Open a fishing map and the contour lines all look equally certain. They are not.
Some contours begin with measurements taken on the water: sonar soundings, agency lake surveys, or hydrographic work. Others are estimates built from a real shoreline and a known maximum depth. Both can help an angler plan a day. Only one can claim that the shape beneath the surface was measured.
We wanted to know how often each kind appears across DepthScout, and whether the national average tells the whole story. So we froze the public catalog on August 13, 2026, counted every water carrying the Survey depth data label, and compared coverage across 63 states, provinces, territories, and countries.
The short answer: measured bathymetry is not spread evenly. It is concentrated in a small number of places, while several of the largest regional catalogs remain almost entirely modeled.
What we found
- 14,343 of 37,672 waters in the sitewide catalog were survey-backed. That is 38.1%.
- The regional split is much sharper than the national average suggests. Eighteen of 63 regional catalogs were at least 90% survey-backed. Sixteen were below 10%.
- The lowest-coverage regions are not a small fringe. Together, those 16 catalogs contained 15,474 regional water placements, 40.8% of all regional placements, but just 93 survey-backed placements.
- Survey-backed coverage is highly concentrated. Ontario, Minnesota, Québec, Finland, and Michigan supplied 9,647 of the 14,482 survey-backed regional placements, or 66.6%.
- Measured does not automatically mean modern or navigation-grade. The label identifies where the contours came from, not how recently or densely the water was surveyed.
The report uses two totals on purpose. The homepage provides the sitewide catalog count. Regional pages describe catalog placements. A lake on a border can appear in more than one region, so adding the regional pages produces a slightly larger number. We use the sitewide total for the national headline and regional placements only when comparing regions.
Figure 1
Regional catalogs by survey-backed coverage band
Number of regional catalogs in each coverage range
View figure data as a table
| Coverage band | Regional catalogs |
|---|---|
| Below 10% | 16 |
| 10–24.9% | 6 |
| 25–49.9% | 7 |
| 50–74.9% | 11 |
| 75–89.9% | 5 |
| 90% or more | 18 |
The average hides two different maps
A national survey-backed rate of 38.1% sounds like a fairly ordinary data gap: more modeled waters than measured ones, but plenty of both. The regional results are less ordinary.
At one end, Québec had survey-backed contours on 1,864 of 1,868 placements. Finland had them on 1,543 of 1,551. Michigan, Maine, and New Hampshire were all above 96%.
At the other end, Sweden had no survey-backed placements among 6,433 waters in the DepthScout catalog. The United Kingdom and Nova Scotia were also at zero. Wisconsin had 25 among 4,318 placements, and British Columbia had 28 among 2,459.
Those zeroes require careful wording. They do not prove that a region has never surveyed a lake, or that no public bathymetric record exists. They mean that no water in that regional DepthScout catalog carried the survey-data label on the snapshot date. A survey may be unavailable, difficult to license, published in a form that has not been integrated, or simply outside the current scope of the product.
| Regional catalog | Placements | Survey-backed | Coverage |
|---|---|---|---|
| Québec | 1,868 | 1,864 | 99.8% |
| Finland | 1,551 | 1,543 | 99.5% |
| New Hampshire | 488 | 483 | 99.0% |
| Maine | 549 | 533 | 97.1% |
| Michigan | 1,404 | 1,353 | 96.4% |
| British Columbia | 2,459 | 28 | 1.1% |
| Wisconsin | 4,318 | 25 | 0.6% |
| Saskatchewan | 349 | 1 | 0.3% |
| Sweden | 6,433 | 0 | 0.0% |
| United Kingdom | 850 | 0 | 0.0% |
Figure 2
Composition of the 12 largest regional catalogs
Share of regional placements that were survey-backed or approximated
View figure data as a table
| Regional catalog | Total | Survey-backed | Approximated | Coverage |
|---|---|---|---|---|
| Ontario | 7,929 | 2,970 | 4,959 | 37.46% |
| Sweden | 6,433 | 0 | 6,433 | 0.00% |
| Wisconsin | 4,318 | 25 | 4,293 | 0.58% |
| Minnesota | 3,774 | 1,917 | 1,857 | 50.79% |
| British Columbia | 2,459 | 28 | 2,431 | 1.14% |
| Québec | 1,868 | 1,864 | 4 | 99.79% |
| Finland | 1,551 | 1,543 | 8 | 99.48% |
| Michigan | 1,404 | 1,353 | 51 | 96.37% |
| United Kingdom | 850 | 0 | 850 | 0.00% |
| Norway | 664 | 515 | 149 | 77.56% |
| Maine | 549 | 533 | 16 | 97.09% |
| New Hampshire | 488 | 483 | 5 | 98.98% |
This is the central finding: the map is not gradually thinning from one place to another. It has hard edges. Cross a jurisdictional line and the share of measured contours can change from nearly everything to almost nothing.
