Long-term vegetation phenology changes and response to multi-scale meteorological drought on the Loess Plateau, China

Investigating the response of vegetation phenology to meteorological drought can provide insight into the feedback mechanisms of ecosystems to environmental and climate change. These key phenological parameters of vegetation are extracted on the China's Loess Plateau (LP) based on TimeSat using MODIS remote sensing data, and their trends are estimated using Sen's slope. Then, response of key physical parameters of vegetation phenology to multi-scale drought was investigated using Standardized Precipitation Evapotranspiration Index (SPEI). The results show that: 1) LP vegetation Start of Season (SOS) and Peak of Season (POS) are delayed from south (low elevation) to north (high elevation). End of Season (EOS) is delayed from northeast to southwest. Length of Season (LOS) show decrease trend with altitude and latitude increasing. Besides, temporal variation of different vegetation phenological periods is highest in farmland and is lowest in the forest. 2) Vegetation SOS and POS are advanced in the eastern part of LP and delayed in the western part. Meanwhile, EOS and LOS are advanced in the southern part of LP. 3) The POS and EOS of LP vegetation responded more significantly to drought. The western Longzhong Plateau of LP was susceptible to the negative effects of preseason drought, and the eastern Northern Shaanxi Plateau and Taihang Mountains were susceptible to the positive effects. Moreover, monthly-and seasonal-scale droughts have significantly effects on vegetation phenology than annual-scale drought. 4) Compared to forest and grassland, the vegetation phenology of farmland is more sensitive to monthly-and seasonal-scale droughts. However, the response of forest to annual-scale drought is more signifi-cant. The results of the study could provide scientific basis for crop yield stabilization and agricultural disaster warning in the LP. Keywords: Vegetation phenology; TimeSat; Drought response; Loess Plateau.