Research Asset

Private Well Statistics by State

By Lakeland Well Pump Repair Research · Dataset version 2026-07-23 ·

Lakeland Well Pump Repair Research is the independent research and reference section of lakelandwellpumprepair.com.

The newest peer-reviewed national estimate is 20,494,449 modeled U.S. housing units self-supplying domestic water in 2020 — 14.6% of all housing units. No federal source publishes a complete current field inventory of active private wells; the primary estimates measure different units, reference years, and methods.

Key statistics

What are the key private well statistics?

Every national-use figure below is an estimate or administrative-record count; none is a field inventory of active wells. Source, year, and unit are attached to each statistic.

  1. An estimated 20,494,449 U.S. housing units self-supplied their domestic water in 2020 — 14.6% of all housing units in the paper's detailed Results section. (Murray, Hall and Riveros-Iregui, Water Research, 2025)
  2. An estimated 45.8 million people relied on self-supplied water in 2020, compared with 37.9 million in the 1990 Census baseline used by the model. (Murray, Hall and Riveros-Iregui, Water Research, 2025)
  3. USGS estimated that 42.5 million people — 13% of the U.S. population — self-supplied their domestic water in 2015. (USGS Circular 1441, published 2018)
  4. Michigan led every state in the reproduced 2020 roll-up, with an estimated 1,445,617 housing units on domestic wells — 7.05% of the national model total. (Lakeland Well Pump Repair Research aggregation of EPA ORD block-group output, verified July 23, 2026)
  5. Pennsylvania led every state in the revised 2015 USGS population series, with an estimated 3.47 million self-supplied residents. (USGS county-level water-use data for 2015, version 2.0)
  6. Maine had the highest estimated self-supplied population share in 2015, at 50%. (USGS county-level water-use data for 2015, version 2.0)
  7. The top 10 states held 46.6% of all U.S. housing units estimated to be on domestic wells in the 2020 roll-up. (Lakeland Well Pump Repair Research aggregation, verified July 23, 2026)
  8. Self-supplied domestic withdrawals totaled an estimated 3,260 million gallons per day in 2015, and more than 98% came from groundwater. (USGS Circular 1441)
  9. The quality and safety of drinking water from private domestic wells are not regulated by the federal government under the Safe Drinking Water Act or by most state governments and laws. (U.S. EPA, page updated June 3, 2026)
  10. Florida ranked 6th in the reproduced 2020 state roll-up with an estimated 887,359 housing units on domestic wells, and 4th in the revised 2015 USGS series with an estimated 2.46 million self-supplied residents. (2020 aggregation verified July 23, 2026; USGS estimate year 2015)
  11. Polk County, Florida, had an estimated 58,822 self-supplied residents in 2015 — 9.05% of its population — and an estimated 19,207 housing units on domestic wells in the 2020 roll-up. (USGS estimate year 2015; Lakeland Well Pump Repair Research aggregation for 2020)
  12. SynDWell-US, published January 14, 2026, contains 5,982,715 digitized domestic-well records spanning 1680 through 2024 — records, not a complete count of active wells. (U.S. Geological Survey, DOI 10.5066/P13JP4JQ)
  13. A USGS study of 2,100 private wells found that water from about one in five contained at least one contaminant above a human-health benchmark. (USGS Circular 1332; summarized by EPA on its page updated June 3, 2026)
  14. Florida's Well Surveillance Program has sampled more than 54,500 wells since 2005; more than 5,700 had chemical concentrations above state and/or federal drinking-water standards. (Florida Department of Health, verified July 23, 2026)
  15. The 2020 model classified 93.14% of held-out Census blocks correctly, with reported sensitivity of 95% for the public-water class and specificity of 81% for the private-water class. (Murray, Hall and Riveros-Iregui, Water Research, 2025)

The two water-quality figures describe sampled wells and surveillance results, not prevalence across every private well. No complete national field inventory of active private wells was identified in the primary federal sources used for this page.

01

What do private well statistics actually count?

Private well statistics can refer to at least five different quantities, and they are not interchangeable. A household is not a housing unit, a housing unit is not a person, a water-withdrawal estimate is not a well count, and an administrative record does not establish that a well is active or unique.

A common error in published private-well coverage is treating one of these quantities as another. The crosswalk below defines the terms used throughout this page.

Table 1. Private well metric crosswalk
TermUnitSource exampleMeasured or estimatedActive-well count?Main limitation
Households relying on private wellsHouseholdsEPA, “more than 23 million”Agency estimateNoPage does not state the underlying reference year or denominator
Housing units self-supplyingHousing unitsWater Research 2025, 20,494,449Modeled estimateNoModel output from Census geography, housing units, service-area info, and well records — not a field inventory
Self-supplied populationPeopleUSGS 2015, 42.5 millionEstimated populationNoDoes not identify individual wells or addresses
Self-supplied withdrawalsMillion gallons per dayUSGS 2015, 3,260 Mgal/dEstimated volumeNoVolume does not establish the number of users or wells
Domestic-well recordsRecordsSynDWell-US, 5,982,715Administrative recordsNoCoverage limited by available state digital sources; a record does not establish that a well is active, unique, or part of a complete inventory

Source: Definitions from U.S. EPA, U.S. Geological Survey, Murray et al. (2025), and SynDWell-US. Compiled by Lakeland Well Pump Repair Research; verified July 23, 2026.

Private well versus domestic self-supplied water

“Domestic self-supplied water” is the federal water-use category used for the complete 2015 state series. USGS defines it as household water withdrawn directly by the user, typically from a private well or captured as rainwater in a cistern. More than 98% of self-supplied domestic withdrawals in 2015 came from groundwater, so the category is a strong private-well proxy — but not a perfect synonym.

