Coverage and gaps
Coverage across countries is genuinely uneven, and pretending otherwise would be the single most misleading thing this site could do. This page reports what Hydrionis holds, what it does not, and where the thin areas are.
What Hydrionis contains
| Observations | 39,613,229 | Each resolving to a registered source. A value without one is refused. |
|---|---|---|
| Of which computed by Hydrionis | 183,302 | Each carrying its method, version and the identifier of every input observation. |
| Distinct series | 124,716 | One metric, one subject, one period resolution. |
| Monitoring locations | 21,227 | River gauges, boreholes and sampling points. These carry most of the depth and, deliberately, none of the URLs. |
| Geographies | 22,145 | Administrative, operational and hydrological units in one tree. |
| Entities | 1,097 | Including 28 utilities, 12 rivers, 5 reservoirs and 20 recorded events. |
| Registered sources | 210 | From 179 distinct publishing organisations. |
| Automated pipelines | 35 | Everything else was entered by hand from a published document and does not refresh on its own. |
Geographical coverage
View the data behind this chart
| Countries and territories in the registry | 205 |
|---|---|
| With at least one core indicator | 199 |
| With all nine | 1 |
| With none | 6 |
| Mean core coverage | 74.4% |
Temporal coverage
Observations by domain
| Domain | Observations | Share |
|---|---|---|
| river-flow | 21,644,330 | 54.6% |
| precipitation | 6,775,770 | 17.1% |
| temperature | 3,436,258 | 8.7% |
| drought | 2,526,276 | 6.4% |
| atmospheric | 1,432,053 | 3.6% |
| snow-ice | 1,188,354 | 3.0% |
| ecosystem | 717,898 | 1.8% |
| groundwater | 584,955 | 1.5% |
| quality | 471,150 | 1.2% |
| reservoir | 292,819 | 0.7% |
| chemistry | 227,748 | 0.6% |
| agriculture | 144,556 | 0.4% |
| contaminant | 60,193 | 0.2% |
| sector-use | 33,245 | 0.1% |
| resources | 22,577 | 0.1% |
| abstraction | 15,378 | 0.0% |
| access | 9,150 | 0.0% |
| drinking-water | 6,999 | 0.0% |
| withdrawal | 6,780 | 0.0% |
| stress | 6,672 | 0.0% |
| sanitation | 3,781 | 0.0% |
| utility-performance | 2,989 | 0.0% |
| leakage | 1,173 | 0.0% |
| tariff | 1,077 | 0.0% |
| infrastructure | 1,043 | 0.0% |
| wastewater | 5 | 0.0% |
Rights composition
What Hydrionis is legally able to do with what it holds. These are properties of the publishers’ licences, not of any Hydrionis decision.
| Observations that may be redistributed | 100.0% |
|---|---|
| Observations that may be used commercially | 99.6% |
| Source registry that is free or open | 98.6% |
| Licence class | Sources |
|---|---|
| Open — attribution required | 108 |
| Public but restricted | 68 |
| Open | 28 |
| Permission required | 4 |
| Proprietary / paid | 2 |
Where Hydrionis is weak
Metrics defined but not populated
186 of 258 metrics carry values. The remaining 72 are defined and empty. They are listed rather than hidden, because a registry showing only what exists implies the rest was never considered.
Geographical concentration
The station-level record is English only. The Environment Agency regulates England; Scotland, Wales and Northern Ireland have their own agencies and their own archives, which are not yet collected. Outside the UK, coverage is national annual statistics.
Domains with thin coverage
Utility performance, leakage, tariffs and wastewater are held as recent regulatory reporting rather than long series. Reservoir storage, soil moisture and abstraction licensing are not collected. Drinking-water compliance is held at the level of published summaries rather than zone-level results.
What is not published anywhere
Some things a reader reasonably wants are not open data at all: facility-level water consumption, most industrial abstraction volumes, and network condition. Hydrionis holds none of these and does not estimate them. Anyone publishing them at scale is modelling, and a model is not an observation.
Acquisition backlog
The panels above say where Hydrionis is thin. This says what would fix it. Each entry names the metrics it would populate, and what is currently held against them is counted from the series index rather than written down — so an entry that has quietly been filled stops claiming to be a gap. Priority follows what an entry unlocks, not how large it is: one missing dataset can block several analyses, and a large one can block none.
| Gap | Priority | Held today | Status | Unlocks |
|---|---|---|---|---|
| Official drought declarations and restrictions | critical | nothing | open | drought event overlays on every hydrology chart; drought against river flow, groundwater and reservoir storage; comparison of conditions between named droughts |
| Authoritative reservoir capacity | critical | nothing | open | percentage full as a derived metric; reservoir-to-reservoir comparison on a common scale; supply-system aggregate storage |
| Reservoir storage published with its capacity | critical | nothing | open | percentage full; reservoir storage compared between reservoirs; drawdown and recovery across historical droughts |
| UK reservoir storage | critical | 187 series · 1905–2026 | blocked | rainfall against reservoir storage; drought events against reservoir storage; river flow against reservoir storage; drawdown and recovery analysis across historical droughts |
| United States groundwater levels | critical | 250 series · 1942–2026 | open | groundwater against official drought status; groundwater against rainfall and snowmelt; recharge following a wet winter; the full water cycle in one country |
| Groundwater abstraction volumes | high | nothing | open | groundwater level against abstraction; catchment water balance; drought attribution between recharge and demand |
| Environment Agency rainfall gauges | high | 591 series · 1827–2026 | open | rainfall against river flow at gauge resolution; rainfall against groundwater level at borehole resolution; catchment rainfall deficit |
| Catchment area for gauging stations | high | 1,687 series · 1874–2025 | open | annual runoff in millimetres; rainfall against river flow as a water balance; flow duration curves compared between rivers; low flow per unit area, which is how abstraction pressure is actually judged |
| Temperature and snow at the newest rainfall stations | high | 121 series · 1814–2026 | blocked | heatwave against drought; snow storage against river flow; temperature against evaporative demand |
| Instantaneous annual maximum flow | high | 2,236 series · 1806–2025 | open | flood frequency and return periods; the calculated daily maximum checked against the true peak, which would measure how much averaging costs on each catchment; flood events against subsequent water quality |
| Basin identifiers for US snow monitoring sites | high | 5,186 series · 1847–2026 | closed | snowpack against river flow; melt timing against the spring flood; snow storage as a forecast of summer low flow |
| Manual snow course records | high | 851 series · 1963–2026 | open | snowpack before the automated era; snow against river flow over a century |
| Soil moisture and evapotranspiration | medium | nothing | open | rainfall against river flow with antecedent conditions; drought development and recovery |
| Which French rivers sharing a name are the same river | medium | 5,896 series · 1847–2026 | open | French river pages that describe a single water; the same catchment-identifier split already applied to United States subbasins |
| The remaining GHCN-Daily long-record stations | medium | 146 series · 1827–2026 | open | rainfall against river flow in more catchments; wider anomaly comparison |
| Snow water equivalent | medium | 851 series · 1963–2026 | open | snow storage against river flow; melt timing shifts |
Update frequency
The 35 automated feeds re-read their publishers, and reading twice changes nothing unless the publisher changed something. What each run did, and what it refused to store, is reported on the data collection page. Values that a publisher has since revised are retained in the Vault rather than overwritten.