How Thick Does Ice Need to Be to Drive a Truck On It? The Agencies Disagree

Search this question and you will be handed a chart. It will look authoritative, it will be unattributed, and it will not tell you that the agency next door publishes a different number for the same vehicle.

Here is what ten agencies and bodies actually publish, kept separate, because the disagreement between them is the useful part.

Every Agency’s Table, Unmerged

For one passenger car, the published minimums run from 7 inches to 12 inches. That is not a rounding difference; it is nearly double.

For a truck, which is what most people asking this question are driving, the spread is wider still. The Army Corps lists 8 inches for a 2-1/2 ton truck and 10 inches for a 7-to-8 ton truck. North Dakota and New York both want 12-15 inches for a medium truck. Minnesota goes to 13 inches for a medium truck, 16-17 for heavy-duty, and 20 or more with a wheelhouse shelter. And if you intend to park that truck rather than drive across, the figure Minnesota publishes for a three-quarter-ton 4x4 is 22 inches — see the stationary table further down, because that is a different question with a much bigger answer.

Three Agencies Push Back on the Question Itself

This is easy to miss, and it is a position rather than an oversight.

Michigan DNR puts it first, above everything else on the page:

“Your safety is your responsibility! There is not a reliable “inch-thickness” to determine if ice is safe.“

It then tells you what to do instead — test with “a spud, needle bar or auger” — and describes what you are looking at rather than counting: “Strongest ice: clear with bluish tint. Weak ice: ice formed by melted and refrozen snow. Appears milky.”

Wisconsin DNR’s ice-safety page carries no table either, and is equally direct about why:

“There really is no sure answer, and no such thing as 100 percent safe ice. You cannot judge the strength of ice by one factor like its appearance, age, thickness, temperature or whether the ice is covered with snow.”

And Ohio DNR does something different again: it publishes a number and then tells you not to act on it if you can avoid it. One line below its own 8-12 inch figure:

“Refrain from driving on ice whenever possible. If you must drive a vehicle, be prepared to leave it in a hurry: keep windows down, unbuckle your seat belt and have a simple emergency plan of action you have discussed with your passengers.”

That is the most protective sentence quoted anywhere on this page, and it sits directly beside the number people take away.

Michigan’s refusal is unambiguous and stated at the top of its page. Wisconsin’s is narrower than we can prove: its ice-safety page publishes no table, but it also links a printed ice-safety brochure that would not open for us, and that brochure may or may not contain figures. What can be said is that neither agency puts a vehicle thickness chart in front of a reader who goes looking on the obvious page — and any article presenting “the ice thickness chart” as settled is quietly overruling at least one agency that deliberately declined to print one.

Why the Numbers Differ: One Equation, Different Margins

The agencies are not disputing the physics. Minnesota publishes the physics outright:

“The thickness guidelines above are based on Gold’s formula [1]: P = Ah², where P is the allowable load in pounds, A is the bearing strength of 50 psi, and H is the total ice thickness in inches. This formula is a starting point for creating ice safety guidelines, but remember ice is NEVER 100% safe.”

Its citation is a real one: Gold, L.W., “Use of ice covers for transportation”, Canadian Geotechnical Journal 8, 170-181 (1971).

Two things follow. First, load scales with the square of thickness, so small changes in thickness matter more than intuition suggests — and a modest change in the assumed bearing strength A shifts every row of the resulting chart at once.

Second, you can watch the agencies diverge, and you do not have to take our word or our arithmetic for it. Minnesota publishes an actual load table alongside the formula, running from 800 lb at 4 inches to 20,000 lb at 20 inches. Every one of its seventeen rows is exactly 50 × h².

Read Minnesota’s table at 7 inches and it gives you 2,450 lb of allowable load.

Now read the Army Corps table at that same 7 inches: “A single passenger automobile.”

Two agencies. One thickness. Two different answers about what it will carry. That is why these charts cannot be laid on top of each other, averaged, or chosen between.

What this does not tell you is why they differ, and it is worth being straight about that. Minnesota shows its working; the Army Corps publishes a bare lookup table with no formula, no constant and no stated assumptions, and the CRREL reference it points to does not supply them either. So the gap could be a different bearing-strength constant, or a different assumed weight for “a single passenger automobile” — 2,450 lb is within range of the lightest subcompacts, and this table’s phrasing reads as a decades-old reference — or the fact that a moving load and a static one are genuinely different engineering problems. No source read for this page discloses which. What is certain is only that the two numbers are not interchangeable.

Minnesota draws the practical conclusion on the same page: “You are responsible for knowing the weight of your vehicle, equipment, and bodies.” That instruction only means something if you have a number like 2,450 lb to compare it against.

The Federal Table Is Measuring Something Else

The Army Corps figures are the lowest here, and they are the ones most often repeated stripped of their conditions. The conditions are on the same page, above the table:

“The following table of safe loads is valid ONLY for ice that is clear and sound, with no flowing water underneath. it is not reliable for stationary loads. When in doubt, stay off the ice”

And the table’s own column heading is “Description of Safe Moving Load.”

So the 7 inches describes a car crossing sound clear ice, not a truck parked over an auger hole for an afternoon. Minnesota addresses that case separately and in the opposite direction: stationary loads longer than two hours require thicker ice than its standard guidelines, which is why it also tells you to “Park at least 50 feet away from other vehicles and move your vehicle every two hours.”

Parked Is Not Driven, and the Difference Is Roughly Double

This is the part most likely to matter to the person actually asking, because the reason to drive onto a lake is usually to stop on it.