Households, housing units, and people

EPA reports households. The 2020 model reports housing units. USGS reports people. A housing unit can be vacant; a household cannot; and one well can serve more than one housing unit. The figures should not be expected to match unless their years, definitions, denominators, and methods also match.

↑ Back to top

02

How many Americans use private wells?

There is no single defensible number because the primary sources measure different units and reference years. EPA's current page reports more than 23 million households and over 43 million people, the newest peer-reviewed model estimates 20,494,449 housing units and 45.8 million people for 2020, and USGS estimates 42.5 million self-supplied people for 2015.

The evidence can be compared, but it cannot be collapsed into one exact count of active wells.

Table 2. Current U.S. private-well evidence, compared
EvidenceReference yearValueUnitStatusWhat limits it
EPA statementPage updated June 2026More than 23 millionHouseholdsAgency estimate, stated as a floorNo stated reference year or denominator
EPA statementPage updated June 2026Over 43 million (“around 15%”)PeopleAgency estimate, floor and rounded shareNo reference year or denominator stated for the pair
Water Research model202020,494,449Housing unitsModeled estimateAbstract says 14.1%; detailed Results say 14.6%
Water Research model202045.8 millionPeopleModeled estimateDerived from modeled housing units, not enumerated
JAWRA model201022.9 million, reported as 23 millionHousing unitsModeled estimateEarlier method; not directly comparable to the 2020 model
JAWRA model201052.6 million (17%)PeopleModeled estimateEarlier method; not directly comparable to the 2020 model
USGS water use201542.5 million (13%)PeopleEstimated populationDifferent unit and method; complete revised state series used here
1990 decennial Census199015.1 millionHousing unitsCensus long-form survey estimateLast decennial Census to include a household water-source question
SynDWell-USRecords through 20245,982,715Digitized domestic-well recordsAdministrative recordsLimited by state digital data; not a complete active-well inventory

Sources: U.S. EPA (page updated June 3, 2026); Murray, Hall and Riveros-Iregui, Water Research 276:123171 (2025); Murray, Hall, Weaver and Kremer, JAWRA 57(5):828–843 (2021); USGS Circular 1441; USGS SynDWell-US (2026). Compiled and verified July 23, 2026.

What is the headline comparison?

The newer 2020 model is lower than the 2010 model: 20,494,449 versus 22.9 million housing units, and 45.8 million versus 52.6 million people — a difference of roughly 2.4 million modeled housing units and 6.8 million modeled people. This does not establish that private-well use fell by those amounts. The 2020 study introduced a new machine-learning classification method, additional explanatory data, community-water-system service-area information, and a different treatment of drilling records.

How should EPA's “around 15 percent” and “over 43 million” be read?

EPA pairs “around 15 percent” with “over 43 million” on its private-wells page, but it does not state the reference year or denominator used for that pair. “Over 43 million” is a lower-bound phrase, not a precise total. This page preserves EPA's wording rather than creating false precision from it.

↑ Back to top

03

What are the private well statistics by state?

The table below puts the reproduced 2020 modeled housing-unit roll-up beside the revised 2015 USGS population and withdrawal series. Michigan leads the 2020 housing-unit measure at 1,445,617, while Pennsylvania leads the 2015 population measure at 3.47 million.

Read the columns as separate estimates. Do not add them together, divide one by the other as a household-size estimate, or present the pair as a measured trend.

Table 3. Private well statistics by state — 2020 modeled housing units and 2015 self-supplied population, all 50 states and D.C.
State2020 est. housing units on domestic wellsRank (2020)Share of U.S. total (2020)2015 est. self-supplied peopleRank (2015)2015 share of state pop.2015 withdrawals (Mgal/d)2015 gal/person/day
Alabama470,165182.29%539,0002511%36.768
Alaska110,500420.54%189,0004026%11.460
Arizona275,359271.34%218,000383%24110
Arkansas59,582480.29%144,000445%12.889
California854,76584.17%1,440,000114%12788
Colorado253,524301.24%286,000365%35.4123
Connecticut201,111340.98%861,0001924%30.836
Delaware105,376430.51%182,0004119%14.580
District of Columbia00.00%0510%0n/a
Florida887,35964.33%2,460,000412%17772
Georgia795,802113.88%1,510,0001015%10469
Hawaii27,536500.13%56,300494%8.1144
Idaho179,953360.88%396,0003124%70.2177
Illinois517,031152.52%1,150,000169%92.180
Indiana609,795122.98%1,680,000725%12776
Iowa297,896251.45%494,0002716%3265
Kansas207,291321.01%149,000435%17.7118
Kentucky141,080390.69%441,0002910%22.150
Louisiana385,440201.88%492,0002811%39.380
Maine358,358211.75%660,0002150%31.648
Maryland480,538162.34%1,420,0001224%11480
Massachusetts271,789281.33%616,000239%35.157
Michigan1,445,61717.05%2,600,000226%18772
Minnesota609,238132.97%1,180,0001421%82.270
Mississippi228,078311.11%435,0003015%48.1111
Missouri561,531142.74%821,0002014%57.570
Montana189,620350.93%305,0003429%23.778
Nebraska126,356400.62%171,000429%19111
Nevada62,254470.30%193,000397%35.8186
New Hampshire271,286291.32%497,0002637%29.760
New Jersey297,309261.45%966,0001811%90.794
New Mexico117,084410.57%292,0003514%24.684
New York876,77474.28%2,500,000313%18775
North Carolina810,90193.96%2,410,000524%16970
North Dakota71,596460.35%46,400506%3.6980
Ohio970,52434.74%1,860,000616%13975
Oklahoma354,036221.73%357,000339%30.385
Oregon344,981231.68%632,0002216%73.9117
Pennsylvania1,064,93825.20%3,470,000127%20860
Rhode Island172,756380.84%113,0004711%6.5758
South Carolina470,631172.30%1,180,0001424%118100
South Dakota98,233440.48%129,0004515%5.5643
Tennessee395,320191.93%594,000249%42.872
Texas947,43144.62%1,320,000135%137104
Utah39,830490.19%61,400482%10.4169
Vermont177,796370.87%244,0003739%1145
Virginia805,583103.93%1,560,000919%12580
Washington315,315241.54%1,020,0001714%110107
West Virginia206,575331.01%392,0003221%31.380
Wisconsin894,09654.36%1,600,000828%76.448
Wyoming78,510450.38%119,0004620%8.9375
United States20,494,449100.00%42,451,10013%3,254.9576.7

