Changes in vegetation phenology in response to climate changes in Northern hemisphere between 1982 and 2006

The research on response and adaptability of vegetation ecosystem to global changes is the scientific frontier of today's global change and ecological researches. While remote sensing phenology is still in development, it is regarded as a key indictor to understanding land-surface processes over large area. In this paper, the Northern Eurasia region is used as study regions. Based on GIMSS AVHRR NDVI composite image data, the key parameters of vegetation phenology, such as start of season (SOS), end of season(EOS), duration of season (DOS) and peak of season (POS) were extracted. The vegetation's response to climate change in different regions was analyzed by using these phenological parameters. Firstly, GIMMS AVHRR NDVI time series images from 1982 to 2006 were smoothed with Hants filter algorithm. Based on the dynamic threshold analysis method, SOS, EOS, DOS and POS were estimated. Secondly, Trends of phenology metrics were derived for each pixel from 1982 to 2006 by calculating the slope of the temporal variation of these phenology metrics in these years. The spatial and temporal variations of vegetation phenology in different regions were analyzed. The results show that: earlier SOS and later EOS are largely spatially distributed in the Northern Hemisphere, which confirm the global warming trend in the Northern Eurasia region. The most significant trend is a longitudinal band of later EOS extending from west Europe to east Russian. The DOS metric variation image shows also a high significant prolongation trend in high latitude areas in the Northern Eurasia region, as well as EOS. A trend toward earlier POS was apparent both regions of Spain, France, Turkey and the southeastern regions of Kazakhstan and North China. No significant trends, including SOS, EOS, DOS and POS, are observed in the Tibetan Plateau.