Minnesota publishes a second table for exactly that case, on the same page as its driving guidance, and it is not a variation of the first — it comes from a third body entirely, the government of Alberta’s Occupational Health and Safety best practice. It covers “stationary loads of more than two hours but less than seven days”:

Set that beside Minnesota’s own driving table and the size of the gap is the point. A vehicle Minnesota clears to drive at 9-10 inches needs 16 inches to sit parked for an afternoon. A three-quarter-ton 4x4 needs 22 inches — more ice than any driving figure on this page, from any agency.

Minnesota’s framing is blunter than any chart: “Frozen lakes are not parking lots.” It adds a field check worth more than a number:

“Tip: Make a hole next to the car. If water starts to overflow the top of the hole, the ice is sinking and it’s time to move the vehicle.”

That is also why the Army Corps’ 7 inches cannot be carried into this section. Its table says in its own words that it “is not reliable for stationary loads”, and this is the case it is disclaiming.

Every Number Above Assumes New, Clear Ice

That qualifier is carrying more weight than the numbers are.

• Minnesota: “When a layer of snow melts and refreezes on top of lake ice, it creates white ice, only about half as strong as new, clear ice. Double the above thickness guidelines when traveling on white ice.”
• Ontario: “clear blue ice is the strongest”; “white or opaque ice is much weaker”; “double the thickness if the ice is white or opaque.”
• New York DEC: “Double the recommendations for white ice.” Plus two more it publishes separately: “Slush ice is about 50 percent weaker” and “Clear ice over running water is about 20 percent weaker.”
• Michigan: “Stay off ice with slush on top. Slush ice is only half as strong as clear ice.”

Double a 12-inch figure and you are looking for two feet of ice. That is the honest shape of the guidance: the headline number is a floor for the best case, and the best case is not what most people are standing on.

North Dakota and Iowa publish no numeric multiplier at all — North Dakota offers only “Look for clear blue ice. New ice is stronger than old ice.” Absence of a multiplier is not permission to skip the adjustment.

Numbers Circulating Under the Wrong Name

Three figures repeated widely could not be traced to the agency they are credited to. One of them is a category error with real consequences.

“The Canadian Red Cross says 20 cm for vehicles.” The Red Cross document says no such thing. Its list, in full, is “15 cm for skating, walking or skiing in small groups”, “20 cm for larger groups, such as skating parties”, and “25 cm for snowmobiles or ATVs”. The 20 cm figure is for a group of people, and the document publishes no row for cars, trucks, or any four-wheeled vehicle. Somewhere in the retelling, a skating-party number became a vehicle number. At roughly 7.9 inches it lands below every state and provincial car figure on this page, and only a whisker above the federal moving-load figure — which is to say it is not a vehicle recommendation from anyone.

A granular Ontario table (10 cm stay off / 12-15 cm snowmobile and ATV / 20-30 cm car or small truck) appears on forums and local news sites attributed to Ontario’s Ministry of Natural Resources. Ontario’s own page publishes two vehicle lines only: snowmobiles at 20 cm, light vehicles at 30 cm.

“Iowa DNR: 8-12 inches for passenger vehicles.” The Iowa DNR ice-fishing page read for this piece publishes no vehicle thickness row at all.

None of these are necessarily wrong as engineering. They are wrong as citations, and on this subject a borrowed authority is worth less than no authority, because it stops the reader checking.

Don’t Confuse These

• A chart is not a measurement. Every agency that publishes numbers also tells you thickness varies across a waterbody. North Dakota: “Ice thickness is not consistent.” The Red Cross: “The ice thickness must be measured every few metres.”
• Moving is not parked. The federal table is explicitly a moving-load table and says it is not reliable for stationary loads.
• Clear is not white. Three agencies say double the figure for white ice; a fourth says slush is half strength.
• Silence is not a low number. Michigan and Wisconsin decline to publish figures; Iowa and the Red Cross publish none for vehicles. None of that means less ice is needed.
• This is not a recovery scenario. A vehicle through the ice is not something a winch solves.

How We Sourced This

Every source listed below was opened and read directly rather than taken from any secondary chart, and each table above was transcribed from the agency’s own page in that agency’s own wording — including where two agencies use different words for what sounds like the same vehicle. No merged or averaged chart appears on this page, deliberately: blending them would erase the only finding that matters, which is that they do not agree.

One gap is worth naming rather than papering over. Wisconsin’s ice-safety page links a printed ice-safety brochure hosted on a document-viewer platform; three retrieval attempts returned the viewer’s own HTML shell rather than the file, so its contents are unknown. Wisconsin is therefore described above only as publishing no table on the page we read, not as publishing none anywhere.

Two agency sites refused automated requests during this research and were retrieved on a second attempt with an ordinary browser request; a 403 is bot mitigation, not evidence that a page is missing. Minnesota’s poster needed rendering, but not for the reason you might assume: it does carry a text layer, and pdftotext returns its words happily. What it does not return is the pairings. The extraction interleaves the poster’s columns — “<4” / 5-7“ / Keep off / 9-12“ / Snowmobile / 4“ / On foot / Car“ — so every thickness is adjacent to the wrong activity. A table whose text extracts cleanly can still extract wrongly, and on this subject a number attached to the wrong vehicle is the whole defect. Its bands looked at first like they contradicted Minnesota’s live page — 9-12 inches for “Car” against 9-10 and 11-12 for two separate categories — but reading both showed the poster is a simplification of the same guidance, not a different one. Both are cited above so you can see the difference yourself.

This page derives nothing. The 2,450 lb figure at 7 inches is a row in Minnesota’s own published table, not a calculation of ours; the only checking we did was confirming that all seventeen of its rows match the formula Minnesota itself prints above them. Full sourcing and rating methodology: review methodology.