Data accompanying the publication: "Sea-level fingerprints separation from altimetry data using deep learning"

This repository provides data and for separating GRD (Gravitation–Rotation–Deformation) sea-level fingerprint contributions from multiple land-ice sources using a multi-decoder U-Net.It includes: (1) synthetic SSH change-rate NetCDF files named syntheticSSHrate_[steric product]_1deg_10d.nc (units: mm per year) on a 1-degree grid (66S to 66N, lon 0 to 359E) at 10-day sampling from 1993-01-01 to 2017-11-01. Each file contains the synthetic field (tssh), the steric component (Steric), the total GRD component (GRD), and nine regional GRD components (GRD_gris plus eight glacier regions). Masked grid cells are set to zero. (2) normalized GRD fingerprint patterns in Normalized_GRD_1deg.nc. (3) the preprocessing mask in mask_1deg.nc. (4) CSV model outputs with DNN-derived regional mass change rate time series (10-day sampling) from satellite altimetry, stored in Mass Change Rate.csv (units: mm yr-1, GMSL equivalent).