Using the NDVI to identify variations in, and responses of, vegetation to climate change on the Tibetan Plateau from 1982 to 2012

We used the third generation Global Inventory Modeling and Mapping Studies normalized difference vegetation index (NDVI) and climate data (temperature and precipitation) to examine recent (1982-2012) spatial and temporal variations in vegetation, and relationships between climate and vegetation for both the growing period and for different seasons, on the Tibetan Plateau (TP). Across the whole plateau, trends calculated by linear regression showed that as temperature and precipitation increased, the growing season (May-September) NDVI values increased at rate of 0.002 decade−1 (p = 0.14) from 1982 to 2012. The ensemble empirical mode decomposition estimation method showed that the rates of increase in the NDVI gradually intensified until the end of the 1990s, and then decreased slightly in the following years. The autumn NDVI increased at a rate of 0.005 decade−1 (p = 0.04) and was a major contributor to the growing season NDVI. The NDVI and temperature were positively correlated at seasonal and monthly timescales during the growing season. The responses of vegetation growth to seasonal and monthly changes in precipitation, however, were complex. The NDVI trends showed obvious spatial heterogeneity and coincided well with regional and seasonal changes in climate. The growing season NDVI increased in 55% of the area of the TP. On a seasonal basis, the largest increase in the NDVI occurred in autumn and affected more than 61% of the TP, while the smallest increase in the NDVI occurred in spring, and affected over 41% of the area. Moreover, there were seasonal and spatial variations in the responses of different vegetation types to temperature and precipitation.