Five regions account for two-thirds of survey-backed placements
The concentration becomes clearer when the question changes from “What percentage of each region is surveyed?” to “Where do the measured maps in the catalog actually come from?”
Five regional catalogs contributed 9,647 survey-backed placements:
- Ontario: 2,970
- Minnesota: 1,917
- Québec: 1,864
- Finland: 1,543
- Michigan: 1,353
Together, they accounted for 66.6% of all survey-backed regional placements in the snapshot. That does not mean those five places have the best bathymetry, the newest surveys, or the most surveyed lakes in the world. It means they currently carry most of the measured coverage visible through DepthScout.
The distinction matters because catalog size and survey coverage are separate things. Ontario had the largest regional catalog and the largest count of survey-backed placements, yet its coverage rate was 37.5%. Québec and Finland were smaller catalogs but almost completely survey-backed. Sweden and Wisconsin were among the largest catalogs overall and among the least survey-backed.
Figure 3
Leading contributors of survey-backed placements
Regional placements labeled “Survey depth data”
View figure data as a table
| Catalog group | Survey-backed placements | Share |
|---|---|---|
| Ontario | 2,970 | 20.51% |
| Minnesota | 1,917 | 13.24% |
| Québec | 1,864 | 12.87% |
| Finland | 1,543 | 10.65% |
| Michigan | 1,353 | 9.34% |
| All other regions | 4,835 | 33.39% |
For an angler, this explains why the DepthScout experience can feel different from one region to another. In one place, nearly every water may show solid, measured contours. In another, the useful information may be the real shoreline, the maximum depth, and an honestly labeled approximation.
“Survey-backed” describes origin, not quality
For this analysis, survey-backed means a water labeled Survey depth data on DepthScout. Its contours are derived from measured bathymetry: government soundings, digitized agency survey maps, or hydrographic surveys.
Approximated contours start with a real waterbody, its shoreline geometry, and a sourced maximum depth, but the individual contour lines are modeled. They are drawn differently and labeled so they are not mistaken for measured bathymetry.
That binary distinction is useful, but it is not a quality score.
A survey-backed map might come from dense modern sonar. It might also come from a paper survey made several decades ago and later digitized. Minnesota’s Department of Natural Resources says the lake maps available through its LakeFinder collection were compiled from the 1930s through the 1990s and makes no guarantee about their accuracy. Those maps are still evidence-based in a way a modeled basin is not; they simply answer a different question than a recent high-density sonar chart.
Several details determine how much confidence a measured map deserves:
- survey year;
- sounding density and spacing;
- positioning method;
- contour interval;
- vertical datum and reference water level;
- changes caused by sediment, dredging, reservoir operations, or shoreline work.
This is why DepthScout treats every chart as fishing guidance, never navigation. The survey label tells you that someone measured the bottom. It does not tell you that every shoal is current, every rock is marked, or the water level matches the day you launch.
Why coverage varies so much
This snapshot measures the outcome of a long data pipeline, not the cause of every regional gap. Still, the public sources show why “Has this lake been mapped?” is not a yes-or-no question.
In Québec, the province’s Géobase des bathymétries de lac du Québec is offered through the public data portal in reusable formats including GeoJSON, GeoPackage, CSV, and file geodatabase. Finland’s environmental institute publishes downloadable lake and river depth areas, contour lines, and sounding points. Those are the kinds of files that can be checked, transformed, and linked back to a named source.
Minnesota illustrates a different reality. Its DNR makes a large archive of lake maps available, but many are historical image files, and commercial reuse requires a license agreement. The information exists; using it responsibly involves conversion, metadata work, quality review, and permission.
Other programs publish PDFs, scanned paper maps, web viewers, point soundings, contour vectors, raster grids, or no central collection at all. Waterbody names and identifiers differ between agencies. Survey dates may be missing. Coordinate systems and depth units change. Licensing language ranges from explicit open-data terms to silence.
The result is an integration gap as much as a surveying gap. A depth survey becomes broadly reusable only after it can be found, understood, licensed, matched to the right waterbody, and converted without stripping away its provenance.
We cannot use this snapshot to assign a cause to an individual region, and we have not tried to. Low coverage in DepthScout should be read as a lead for further investigation, not a verdict on the work of that region’s fisheries or mapping agencies.