Sources: The 2020 housing-unit columns were aggregated by Lakeland Well Pump Repair Research from the block-group output released with Murray, Hall and Riveros-Iregui, Water Research 276:123171 (2025), in the USEPA/ORD_Water_Source_2020 repository file Analysis/BG_Wells_Est_2020.dbf. The reproduced national sum equals the paper's detailed Results total of 20,494,449. The 2015 columns come from the revised USGS Estimated Use of Water in the United States County-Level Data for 2015, version 2.0 (DOI 10.5066/F7TB15V5), and USGS Circular 1441. Published state values are rounded, so displayed state sums differ slightly from the published national totals of 42.5 million people and 3,260 Mgal/d. Verified July 23, 2026.

Download the state dataset (CSV) — 51 rows, one per state and D.C., with state USPS and FIPS codes, 2020 modeled housing units, 2020 rank and share of the national model total, revised 2015 self-supplied population, 2015 rank and population share, withdrawals, per-person use, verification tier, source, and verification date.

What did the roll-up show?

  • The top five states held 26.0% of the national model total. Michigan, Pennsylvania, Ohio, Texas, and Wisconsin accounted for 5,322,606 modeled housing units.
  • The top 10 states held 46.6% of the total; the bottom 10 held 770,501 housing units — 3.8%. Among the 50 states, the two middle values were New Jersey at 297,309 and Iowa at 297,896.
  • The released table contained 92,680 Census block groups with a nonzero estimate. Every state had a nonzero roll-up; the District of Columbia had zero, matching its zero 2015 USGS self-supplied domestic estimate.
  • Rank order changed between the two measures. Connecticut ranked 19th on the 2015 population measure and 34th on the 2020 housing-unit measure; Texas ranked 13th and 4th, respectively. Oklahoma, Kansas, and Arizona each moved 11 places higher on the 2020 measure.

↑ Back to top

04

Which states have the highest private well reliance?

Reliance can be ranked by total estimated population, share of population, or modeled housing-unit count — and the measures produce different lists. Pennsylvania led the revised 2015 population series, Maine led the 2015 population-share series, and Michigan led the reproduced 2020 housing-unit roll-up. Each ranking belongs to its stated year and metric.

Table 4. States with the largest estimated self-supplied populations, 2015
RankStateEstimated self-supplied peopleShare of state population
1Pennsylvania3,470,00027%
2Michigan2,600,00026%
3New York2,500,00013%
4Florida2,460,00012%
5North Carolina2,410,00024%
6Ohio1,860,00016%
7Indiana1,680,00025%
8Wisconsin1,600,00028%
9Virginia1,560,00019%
10Georgia1,510,00015%

Source: Revised USGS county-level water-use data for 2015, version 2.0. Ranking derived by Lakeland Well Pump Repair Research; verified July 23, 2026.

Table 5. States with the highest self-supplied population share, 2015
RankStateShare of state population
1Maine50%
2Vermont39%
3New Hampshire37%
4Montana29%
5Wisconsin28%
6Pennsylvania27%
7Alaska26%
7Michigan26%
9Indiana25%
10Connecticut, Idaho, Maryland, North Carolina, and South Carolina24%

Source: Revised USGS county-level water-use data for 2015, version 2.0. Competition ranking preserves ties; verified July 23, 2026.

Table 6. States with the largest estimated 2020 housing-unit counts
RankStateEstimated housing units on domestic wells, 2020Share of U.S. total
1Michigan1,445,6177.05%
2Pennsylvania1,064,9385.20%
3Ohio970,5244.74%
4Texas947,4314.62%
5Wisconsin894,0964.36%
6Florida887,3594.33%
7New York876,7744.28%
8California854,7654.17%
9North Carolina810,9013.96%
10Virginia805,5833.93%

Source: Lakeland Well Pump Repair Research aggregation of the EPA ORD block-group output released with Murray, Hall and Riveros-Iregui (2025). National sum reproduced exactly; verified July 23, 2026.

Per-person use is a separate water-use-intensity measure, not a count or quality measure. In the revised version 2.0 data used here, Nevada had the highest 2015 value at 186 gallons per person per day, followed by Idaho at 177 and Utah at 169; Connecticut had the lowest reported state value at 36. These values do not measure well-water quality, reliability, or system condition.

↑ Back to top

05

Why do private well estimates disagree?

The estimates differ because they use different units, reference years, models, source versions, and rounding conventions. Where primary sources genuinely conflict, this page keeps the conflict visible instead of smoothing it into a single number.

