Methodology

This page explains exactly how a public agency record becomes a number on this site. If something here doesn’t match what you see, that’s a bug. Please tell us.

Sources

Every record comes from a public well-construction layer published by a Florida water management district. We read those layers from a statewide snapshot (v2026-10-25): non-personal columns only, rebuilt about monthly. We don’t buy data or scrape other sites. Each well still links to its district record. South Florida household wells are mostly permitted by counties and cities, and those permits are not in the district layers.

We don’t use U.S. Geological Survey groundwater sites: they’re mostly observation wells with no use recorded. Boundaries for counties, ZIP codes and towns are U.S. Census Bureau 2020 TIGER/Line shapes from TIGERweb. Addresses are matched with the free U.S. Census Bureau Geocoder. This product uses the Census Bureau Data API but is not endorsed or certified by the Census Bureau.

How records reach a county

1. The snapshot carries the fields we need for depth (depth, casing, use, date, location method, record id) and never owner names, street addresses or contractor names.
2. County by location, not by label. Most snapshot records don’t name a county, so we place every record in a county from its map point and the Census county boundaries. Of 1,715,247 records, 1,714,543 had no county in the snapshot. 1,714,869 fell inside a Florida county; 378 (offshore or across the state line) were dropped. Where the snapshot did name a county, the boundary agreed 704 times and disagreed 0.
3. Records are filtered and normalized (below). Each one is then placed in a ZIP code area and a town the same way, from Census boundaries. The snapshot’s own ZIP column is empty in this version, and we wouldn’t rely on it anyway.
4. Every county with at least one record left after filtering gets a page. Fewer than 100 water-supply wells is marked thin coverage. The coverage page lists each county’s counts.

What counts as a well here

We keep records that have a reported depth and a usable location, and whose use is a water supply: household (domestic), irrigation, public supply, agricultural (livestock, groves, aquaculture), other supply (fire protection, industrial) or use not reported. The districts record use as free text; the snapshot maps it to a short list (domestic, irrigation, public supply, agricultural, livestock, monitoring, injection, other, unknown), and we map that list to ours with published, tested rules (scripts/ingest/snapshot.ts). Livestock counts as agricultural. “Other” (which includes HVAC supply and dewatering wells) is stored as other supply and never counted in typical depth.

Of 1,714,386 agency records inside the counties with pages, these were left out:

Wells never built. Stations the SJRWMD, NWFWMD and SRWMD layers mark as “proposed” have a permit but no completion report, so their depth is a plan, not a well. We leave them out, as we do SWFWMD sites that were cancelled or denied.

SWFWMD records with no use. The snapshot files SWFWMD abandonment, plugging and cancelled permits under “use unknown” and doesn’t carry the site status that would tell them apart from built wells. We checked against the full district records for Citrus and Hernando: fewer than 1 in 20 of these were real wells. So we leave the whole group out. It never counted toward typical depth anyway.

Duplicates. SWFWMD files a new permit for each repair or replacement on a well site. The snapshot doesn’t carry the site id, so a well with several permits can appear more than once. Each permit is a real filing with its own depth. Records that repeat exactly are kept once. The other districts list one record per station. We don’t merge records across districts; near district lines one well could appear in two, though this is rare.

Typical depth

• Median, not average. Half of the wells are shallower than the median and half are deeper. A few very deep public-supply wells can’t drag it around the way they would an average.
• Middle half is the range from the 25th to the 75th percentile. Percentiles use linear interpolation between ranks (the same method as spreadsheet PERCENTILE.INC).
• Household wells first. The headline figure uses domestic wells when there are at least 20 of them, because that’s what most people planning a well need. If there are fewer, it uses all water-supply wells when those reach 20, so a handful of household wells can’t outvote hundreds of others. Below that, the same order applies with a minimum of 5. Irrigation wells are often much shallower; in Flagler County, for example, most records are shallow lawn-irrigation wells. “All water-supply wells” means domestic, irrigation, public supply and agricultural. Wells with no reported use are listed but never counted in the headline.
• Small samples. Below 5 wells we don’t state a typical depth. Below 20 we label the figure “limited data”.
• Rounding. Summary depths of 100 ft or more are rounded to the nearest 5 ft; individual wells show the depth as reported (to the foot).
• Histograms use round depth bands; the deepest 5% or so are folded into one “or deeper” band so the chart stays readable.

Area pages

County pages cover every county on the coverage page. ZIP code and town pages are published only when they have at least30 water-supply wells; smaller areas are listed on their county page without their own page, so we never publish a page with nothing useful on it. ZIP codes are Census ZIP Code Tabulation Areas (ZCTAs), which closely match but aren’t identical to USPS ZIP codes. Towns are Census incorporated places and census-designated places. When part of a county, ZIP or town lies outside our coverage (in a county without a page, or in the South Florida district), its page says so.

Location precision

Distances in the lookup are straight-line distances from your search point. How precisely a well is placed depends on the source:

• SWFWMD: the layer doesn’t say how each point was located.
• SJRWMD: the snapshot says when a point came from GPS, an address match or a map. Other methods (surveys, section centers) show as “method not reported”.
• NWFWMD: the snapshot says when a point came from GPS, an address match or a map. Other methods (surveys, section centers) show as “method not reported”.
• SRWMD: the snapshot says when a point came from GPS, an address match or a map. Other methods (surveys, section centers) show as “method not reported”.
• SFWMD: the layer doesn’t say how each point was located.
• Records placed only at the center of a survey section (Section centroid (approx. ±0.5 mi)) are marked “≈” in the lookup when the source says so. The snapshot doesn’t, so most section centers are caught by the shared-point rule below instead.
• Shared points. Many records sit on exactly the same coordinates as other records from the same source. When 5 or more share a point, we treat the location as approximate: it’s probably a section, subdivision or parcel-group center, not the well. The lookup marks these “≈” and draws them as one larger dot. Share of records on shared points, by county: from 0% (Broward) to 58% (Citrus), median 20%.

Because of this, the distance to the single nearest well is often less reliable than it looks. The typical depth across a 1-mile or wider radius holds up much better, which is why the lookup leads with it.

Known limits

• Not every well is here. Wells built before permitting, or never reported, are missing. Older wells are under-represented.
• Depth as reported. We show the depth the agency record carries. We can’t verify it in the field.
• No dates outside SWFWMD. The SJRWMD, NWFWMD and SRWMD layers have no construction or permit date, so we can’t say how recent those wells are.
• South Florida is mostly missing. Outside Monroe, Charlotte and large Broward wells, South Florida well permits are issued by counties and cities, and those records aren’t in the district layer. Counties there either have no page or carry a partial-coverage notice.
• Coverage is partial. Only the counties on the coverage page have pages.
• Snapshots. Data is as of October 25, 2026, the day the snapshot was pulled from the district layers. The changelog records each refresh.

Not covered by this site

We report depth only. We don’t estimate drilling cost, yield or water quality, and a depth record is not advice about your property. Talk to a Florida-licensed water well contractor and your water management district.