The freezing-point depression math, the radiative-cooling correction that turns a thirty-six-degree morning into a thirty-degree panel, the orientation table that explains why a south-facing February panel is a different planet from a north-facing one, and the six standard winter solution mixes from pure water through methanol antifreeze. The Twin Cities and Upper Midwest cold-weather route year, ported to software, and the symmetric counterpart to Easton's warm-weather Drying Window Predictor.
What the Cold-Weather Window Calculator does, in five points:
The math is the answer to the wrong question. The right question is not 'is it too cold to clean windows today' — it is 'on which elevations of which structures, in which conditions, with which solution, at which point in the day, is the outdoor work still feasible right now.' The tool exists so the operator does not have to carry the freezing-point depression table and the radiative-cooling correction in their head at 7:15 a.m. on a 28°F February morning with the route to plan. It does the math, and it returns the trade decision in a form the operator can act on in the time it takes the windshield to defrost.
There is a particular January morning I think of when I open up the route schedule on a cold day in Saint Paul, and it is a morning in 2012 — the third or fourth winter of the operation, before I had calibrated the cold-weather mix the way I run it now. The forecast was twenty-two degrees, light wind, partly cloudy. The truck thermometer agreed when I parked at the first house. I had a bucket of the standard mix on me — Dawn and distilled, the warm-weather house standard — because the morning had been the marginal kind where I thought I could get away with it. The first panel was a north-facing sash on the porch side of a Highland Park bungalow, in deep shade, with the sky clear above the trees. I applied the solution and the squeegee skipped. The film was already gelling at the perimeter beads before I had the squeegee in hand, and what came off was not a clean panel but a milky streak that I had to redo with a heat pad on the bucket and a stronger mix.
The lesson from that morning is the lesson the Cold-Weather Window Calculator is built around. The naive cold-weather rule is "above thirty-two degrees, work; below thirty-two degrees, do not." It is wrong in both directions. The thermometer reading on the truck dashboard is not the constraint. The glass is. And the glass on a north-facing panel under a clear cold sky at twenty-two degrees ambient is not at twenty-two degrees on the surface — it is at seventeen or eighteen, because the glass is radiating heat to a sky that is colder than the air around it, and the standard cleaning solution freezes there.
This piece is the methodology behind the trade decision the tool makes. The freezing-point depression math that turns each solution mix into a specific working margin, the radiative-cooling and solar-gain corrections that turn ambient air temperature into glass surface temperature, the application-method math that explains why a warmed bucket out of the cab cooler buys you four degrees of working temperature, and the verdict logic that maps the freeze-to-clean-time ratio onto a GO / CAUTION / STOP answer the operator can act on. The companion piece for warm-weather work is Easton Giordano's Drying Window Predictor methodology; the two tools are designed to read as a paired set, with the same verdict grammar and the same UI idiom, because the seasonal divide is a real one in the trade and the operator is making the same kind of decision on either side of it.
Below freezing, the constraint is the solution, not the technique. The most experienced squeegee handler in the trade cannot draw a clean line across a panel where the film is gelling under the blade. The film either stays liquid through the pass or it does not, and whether it does is set by two numbers: the freezing point of the solution and the working temperature of the film when it is on the panel. The technique decisions — stroke pattern, blade angle, how to handle the perimeter beads — all of them assume the film is liquid. If it is not, no technique recovers.
This is the rule the tool puts in front of the operator: the cold-weather trade decision is a chemistry-and-physics decision before it is a technique decision. The right answer is to raise the alcohol fraction, warm the bucket in the cab, or defer the elevation to interior work — not to try to clean faster through a freezing film.
The first piece of physics the tool needs is the freezing point of each of the standard winter solution mixes. The values are not theoretical — they are published in the CRC Handbook of Chemistry and Physics, Table 8-72 (binary mixtures of water with common solvents), and they have been refined over a century of laboratory measurement. The relevant values for the trade are:
| Solution mix | Freezing point |
|---|---|
| Pure water (DI / RO / distilled) | 32°F (0°C) |
| House standard (Dawn + distilled) | 31°F (−0.5°C) |
| 10% isopropyl alcohol mix | 19°F (−7°C) |
| 25% isopropyl alcohol mix (reference) | 5°F (−15°C) |
| 50% isopropyl alcohol mix | −18°F (−28°C) |
| Methanol antifreeze blend (commercial) | −40°F (−40°C) |
The progression is the progression any working cleaner in a cold-winter market knows by heart. The 10% IPA mix is the shoulder-season solution — late October, late March, the days when the morning starts cold but the afternoon comes up. The 25% IPA mix is the core-winter standard for residential work in the Twin Cities and the broader Upper Midwest; it covers most of December through February on south-facing elevations and most days on shaded elevations down to roughly 10°F ambient. The 50% IPA mix is the deep-winter solution, used the few weeks per year when the 25% will not hold. The methanol blend is a commercial deep-cold product — substantially more hazardous to handle than IPA, requiring nitrile gloves under the bucket sleeve, and not used in residential work for any operator I know.
