What the Water Stress Service Returns
One coordinate goes in and one record comes out, in the same request, through the Water Stress API. The service finds the piece of land the coordinate sits on in the WRI Aqueduct 4.0 Water Risk Atlas and returns what the dataset publishes for that piece: 13 baseline indicators, 10 industry profiles, 12 monthly values, and projected values for three pathways at 2030, 2050 and 2080.
Every number is copied as published. Each value has four parts: the raw figure, a 0 to 5 score, a category number and a text label. The service adds one field per value, a flag, saying whether the value is a published figure, a code, a placeholder or missing. It applies no rating, maps no pathway onto another scenario set, and adds no horizon.
Data Source
- Dataset: WRI Aqueduct 4.0 Water Risk Atlas, World Resources Institute, version Y2023M07D05, pinned. Three tables (baseline annual, baseline monthly, future annual) plus the basin geometry.
- Licence: Creative Commons Attribution 4.0 International (CC BY 4.0). Every response carries the attribution line “WRI Aqueduct 4.0 Water Risk Atlas, World Resources Institute, CC BY 4.0”.
- Citation: Kuzma, S., Bierkens, M.F.P., Lakshman, S., Luo, T., Saccoccia, L., Sutanudjaja, E.H., Van Beek, R. (2023). Aqueduct 4.0: Updated decision-relevant global water risk indicators. Technical Note. World Resources Institute, Washington DC. doi.org/10.46830/writn.23.00061.
- Basins: Lehner, B., Grill, G. (2013). Global river hydrography and network routing. Hydrological Processes 27(15): 2171 to 2186 (HydroBASINS).
The industry profiles weight in a country ESG risk index published by RepRisk, carried as the dataset publishes it.
The Basin Unit and How a Coordinate Is Matched
The unit is an Aqueduct piece: a HydroBASINS level 6 river basin, cut where it crosses a province or aquifer boundary. The dataset holds 68,506 pieces. The median piece covers 368 km²; the example site on this page (label “site-01”) sits in one of about 412 km². Water stress, depletion, variability and the monthly and projected tables are constant across the basin. Groundwater table decline follows the aquifer. The regulatory and reputational indicators follow the country. No value is measured at the site.
| Case | Result |
|---|---|
| One basin-bearing piece contains the coordinate | Method contains, gap 0 m. The ordinary case. |
| None contains it; one lies within 2,000 m | Method nearest, with the gap in metres. Jetties, marinas, reclaimed land, water close to shore. |
| Nothing basin-bearing within 2,000 m | Status FAILED, code NOT_COVERED. The usual answer for a coordinate in open ocean. |
The dataset also holds 5,159 placeholder pieces, WRI’s “no basin here” tiles, mostly coastal strips a few hundred metres wide. A placeholder publishes no data, so a coordinate inside one is matched to the nearest basin-bearing piece within the cap and the response names the strip. A basin whose indicators read No Data is a real basin and is matched by contains.
The 13 Baseline Indicators and the 10 Industry Profiles
Band limits are written inside each label and differ per indicator. The table shows the labels the example site received.
| Code | Indicator | What it describes | Example label (band limits as published) |
|---|---|---|---|
| bws | Baseline water stress | Water withdrawn divided by water available, per year | Extremely High (above 80%) |
| bwd | Baseline water depletion | Share of available water consumed | Medium to High (25 to 50%) |
| iav | Interannual variability | Year to year swing in supply | Medium to High (0.50 to 0.75) |
| sev | Seasonal variability | Swing in supply within the year | High (1.00 to 1.33) |
| gtd | Groundwater table decline | Fall of the water table, cm per year | Extremely High (above 8 cm/y) |
| drr | Drought risk | Drought hazard, exposure and vulnerability | High (0.8 to 1.0) |
| rfr | Riverine flood risk | Share of the population expected to be affected by river flooding in an average year | High (6 in 1,000 to 1 in 100) |
| cfr | Coastal flood risk | Share of the population expected to be affected by coastal flooding in an average year | No Risk (placeholder) |
| ucw | Untreated connected wastewater | Share of collected wastewater left untreated | Extremely High (100%) |
| cep | Coastal eutrophication potential | Potential for river nutrient loads to drive harmful algal blooms in coastal waters | Extremely High (above 5) |
| udw | Unimproved / no drinking water | Share of people without improved drinking water | Medium to High (5 to 10%) |
| usa | Unimproved / no sanitation | Share of people without improved sanitation | Extremely High (above 20%) |
| rri | Peak RepRisk country ESG risk index | Business conduct risk exposure to ESG issues in the country | Medium to High (50 to 60%) |
The 10 profiles are Default, Agriculture, Chemicals, Construction materials, Electric power, Food and beverage, Mining, Oil and gas, Semiconductor and Textile. Each applies WRI’s industry weights (in the WRI technical note) and gives four results: physical risk quantity, physical risk quality, regulatory and reputational risk, and overall water risk. At the example site the Default overall score is 4.45 (Extremely High) and the Oil and gas overall score is 3.98 (High): one basin, weighted for different industries.