What better public bathymetry looks like
If agencies and data stewards want their lake surveys to travel farther, five publishing choices make an outsized difference.
- Release reusable geometry, not only a picture of the map. Sounding points, contour vectors, and documented rasters preserve more information than a scanned PDF.
- Attach a stable waterbody identifier. Names repeat. A durable ID makes it possible to join bathymetry to fisheries records, access points, regulations, and later surveys without relying on fuzzy name matching.
- Keep the survey metadata with the data. Year, method, units, datum, contour interval, and responsible agency should survive every download.
- State the reuse terms plainly. A clear license is more useful than a dataset whose legal status has to be guessed.
- Preserve the distinction between measured and modeled. An estimated contour can be useful. It becomes misleading only when its origin disappears.
For mapping platforms, the obligation runs the other direction: keep the source visible, retain important caveats, never merge modeled and measured contours under one label, and provide a correction path when the public record is wrong or outdated.
What this means when you are choosing a water
The practical lesson is simple: check the data label before reading the contour lines too closely.
On a survey-backed water, use the map to study the measured shape of the basin (points, breaks, holes, humps, channels) while remembering that survey age and water level still matter.
On a water with approximated contours, use the chart at a broader scale. It can help you understand the shoreline, depth range, and likely basin shape. Do not treat a specific underwater turn or isolated hump as a measured feature.
Neither chart belongs in the navigation category. Slow down, use current official navigation products where required, watch the water in front of you, and follow local rules.
Methodology
We captured the public DepthScout catalog on August 13, 2026.
The national figures came from the DepthScout homepage, which reported 37,672 lakes and rivers in the sitewide catalog and 14,343 waters with real survey depth data. We divided the survey-backed count by the sitewide count to calculate the national rate of 38.1%.
For regional analysis, we collected the total and survey-backed counts published on each of 63 regional fishing-map pages. For every region we calculated:
survey-backed coverage = survey-backed placements ÷ total catalog placements
The regional pages summed to 37,905 placements, including 14,482 survey-backed placements. These totals are slightly higher than the national figures because regional catalogs are not a deduplicated list of waters. Shared and border waters can appear in more than one regional catalog. We therefore use “placements” whenever regional rows are added together.
Coverage bands were defined before comparing regions:
- high coverage: 90% or more;
- low coverage: below 10%;
- all remaining regions: 10% through 89.9%.
All percentages in the article are rounded for readability. Calculations were made from the unrounded counts.
Limitations
This report is a census of the public DepthScout catalog at one point in time. It is not a census of every lake, every bathymetric survey, or every file held by a government agency.
The analysis does not measure survey age, sounding density, positional accuracy, contour interval, licensing difficulty, or whether an unintegrated public survey exists. It also does not rank fishing quality. A region with a high survey-backed rate is not necessarily a better fishing destination, and a survey-backed chart is not automatically more current than an approximation.
DepthScout’s catalog changes as waters and source data are added or corrected. The downloadable dataset is frozen so this report remains reproducible even when live totals move.
Frozen dataset
All 63 regional catalogs
Choose any column heading to sort. Regional totals are placements, not unique waters.
| Source | |||||
|---|---|---|---|---|---|
| Ontario | 7,929 | 2,970 | 4,959 | 37.46% | Regional page |
| Sweden | 6,433 | 0 | 6,433 | 0.00% | Regional page |
| Wisconsin | 4,318 | 25 | 4,293 | 0.58% | Regional page |
| Minnesota | 3,774 | 1,917 | 1,857 | 50.79% | Regional page |
