Effects of data temporal resolution on phenology extractions from the alpine grasslands of the Tibetan Plateau

The vegetation phenology is a commonly used indicator signaling vegetation responses to global changes. Monitoring vegetation phenology at a regional and global scale needs to rely to remote sensing data, for which multiple sources of datasets and extraction methods have been developed. To be efficient, remote sensing data with coarse temporal resolution is conventionally preferred in exploring vegetation phenology patterns at the continental or global scale. As fine temporal resolution data is increasingly available, effects of their temporal resolution on our analysis are still elusive. In this study, we applied several commonly utilized vegetation phenology extraction methods on two different temporal resolution MODIS NDVI data and compared their performances on the Tibetan Plateau (TP). The results showed there were certain discrepancies in the magnitude, trends and spatial patterns of extracted phenological parameters between the two datasets and among the different extraction methods. Generally, the phenological parameters derived from fine temporal resolution NDVI (MOD09A1) displayed later start of growing season (SOS), earlier end of growing season (EOS), shorter length of growing season (LOS), and were more accurate in capturing vegetation SOS compared with the coarse temporal resolution NDVI data (MOD13A2). The double logistic method can minimize the differences of extracted SOS or EOS between the two temporal resolution datasets. The findings of this study would improve accuracies of applying remote sensing data on monitoring vegetation dynamics and advance our understanding on vegetation responses to climate changes.