Table 7. Documented differences and conflicts in the primary sources
TopicSource ASource BTreatment on this page
2020 share of housing unitsWater Research 2025 abstract: 14.1%Same paper's detailed Results: 20,494,449 housing units and 14.6%Publish 20,494,449 and 14.6% as arithmetically consistent with the reported model totals; disclose 14.1% in the abstract
2015 self-supplied populationUSGS OFR 2017–1131 uses 42.5 million in one summarySame report uses 42.6 million in a trend passageTreat as rounding within the source; use 42.5 million from the final USGS Circular 1441
Highest 2015 per-person useSuperseded OFR 2017–1131: Utah at 220 gal/person/dayRevised county dataset version 2.0: Utah 169 and Nevada 186Use revised version 2.0 and Circular 1441 data
EPA population wordingEPA: “around 15 percent”EPA: “over 43 million”Preserve both as EPA states them; do not infer a precise denominator or reference year that EPA does not publish on the page
Households versus housing unitsEPA: more than 23 million householdsWater Research 2025: 20.5 million housing units in 2020Do not reconcile; different unit and different stated reference year

Sources: U.S. EPA; Murray et al. (2025); USGS OFR 2017–1131; USGS Circular 1441; revised USGS county-level data for 2015, version 2.0. Verified July 23, 2026.

How do reference years differ?

The complete revised state population series is from 2015. The housing-unit model is for 2020. SynDWell-US records extend through 2024. EPA's private-wells page was updated in June 2026, but the page does not assign a reference year to its national household and population estimates.

How do the methods differ?

The 2015 USGS figures are water-use estimates assembled at county level and aggregated to states. The 2020 figure comes from a machine-learning classification model using Census geography, housing units, 1990 water-source information, public-water-system service-area information, and domestic-well records. The 1990 value came from a decennial Census long-form household water-source question.

What is the defensible way to read the series?

Table 8. Timeline of U.S. private-well evidence
YearValueUnitMethodDirectly comparable to the row above?
199015.1 millionHousing unitsDecennial Census long-form survey estimateBaseline
201022.9 million housing units; 52.6 million peopleHousing units and peopleHousing-unit-change model using 1990 as baselineNo — modeled update rather than a repeat Census measurement
201542.5 million peoplePeopleUSGS county and state water-use estimatesNo — different unit and method
202020,494,449 housing units; 45.8 million peopleHousing units and peopleMachine-learning Census-block classificationNo — different model and reference year
Records through 20245,982,715Digitized domestic-well recordsAdministrative-record synthesisNo — records, not users or active wells

Sources: 1990 and 2010 values from Murray et al. (2021); 2015 from USGS Circular 1441; 2020 from Murray et al. (2025); records through 2024 from SynDWell-US. Compiled and verified July 23, 2026.

Do not plot these rows as one continuous trend without preserving their different units and methods. They are a chronology of evidence products, not five equivalent measurements.

↑ Back to top

06

Why aren't well records a count of active wells?

SynDWell-US contains 5,982,715 digitized domestic-well records spanning 1680 through 2024, and the United States Groundwater Well Database contains 14,260,752 records for wells of all purposes spanning 1763 through 2023. Both products describe records assembled from available sources; neither is a complete field inventory of active private wells. Records inherit each state's reporting history, update practices, coverage, status labels, and duplicate-handling problems.

Table 9. Modeled estimates versus administrative records
DatasetValueTime coverageWhat the value representsWhat cannot be inferred
USGS water use, 201542.5 million peopleCalendar year 2015Estimated self-supplied domestic populationNumber or location of active wells
Water Research model, 202020,494,449 housing unitsDecennial year 2020Modeled housing units classified as self-supplying domestic waterAddress-level status or an exact active-well count
USGS domestic self-supply model releaseRaster grids and modeled estimatesDecennial year 2020Modeled household counts, locations, and withdrawals for the conterminous U.S.Parcel-level certainty or field-enumerated well counts
SynDWell-US5,982,715 domestic-well records1680–2024Digitized domestic-well records from states with available digital sourcesThat each record is active, unique, or part of a complete state inventory
USGWD14,260,752 records of all well purposes1763–2023Standardized state and federal groundwater-well recordsThat all records are domestic, active, unique, current, or complete

Sources: USGS Circular 1441; Murray et al. (2025); USGS DOI 10.5066/P13BILQT; USGS SynDWell-US, DOI 10.5066/P13JP4JQ; Lin et al., Scientific Data 11:335 (2024). Compiled and verified July 23, 2026.

USGWD reports 8,338,041 records as active, 1,411,512 as inactive, and 4,511,199 — 31.6% — as unknown. The same study flags 1,421,235 of its 14,260,752 records as possible duplicates and states that coverage, duration, completeness, quality, and update practices vary by state.

↑ Back to top

07

Are private wells regulated?

Private household wells are not regulated by the federal government under the Safe Drinking Water Act. Federal drinking-water standards apply to public water systems, generally systems that provide water for human consumption through constructed conveyances to at least 15 service connections or regularly serve at least 25 people for at least 60 days per year.

State and local rules can still govern well contracting, construction, permitting, location, abandonment, and testing. In Florida, the Department of Environmental Protection regulates water-well contracting, permitting, and construction under Chapter 373, Part III, Florida Statutes, and Chapters 62-531 and 62-532, Florida Administrative Code. A person engaged in constructing, repairing, or abandoning wells as a business in Florida must be a Florida-licensed water-well contractor.

Water quality remains the practical consequence of the regulatory split. EPA cites a USGS study of 2,100 private wells in which water from about one in five contained at least one contaminant above a human-health benchmark. Florida's Department of Health recommends that private-well owners test for bacteria and nitrate at least once each year. Well construction, repair, abandonment, disinfection, confined-space work, pressurized-system work, and electrical work should be handled under the applicable rules by qualified professionals.

