Abstract:Land surface phenology (LSP) is a sensitive indicator of environmental change in ecosystems. The LSP in alpine wetlands can reflect the changing patterns of naturally seasonal phenomena in alpine wetland ecosystems and their response and adaptation to environmental changes, which has become one of the research hotspots in the field of global change. To explore the relationship between spatio-temporal changes of phenology and ecological patterns and processes of wetland landscapes in Zoige alpine wetlands, based on the GIMMS3g NDVI and MODIS NDVI datasets and 7-phase Landsat TM/OLI satellite remote sensing data from 1990 to 2020, the surface phenological parameters were extracted by using the threshold method and the land use data were interpreted by using object-oriented classification method. The Positive and Negative Transformation Index (PNTI) model was used to portray the dynamic change process of alpine wetlands. The results showed that (1) the wetland area in the study area from 1990 to 2020 presented the characteristics of first decreasing and then stabilizing. According to the evolution path and intensity of land use types, it was divided into dynamic equilibrium area, degradation evolution area, and restoration evolution area, accounting for 56.84%, 28.14%, and 15.02%, respectively. (2) From the spatial distribution characteristics, the start of the growing season (SOS) and the peak of the growing season (POS) were gradually delayed from south to north, while the end of the growing season (EOS) was gradually delayed from the middle to the periphery. The length of the growing season (LOS) was short in the middle and long in the periphery, short in the northwest and long in the southeast. The SOS ranged mainly from 96 to 149 Julian days, EOS ranged mainly from 249 to 284 Julian days, POS ranged mainly from 179 to 209 Julian days, and LOS was concentrated within 125 to 173 days. In terms of interannual variation, the SOS and POS delayed first and then advanced, while EOS advanced first and then delayed. The LOS shortened first and then extended, with a shortening rate of 10.8 d/10a from 1990 to 2005 and an extension rate of 2.4 d/10a from 2000 to 2020. (3) The SOS and EOS were sensitive to the response of evolutionary processes of landscape patterns in alpine wetlands with different directions of evolutionary paths (i.e., restoration and degradation), while they were not sensitive to the response of evolutionary paths with the same direction but different intensity levels. The change rates of the SOS and EOS in the degradation evolution area were overall higher than those of SOS and EOS in the restoration evolution area and dynamic equilibrium area. The advance of SOS and delay of EOS can be used as indicators of alpine wetland degradation and has an ecological early warning function. Alpine wetland landscape evolution had no significant effect on LOS and POS at P=0.05 level.