What the number means, where it comes from and what it does not tell you.
Data updated October 24, 2026Reviewed by Steve Saper, founder, on October 25, 2026
The score starts at 100. It goes down 8 for each contaminant above a cited health guideline whose source states a cancer basis, 5 for each other contaminant above its cited guideline, 5 for each of PFOA and PFOS averaging above the 4.0 ppt limit in EPA's 2024 PFAS rule, 10 for each contaminant whose average of all samples is above EPA's legal limit (instead of its guideline step), and 1 for each contaminant with a cited cancer-based guideline that is detected below it. Only guidelines with a cited document count. The lowest score is 5. Higher is better. Every contaminant is compared with a published health guideline, and every number on a report carries its source and years.
The calculation
Start at 100.
Subtract 8 for each detected contaminant above a published health guideline whose cited document states it is based on cancer risk.
Subtract 5 for each other detected contaminant above its published health guideline.
Subtract 5 for each of PFOA and PFOS measured above the 4.0 ppt limit in EPA's 2024 PFAS rule. Under EPA's 2024 PFAS rule, water systems must meet the PFOA and PFOS limits (4.0 ppt each) by April 2029; in May 2026 EPA proposed allowing extensions to April 2031. Results before then are monitoring data, not violations. For this step, the level is the average of all of the utility's UCMR 5 results, with results below EPA's reporting level counted as zero, and the step applies when that average is above 4.0 ppt. Action levels are never compared with a single result.
A contaminant whose average of all samples is above EPA's legal limit subtracts 10. EPA judges compliance on a running annual average at each sampling location, so this is a guide, not a compliance finding. A total such as TTHMs is scored against its legal limit, and the chemicals in it are scored against their own health guidelines. The step counts only an average of all samples that we have checked against the source results and against EPA's compliance records (SDWIS).
Lead is scored only when a utility's 90th-percentile tap result is above EPA's action level, because lead samples come from homes selected for likely lead plumbing. When it is, subtract 5.
Subtract 1 for each contaminant with a cited cancer-based guideline that is detected below its guideline.
Only guidelines with a cited document, year and link count. Results more than 10 years older than the data set are not shown or scored.
Florida DEP results include samples of untreated source water; we leave those out, because they are taken before treatment.
A detected result more than 50 times the utility's median detected result for that contaminant, and above 100, is checked against the source before we use it. Until it is checked, it is left out, and we list each one we leave out.
The lowest score is 5. Higher is better.
PFOA and PFOS under EPA's 2024 rule are scored separately (5 points each when the all-samples average is above 4.0 ppt) and are not counted among the "above health guidelines" results.
Bands, shown in words beside every score: Low (0 to 39), Weak (40 to 59), Fair (60 to 79) and Strong (80 to 100).
What counts as above a guideline
A detected level is compared with the published health guideline for that contaminant. Guidelines come from California OEHHA (public health goals), U.S. EPA (PFBS health advisory, maximum contaminant levels and action levels). Each contaminant card names the guideline source and year. Where a federal legal limit exists, it is shown next to the guideline. The guideline is a published reference value and not a verdict about anyone's water.
A legal limit and a health guideline are different
A legal limit is an enforceable standard, set by the EPA under the Safe Drinking Water Act. Limits weigh health research together with the cost and feasibility of treatment. A health guideline is published by public-health sources and based on health research alone. A utility can meet every legal limit and still be above a published health guideline.
How a ZIP code is matched to a utility
The ZIP code areas come from the Census Bureau and the utility service areas from the EPA Water System Boundaries layer. Where they overlap, the share of the ZIP code area inside each utility area is measured.
Every utility with at least 10% of the area is kept. The utility with the largest share is the main utility, unless a retail utility with a close share buys its water from the leader, in which case the retail utility, the one on the bill, is the main utility.
ZIP codes without a Census area (PO box and unique ZIP codes) borrow the candidates of the three nearest ZIP codes. These are always shown as low-confidence matches.
Confidence: high means one utility covers at least 80% of the area and the boundary was reported by the state, county, water district or utility. Medium means 50% to 80%, or a modeled boundary. Low means the main utility covers under 50%, or the match was borrowed.
ZIP -> utility candidates by area overlay (>= 10% of ZIP area); primary = highest share, retail buyer preferred over its wholesaler when within 10 points; ZIPs without a ZCTA inherit candidates from the nearest ZCTA ZIPs (low confidence).
Purchased water
When EPA SDWA records show a utility buys water from another system, results measured by the supplier are used for the buyer and marked with an asterisk, and the pages show the chain from supplier to utility to tap.
What the score is not
It is not a test of your tap. It is based on public data from EPA SDWA, EPA UCMR and Florida DEP data, and the EPA compliance records come from the July to September 2026 extract. Data may be one to three years old and may not reflect water at your tap. Water Scores and reports are educational estimates from public data, not medical advice.
Water Score rankings
Formula: a ranking uses the Water Score described above, unchanged, for every row. A higher score means fewer findings in public records. Scores run from 5 to 100.
