Data on Leaf Area Index After Filling in Missing Values

This study used the MODIS Leaf Area Index (LAI) product to characterize vegetation growth and constructed a long-term time series using 8-day-resolution LAI data from 2001 to 2024. LAI can characterize the leaf area and structural features of the vegetation canopy. Compared to NDVI, it is less prone to significant spectral saturation in areas with medium to high vegetation cover and is therefore more suitable for analyzing long-term vegetation changes and temporal stability. However, due to factors such as surface reflectance characteristics, observation conditions, and inversion quality, the raw LAI product contains a large number of missing values in areas of low vegetation cover in the northwestern inland region; directly removing these values may result in a loss of spatial information in this region. To improve the spatial completeness of LAI data, this study retained the original valid LAI data and employed a random forest model, using MODIS NDVI, annual precipitation, annual mean temperature, and DEM as auxiliary variables, to perform local reconstruction of missing pixels in low-LAI areas of the northwestern inland region. After undergoing projection transformation, spatial consistency checks, and scale normalization, the reconstructed LAI data ultimately form a dataset that can be used for subsequent analyses of vegetation change and temporal stability.