The house standard and pure water are warm-weather solutions; they are in the table for completeness, because the tool's job is to flag the cases where the operator is reaching for the warm-weather bucket on a cold-weather day. That failure mode is the one I described from the January 2012 morning, and it is the one the tool is most worried about preventing.
The implication for the operator: the freezing point of the solution is not the threshold below which the cleaning fails. It is the threshold below which the cleaning becomes impossible. The cleaning starts to fail at the margin between the working film temperature and the freezing point — the perimeter beads gel first, then the bulk film, and the technique tolerances tighten progressively as the margin shrinks. The tool computes the margin and uses it to drive the freeze window.
The second piece of physics, and the one that catches new operators in cold markets, is the radiative-cooling correction.
Glass is a near-blackbody emitter in the thermal infrared. On a clear cold night or a clear cold morning, the glass surface radiates heat to a sky that is, in radiative-temperature terms, very cold — the effective sky temperature under a clear winter sky in the Upper Midwest is roughly 250 Kelvin (−23°C, −10°F). The glass loses heat to that sky faster than it gains heat from the surrounding air, and the surface temperature stabilizes a few degrees below ambient. The standard ASHRAE figure for this effect is three to six degrees Celsius below ambient under broken sky cover; under fully clear sky on a still morning, the surface can run as much as eight degrees below ambient on a horizontal pane, less on a vertical pane.
The tool applies a five-degree-Celsius radiative-cooling correction for the "clear cold sky" case, which is the standard pre-dawn and post-dusk winter condition. The correction is set to zero when the panel is in direct sun (the solar gain is overwhelming the radiative loss) and to zero when the sky is fully overcast (the cloud cover radiates back at near-ambient temperatures, and the radiative loss to the cloud layer is negligible).
The implication for the operator: a north-facing panel at six in the morning in late February, under a clear sky, with the thermometer reading 34°F, is not a 34°F panel. It is a 29°F panel, and pure water will freeze on it. The tool flags this case with an adjustment specifically recommending waiting for the sun to come up on the elevation, because once the sun is on the panel — even indirectly, through a thin overcast — the radiative loss goes to zero and the panel comes back to ambient or above within fifteen or twenty minutes.
The radiative-cooling correction is the largest cold-weather correction the tool makes that is not visible on a truck thermometer, and it is the correction the naive forecast-only approach to scheduling consistently gets wrong.
The third piece of physics is the conversion from ambient air temperature to working glass surface temperature on the panel via the solar-gain bump. This is the symmetric counterpart to the surface-temperature bump that Easton's Drying Window Predictor uses for warm-weather work, with the values adjusted for winter sun angles.
Winter sun is lower in the sky than summer sun, the angle of incidence on a vertical pane is steeper, and the absorbed irradiance is correspondingly higher per unit area. A south-facing panel in direct February sun in the temperate latitudes (30°–45° N, essentially the entire continental United States outside Alaska) absorbs more direct solar energy per square foot per minute than the same panel does in June. The bump on the working surface temperature reflects this:
| Orientation | Direct winter sun bump | Partial sun bump |
|---|---|---|
| Full shade | 0°C | 0°C |
| North-facing | +3°C | +1°C |
| East-facing | +9°C | +5°C |
| West-facing | +11°C | +6°C |
| South-facing | +14°C | +7°C |
The south-facing bump in winter is large enough that it is the dominant correction the tool makes for a sun-on-panel case. A south-facing panel in direct winter sun at 28°F ambient is sitting at roughly 53°F on the surface — a comfortable working temperature for any reasonable solution. The operator who walks past the south-facing elevation because the truck thermometer says 28°F is walking past the friendliest panel of the day.
The east-facing and west-facing bumps work as you would expect from the time-of-day pattern: east-facing peaks in mid-morning, west-facing peaks in early afternoon, and the working window on each is roughly two hours wide around the peak. A Tuesday morning route that hits the east-facing elevations first, then transitions to the south-facing elevations through midday, then catches the west-facing elevations on the way back is doing the work the route schedule should be doing in winter.
