How Riparia scores water risk
The data, formulas, sources and limits behind every Riparia score — explained in plain language.
Every address you check on Riparia returns a 0–100 composite score and a five-tier rating. This page explains what that number actually means, what data sits underneath it, how often we refresh it, and — just as importantly — what it cannot tell you. We score on basin-scale public hydrology, not on your specific roof. Be sceptical of any product that claims otherwise.
The 0–100 score and the five tiers
Riparia returns a single composite score from 0 (lowest stress) to 99 (highest), mapped to five named tiers: low, moderate, elevated, high, extreme. The ladder is fixed — the same thresholds apply to every basin in Europe, so two addresses are directly comparable.
Thresholds: low < 25 · moderate 25–44 · elevated 45–64 · high 65–81 · extreme ≥ 82. These cut-points are locked at the formula level (currently v2) and are visible in the per-report freshness footer next to the score.
The four dimensions we measure
The composite blends four independent dimensions, each scored 0–100 from public datasets:
Drought & water stress — multi-decade rainfall trends, basin withdrawals against natural recharge, and projected shortfall under IPCC AR6 climate scenarios. Driven by WRI Aqueduct 4.0 (BWS, BWD) and Copernicus EDO SPEI-3 for current conditions.
Flood exposure — river, surface and coastal flood probability for the 100-year event, plus how it shifts with sea-level rise where coastal. Driven by Aqueduct 4.0 (RFR, CFR) and JRC EFAS where coverage is published.
Tap-water quality — public sampling for nitrates, lead, microplastics and pharmaceutical residues. Driven by EEA WISE annual reporting under the EU Drinking Water Directive, with country-level coverage indicators from Eurostat.
Supply infrastructure — basin-level stress (withdrawals vs. renewable supply) plus the resilience of the network feeding the address. Driven by Aqueduct 4.0 (BWS) and, where available, local utility outage records.
How the composite is built
The four dimension scores are blended into the composite via a weighted average, with weights chosen to reflect the typical impact on a household. Drought and supply carry the largest weight; flood and tap-water quality modulate the result based on local hydrology and utility data.
When a dimension has no published data for a basin (e.g. EFAS not yet covering a country), Riparia marks it explicitly as unavailable rather than imputing a value. Composite weights are renormalised over the dimensions that do have data, and the affected dimension renders with a clear "no data" note in the report.
The exact weights and formula version powering your report are printed in the methodology footer of every full PDF, so two reports computed under different formula versions are always distinguishable.
Data sources and refresh cadence
WRI Aqueduct 4.0 — global water-stress baseline and projections. CC BY 4.0. Major release roughly every 4 years; the current vintage is 2023.
Copernicus EDO — European Drought Observatory, SPEI-3 index. Copernicus License. Refreshed weekly; we surface the current value in the drought dimension where the basin falls inside the European coverage area.
JRC EFAS — European Flood Awareness System. Where coverage is published, it provides high-resolution riverine flood probability. We use the latest released version.
EEA WISE — European Environment Agency water-quality reporting under the EU Drinking Water Directive. CC BY 2.5 DK. Annual reporting cycle.
Eurostat — national indicators on population connected to public supply and wastewater treatment. CC BY 4.0. Annual.
OpenStreetMap — reverse geocoding of input coordinates to addresses. ODbL.
Country-specific basin data and utility reports — used where available to refine the supply dimension. National regulators include ANRSC (RO), BDEW (DE), SISPEA (FR), AEAS (ES), Ofwat (UK).
What this score cannot tell you
Riparia scores at the basin and country level, not the individual building. The composite is a probabilistic indicator of the hydrological context surrounding your address — what is typical for the basin, not what is certain for your roof.
Whether your specific house will flood depends on local elevation, drainage, foundation height and building-level mitigation that no basin-scale dataset can capture. A high flood-dimension score is a signal to commission a site-level survey; it is not a verdict.
Similarly, tap-water quality data reflects compliance reporting at the water-utility level. Your actual exposure depends on the pipework between the utility distribution point and your tap, which only your local utility and an in-home test can verify.
Riparia is independent and is not shared with insurers, banks, or estate agents. We have no commercial relationship with the property market — the score is for your own decision-making, not for anyone else's pricing decision.
Versioning, audit and freshness
Every full PDF report includes a freshness footer with: the formula version used (e.g. v2), the timestamp at which the score was computed, and the per-source data vintages. If we re-issue a report for the same address six months later with a newer Aqueduct vintage, the difference will be transparently visible in those footer fields.
On the live globe, the country-level fill colors reflect the same tier system as individual addresses, aggregated to the national-average basin stress. The legend in the top-right of the map is the same five-tier ladder used everywhere in the product.
If you spot a number that looks wrong or a source description that has drifted, write to hello@riparia.live — we publish corrections rather than hide them.