MST NDSI Collection: A Cloud-free MODIS NDSI Dataset (2001–2022) for Tibetan Plateau Generated by a LightGBM-Based Method Using Multivariate Features

1. Cloud-free MODIS normalized difference snow index (NDSI) dataset for Tibetan Plateau from 2001 to 2022 is generated. 2. This dataset is derived from daily 500-m MODIS NDSI datasets (MOD10A1 and MYD10A1). 3. This dataset is provided using a WGS_1984_UTM_45N projection, with the data format of TIFF images. 4. The NDSI value ranges from -10,000-10,000, corresponding to the raw NDSI band of MOD10A1 and MYD10A1. The fill value is set to -32768. 5. Due to the large amount of data, the data for each year has been divided into several split archives (YYYY.partX.rar). 6. Each RAR contains NDSI data for one year (January to December). After uncompressing into the TIFF format, the files are named as "NDSI_Daily_YYYY_mm_dd.tif" for NDSI data and "NDSI_Daily_YYYY_mm_dd.tif". 7. The accuracy of this collection has been well validated by simulation experiments. 8. This version includes data from 2005 to 2010 (https://doi.org/10.5281/zenodo.14038846). Version 1 includes data from 2001 to 2004 (https://doi.org/10.5281/zenodo.14027672). Version3 includes data from 2011 to 2016 (https://doi.org/10.5281/zenodo.14089925). Version 4 includes data from 2017 to 2022 (https://doi.org/10.5281/zenodo.14177009). 9.For the specific processing steps and further details, please refer to our published paper: A novel approach for cloud-free MODIS NDSI reconstruction on the Tibetan Plateau combining spatiotemporal cube and environmental features, IEEE Transactions on Geoscience and Remote Sensing, doi:10.1109/TGRS.2025.3542095.