Abstract:

Aim Estimating regional variation in vegetation phenology from time-series remote sensing data is important in global climate change studies. However, there are few studies on vegetation phenology for the Qinghai-Tibet Plateau and most are based on field records of stations.

Methods We utilized the dynamic threshold method to explore vegetation phenological metrics (greenup date, length of season and senescence date) of typical grassland in the Northern Tibetan Plateau. We used time-series TERRA/MODIS EVI data for 2001-2010 reconstructed by the asymmetric Gaussian function fitting method to analyze spatial pattern and differentiation of vegetation phenology and its inter-annual variation and to examine the relationship between phenological variation and climate changes.

Important findings The spatial pattern of date of vegetation greenup was embodied by transition from southeast to northwest and vertical zonation in the mountainous topography of the southeast. The vegetation greenup date in approximately sixty percent of the northern Tibetan Plateau had advanced, especially in high mountains. Inter- annual variation of vegetation senescence date was not obvious, and most of the region had natural inter-annual fluctuations. The variation of growing season length is influenced by greenup and senescence dates, but was chiefly affected by advanced greenup date lengthening the growing season. Among the four different climatic zones in the study area, the mountain and valley Nagqu sub-arctic and sub-humid zone and the southern Qinghai sub-arctic and semi-arid zone had the most apparent advanced greenup date and prolonged growing season. Based on measured data from weather stations, increased temperature appears to be a critical factor contributing to earlier greenup and prolonged growing season; however, the relationship between precipitation fluctuations and phenological variation was unclear.

Key words: dynamic threshold, MODIS, Northern Tibetan Plateau, phenology, spatio-temporal change