↑ Back to top

08

What are Florida private well statistics by county?

The revised 2015 USGS county rows sum to an estimated 2,460,295 self-supplied Florida residents — 12.14% of the state population — and 176.92 million gallons per day. The reproduced 2020 roll-up totals 887,359 modeled housing units on domestic wells. Gilchrist, Union, and Lafayette counties had the three highest revised 2015 self-supplied population shares. These are estimates from two different federal evidence products, not current county permit or active-well counts.

What does Florida's own agency say?

The Florida Department of Health's statewide page says about 88% of Florida residents are served by public water systems covered by the Safe Drinking Water Act, while the other 12% receive water from the combined category of limited-use public water systems and private wells. The Polk County page separately says about 20% receive water from that combined category. Because both statements combine limited-use public systems with private wells, neither can be read as a private-well-only count.

The revised 2015 USGS self-supplied estimate is 12.14%, numerically close to the statewide Florida page but based on a different federal water-use definition and year. The figures should remain labeled rather than treated as interchangeable.

Table 10. Florida private well statistics by county — 2015 population and 2020 modeled housing units, all 67 counties
RankCounty2015 est. self-supplied people2015 share of county pop.2015 withdrawals (Mgal/d)2015 gal/person/day2020 est. housing units on domestic wells
1Lee175,83725.05%11.466510,605
2Duval144,32215.81%12.738873,440
3Marion141,16141.12%11.868464,025
4Hillsborough126,5239.38%7.115630,568
5Palm Beach120,4988.47%6.25528,222
6Osceola89,86127.74%6675,157
7Pasco80,20616.11%4.685826,208
8Lake74,38722.83%9.5512830,808
9Clay73,69936.13%6.869327,988
10Volusia71,89413.88%6.659239,953
11Orange70,5535.48%3.625121,715
12Miami-Dade64,3722.39%2.073216,959
13Collier61,71017.27%3.966410,225
14Polk58,8229.05%3.225519,207
15St. Johns54,44524.02%2.885319,040
16Alachua52,32420.13%4.127919,541
17Putnam51,50271.51%2.474824,557
18Brevard47,9298.44%2.795845,865
19Columbia47,28469.18%3.727912,925
20Leon47,19816.49%5.4211514,423
21Citrus44,26931.38%4.169428,521
22Sarasota42,60010.50%1.52367,846
23Hernando38,85521.77%3.2829,510
24Suwannee35,76881.74%2.747710,678
25Levy31,66779.50%1.966210,727
26Jackson31,44364.70%3.129913,568
27Nassau31,23539.82%3.8612415,978
28St. Lucie29,3729.84%1.52524,293
29Seminole25,2675.63%1.997920,187
30Bay23,92313.17%2.39612,433
31Manatee23,6766.52%0.6252,467
32Baker20,99076.55%2.121015,644
33Santa Rosa20,40712.22%0.49247,771
34Escambia20,2706.52%0.87433,814
35Bradford19,74673.33%1.981006,175
36DeSoto19,50855.02%1.47756,222
37Sumter19,30016.23%2.511308,398
38Wakulla18,24057.84%0.86476,563
39Highlands17,29617.38%1.21709,093
40Okaloosa16,3618.24%0.5315,805
41Okeechobee16,14240.90%1.39865,494
42Washington16,07165.10%1.661037,586
43Gilchrist15,23488.57%1.29854,472
44Gadsden14,63131.78%1.471008,154
45Charlotte14,4078.32%0.42291,384
46Hendry13,93935.63%1.06764,851
47Madison13,47173.18%1744,655
48Dixie13,40582.73%0.98735,012
49Union13,31487.40%1.16872,589
50Holmes13,25468.59%1.28976,529
51Hardee11,65142.36%0.5434,012
52Martin11,6457.45%0.45395,182
53Taylor11,60151.58%0.91784,931
54Calhoun10,09769.82%0.91904,045
55Flagler9,8539.35%0.3304,155
56Hamilton9,64367.46%0.74772,634
57Pinellas9,4250.99%0.242592
58Indian River9,0986.15%0.22216,685
59Jefferson8,81162.57%0.81924,805
60Broward7,8530.41%0.64813,893
61Lafayette7,44785.96%0.65872,230
62Walton7,37311.61%0.9212511,946
63Glades6,68548.90%0.48721,795
64Liberty4,85258.24%0.46952,201
65Monroe3,2764.23%0046,444
66Gulf1,58810.01%0.483023,052
67Franklin8096.88%0.121481,407
Florida total2,460,29512.14%176.9271.9887,359

Sources: The 2015 columns come from the revised USGS Estimated Use of Water in the United States County-Level Data for 2015, version 2.0 (DOI 10.5066/F7TB15V5). The 2020 column was aggregated by Lakeland Well Pump Repair Research from EPA ORD block-group output released with Murray et al. (2025). Counties are ranked by revised 2015 estimated self-supplied population. Polk County is highlighted. Verified July 23, 2026.

Download the Florida county dataset (CSV) — 67 rows with county FIPS, 2015 total population, estimated self-supplied population and share, withdrawals, per-person use, four Florida rankings, 2020 modeled housing units, verification tier, source, and verification date.

What does the cross-dataset diagnostic ratio show?

For each Florida county, we divided the revised 2015 self-supplied population estimate by the 2020 modeled housing-unit estimate. The median of the 67 ratios is 2.82. Because the numerator and denominator come from different years and methods, the ratio is a diagnostic for locating divergence — not a household-size estimate, validation statistic, or proof that one dataset is correct.