Monthly Values and Projected Values
The monthly table carries water stress, depletion and interannual variability for each calendar month over the baseline period, 1979 to 2019, which the dataset calls the long-term, chronic trend. At the example site the annual stress ratio is 0.98 (about 98% of supply withdrawn); month 5 reads 4.20 (420%, score 5) and month 8 reads 0.008 (under 1%, score 0).
The projected table carries three pathways under the dataset’s own names: Optimistic (SSP1 RCP2.6), Business as usual (SSP3 RCP7.0) and Pessimistic (SSP5 RCP8.5). SSP and RCP name the population storyline and the emissions pathway a projection runs on. Each pathway has three horizons, each a window of years: 2030 is 2015 to 2045, 2050 is 2035 to 2065, and 2080 is 2065 to 2095. Each value is the median across 5 global climate models. Water stress, depletion, interannual and seasonal variability are projected, with available supply and gross demand in cm per year. At the example site, 2050 water stress reads 0.77 (High) under Optimistic, 0.93 under Business as usual and 0.83 under Pessimistic.
Scores, Categories and Flags
The score runs from 0 to 5 and the category from 0 to 4. For baseline water stress the bands are:
| Category | Label | Withdrawals as a share of supply |
|---|---|---|
| 0 | Low | under 10% |
| 1 | Low to Medium | 10 to 20% |
| 2 | Medium to High | 20 to 40% |
| 3 | High | 40 to 80% |
| 4 | Extremely High | above 80%, open-ended |
The dataset’s water stress score formula is score = clamp(log2(raw / 0.1) + 1, 0, 5). The score is 1 when withdrawals reach 10% of supply, and each doubling adds a point: 2 at 20%, 3 at 40%, 4 at 80%, 5 at 160% and above. Below 5% the formula goes negative and the published score is held at 0; a score of 0 is a value. For the example site, 0.98 gives log2(9.8) + 1 = 4.29, the published score. The audit block reproduces every published score this way as a check; the service never applies the formula itself.
The flag is the only field the service adds to a cell. Everything else is copied.
| Flag | Source state | What the cell carries |
|---|---|---|
| (no flag) | A published value | raw, score, category and label as published |
| no_data | Source value -9999 | raw, score and category empty; label “No Data” |
| severe_scarcity_code | Source value 9999 | raw 9999, score 5, category 4, Extremely High. A code for severe scarcity with no meaning as a ratio or percentage |
| score_not_published | Real raw, score and category -9999 | Groundwater table decline, “Insignificant Trend”: raw published, score and category empty |
| arid_low_water_use | Category -1 on water stress or depletion | raw 1.0 and score 5.0 are placeholders for an arid, low water use basin |
| no_risk | Category -1 on coastal flood risk | raw and score 0.0 placeholders |
| no_wastewater_collected | Category -1 on untreated connected wastewater | raw -1.0 and score 5.0 placeholders |
9999 is severe scarcity and -9999 is no data; the two are never folded together. Reading an arid basin’s score of 5 as Extremely High water stress misreads a placeholder.
Limitations
- Every value describes a basin piece, an aquifer or a country. Nothing is measured at the site; two coordinates in the same basin get the same figures.
- WRI’s own cautions travel with the record: “Aqueduct risk indicators cannot be measured directly and therefore have not been validated.” “Aqueduct does not model interbasin transfers.” “Projected water demand contains multiple assumptions based on data limitations.”
- The pathways are the dataset’s, under their own names, with no mapping onto any other scenario set.
- There is no 2040 horizon. The dataset publishes 2030, 2050 and 2080.
- Codes and placeholders are returned as the dataset holds them, with a flag. In a random sample of land points, about 8.7% fell inside a placeholder piece and 10.1% inside a basin whose stress indicator is No Data.
- The 2,000 m gap is measured on a flat approximation of the Earth’s surface and can differ from the true distance by up to about 7%.
Frequently Asked Questions
Why does the water stress rating in a Climate Risk report differ from this record?
The Climate Risk module reads the same dataset with its own thresholds and arid exclusion to produce a rating. This service carries the full record and rates nothing.
Why is there no 2040 horizon?
The dataset publishes 2030, 2050 and 2080. The service adds nothing between them.
Can I read the Business as usual pathway as SSP2-4.5?
No. The dataset’s Business as usual pathway is SSP3 RCP7.0, and the record names it that way. Cite each pathway under its own name.
My coordinate came back with method nearest and a gap of a few hundred metres. Is that an error?
No. The coordinate fell on water or inside a placeholder strip, and the service matched it to the nearest basin-bearing piece within 2,000 m. The gap and the strip are both in the record.
Appropriate use: Portfolio water risk screening, comparison of sites and basins on one published scale, seasonal exposure checks from the monthly table, and disclosure support where the source, its version and its cautions are cited.
Not suitable for: Site level water supply assessment, engineering or abstraction licensing decisions, any use that treats a basin figure as a measurement at the property, or any reading of 9999, arid or No Data cells as measured values.