The fourth correction is the wind contribution. Wind on glass is different from wind on human skin: there is no metabolic heat to lose, so the "wind chill" terminology of weather forecasts does not transfer directly. What wind does to a glass panel is drive its surface temperature toward ambient air temperature: if the panel is sun-warmed above ambient, wind cools it back toward ambient; if the panel is radiatively cooled below ambient, wind warms it back toward ambient.
The magnitude of the effect scales with wind speed in m/s. The tool uses a saturating-curve approximation: at 0 m/s (calm), zero coupling; at 5 m/s (roughly 11 mph), about 55% of the deviation from ambient is damped out; at 10 m/s, about 70% is damped. The curve saturates at roughly 85% — wind cannot fully overcome a strong solar gain or radiative loss, just as wind cannot pull a panel temperature below ambient air temperature.
The implication for the operator: wind on a sun-warmed south-facing panel reduces the working margin (the panel is colder than it would be in calm air). Wind on a radiatively-cooled shaded panel raises the working margin (the panel is warmer than it would be in calm air). The two effects are symmetric and the tool computes them in the same step. The adjustments are different, though: in the first case, the right move is to wait for calmer conditions; in the second case, wind is actually helping the operator and no adjustment is needed.
The fifth correction is the application-method bump. The cleaning film does not arrive on the panel at the glass surface temperature; it arrives at the temperature of the solution as it was applied, and then cools toward the surface temperature over the seconds that the squeegee is working it.
Three application methods, with the corresponding initial film temperatures the tool assumes:
The tool computes an effective film temperature as a weighted average of the initial film temperature and the glass surface temperature (0.3 × initial + 0.7 × surface), because most of the film's working life on the panel is spent at near-surface temperature. The warmed-bucket bump on the effective film temperature is roughly four degrees Celsius for a typical cold-weather case, which is enough to move many CAUTION verdicts to GO.
The reference freeze window the tool is calibrated against is 75 seconds: the time from application to the slush-and-skip threshold for a 25% IPA solution at −1°C effective film temperature on a medium panel in calm air, applied by bucket-and-sleeve with ambient solution. The 14°C margin between the effective film temperature and the solution's freezing point (−15°C) is the calibration anchor, and the freeze window scales linearly with the margin with a small diminishing-returns rolloff at large margins.
The 75-second figure was set against fifteen years of Twin Cities residential field-cleaning data on the standard 25% IPA mix, working the marginal-condition cases — late-fall north-facing residential work in early November, early-spring east-facing residential work in mid-March — where the operator was timing the freeze threshold deliberately to calibrate against. The figure is conservative; in practice, the experienced operator on a familiar panel can sometimes draw a clean line at a slightly tighter margin than the tool's 75-second anchor would predict, but the conservative anchor is the right one for the calculator because the failure mode (a frozen film and a panel that needs to be redone with a heat pad and a stronger mix) is more costly than the false-caution case.
The diminishing-returns rolloff at large margins reflects an operational fact: above roughly a 20°C margin, the freeze window is no longer the binding constraint on the work; the warm-weather dwell window (which is what the Drying Window Predictor calculates) becomes the constraint instead. The rolloff begins immediately above the reference margin and slows the growth of the freeze window to half the linear rate — at a margin of 28°C (twice the reference), the freeze window is roughly 150 seconds rather than the 150 the linear extrapolation would give from a doubled margin. The intent is operational: at large margins, the precise freeze-window number stops mattering, and the operator should be reading the Drying Window Predictor's dwell-window output instead of this tool's.
The verdict bands map onto the freeze-to-clean-time ratio in the same way that the Drying Window Predictor's verdict bands map onto the dwell-to-clean-time ratio:
GO (ratio ≥ 2.2×). You have at least twice the freeze window the panel needs. Standard application, standard squeegee, no special handling. The film will stay liquid through the pass; the squeegee will draw a clean line. The redo risk from the freeze constraint is essentially zero.
CAUTION (ratio 1.0× to 2.2×). The film will stay liquid through one good squeegee pass, but the tolerances are tight. A missed corner, a slow stroke, or a perimeter bead left too long will gel before the next pass picks it up. The verdict exists to flag conditions where a small adjustment — raising the alcohol fraction, warming the solution, chasing the sun — will move the verdict to GO.
