The maximum-value composites, linear regression method, and partial correlation analysis were used to investigate the spatial distribution and changes in the NDVI and the relationships between the NDVI and precipitation and temperature in the Loess Plateau. The results showed that the average NDVI value for the Loess Plateau was 0.31 from 1999 to 2013. The higher NDVI values were mainly distributed in the southeast part of the Loess Plateau. In contrast, the northwest part of the Loess Plateau had lower NDVIs. The NDVI increased significantly from 1999 to 2013 (P < 0.01), and the slope of the trend line for NDVI was 0.0099. This indicated that the implementation of GGP improved vegetation coverage in the Loess Plateau and the vegetation construction achievements were quite noticeable. The seasonal NDVIs (spring, summer, autumn, and winter) and the growing season NDVI also showed significantly increased. Summer has the fastest NDVI growth rate (0.0087), followed by autumn (0.0081), spring (0.0064), and winter (0.0018). The growing season NDVI had the same growth rate as summer. From 1998 to 2013, precipitation and temperature in the Loess Plateau increased and decreased, respectively, but neither change was significant. The partial correlation analysis showed that NDVI correlated significantly with temperature and precipitation at the monthly scale. However, no significant correlations were noted between annual (or growing season and seasonal) NDVI and annual (or growing season and seasonal) temperature and precipitation. Temperature change had a lagging effect on vegetation growth, but precipitation change did not.