广西南流江流域1986—2020年植被覆盖度时空变化及预测

为探究我国南方丘陵山地流域植被覆盖度长时序变化情况，本研究基于1986—2020年Landsat系列影像数据，运用Google Earth Engine平台计算植被覆盖度FVC，利用Theil-Sen median趋势分析和Mann-Kendall 以及Hurst指数方法分析南流江流域植被覆盖度时空变化趋势特征。结果表明：南流江流域植被覆盖度高，植被覆盖度变化呈现明显的上升趋势。1986—2020年间，流域植被改善面积（58.23%）远远大于植被退化面积（8.29%），植被改善区域在流域中、下游表现最为显著，退化区域自流域上游而下依次减少。未来南流江植被覆盖度变化将呈现波动变化趋势，其中持续性改善面积占14.83%，持续性不变面积占12.25%，持续性退化面积占3.48%，余下69.44%面积为波动变化区域，流域植被覆盖变化与造林绿化工程及城市扩张发展息息相关。Google Earth Engine平台在遥感影像处理中有明显优势，它能在长时序、大尺度植被监测研究中发挥重要作用。

Dynamic Changes and Predictions of Vegetation Voverage in the Nanliu River Basin in Guangxi from 1986 to 2020

In order to explore the long-term changes of vegetation coverage in the hilly and mountainous watersheds in southern my country, this study is based on the Landsat series of image data from 1986 to 2020, and the Google Earth Engine platform was used to calculate the vegetation coverage FVC, also Theil-Sen median trend analysis and Mann-Kendall and The Hurst index method was used to analyze the temporal and spatial trend characteristics of vegetation coverage in the Nanliu River Basin. The results showed that the vegetation coverage in the Nanliu River Basin was high, and the change of vegetation coverage showed an obvious upward trend. From 1986 to 2020, the area of vegetation improvement in the basin (58.23%) was much larger than the area of vegetation degradation (8.29%). The area of vegetation improvement was most significant in the middle and lower reaches of the basin, and the degraded areas decreased sequentially from the upper reaches of the basin to the lower. It is concluded that the vegetation coverage changes of the Nanliu River will show a fluctuating trend in the near future, of which 14.83% are continuously improved, 12.25% are persistently unchanged, 3.48% are persistently degraded, and the remaining 69.44% are fluctuating areas. Vegetation coverage changes in watersheds are closely related to afforestation projects and urban expansion and development. The Google Earth Engine platform has obvious advantages in remote sensing image processing, and it can play an important role in long-term, large-scale vegetation monitoring research.