Data years: each row shows the years of the records it uses (the years of the utility's results and the years of its UCMR 5 PFAS results). The page's "as of" date is the date of the latest data update.
ZIP codes served by more than one utility: a ZIP code is scored on its main utility, the one with the largest share of the area (or the retail utility on the bill, as in the matching steps above). The other utilities are named in the row.
No data, no rank: a ZIP code whose main utility has no PFAS result (EPA UCMR 5) or no lead 90th percentile is not ranked. It is listed separately under "Not enough public data to rank".
Ranks: equal scores share a rank. Within a score, rows are listed by population served, largest first. No row is moved by hand and there is no paid placement.
Outliers: a row whose lead 90th percentile is above 100 ppb stays out of headlines, titles and the ten highest and lowest lists unless the utility confirms the figure. It can appear in the full table, with a note.
PFOA or PFOS: the score subtracts points when a utility's UCMR 5 results averaged above 4.0 ppt, the level EPA set in 2024, not yet enforceable. Health guidelines aren't legal limits.
The U.S. cities page has no score and no ranking: it lists utilities whose UCMR 5 PFOA or PFOS average, or lead 90th percentile, is above a level, from EPA records, sorted by state and city.
ScoreYourWater also sells water treatment systems.
October 23, 2026: EPA's compliance records now come from its July to September 2026 extract, which replaces the April to June 2026 extract. The new extract no longer lists 11 health-based "LSL Inventory" violations from 2025 (the lead service line inventory requirement) at 11 water systems, so their pages no longer show them. The new extract doesn't show them as resolved; they are no longer listed. It also lists 3 total trihalomethanes violations at Sunshine Meadows, for compliance periods from April to December 2025, that the earlier extract didn't. The latest lead 90th-percentile result changed at 2 systems. Health-based violations are now listed for compliance periods starting since July 1, 2021, so violations from April to June 2021 are no longer listed. No Water Score changed.
October 22, 2026: Florida DEP results now leave out samples of untreated source water, taken from wells or raw-water taps before treatment, because they do not show the water that is delivered. On utility pages, 31 Florida DEP values went down, 18 went up, and 4 are no longer shown because every sample behind them was untreated water. No Water Score changed.
October 22, 2026: This replaces the "average of detected results" note added earlier on October 22, 2026. Florida DEP results from 2018 to 2022 are now the average of all of a utility's samples, as the UCMR levels are: a result DEP reports as zero (not detected) counts as zero, and a row that DEP's files repeat exactly is counted once. Before, those levels averaged only the results in which the contaminant was detected, which left out every non-detect, and they counted repeated rows again: DEP's 2022 file repeats 19,267 of its 150,308 rows exactly, and its 2018 to 2020 files repeat 174 more. The note "average of detected results" is replaced by the highest single result and the number of samples in which the contaminant was detected. On utility pages, 595 Florida DEP values went down and 120 went up. Bromoform at the City of Miami Beach and the City of Opa-locka is now below its health guideline, so their Water Scores went up by 7 (Miami Beach from 55 to 62, Opa-locka from 42 to 49); no other Water Score on a ZIP code page changed. Dade Juvenile Residential Facility's total trihalomethanes (TTHMs) and haloacetic acids (HAA5) now pass the checks for the "Average above EPA's legal limit" group, so the facility now has its own page, with a Water Score of 48.
October 22, 2026: Florida DEP results from 2018 to 2022 now carry the note "average of detected results": each of those levels averages only the results in which the contaminant was detected. A new group, "Average above EPA's legal limit", lists a contaminant only when its average of all samples is above EPA's legal limit and has been checked against the source results and EPA's compliance records (SDWIS); each one subtracts 10 from the Water Score. No page shows the group yet, and no Water Score changed. Where a PFOA or PFOS step applies, the score arithmetic now names it.
October 22, 2026: Every UCMR level (HAA9, manganese and the other unregulated contaminants, not only PFAS) is now the average of all of a utility's samples, with results below EPA's reporting level and results reported as zero counted as zero; the highest single result and the number of samples in which it was detected are shown with each level. Before, those levels averaged only the detected results, and a result reported as zero was left out. Values went down for 15 rows on utility pages (for example the City of Hialeah's HAA9, from 39.7 to 37.4 ppb); no Water Score changed. Utilities that are not the main utility for any ZIP code now show their full results, and water systems the boundary data does not match to a ZIP code now have their own pages. The years each source covers are now taken from the results shown.
October 22, 2026: PFAS levels are now the average of all of a utility's samples, with results below EPA's reporting level counted as zero; the PFAS group and its score step use that average. Before, the average left out results below the reporting level. Lead now shows each utility's latest 90th-percentile tap result. Before, a date-sorting error picked results from 2001 to 2004, which were then hidden as too old, so no lead result was shown. Lead is scored only when it is above EPA's action level, and it is never counted as above a health guideline. HAA9 now shows the value measured under UCMR 4; a row that repeated the HAA5 value was removed. Some Water Scores went up by 5 as a result; none went down.
Our claims cover three results: chlorine and hardness at your kitchen tap, and dissolved solids at your drinking-water faucet. Guarantee terms ›