Monthly 30-m Surface Water Dataset of the Tibetan Plateau from 2000 to 2021

This repository contains the relevant data for the paper Landsat-Derived Gap-Free Monthly 30-m Dataset (2000–2021) Unraveling Intra-Annual Surface Water Dynamics on the Tibetan Plateau. The Tibetan Plateau Monthly Water History (TP MWH) dataset (2000–2021) is archived in the file “TP2000_2021_MonthlyWaterHistory_afterSGF.zip”. It contains 264 standard 30 m-resolution GeoTIFF images with a unified EPSG:4326 coordinate system, classifying each pixel as water (value = 3) or not-water (value = 1), and is compatible with mainstream GIS and remote sensing software. The stratified random sampling validation data are stored in “TP_validation_samples.csv”, including 12,141 representative sample records. Each entry covers precise eight-decimal geographic coordinates, Google Earth Pro high-resolution imagery observation dates, manually interpreted ground-truth land cover types, and pixel classification results and validation records for the TP MWH, JRC MWH, and GLAD datasets, supporting comprehensive and reproducible quantitative accuracy assessment. File name Detailed definition Description TP MWH Dataset Output Files: TP2000_2021_MonthlyWater History_afterSGF.zip File composition The core dataset comprises 264 standardized monthly surface water distribution maps covering the entire Tibetan Plateau from January 2000 to December 2021. File format GeoTIFF raster format, compatible with mainstream remote sensing and GIS software platforms. Spatial reference WGS84 geographic coordinate system (EPSG:4326), with a spatial resolution of 30 m, consistent with the original Landsat observation data. Pixel classification rules Binary classification system, where pixel value = 3 represents surface water body, and pixel value = 1 represents not-water; no invalid gap pixels exist in the final reconstructed dataset. Naming convention TP_YYYY_MMafterSGF.tif, where YYYY denotes the 4-digit year and MM denotes the 2-digit month, enabling straightforward temporal indexing and batch processing. Validation Auxiliary File: TP_validation_samples.csv This CSV file records complete attribute information for all 12,141 stratified random validation samples used in this study, supporting reproducible quantitative accuracy evaluation and cross-dataset comparison. All geographic coordinates are recorded to 8 decimal places to eliminate spatial matching errors. samples Unique numbering sequence identifier for each validation sample point. Continuous integer from 1 to 12141. longitude Longitude coordinate of the sample point in the WGS84 geographic coordinate system. Numeric value recorded to 8 decimal places. latitude Latitude coordinate of the sample point in the WGS84 geographic coordinate system. Numeric value recorded to 8 decimal places. GE_observation_date Acquisition date of the high-resolution historical imagery from Google Earth Pro for the sample point, retrieved via the built-in time-slider function to match the dataset’s temporal coverage. Date format: YYYY/MM/DD actual_classification Ground truth classification of the sample point, determined through manual visual interpretation of Google Earth Pro high-resolution historical imagery. 1 = not-water; 3 = water. land_cover_type Refined 9-category land cover classification of the sample point, supporting stratified accuracy evaluation for typical plateau surface features. 9 = visualization plotting placeholder; 8 = Frozen Water Surface; 7 = Snow-Covered Water Surface; 6 = Cloud and Terrain Shadow Occlusion; 5 = River Water; 4 = Saline Shallow Water and Saline Land; 3 = Natural Open Water; 2 = Vegetation; 1 = Terrestrial Land TPMWH_classification Classification result of the sample point extracted from the corresponding year-month TP MWH dataset image. 1 = not-water; 3 = water. TPMWH_validation_results Binary validation result indicating the matching status between TP MWH classification and ground truth actual_classification. 1 = correct prediction; 0 = incorrect prediction. JRC_classification Classification result of the sample point extracted from the corresponding year-month original JRC MWH dataset image. 1 = not-water; 3 = water; 0 = invalid gap pixel. JRC_validation_results Binary validation result indicating the matching status between JRC MWH classification and ground truth actual_classification 1 = correct prediction; 0 = incorrect prediction. GLAD_classification Classification result of the sample point extracted from the corresponding year-month GLAD Global Surface Water Dynamics dataset image. 1 = not-water; 3 = water; 255 = invalid gap pixel. GLAD_validation_results Binary validation result indicating the matching status between GLAD dataset classification and ground truth actual_classification. 1 = correct prediction; 0 = incorrect prediction.