Data and code for: A greening surface over a falling water store in Rajasthan, India: an integrated satellite and well-census water account with implications for Sustainable Development Goal 6.4

Data and analysis code for the paper "A greening surface over a falling water store in Rajasthan, India: an integrated satellite and well-census water account with implications for Sustainable Development Goal 6.4" (De and Dey, revised manuscript GSD-D-26-01027R1, Groundwater for Sustainable Development). The deposit regenerates every number, table and figure in the paper from public archives; nothing in it is hand-entered. Version 2 (2 November 2026) adds the folder revision_2026_11, prepared in response to peer review. It corrects the storage record: the series used in version 1 is the JPL GRACE/GRACE-FO RL06 mascon solution (version 04), not the CSR-GFZ-JPL ensemble described there, and each monthly solution is now dated by its window mid-point. The corrected series supersedes RAJ_GRACE_monthly.csv and gws.json, which are kept unchanged for the record of version 1. The folder also holds the uncertainty budget (product, inter-centre spread, formal error, leakage, surface and canopy water), the monsoon storage-gain recharge test, the district rainfall and irrigation-source attribution, the first-order abstraction-cut sensitivity, the Google Earth Engine export script for the revision series, the figure builders and the manuscript build. Read revision_2026_11/README_revision.md first. Version 1 contents, retained: computed metric files in JSON; derived annual and monthly series in CSV (state and district NDVI, GRACE terrestrial water storage, GLDAS soil moisture, CMIP6 projections); per-district Theil-Sen greening rates; Google Earth Engine export scripts; figure builders and district-map helpers; and the version 1 manuscript source. Environment: Python 3.11 with numpy, pandas, scipy, statsmodels, pymannkendall, matplotlib, adjustText, python-docx, openpyxl and pillow. Satellite processing runs in Google Earth Engine. Trends are Theil-Sen slopes with Mann-Kendall significance and the Hamed-Rao variance correction. Licence: code MIT; data CC BY 4.0. Source archives retain their own licences.