YourWaterReport.com aggregates groundwater quality data from public federal and state databases to help private well owners understand what contaminants may be present in their area. This page documents exactly where our data comes from, how it is processed, and where its limitations lie.
Data sources
All contaminant occurrence data displayed on this site comes from publicly available government databases. We do not generate, sample, or certify any water quality data. Our role is aggregation, normalization, and presentation.
Water Quality Portal (WQP)
The Water Quality Portal (waterqualitydata.us) is a cooperative service operated by the USGS, EPA, and USDA that integrates water quality data from over 400 federal, state, tribal, and local agencies. It contains hundreds of millions of records spanning decades of groundwater monitoring. WQP is our primary source for groundwater analyte measurements including arsenic, nitrate, iron, manganese, uranium, fluoride, lead, and PFAS.
We query WQP by state and analyte, then resolve each monitoring location to a county FIPS code using WQP's station endpoint. Results include sample date, monitoring location identifier, analytical method, and measured value.
State PFAS databases
Several states operate dedicated PFAS residential well sampling programs that supplement WQP records. We load data directly from:
- Michigan EGLE — residential well sampling 2018–present
- New Jersey NJDEP — ambient groundwater monitoring network (149 wells, 2020–2022)
- Minnesota MPCA — Atlas statewide monitoring database
- Pennsylvania DEP — quarterly PFAS sampling program
- Wisconsin, Ohio, New York — state data submitted to and retrieved via WQP
EPA UCMR5 (Unregulated Contaminant Monitoring Rule, Round 5)
UCMR5 covers PFAS testing of municipal public water systems 2021–2025. We include this data for contextual county-level PFAS indicators but label it explicitly as municipal water system data. It is never merged with private well statistics.
USGS Aquifer Classification
We use the USGS National Aquifer Code dataset to assign principal aquifer type to each county. Aquifer type (carbonate, unconsolidated sand and gravel, crystalline rock, etc.) provides geological context for why certain contaminants are more prevalent in some areas.
U.S. Census Bureau (American Community Survey)
ACS 5-year estimates provide county-level housing age data used as a proxy for lead exposure risk. Housing units built before 1986 may contain lead solder in plumbing, which was federally banned that year.
How we process measurements
Normalization
All measurements from all sources are normalized to common units (µg/L for metals and organics) and mapped to canonical analyte names before storage. Variant names from different databases — for example, "Arsenic, dissolved," "Arsenic, total recoverable," and "As" — are all mapped to a single "arsenic" record.
Deduplication
Each sample is stored once. We prevent duplicates at two levels:
- Within-source: each data source provides a unique sample identifier (e.g., WQP's ActivityIdentifier). A second load of the same source is a no-op for records already in our database.
- Cross-source: state agencies commonly submit their data to WQP, so the same physical sample may appear in both a state loader and the WQP loader. When both sources share the same monitoring location identifier, we keep only one record.
Non-detect handling
A "not detected" lab result means the measured concentration fell below the method's detection limit — it is a real measurement, not missing data. We store non-detects explicitly and treat them differently depending on which statistic we are computing:
- Concentration statistics (median, mean, 90th percentile) are computed from detected samples only. This answers the question "when this contaminant is found, how much is typically present?" Including zeros from non-detects would suppress the median to near zero in counties with low detection rates, making the number misleading.
- Prevalence statistics (detection rate, % of wells exceeding the limit) use all samples as the denominator. A non-detect clearly does not exceed any limit, and the full sample count is needed to accurately describe how often contamination occurs.
- Distribution bands (Low / Moderate / High relative to the EPA limit) also use all samples. Non-detects correctly fall into the "Low" band.
This approach is consistent with EPA guidance for handling left-censored environmental data.
Geographic aggregation
All data is aggregated at the county level using 5-digit FIPS codes. Geographic assignments are made by associating each monitoring location's coordinates or reported county code with the corresponding county FIPS.
For each county and analyte we report:
- Median concentration — 50th percentile of detected measurements
- 90th percentile concentration — upper range of detected measurements
- Detection rate — percentage of all samples with a detected result
- % exceeding EPA limit — percentage of all samples above the MCL (or action level)
- Sample count — total measurements underlying the statistics
- Trend data — year-by-year median concentration and exceedance rate where sufficient data exists
Analytes covered
| Analyte | EPA Limit | Units | Limit Type |
|---|---|---|---|
| Arsenic | 10 | µg/L | Primary MCL |
| Nitrate | 10 | mg/L | Primary MCL |
| Lead | 15 | µg/L | Action level |
| Uranium | 30 | µg/L | Primary MCL |
| Fluoride | 4.0 | mg/L | Primary MCL |
| Iron | 300 | µg/L | Secondary MCL (aesthetic) |
| Manganese | 50 | µg/L | Secondary MCL (aesthetic) |
| Hardness | — | mg/L as CaCO₃ | No federal limit |
| PFOA | 0.004 | µg/L | Primary MCL (EPA 2024) |
| PFOS | 0.004 | µg/L | Primary MCL (EPA 2024) |
| PFNA | 0.010 | µg/L | Primary MCL (EPA 2024) |
| PFHxS | 0.010 | µg/L | Primary MCL (EPA 2024) |
| HFPO-DA (GenX) | 0.010 | µg/L | Primary MCL (EPA 2024) |
| Radon | — | pCi/L | No federal MCL (proposed 300 pCi/L) |
| E. coli / Coliform | 0 | presence/absence | Primary MCL |
Secondary MCLs (iron, manganese) are not enforceable for public systems and do not legally apply to private wells. We use them as benchmarks for nuisance-level concerns and label them as such.
Data limitations
- Monitoring wells are not private wells. Many records in public databases come from municipal, research, or regulatory monitoring wells. These may not represent the specific aquifer your private well draws from.
- County averages mask within-county variation. Two wells in the same county can have very different water quality depending on depth, aquifer, and local geology. County data is a regional indicator, not a prediction for any individual well.
- Not all private wells are sampled. Private wells are not required to report results to any public database. The records we aggregate represent wells sampled for regulatory, research, or remediation purposes — a biased subset, not a random sample.
- Detection limits vary. A non-detect result means concentration was below the lab's detection threshold for that method — not necessarily that the contaminant is absent. Detection limits vary across labs, methods, and decades.
- Data recency varies by analyte and county. PFAS data is largely post-2018. Arsenic and nitrate records in many counties span decades. We display the date range on each county page.
Update schedule
County-level aggregates are refreshed quarterly. Data is re-fetched from source APIs, re-normalized, deduplicated against existing records, and the county statistics are recomputed. The most recent update date is shown on each county page.
Housing age data from the U.S. Census Bureau (ACS 5-year estimates) is updated on the ACS release schedule, approximately annually.
Contaminant content standards
All health and regulatory information on contaminant guide pages is sourced from:
- U.S. Environmental Protection Agency (EPA) drinking water standards and health advisories
- IARC (International Agency for Research on Cancer) carcinogen classifications
- Peer-reviewed literature cited at the bottom of each contaminant guide page
- WHO drinking water quality guidelines
We do not accept payment to modify or favor any contaminant risk assessment. Our content does not reflect the interests of the water treatment industry.
Questions or corrections
If you find an error in our data or methodology, please use our contact page to get in touch.