Spatial Features of the Coupling between Spring NDVI and Temperature over Northern Hemisphere
北半球春季植被NDVI对温度变化响应的区域差异

There is increasing attention focused on the variations in global vegetation condition due to its importance in the global carbon cycle. The vegetation variability arises from many causes. It is well known that climate drives ecosystems on both local and global scales. How and to what extent the vegetation responds to the large-scale climate change is a challenging subject in global change study. In the northern mid- to high-latitudes there experienced dramatic temperature variations, as well as the significant changes in vegetation conditions during the last about two decades. However, both temperature and vegetation variations are not uniform in geographical distribution. In the present study, the authors analyze the spatial features in Pathfinder AVHRR-NDVI/temperature relationship over northern hemisphere in spring for the period from 1982 to 2000. A singular value decomposition analysis is utilized to the covariance matrix of NDVI and temperature. Most of the squared covariance is captured by the first several paired-modes. The first seven modes account for 91.6%. This implies that the temperature is a very important factor influencing vegetation activity. The NDVI changes in response to temperature fluctuations on the interannual time scale show well-defined large-scale and consistent patterns. The first paired-modes, which explains 42.64% of the squared-covariance, indicate the strongest coupling between vegetation and temperature appears in western Siberia. The large-scale atmospheric system, Eurasian pattern (EU), plays a dominant role for that. The relationship between NDVI/temperature and nine large-scale atmospheric circulation systems is analyzed. Results show that much of the NDVI/temperature covariance can be attributed to the fluctuations in these circulation indices. Averaging over the mid- to high-latitude northern hemisphere, 55.6% of the satellite-sensed NDVI variance is explained. The nine climate indices can account for a large portion of the long-term trends in NDVI too, particularly in the northwestern North America, southeastern North America, most of Europe, Siberia, and East Asia. This implies that the regional response of vegetation to climate fluctuations under future climate change scenarios would differ from region to region. Some areas related to the important circulation systems would experience higher sensitivity and predominant changes than other regions.