Table 10A. Selected Florida cross-dataset diagnostic ratios
County2015 self-supplied people2020 est. housing unitsRatioDescriptive result
Pinellas9,42592102.4Very large divergence; 2015 people are high relative to the 2020 housing-unit estimate
Osceola89,8615,15717.4Large divergence in the same direction
Lee175,83710,60516.6Large divergence in the same direction
Palm Beach120,4988,22214.7Large divergence in the same direction
Polk58,82219,2073.06Near the statewide median ratio of 2.82
Indian River9,09816,6850.552020 housing units are high relative to the 2015 population estimate
Monroe3,27646,4440.07Extreme divergence in the reverse direction

Source: Ratios calculated by Lakeland Well Pump Repair Research from the revised 2015 USGS county rows and the reproduced 2020 EPA ORD roll-up. The ratio is descriptive only; verified July 23, 2026.

The ratio identifies counties where the two products diverge sharply. It does not identify the cause, which could involve source definitions, model classification, administrative boundaries, public-supply reporting, actual change between 2015 and 2020, or more than one factor.

↑ Back to top

09

What are Polk County private well statistics?

Polk County had an estimated 58,822 self-supplied residents in 2015 — 9.05% of its 650,092 residents — using an estimated 3.22 million gallons per day, or 55 gallons per person per day. The reproduced 2020 roll-up contains an estimated 19,207 housing units on domestic wells. Their cross-dataset ratio is 3.06, near the Florida median of 2.82, but the ratio remains descriptive because the values come from different years and methods.

Polk ranks 14th among Florida's 67 counties by revised 2015 self-supplied population and 52nd by share.

Table 11. Polk County, Florida — private-well and self-supplied-water statistics
MetricValueYearRank among Florida's 67 counties
Total population650,0922015
Estimated self-supplied population58,822201514th
Share of county population self-supplied9.05%201552nd
Self-supplied domestic withdrawals3.22 Mgal/d201518th
Self-supplied use per person55 gal/day201548th
Estimated housing units on domestic wells19,207202015th

Sources: Revised USGS county-level water-use data for 2015, version 2.0; 2020 figure aggregated by Lakeland Well Pump Repair Research from EPA ORD block-group output. Verified July 23, 2026.

Polk ranks much higher by total self-supplied population than by population share because it is a populous county in which most residents are estimated to use public supply. Both measures are necessary to describe the county accurately.

How does Polk County well permitting work?

The Florida Department of Health in Polk County describes well-construction permitting as a two-part process involving the health department and the Southwest Florida Water Management District. Its page directs applicants to schedule a pre-construction site inspection before drilling a public drinking-water well so the proposed location can be checked against setback requirements. It also states that well drilling in eastern Polk County is handled by the Florida Department of Health in Osceola County.

The DOH-Polk Environmental Public Health office is at 2090 E. Clower St., Bartow, Florida 33830. None of the statistical figures on this page is a current count of permitted or active Polk County wells, and no current official active-well inventory was identified in the primary sources reviewed for this edition.

For Polk County well service, see Lakeland and Polk County service area and well repair and casing work.

↑ Back to top

10

How were these private well statistics compiled?

The source review and data build covered July 19 through July 23, 2026. Consequential figures were taken from issuing agencies, official data releases, original research papers, and calculations reproduced from those source files; no numerical claim is sourced to a commercial blog or general-interest aggregator. The methodology keeps estimates, survey results, model output, water volumes, and administrative records in separate fields.

How was the 2020 state and county roll-up produced?

The 2020 study models water source at Census-block level, and its public EPA ORD repository includes a block-group result table at Analysis/BG_Wells_Est_2020.dbf. Lakeland Well Pump Repair Research read the block-group table, assigned each GEOID to state and county using the first two and five digits, and summed the Wells estimate within each geography.

The source table contains 92,680 block groups with a nonzero estimate. The resulting 51-jurisdiction sum is 20,494,449 modeled housing units, exactly matching the total printed in the paper's detailed Results section. That exact match is the integrity gate for the 2020 state and county values on this page.

The paper also reports 120,003,285 modeled public-supply housing units. The private and public model totals sum to 140,497,734, while the 2020 Census counted 140,498,736 housing units — a difference of 1,002 units, or about 0.0007%. This is an additional aggregate check, not proof that every state or county estimate is correct.

This release does not publish a modeled 2020 percentage of all housing units by state because state housing-unit denominators were not joined to the roll-up. It publishes the modeled count, rank, and share of the national model total.

How was the revised 2015 benchmark checked?

The 2015 values come from the USGS county-level water-use release for 2015, revised June 19, 2018 as version 2.0, and USGS Circular 1441. USGS states that some public-supply and domestic values were updated from version 1.0. The 51 displayed state rows sum to 42,451,100 people, compared with the published rounded national estimate of 42.5 million. The withdrawal rows sum to 3,254.95 Mgal/d, compared with the published rounded national estimate of 3,260 Mgal/d. The 67 Florida county rows sum to 2,460,295 people and 176.92 Mgal/d. All displayed state and county ranks, shares, sums, and per-person arithmetic were recalculated during this audit.

What do the verification tiers mean?

  • Read directly against a primary publication or official output, or reproduced by aggregating official source output and validating the published national total. This applies to the 2020 housing-unit roll-up, current agency-page statements, and figures quoted directly from the original research and official data releases.
  • Compiled from the revised official USGS version 2.0 series and validated against official national and state totals, anchor values, row arithmetic, and recalculated rankings. This applies to the 2015 state and Florida county rows.

How were derived values calculated?

