30 m Normalized Difference Vegetation Index Maps of Pure Pixels over China for Estimation of Fractional Vegetation Cover (2014, 2018, 2022)

Using multi-angle remote sensing data, we generated 30-m maps for the normalized difference vegetation index (NDVI) of fully-covered vegetation (Vv) and bare soils (Vs) across China in 2014, 2018 and 2022. These pixel-wise Vv and Vs maps can be integrated with the vegetation index (VI)-based model to facilitate the accurate and rapid estimation of fractional vegetation cover (FVC) across various spatial resolutions and large scales. The products were produced using a multi-angle algorithm (MultiVI), which effectively addressed the spatial variability inherent in Vv and Vs and enhanced the accuracy of FVC estimations in comparison to traditional statistical methods. The estimated FVC demonstrated a root mean square deviation (RMSD) of approximately 0.1 when evaluated against field-measured FVC across different experimental sites.本研究利用多角度遥感数据，生成了2014年、2018年及2022年中国全域全覆被植被（Vv）与裸土（Vs）的归一化差值植被指数（Normalized Difference Vegetation Index, NDVI）30米分辨率影像。上述逐像素Vv与Vs影像可与基于植被指数（Vegetation Index, VI）的模型相结合，以实现在不同空间分辨率与大尺度下对植被覆盖度（Fractional Vegetation Cover, FVC）的精准快速估算。本数据集采用多角度算法（MultiVI）生产，有效解决了Vv与Vs固有的空间变异性问题，相较传统统计方法显著提升了FVC估算精度。经多试验站点的野外实测FVC验证，本次估算得到的FVC的均方根偏差（Root Mean Square Deviation, RMSD）约为0.1。