Contaminant Data — All Analytes
38 well testing events on record, covering
19 analytes.
Data spans 1970 to 2022.
| Contaminant |
Samples |
Years |
Detection Rate |
Distribution LowModHigh |
vs. Limit |
vs. MI Avg |
|
Manganese
|
13 |
1978–2022
|
100% |
|
9% of limit
|
↓ 93% below
|
|
Iron
|
13 |
1978–2022
|
92% |
|
29% of limit
|
↓ 86% below
|
|
Arsenic
|
5 |
1978–2021
|
100% |
|
10% of limit
|
↓ 69% below
|
|
Chloride
|
13 |
1978–2022
|
100% |
|
2% of limit
|
↓ 81% below
|
|
Fluoride
|
13 |
1978–2022
|
100% |
|
2% of limit
|
↓ 75% below
|
|
HFPO-DA (GenX)
municipal
|
38 |
2023–2025
|
0% |
|
— |
— |
|
Lead
|
5 |
1978–2021
|
60% |
|
0% of limit
|
↓ 97% below
|
|
Nitrate
|
13 |
1978–2022
|
38% |
|
6% of limit
|
↑ 53% above
|
|
Nitrite
|
13 |
1978–2022
|
54% |
|
0% of limit
|
~ typical
|
|
PFHxS
municipal
|
38 |
2023–2025
|
0% |
|
— |
— |
|
PFNA
municipal
|
40 |
2008–2025
|
0% |
|
— |
— |
|
PFOA
municipal
|
42 |
2008–2025
|
0% |
|
— |
— |
|
PFOS
municipal
|
38 |
2023–2025
|
0% |
|
— |
— |
|
Sulfate
|
13 |
1978–2022
|
100% |
|
5% of limit
|
↓ 72% below
|
|
Uranium
|
1 |
1978
|
100% |
|
0% of limit
|
↓ 82% below
|
|
Sodium
|
13 |
1978–2022
|
100% |
|
— |
↓ 29% below
|
|
pH
|
13 |
1978–2022
|
100% |
|
— |
~ typical
|
|
PFBS
municipal
|
38 |
2023–2025
|
0% |
|
— |
— |
|
Hardness
|
13 |
1978–2022
|
100% |
|
— |
↑ 179% above
|
Distribution shows the share of samples in each concentration band relative to the EPA Maximum Contaminant Level (MCL): Low = below half the MCL, Moderate = between half and the MCL, High = above the MCL. Analytes without an MCL (e.g. sodium, pH) show — in the limit columns. State average is based on county median values across MI.
Public vs. Private Water in Berrien County
210
Active public water systems
133,064
Residents on public water
14%
Households on private wells
Public water systems in Berrien County are regulated by the EPA and must test and report contaminant levels. Private well owners are responsible for their own testing — there is no routine monitoring of private wells by any government agency.
Data Sources
This report aggregates data from the following public databases:
- Water Quality Portal (WQP) — groundwater sample results for Berrien County from EPA, USGS, and state agencies. Browse WQP data →
- EPA Safe Drinking Water Information System (SDWIS) — public water system inventory used to estimate the share of residents on private wells. Browse SDWIS →
- Census ACS 5-Year Estimates — well and cistern household counts (variable DP04_0074E) used to calculate the private-well percentage for Berrien County. Census data explorer →
- USGS Aquifer Atlas — principal aquifer type underlying Berrien County, which informs which contaminants are geologically likely. USGS Aquifer Atlas →
- EPA UCMR 5 (2023–2025) — PFAS occurrence data from public water systems, used as a proxy indicator for aquifer PFAS exposure. Rows marked "municipal" in the table above come from this source. About UCMR 5 →
- EPA Toxics Release Inventory (TRI) — industrial facility chemical releases near Berrien County, used in contamination risk scoring. TRI Explorer →
- EPA Superfund NPL — National Priorities List sites in or near Berrien County, used in contamination risk scoring. Superfund site search →
- CDC PLACES — county-level health outcome prevalence (kidney disease, cancer, heart disease) correlated with detected contaminants. About CDC PLACES →
Methodology
Raw records are downloaded from the Water Quality Portal and normalized to µg/L (ppb). Records are deduplicated by sample ID and date, and certified outliers are excluded. Analyte names are mapped to EPA canonical forms. Detection rates, distribution bands, and MCL comparisons are computed from the normalized dataset.
Distribution bands use the EPA Maximum Contaminant Level as the threshold: concentrations below 50% of the MCL are classed as Low, between 50% and 100% as Moderate, and above 100% as High. For analytes without an MCL (sodium, hardness, pH), distribution is not computed.
State comparison uses the median of county median values across all counties in MI with at least one sample for that analyte.
Last updated: 2026-07-15
Full methodology →