There is no single global database of reservoir levels. Every country publishes its own, through its own agency, in its own units — and the gap between the best and the worst is enormous: some governments hand you a clean open API you can query from anywhere in the world, while others put the same numbers behind a weekly PDF, a geo-block, and a login. This is the directory we wish had existed when we started: for each of the eleven countries we track, the official source of reservoir data, what it publishes, how often, and how open it actually is. It is also the raw material behind our Reservoir Transparency Index, which scores countries on exactly these questions.
Key takeaways
- No global reservoir database exists. Reservoir data is published country by country by a national agency — usually a water ministry, sometimes an electricity-grid operator, sometimes the weather bureau — and you have to go to each one separately.
- The most open sources are Australia (BOM), Taiwan (WRA), the United States, and Spain (MITECO). Each offers either an open API or a downloadable dataset you can pull from anywhere, no key or country needed.
- The hardest is India (CWC): a weekly PDF bulletin whose live system is geo-restricted to Indian IP addresses, encrypted, and behind a login — see how we get India's data.
- Two countries publish reservoirs as energy, not water. Colombia (XM) and Norway (NVE) report stored hydropower in gigawatt- and terawatt-hours, because their reservoirs exist to generate electricity.
- Update frequency ranges from daily to monthly. Daily: United States, Australia, Brazil, Taiwan, Colombia, Pakistan. Weekly: Spain, India, South Africa, Norway. Monthly: Portugal.
The directory: official reservoir data, country by country
| Country | Official source | Publishes | Frequency | Access |
|---|---|---|---|---|
| Australia | Bureau of Meteorology (BOM) — Water Storage / Water Data Online | Volume + % full, per storage | Daily | Open API. No key. Gold standard. |
| Brazil | ONS — Operador Nacional do Sistema Elétrico (open-data portal) | Hydropower reservoir storage | Daily | Open CSV. Downloadable, no key. |
| Colombia | XM — market & grid operator (API) | Stored energy (GWh), not volume | Daily | Open API. |
| India | Central Water Commission (CWC) — Weekly Reservoir Storage Bulletin | ~166 major reservoirs, national | Weekly | Closed. PDF; live system geo-blocked + encrypted + login. |
| Norway | NVE — Norges vassdrags- og energidirektorat (Magasinstatistikk) | Stored energy (TWh) by price area, not individual dams | Weekly | Open API. |
| Pakistan | IRSA — Indus River System Authority (with WAPDA) | Tarbela, Mangla, Chashma levels & storage | Daily | Open, but published as a document, not a feed. |
| Portugal | SNIRH — Sistema Nacional de Informação de Recursos Hídricos (APA) | Reservoir storage & % full | Monthly | Open web portal; no clean API (we scrape it). |
| South Africa | DWS — Department of Water and Sanitation (Weekly State of Reservoirs) | Provincial + per-dam storage | Weekly | Open web page, but fragile — publication has stalled at times in 2026. |
| Spain | MITECO — Boletín Hidrológico Semanal + historical database | ~370 reservoirs, national | Weekly | Open. PDF bulletin and a downloadable database with history back to 1988. |
| Taiwan | WRA — Water Resources Agency (Open Data platform) | Per-reservoir storage & % full | Daily | Open JSON API. No key. |
| United States | USGS, US Army Corps (CWMS), Bureau of Reclamation (RISE), California's CDEC | Per-reservoir storage | Daily | Open, but scattered across four+ federal agencies — see how we get US data. |
The open front-runners
Four sources set the bar for what open reservoir data should be. Australia's Bureau of Meteorology is the cleanest of all: a national water-storage service with an open API, updated daily, that answers a request from anywhere in the world without a key. Taiwan's Water Resources Agency is the same story in miniature — a documented JSON API on a national open-data platform, no registration required. The United States is open but scattered: no single agency holds the country's reservoirs, so the picture has to be stitched from the USGS, the Army Corps of Engineers, the Bureau of Reclamation and California's CDEC, each with its own units and quirks. And Spain's MITECO does the thing almost no one else does — it publishes not just the current weekly bulletin but a downloadable historical database with readings back to 1988, which is why our Spanish coverage has decades of context behind every current number.
Reservoirs measured as energy, not water
Two of our sources never mention cubic metres at all. Colombia (XM) and Norway (NVE) are electricity-system operators, and in both countries the reservoirs exist primarily to generate hydropower. So they report what the grid cares about: stored energy — gigawatt-hours in Colombia, terawatt-hours by price area (NO1–NO5) in Norway — rather than water volume. Norway does not even publish individual reservoirs; it aggregates to the regional level. We keep these in their native energy units on the site rather than inventing a volume conversion that the source never made, because a fabricated cubic-metre figure would be less honest than the real gigawatt-hour one.
The hard cases
Not every government makes this easy. India is the extreme: the Central Water Commission publishes a genuinely useful national bulletin, but only as a weekly PDF, and its live bulletin system is restricted to Indian IP addresses, wraps its responses in encryption, and sits behind a login — the whole difficulty is getting in, which is why India needed its own dedicated pipeline. Portugal's SNIRH is open but slow and unstructured — monthly updates through a web portal with no clean API. Pakistan's IRSA publishes daily, but as a document rather than a machine-readable feed. And South Africa's DWS runs a good weekly report that has proven fragile in practice: publication stalled for stretches of 2026, and when the source goes quiet there is no API to fall back on.
What a "perfect" reservoir data source looks like
Reading across eleven countries, the same handful of properties separate the sources we can trust from the ones we have to fight. A government publishing reservoir data well would:
- Update on a fixed, frequent schedule — daily where possible, weekly at worst — so a reading is never stale by accident.
- Publish machine-readable data (an open API, CSV or JSON), not a PDF or an HTML table a human has to retype.
- Cover capacity and live storage, so a percentage-full can be computed rather than guessed.
- State its method and datum — what "full" means, which storage it counts — so two sources can be reconciled.
- Keep the full time series, not just today's snapshot, so seasonal and multi-year context survives.
- Serve it under an open licence, from a stable URL, with no geo-block — reachable from anywhere, by anyone.
That checklist is not hypothetical: it is close to the rubric behind our Reservoir Transparency Index, which grades countries on how openly they publish, and what a top score requires. Australia would sit near the top; India, on access alone, near the bottom — despite collecting excellent data.
Frequently asked questions
Is there a single global database of reservoir levels? No. There is no worldwide reservoir database. Each country publishes its own data through a national agency, in its own format and units, and they must be gathered and normalised one at a time — which is the work behind this site.
What is the official source of reservoir levels for a given country? It depends on who runs the reservoirs. It is usually the national water agency (Spain's MITECO, India's CWC, South Africa's DWS, Portugal's SNIRH), but where reservoirs are mainly for hydropower it is the electricity-grid operator (Colombia's XM, Norway's NVE, Brazil's ONS), and in Australia it is the national weather bureau (BOM).
Why do two official sources report different levels for the same reservoir? Usually units, timing, or definitions: agencies use acre-feet, cubic metres, or billion gallons; a weekly bulletin lags a daily portal; and "percent full" depends on which storage (live, gross, dead) and which datum the source counts. We normalise everything to a common unit before comparing — how we handle it.
Which country publishes the most open reservoir data? Among the countries we track, Australia (BOM) and Taiwan (WRA) are the most open — documented APIs, updated daily, reachable worldwide with no key — with the United States and Spain close behind.
This directory covers the eleven countries currently in our data set and is updated as we add more. For how the numbers are collected and cleaned, see our "How We Get Our Data" series on India and the United States.