Ground thaws from the top down. That one directional fact explains most of what happens to a northern road in March, because it means meltwater has nowhere to go. FHWA’s geotechnical reference manual for pavements states the consequence: “Water is also often trapped above frozen soil during the thaw, which occurs from the top down, creating the potential for long-term saturated conditions.” A layer that spent the winter behaving like rock spends several weeks behaving like a sponge with a lid on it, and the manual adds that bearing capacity “may be reduced substantially during mid-winter thawing periods.”
Traffic does not pause for that. The structural weakness arrives while the freight schedule is unchanged, and a pothole is the visible receipt.
Three conditions, and any one of them is a defense
Frost damage is often narrated as water in a crack expanding as it freezes. That happens, and it is not where the structural problem originates. The reference manual locates heave in the subgrade: “three conditions must be present to cause frost heaving and associated frost action problems: frost-susceptible soils; subfreezing temperatures in the soil; and, source of water.” Frost heave results from the crystallisation of ice lenses in the voids of soils containing fine particles, which is a process that draws additional water toward the freezing front rather than merely expanding the water already present.
Soil type does most of the sorting. The manual reports that “silts are highly frost-susceptible,” attributing it to “The condition of relatively small voids, high capillary potential/action, and relatively good permeability of these soils,” and that “Little to no frost action occurs in clean, free draining sands, gravels, crushed rock, and similar granular materials, under normal freezing conditions.” That last qualifier is doing real work and should not be dropped when the sentence is quoted. The manual’s own classification runs from F1 to F4, keyed to the percentage of fines passing the 0.075 millimeter sieve.
The list is also a design menu, because a road only needs one of the three broken. Non-frost-susceptible material to the depth of frost removes the first. Insulation or depth removes the second. Subdrainage removes the third, and it is the cheapest to retrofit and the easiest to leave clogged.
Frost action can cause differential heaving, surface roughness and cracking, blocked drainage, and a reduction in bearing capacity during thaw periods.
Blocked drainage belongs on that list and rarely gets there. A heaved shoulder or a lifted gutter line stops water leaving the pavement, which supplies the third of the three conditions for the following winter.
The worst of it is not the far north
Cycle count matters more than depth of cold, and it puts the damage somewhere most people would not guess. The same manual reports that “frost heave occurs in over half of the states in the U.S. and damage may be most severe in the central states, where many more frost cycles occur than in the most-northern states.” A place that freezes hard once and stays frozen runs one cycle. A place that crosses freezing four nights a week for two months runs dozens.
What the surveyor is allowed to call a pothole
Federal distress definitions are narrower than the word. In FHWA’s LTPP manual a pothole is “A bowl-shaped depression in the pavement surface.” Nothing in the entry mentions frost, water or traffic. It is a shape.
The causal chain sits in a different federal document, grouped by climate rather than by season. The federal state-of-the-practice report on pavement resilience, issued in 2023, tabulates the performance indicators affected by climate, and on the asphalt side lists “Fatigue cracking and potholes” as a single indicator, beside “Rutting of subgrade and unbound base” and “Stripping.” The two belong together because a pothole is what fatigue cracking becomes once water reaches an unbound layer that has lost its support. Which failure a road is prone to depends partly on surface type, the argument in asphalt versus concrete.
Minnesota publishes the calendar, which is the evidence
The clearest public record of thaw weakening is not a research paper. It is a load restriction schedule. Minnesota’s Department of Transportation divides the state into six zones and publishes two sets of dates for each: a winter load increase, when a frozen road may legally carry more than normal, and a spring load restriction, when it may carry less.
For 2026 the northern zone’s winter increase ran from 12 December 2025 to 20 March 2026, and its spring restriction from 20 March to 15 May. Further south the whole calendar shifts earlier, with the metropolitan, south and southeast zones restricting from 3 March to 20 April. In four of the six zones, including the metro zone, the winter increase ends on the exact date the spring restriction begins. In the south and southeast zones it does not. There the winter increase ends 24 February, a week before the 3 March restriction takes effect, and for those seven days the road carries neither rating. A Minnesota trunk highway in the north is officially either stronger or weaker than standard for roughly five months of the year, from 12 December to 15 May; in the south and southeast the same span runs about three and a half months, from 29 December to 20 April, with that week in the middle when neither regime applies.
