Abstract: Mount Fanjing has a unique subtropical isolated island ecosystem and rich biodiversity. It is of great significance to study the long-term temporal and spatial evolution and change trends of its vegetation coverage for the ecological management of the reserve. Based on the GEE cloud platform, this paper uses the Landsat SR dataset for cloud removal and fusion processing, and uses the maximum composite method to synthesize the annual NDVI dataset. The vegetation coverage was inverted by the pixel dichotomy model, and the temporal and spatial evolution and trend of FVC in Fanjingshan National Nature Reserve were analyzed pixel by pixel by the linear regression trend analysis method and Sen+Mann-Kendall trend analysis method, and compared the difference between the two trend analyses. The results show that: ①In the past 30 years, the vegetation coverage of Fanjingshan National Nature Reserve showed a “U”-shaped growth trend, which first decreased and then increased, with a growth rate of 3.46%, and the overall vegetation coverage improved significantly.②In terms of spatial distribution, the vegetation coverage of Fanjingshan National Nature Reserve is high in the middle and low in the surrounding areas. The area of vegetation improvement (accounting for 94.80%) is much larger than the area of vegetation degradation (accounting for 5.20%). ③The change trends of vegetation coverage in Fanjingshan National Nature Reserve obtained by the two trend analysis methods are basically similar, but the Sen+Mann-Kendall trend analysis method is 2.18% more sensitive to significant changes.

Key words: Google Earth Engine, fractional vegetation coverage, Sen+Mann-Kendall, univariate linear regression, Mount Fanjing