AquaStateWater & climate intelligence
Updated 13:50 UTC · 249 countries
↑ ↓ moveEnter openEsc close

Source ledger

Every value has a timestamp.

Keyless public sources, honest labels, and a readable failure path. Live means a fetch succeeded; cached means it is the last good copy; modeled and proxy are named as such.

Live source status

live NASA EONET eonet
2026-11-1913:50 UTC 200 records fresh
live NASA POWER / MERRA-2 nasa_power
2026-11-1913:50 UTC 32121 records fresh
live Open-Meteo GloFAS open_meteo_flood
2026-11-1913:50 UTC 18924 records fresh
error Open-Meteo forecast open_meteo_forecast
no successful fetch 0 records due

https://api.open-meteo.com/v1/forecast: 429 Too Many Requests

live World Bank world_bank
2026-11-1914:16 UTC 5 records fresh

Current snapshot wcri-20261119135058 · 2026-11-19 13:50:58.563391+00:00 UTC · 249 scored countries · 200 active events.

What is fetched

SourceUsed forKindCadence
NASA POWER / MERRA-2 monthly10-year rainfall, temperature, soil wetness, PET inputslive24 h
NASA POWER / MERRA-2 dailyrecent rain and soil at a hydrologic centroidlive6 h
Open-Meteo forecast16-day rain, ET0, temperature, humidity, wind; 92-day fire-weather baselinelive30 min
Open-Meteo GloFAS flood60-day past + 16-day forecast river dischargelive60 min
NASA EONETopen drought, flood, wildfire, dust and temperature eventslive30 min
World Bankpopulation, renewable freshwater per capita, withdrawal, forest and arable landreference7 days
Wikidata + Wikipedia sitelinksglobal notable-dam inventory (names, coordinates, country, height, power)reference30 days
Natural Earth 50mcountry geometry for the mapreferencestatic
Mass-balance / screening modelreservoir storage, soil and index calculationsmodeledrecomputed

Refresh

The world layer is designed to be refreshed without a page being open. The UI also has a Refresh live data button that starts a single-flight background refresh.

python manage.py bootstrap_world              # fetch + recompute + snapshot
python manage.py bootstrap_world --offline    # use cached/source files only
python manage.py bootstrap_world --limit 10   # smoke run
python manage.py load_dam_atlas --reset       # rebuild the dam catalogue
python manage.py dedupe_dams                  # merge duplicate monitored rows

For unattended operation, run python manage.py refresh_world --loop --interval 900 in a second terminal. The page polls /api/world/meta/ every 60 seconds and updates the map when the snapshot version changes.

Honesty rules

  • Live is a successful fetch with a stored timestamp, not a label of intent.
  • Cached is the last successful copy; the UI never calls it live.
  • Modeled storage is a mass-balance/reference estimate, not SCADA telemetry.
  • Proxy groundwater uses root-zone wetness when no aquifer inventory exists.
  • Missing data is neutral with low confidence; it is never a hidden zero or a fake 50 in the score.
  • Scenario values are simulated, tagged, and never overwrite an observed snapshot.
  • “Not for operational decisions” appears in the footer and in the API language.

Coverage

Countries with a profile249 / 249
Catalogued dams82,394
Monitored dams33
EONET events205

The dam atlas is built from notable dams with an English Wikipedia article and a Wikidata coordinate, supplemented by curated public inventories. It is a catalogue, not a complete census. Country geometry covers 236 polygons and 13 point markers; boundaries are Natural Earth, not legal.

Attribution: NASA POWER / MERRA-2, Open-Meteo and GloFAS, NASA EONET, World Bank, Natural Earth, Wikidata and Wikipedia contributors, OpenStreetMap contributors, IBM Plex (SIL OFL), Leaflet (BSD-2-Clause).