Abstract: Vegetation phenology is the best indicator in monitoring the influence of climate on vegetation. More and more remote-sensing image time series are now available to monitor vegetation phenology dynamics over large areas. Based on the Asymmetric Gaussians Filtering and Dynamic Threshold Method，the spatial-temporal pattern variations of alpine grassland phenology in source region of the Yellow River were analyzed using EOS-MODIS [NDVI] data from 2000 to 2012. The results showed that alpine grassland phenology in the study area generally started from late April to early May，and ended in late September to early October. The spatial patterns of alpine grassland phenology were significantly different. Start of growing season（SOG） of alpine grassland phenology in the southeast started earlier than the northwest，and end of growing season （EOG） ended later. Therefore，long of growing season（LOG） in the southeast was longer than those in the northwest influenced by variations of thermal and hydrological conditions. Alpine grassland phenology was closely related with altitude. SOG delayed，EOG advanced，LOG prolonged with altitude increasing．Inter-annual linear tendency of SOG，EOG，LOG in the region from 2000 to 2012 was -10.38 d/10 a（R=-0.65，p=0.017），0.07 d/10 a（R=0.01，p=0.95），10.46 d/10 a（R=0.72，p=0.005）. The trends of alpine grassland phenology had certain spatial discrepancy，which expressed that the LOG prolonged in northwest，southwest areas and shorten in central areas covering about 82.77% and 17.23% of total areas respectively．The spatial patterns of alpine grassland trends were altitude dependent. The magnitude of the trend in SOG and LOG increased in higher altitudes contrasted with lower altitudes，while EOG had no significant change. It was inferred that alpine grassland in higher altitudes was more sensitive to climate change. Increasing in spring and fall temperature mainly caused SOG advancing and LOG extending of alpine grassland in source region of the Yellow River. The phenology of alpine grassland extracted using EOS-MODIS [NDVI] datasets in this paper was generally concordant with many reported researches. This paper can provide a reference for the evaluation and protection of ecological environment in source region of the Yellow River.

Key words: the source region of the Yellow River, phenology, spatial-temporal features, altitude, influencing factors