Rankings use competition ranking. Florida county shares equal estimated self-supplied population divided by total population. The Florida diagnostic ratio equals the revised 2015 estimated self-supplied population divided by the reproduced 2020 modeled housing-unit count. No conflicting national figures were averaged together. Source conflicts remain visible with their original units, reference years, and status.

What does the 2020 model actually do?

The study uses a decision-tree classifier trained with complete community-water-system service-area information from California, Connecticut, and New Jersey and applies additional public- and private-water evidence nationally. It uses the 1990 Census water-source baseline, 2020 housing units, service-area boundaries, and approximately 3.2 million geolocated domestic-well records from 33 states for wells drilled between 1990 and 2020.

The paper reports 93.14% overall held-out-block accuracy, 95% sensitivity for the public class, and 81% specificity for the private class. The EPA dataset description states that the output should not be used to convey exact counts of well or public-water users. It is model output, not a physical census of wells.

↑ Back to top

11

What does this data show — and not show?

The page compares primary-source estimates and administrative records; it does not create a national inventory of active wells. Its strongest use is state and county comparison with the unit, reference year, method, and limitation kept visible.

  1. “Private well” and “domestic self-supplied water” are closely related but not identical. Self-supplied water can include cisterns and other uncommon private sources.
  2. A household, housing unit, person, water-withdrawal estimate, and well record are different units. One well can serve multiple housing units.
  3. The complete revised state population series is from 2015.
  4. The 2020 figures are model output. The study reports 81% specificity for the private-water class and the official dataset warns against presenting the results as exact user counts.
  5. Administrative-record counts inherit historic, inactive, unknown-status, duplicate, incomplete, and unevenly digitized state data.
  6. Rankings belong to the stated year and metric. A state's position can change when the unit, method, or reference year changes.
  7. None of the state or county data measures water quality, well condition, pump condition, permit status, or regulatory compliance.
  8. The Florida cross-dataset ratio is descriptive. It cannot establish household size, validate either source, or explain why two county estimates diverge.
Table 12. Figures and claims this page does not publish as established facts
Figure or claimPublication treatmentReason
Current exact number of active private wells in the United StatesNot publishedNo complete current federal field inventory was identified
Exact number of active wells in any state or Florida countyNot publishedThe datasets measure modeled housing units, estimated people, withdrawals, or administrative records
2020 percentage of all housing units on wells by stateNot includedState housing-unit denominators were not joined to this release's roll-up
A precise national trend from 1990 through 2020Not publishedThe evidence products use different survey and model methods
Causes of county-level divergence between the 2015 and 2020 estimatesNot inferredThe descriptive ratio cannot identify causation
SynDWell-US domestic-well record counts by stateNot includedThis edition uses the verified national release total and did not aggregate its state records
Current active or permitted well count for Polk CountyNot publishedNo current official inventory was identified in the reviewed primary sources
"500,000 new wells drilled annually" or "2.5% annual growth"RejectedNo authoritative primary source was located for the repeated claims
"First," "only," or "most complete" description of this page's datasetRejectedThe superlatives were not established and are not needed

Source: Publication decisions recorded by Lakeland Well Pump Repair Research after primary-source review and data validation, July 23, 2026.

↑ Back to top

12

Where can the dataset be downloaded?

The three files below are the data products described on this page, versioned 2026-07-23.

Download the dataset

License: CC BY 4.0 for the original aggregation, rankings, and explanatory text; underlying federal data and research are public-domain or CC0 government products.

Table 13. Dataset file checksums
FileRows or contentsSHA-256
private-well-statistics-by-state.csv51 jurisdiction rowse66839ee509d2863c245cbc29335de0ee1b53a76cdf8220032e78b3d6e9974d8
private-well-statistics-florida-counties.csv67 Florida county rows12de2e6e826c12979acba8115e17df0adee85a810d2fd1e1cc6df446309c0919
private-well-evidence-matrix.jsonEvidence matrix and source registry3b0dc14b711530770da6713731ff7c253112207abf460f3990d2c9a81c6e8705

Source: SHA-256 hashes generated from the files distributed with dataset version 2026-07-23.

↑ Back to top

13

How should this page be cited?

This block supplies neutral publication metadata for editors, researchers, and reference managers. It does not alter the citation requirements of the underlying sources, which are listed separately below.

Publication: Lakeland Well Pump Repair Research

Page title: Private Well Statistics by State

URL: https://lakelandwellpumprepair.com/research/private-well-statistics-by-state/

Dataset version: 2026-07-23

Last verified:

Reference format:

Lakeland Well Pump Repair Research. “Private Well Statistics by State.” Dataset version 2026-07-23. Last verified July 23, 2026.

Primary underlying sources: U.S. EPA; Murray, Hall and Riveros-Iregui, Water Research 276:123171 (2025); U.S. Geological Survey Circular 1441; USGS county-level water-use data for 2015 (DOI 10.5066/F7TB15V5); and the agencies, data releases, and publications listed in the primary sources section.

↑ Back to top

14

What are the most common questions about private well statistics?

These answers use the same definitions, years, and source boundaries as the tables above. None converts a household, housing unit, person, withdrawal, or record into an exact active-well count.

How many private wells are there in the United States?

No federal source used for this page publishes a current, complete field inventory of active private wells. EPA reports more than 23 million households relying on private wells; a 2025 peer-reviewed model estimated 20,494,449 housing units self-supplying water in 2020; and SynDWell-US contains 5,982,715 digitized domestic-well records. Those are different quantities, and none is an exact count of active wells.

What percentage of Americans use private wells?

EPA's page says around 15 percent of the U.S. population — over 43 million people — rely on private wells. USGS estimated 13% of the population was self-supplied in 2015, while the 2025 model estimated 14.6% of housing units and 45.8 million people were self-supplied in 2020. The figures use different years, units, denominators, and methods.