STOP (ratio < 1.0×). The math says the film will start to freeze before the squeegee reaches the bottom of the panel. The squeegee will skip and tear; the panel will need to be redone with a stronger mix and likely a heat-pad treatment on the bucket. The verdict exists to tell the operator that the cost of working the panel right now is the redo plus the wear-and-tear on the squeegee rubber, and the adjustments below are the route around the cost.
The adjustments the tool surfaces in CAUTION and STOP cases are not generic. They are ordered by effectiveness for the specific input combination, and they reflect the standard cold-weather route-management moves that experienced Upper Midwest operators have developed over decades.
Raise the alcohol fraction. The single highest-leverage adjustment when the verdict is driven by the solution. Pure water → house standard → 10% IPA → 25% IPA → 50% IPA is a ladder, and each step lowers the freezing point of the solution by enough to move the verdict roughly one band. The adjustment is the one I default to first on the route when the morning is colder than I expected.
Warm the solution in the truck. A cab-warmed bucket lifts the initial film temperature by four to six degrees Celsius and the effective film temperature by three to four degrees Celsius. That lift translates to roughly thirty to fifty percent more freeze window on a cold elevation. The cost is a $35 thermal bucket and a route-truck dashboard cup-warmer; the payback is one or two retainable winter clients per season.
Chase the sun. Move to a south-facing or west-facing elevation in direct sun. The fourteen-degree-Celsius solar-gain bump on a south-facing winter panel is the largest correction the tool computes; it is large enough to move most STOP verdicts to GO without any other change. The route schedule that absorbs the rotation is the route that does not redo work in winter.
Wait for the sun to come up on this elevation. The radiative-cooling correction goes to zero once the sun is on the panel. A north-facing panel that is a STOP at 6:30 a.m. under a clear cold sky may be a GO at 9:15 a.m. when the diffuse sky-light contribution has neutralized the radiative loss. The adjustment is the one that absorbs the cost of starting the route earlier than the schedule strictly requires.
The wind is part of the problem. Flagged when wind is dragging a sun-warmed panel back toward ambient. The response — work the sheltered face of the structure, save the wind-exposed face for a calmer day — is different from the response to a radiatively-cooled panel and worth surfacing as its own adjustment.
Work the panel in sections. The freeze window is the time-to-gel for the whole film; sectioning halves the effective clean time per application. The adjustment matters most on large picture-window or storefront panels in marginal conditions.
Today is an interior-only day. The final adjustment, surfaced when the ambient is deep enough that no reasonable adjustment will move the verdict to GO. The Upper Midwest cold-weather route year is built around this case: from mid-December through mid-February, the deepest weeks of the winter are genuinely interior-only, and the calendar that absorbs them is the seasonal commercial interior book the operation maintains specifically for these weeks. The dictum on the contributor page applies — the winter is genuinely interior-only, and plan accordingly.
The Cold-Weather Window Calculator is a tool for cold-weather outdoor residential and small-commercial decision-making. It is not a tool for warm-weather work (use the Drying Window Predictor for that), not a tool for interior cleaning where ambient room conditions are stable, not a tool for ice-dam meltwater residue removal (which is a percarbonate-citric ladder problem, not a freezing-point problem), and not a tool for the deep-cold commercial work where methanol-blend solutions and hazmat handling become the operational concern.
The piece of the trade the tool is built for is residential and small-commercial cold-weather routes in the temperate-to-cold-winter markets of the continental United States — roughly the band from Boston through Cleveland through Minneapolis through Denver. Markets that work outdoors year-round (Florida, the Gulf Coast, Southern California, the desert Southwest below roughly 4,000 feet of elevation) will rarely need the tool. Markets that work essentially no outdoor windows from December through March (Anchorage, International Falls, the upper plains in genuine deep-cold weeks) will find the tool useful mainly in the shoulder seasons.
The case the tool exists for is the case I built it from: the marginal late-fall and early-spring days, and the friendlier mid-winter days on south-facing elevations, where the outdoor work is possible if the operator gets the solution and the elevation and the timing right. Those are the days that determine whether the winter route is profitable or a wash. The math is the difference between a route that hits its monthly target through February and a route that does not.
— Linnea Jorgensen, for the editorial team
Linnea Jorgensen is a regional contributor based in Saint Paul, Minnesota, covering the Upper Midwest and Twin Cities residential market for Window Washing Guide. Sixteen years on Twin Cities routes plus a seasonal lake-country cabin book, and an eight-year background in cleaning-supply distribution before going independent in 2010. The cold-weather expertise comes from running a Saint Paul route through fifteen winters and from the years writing technical sheets for a Minneapolis distributor's house-brand chemistry line.