TC-GROUND preview: ground-radar dual-Doppler analyses of Debby (2024), Sally (2020) and Isaias (2020)

TC-GROUND (Tropical Cyclone Ground-radar Observations, Unified multi-agency Near-core Doppler analyses) is a dataset of storm-centred, three-dimensional reflectivity and dual-Doppler wind analyses of tropical cyclones observed by coastal ground-based radars (NOAA WSR-88D and Météo-France). The full dataset (38 storms, 2008–2025, 28 radars, about 5,200 analyses) is described in a manuscript in preparation (Alvey et al., Part I). This record is an early preview with three WSR-88D cases: Debby (2024, 161 analyses), Sally (2020, 344 analyses) and Isaias (2020, 325 analyses). More storms will be added as new versions of this record. Values may change before the v1.0 release; please contact the authors before using these files in a publication. Each storm is one netCDF-4 file on a storm-centred grid (2-km spacing, ±324 km, 0.5–12 km height every 0.5 km). Fields include the reflectivity mosaic, earth-relative U and V and vertical velocity W from the FRACTL dual-Doppler retrieval, beam-geometry quality fields and coverage masks, vortex centres at 2–8 km, storm motion, best-track intensity and SHIPS environmental predictors. Retained winds agree with coincident NOAA P-3 tail Doppler radar analyses (TC-RADAR v4.0) to a bias of −0.75 m/s and RMSE of 3.72 m/s over eight WSR-88D storms. README.md describes the grid, variables, recommended masking and the aircraft comparison. tcground_quicklook.py is an example reader that makes the included quicklook figures. MANIFEST.json lists checksums; croissant.json gives Croissant metadata.