Keyhole Reconnaissance Imagery Spatio-Temporal coverage（KRIST）

This project provides the <i>Keyhole Imagery Global Coverage Dataset (1960–1984)</i>, a globally consistent framework for evaluating the spatio-temporal availability of declassified U.S. Keyhole reconnaissance imagery. The dataset is built on a global equal-area sinusoidal grid of ~25 million points with ~5.6 km spacing and integrates multiple resolution classes (C1: meter-level, C2: five-meter, C3: ten-meter). Standardized indicators are provided, including cover count, number of unique observation dates, first/last observation year, temporal span, peak observation year, and multi-period coverage classification.The project includes fololwing components:<b>global_keyhole.zip</b>: Including two gpkg files as global_keyhole_free.gpkg and global_keyhole_full.gpkg. Each contains three layers as Declass1, Declass2, and Declass3 downloaded from USGS explore.<b>classification.zip</b>: Including two gpkg files as classification_free.gpkg and classification_full.gpkg. Each contains four layers as C1 (meter-level images), C2 (three-meter images), C3 (ten-meter images), as well as the combined layer C0 layer.<b>global_grid.zip</b>: a global equal-area sinusoidal grid of ~25 million points with ~5.6 km spacing.<b>nc+h5.zip</b>: The netCDF file stores numeric variables (e.g., <i>cover_count, unique_date_count</i>). The HDF5 file stores the complete list of metadata URLs for each grid point at that resolution, which could enable users to quickly check cloud conditions through thumbnails and locate download links within the selected area of interest. Both includes eight files as free/full_C{0–3}.<b>Code</b>: Python scripts implementing the workflow as merging shapefiles to gpkg, imagery classification, global grid generation, and point-based property calculation, as well as, querying the complete list of metadata URLs on a specific grid point.The dataset supports reproducible Earth system research, historical land use reconstruction, and environmental change studies. Users are encouraged to cite the associated publication and dataset DOI when using these resources.