| British Columbia | 2,459 | 28 | 2,431 | 1.14% | Regional page |
| Québec | 1,868 | 1,864 | 4 | 99.79% | Regional page |
| Finland | 1,551 | 1,543 | 8 | 99.48% | Regional page |
| Michigan | 1,404 | 1,353 | 51 | 96.37% | Regional page |
| United Kingdom | 850 | 0 | 850 | 0.00% | Regional page |
| Norway | 664 | 515 | 149 | 77.56% | Regional page |
| Maine | 549 | 533 | 16 | 97.09% | Regional page |
| New Hampshire | 488 | 483 | 5 | 98.98% | Regional page |
| Montana | 366 | 350 | 16 | 95.63% | Regional page |
| Saskatchewan | 349 | 1 | 348 | 0.29% | Regional page |
| Wyoming | 320 | 5 | 315 | 1.56% | Regional page |
| Massachusetts | 262 | 255 | 7 | 97.33% | Regional page |
| New Brunswick | 261 | 180 | 81 | 68.97% | Regional page |
| Washington | 252 | 233 | 19 | 92.46% | Regional page |
| Indiana | 214 | 155 | 59 | 72.43% | Regional page |
| North Dakota | 212 | 202 | 10 | 95.28% | Regional page |
| Iowa | 192 | 181 | 11 | 94.27% | Regional page |
| Texas | 191 | 105 | 86 | 54.97% | Regional page |
| Ohio | 177 | 158 | 19 | 89.27% | Regional page |
| Alberta | 171 | 164 | 7 | 95.91% | Regional page |
| Florida | 167 | 154 | 13 | 92.22% | Regional page |
| New York | 152 | 139 | 13 | 91.45% | Regional page |
| Connecticut | 146 | 139 | 7 | 95.21% | Regional page |
| Nebraska | 141 | 134 | 7 | 95.04% | Regional page |
| Missouri | 115 | 7 | 108 | 6.09% | Regional page |
| Vermont | 111 | 24 | 87 | 21.62% | Regional page |
| California | 100 | 16 | 84 | 16.00% | Regional page |
| Oklahoma | 96 | 88 | 8 | 91.67% | Regional page |
| South Dakota | 94 | 5 | 89 | 5.32% | Regional page |
| Utah | 93 | 2 | 91 | 2.15% | Regional page |
| Arizona | 88 | 4 | 84 | 4.55% | Regional page |
| Colorado | 86 | 3 | 83 | 3.49% | Regional page |
| Alaska | 83 | 56 | 27 | 67.47% | Regional page |
| Illinois | 83 | 27 | 56 | 32.53% | Regional page |
| New Mexico | 77 | 4 | 73 | 5.19% | Regional page |
| Nova Scotia | 74 | 0 | 74 | 0.00% | Regional page |
| Kansas | 73 | 66 | 7 | 90.41% | Regional page |
| West Virginia | 63 | 56 | 7 | 88.89% | Regional page |
| Idaho | 62 | 6 | 56 | 9.68% | Regional page |
| Mississippi | 62 | 9 | 53 | 14.52% | Regional page |
| Kentucky | 54 | 27 | 27 | 50.00% | Regional page |
| New Jersey | 54 | 46 | 8 | 85.19% | Regional page |
| Alabama | 50 | 29 | 21 | 58.00% | Regional page |
| Oregon | 50 | 31 | 19 | 62.00% | Regional page |
| Tennessee | 48 | 10 | 38 | 20.83% | Regional page |
| Georgia | 43 | 16 | 27 | 37.21% | Regional page |
| Delaware | 39 | 28 | 11 | 71.79% | Regional page |
| Manitoba | 39 | 32 | 7 | 82.05% | Regional page |
| Nevada | 39 | 3 | 36 | 7.69% | Regional page |
| Virginia | 38 | 22 | 16 | 57.89% | Regional page |
| North Carolina | 37 | 12 | 25 | 32.43% | Regional page |
| Louisiana | 35 | 11 | 24 | 31.43% | Regional page |
| Rhode Island | 29 | 8 | 21 | 27.59% | Regional page |
| Pennsylvania | 27 | 6 | 21 | 22.22% | Regional page |
| South Carolina | 27 | 13 | 14 | 48.15% | Regional page |
| Arkansas | 25 | 6 | 19 | 24.00% | Regional page |
| Maryland | 22 | 11 | 11 | 50.00% | Regional page |
| Hawaii | 17 | 0 | 17 | 0.00% | Regional page |
| Puerto Rico | 12 | 12 | 0 | 100.00% | Regional page |
Data, sources, and citation
Download the frozen CSV dataset and its data dictionary and limitations.
Download the regional survey-coverage snapshot. It includes the counts, calculated coverage rate, and source URL for all 63 regional catalogs.
Primary references:
- DepthScout catalog snapshot
- DepthScout data and methodology
- Québec lake bathymetry database (GBLQ)
- Finnish Environment Institute downloadable spatial datasets
- Minnesota DNR lake depth maps
Suggested citation:
DepthScout Research. “The Uneven Map Beneath the Water: What 37,672 Public Waters Reveal About Survey-Backed Depth Coverage.” August 13, 2026. https://www.depthscout.com/research/public-lake-depth-mapping-2026
What we want to measure next
Coverage is only the first layer. The next useful questions are about quality and change: How old are the surveys anglers can actually access? Which public programs publish soundings or vector contours instead of static images? How quickly is measured coverage expanding, and where are the largest tractable gaps?
Those questions need their own datasets. They should not be guessed from this one.