Which state has the most private well users?

It depends on the measure. Pennsylvania had the largest estimated self-supplied population in the revised 2015 USGS series, at 3.47 million people. Michigan had the largest estimated number of housing units on domestic wells in the 2020 roll-up, at 1,445,617.

Which state has the highest percentage of residents on well water?

Maine had the highest estimated self-supplied population share in the revised 2015 USGS series, at 50%. Vermont followed at 39% and New Hampshire at 37%. These are self-supplied-water estimates for 2015, not current field counts of private-well users.

Why does the page include both 2015 and 2020 state data?

The revised 2015 USGS series provides the complete state population and water-withdrawal benchmark used here. The newer 2020 model provides estimated housing units rather than a directly comparable population series. Showing both preserves the strongest available state-level evidence without treating unlike measures as one trend.

Does one well record equal one active well?

No. A record can describe a construction, status, or other administrative event, and national compilations inherit differences in state coverage, updates, completeness, and duplicate handling. In USGWD, 1,421,235 of 14,260,752 records were flagged as possible duplicates and 4,511,199 records had unknown activity status.

Are private wells tested for contaminants?

Private household wells are not covered by routine federal Safe Drinking Water Act monitoring. EPA cites a USGS study of 2,100 private wells in which water from about one in five contained at least one contaminant above a human-health benchmark. In Florida, the Department of Health recommends testing private-well water for bacteria and nitrate at least once each year.

How many people in Florida use private wells?

The revised 2015 USGS county data sum to an estimated 2,460,295 self-supplied Florida residents, or 12.14% of the state's population. Self-supplied domestic water is a strong private-well proxy, but it is not a private-well-only field count. Florida's Department of Health separately says 12% of residents receive water from the combined category of limited-use public systems and private wells.

↑ Back to top

15

What are the primary sources?

Every consequential external figure on this page traces to an issuing agency, official data release, statute, or original research publication.

  1. U.S. Environmental Protection Agency. Private Drinking Water Wells. Page updated June 3, 2026; accessed July 23, 2026.
  2. Murray, A., Hall, A., and Riveros-Iregui, D. (2025). A machine learning approach to estimate domestic use of public and private water sources in the United States. Water Research, 276, 123171.
  3. U.S. EPA Office of Research and Development. ORD_Water_Source_2020 code and output repository. Block-group output used for aggregation; accessed July 23, 2026.
  4. U.S. EPA. United States Estimated Water Sources — 2020 Census Blocks. EPA dataset record, DOI 10.23719/1531806; accessed July 23, 2026.
  5. Murray, A., Hall, A., Weaver, J., and Kremer, F. (2021). Methods for Estimating Locations of Housing Units Served by Private Domestic Wells in the United States Applied to 2010. Journal of the American Water Resources Association, 57(5), 828–843.
  6. Dieter, C.A., Maupin, M.A., Caldwell, R.R., Harris, M.A., Ivahnenko, T.I., Lovelace, J.K., Barber, N.L., and Linsey, K.S. (2018). Estimated Use of Water in the United States in 2015. USGS Circular 1441.
  7. Dieter, C.A., and Linsey, K.S. (2018). Estimated Use of Water in the United States County-Level Data for 2015. Revised June 19, 2018, version 2.0.
  8. Dieter, C.A., and Maupin, M.A. (2017). Public Supply and Domestic Water Use in the United States, 2015. USGS OFR 2017–1131. Superseded by Circular 1441; retained to document version and rounding differences.
  9. McCarthy, P.M., Larsen, J.D., Ransom, R.K., Dieter, C.A., and Niswonger, R. (2026). Synthesized Domestic Well Database for the Conterminous United States (SynDWell-US). USGS data release, published January 14, 2026.
  10. Ely, C.P., Niswonger, R., Dieter, C.A., Ransom, R.K., McCarthy, P.M., and Larsen, J.D. (2025). Domestic self-supplied household locations, counts, and withdrawal estimates in the conterminous United States. USGS model and dataset for decadal year 2020, published September 30, 2025.
  11. Lin, C.-Y., Miller, A., Waqar, M., and Marston, L.T. (2024). A database of groundwater wells in the United States. Scientific Data, 11, 335.
  12. U.S. Census Bureau. Housing Characteristics: 2020 and 2020 Census population and housing release. National housing-unit total checked July 23, 2026.
  13. U.S. Geological Survey. Domestic Water Use. Definitions and 2015 national estimates; accessed July 23, 2026.
  14. Florida Department of Health. Private Well Water Testing. Statewide page; accessed July 23, 2026.
  15. Florida Department of Environmental Protection. Water Well Construction Program. Page modified July 8, 2026; accessed July 23, 2026.
  16. Florida Department of Health in Polk County. Well Permitting. Permitting roles and Bartow office address; accessed July 23, 2026.
  17. Florida Department of Health. Well Surveillance Program. Survey, sampling, and exceedance totals; accessed July 23, 2026.
  18. DeSimone, L.A., Hamilton, P.A., and Gilliom, R.J. (2009). Quality of Water from Domestic Wells in Principal Aquifers of the United States, 1991–2004. USGS Circular 1332.
  19. U.S. Environmental Protection Agency. Information about Public Water Systems, and U.S. Code, 42 U.S.C. §300f. Safe Drinking Water Act definitions; accessed July 23, 2026.

Last verified: July 23, 2026 · Byline: Lakeland Well Pump Repair Research

Lakeland Well Pump Repair Research is the independent research and reference section of lakelandwellpumprepair.com.

↑ Back to top