Extension of GIMMS3g on MODIS NDVI for long-term global vegetation trend analysis

作者：Kai Di, Wenjing Chen, Haiyang Pang, Zaid Khan, Weirong Zhang, Yue Jiao, Zehao Fan, Chuan Jin · 发表于：International Journal of Remote Sensing · 年份：2025 · DOI：10.1080/01431161.2025.2561131 · 被引用次数：2 · 研究领域：Remote Sensing in Agriculture、Remote Sensing and Land Use、Environmental and Agricultural Sciences

Long-term trends in global vegetation have become a hot topic in recent years. However, discrepancies among different remote sensing products, due to sensor differences, often lead to variations in the observed magnitude and trends of vegetation changes. To generate ultra-long-term NDVI products, this study integrated GIMMS3g NDVI and MODIS NDVI for overlapping years. Using an 8-day compositing method, MOD13A1 and MYD13A1 were combined to produce 8-day MODIS NDVI data. These MODIS NDVI datasets were further interpolated into daily-scale data using three methods: spline, linear and moving average. The newly generated high-quality daily-scale data corresponding to GIMMS NDVI observation dates were then used as a baseline to correct GIMMS3g NDVI. Results indicated that the moving average method provided the best interpolation performance for daily-scale MODIS NDVI. While the trends of MODIS moving average NDVI were consistent with those of the original MODIS NDVI, the corrected GIMMS NDVI values were slightly lower, potentially due to differences in defining vegetation cover in high-latitude regions of the Northern Hemisphere between the two products. Overall, the corrected GIMMS NDVI showed consistent variations and trends with multiple fused NDVI datasets. The MODIS moving average NDVI also proved to be a reliable alternative to the original MODIS NDVI for analysing vegetation dynamics in middle and low latitude regions. This study provides a new dataset to support future anal...