The asymmetry is more informative than the dates. One pavement is rated above its normal capacity in January and below it in April, which is a formal admission that the structure’s strength depends on the state of the water in it rather than on how it was built. The notice period says something too. MnDOT states that “All starting and ending dates begin at 12:01 a.m. and are preceded by at least a three-day advance notice,” and three days is what an agency gives when it is reading conditions rather than following a diary. Why axle loads carry this much weight in the calculation is the subject of how trucking shapes highway design.
FHWA’s 2023 review confirms that the timing problem is unsolved rather than merely local. Among the concerns raised at its 2020 peer exchanges were “fluctuations in the date to apply spring load restrictions and winter load premiums when the pavements are frozen as few tools are available to identify the timings.” Participants also raised the possibility that “an increase in the number of freeze-thaw cycles in areas that normally receive a hard freeze may increase damage to the pavement,” which is the counterintuitive direction of the climate signal here. Warmer winters do not simply reduce frost damage. The same review notes projections of fewer days below freezing, shallower frost penetration and reduced risk of frost heave, alongside “potential changes in the number of freeze-thaw cycles experienced.” A region that loses its hard freeze can trade one deep cycle for many shallow ones, and the hedge in that word “potential” is the honest state of the science. The opposite end of the same problem, where historical temperature records set a design input that is no longer descriptive, runs through extreme heat and roads and across climate resilience coverage more broadly.
The largest pothole repair experiment ever run, and what it settled
Between March 1991 and February 1992, 1,250 pothole patches were placed at eight test sites across the United States and Canada as part of the Strategic Highway Research Program’s H-106 project, installed by local maintenance forces from six state departments of transportation, one Canadian province and one city public works department, across four climatic regions. FHWA continued monitoring the sites and published the results in 1998.
Four procedures were compared. Throw-and-roll is as crude as it sounds and was specified that way: material is placed into the pothole “without any preparation or removal of water and debris prior to material placement,” the patch is compacted under truck tires in four to eight passes, and a slight crown is checked for before the crew moves on. Edge seal adds a band of bituminous tack and sand around the patch a day later. Semipermanent, which the report calls the “do-it-right” method, removes all water and debris, squares the sides of the hole to vertical with a jackhammer or a saw, then compacts from the centre outward. Spray injection uses a dedicated machine.
FHWA’s summary of the key finding is uncomfortable for anyone who assumes effort maps to durability: “the throw-and-roll technique proved as effective as the semipermanent procedure in most situations and is more cost-effective, making it a good patching choice.” The agency also reported that “if quality materials are used, pothole patches can remain in service for several years, even though the patches are often intended as only temporary repairs,” and that 56 percent of all patches survived until the final round of performance monitoring in 1995.
Underneath that headline sits a result that reframes the whole subject. The report states that 86 percent of the experimental repair types performed no differently from the control, a result it attributes to an earlier screening project having already eliminated the poor materials. Table 15 lists 11 comparisons that differed significantly at the 0.10 level, out of what works out to 80 total comparisons. But control patches at the same site did differ from one another, and the report is explicit about why that is interesting: the material, placement procedure and compaction effort “were the same for each set,” so it attributes the spread to “site-specific factors, such as underlying support and drainage.”
Where the money actually goes
The 1998 report also priced the trade-off, in an illustrative calculation built on its own assumptions: a 200 metric tonne job, a crew placing 5.0 tonnes a day, $400 a day for labor, $50 for equipment and $500 for traffic control. A material at $80 per tonne with a 30-month mean survival life produced $64,800 across a three-year analysis period. A cheaper material at $25 per tonne with a six-month mean survival produced $258,000. Those are 1998 dollars and one hypothetical job, so the ratio is the durable part and the totals are not. The recommendation has aged well: “The cost of patching the same potholes over and over because of poor-quality material quickly offsets the savings from purchasing a less expensive cold mix.” The report also urges agencies to count reduced user delay and “the improved safety conditions that less crew time in traffic will allow.”
So there are two clocks running on a spring pothole and only one of them is in the weather record. The calendar names the week, and Minnesota publishes it three days at a time. The subsurface names the mile, and the 1998 result says it does so through underlying support and drainage, on segments that will produce the same holes next April whatever cold mix goes into them this one.