A Question About Frost

It’s November now, which means we’re edging into winter, and my morning ritual has been expanded to include scraping the frost off the cars when I get back from walking the dog. I’ve had to do this half a dozen times already, and I’ve noticed a puzzling pattern.

Our driveway is aligned almost exactly east-west, with the cars facing east when they’re pulled in at night. This means that one set of side windows faces north, and the other south. And here’s the thing that puzzles me: the frost layer is significantly thicker on the south-facing side windows than on the north-facing side windows. I have to actively scrape the south-facing windows, but the north-facing ones I can just about wipe off with my hand.

I have no idea why that should be the case (the cars are symmetric under reflection about the axis, after all), but it’s very repeatable. But then, I’m not all that clear on the science of frost formation in the first place. I’m sure that some of my wise and worldly readers know more about weather-related matters than I do, though, so maybe one of you can explain it to me. Why do the south-facing car windows collect more frost than the north-facing car windows?

30 thoughts on “A Question About Frost

  1. I’m not sure about this, but it may be a lee effect. The prevailing breeze in cold weather is from the northwest in these parts, so presumably the breeze is moving moisture off of the north side of vehicles.

  2. Another question: do you have a carport, is the car next to the side of the house (and which side), or do you park with the car completely in the open? Proximity to structures, especially if heat is escaping from them, should have some effect on frost formation. If you park in the open, then the lee effect is the most likely explanation. You could test the latter theory in January when you get sub-freezing mornings with a southwesterly breeze.

  3. I have a related question, in that it deals with frost. The temperature this morning was 36F. My rear window, SE facing, had frost on it while my NW facing windshield had water. When I turned my wipers on, the water on the windshield froze… Is there such a thing as a super cooled liquid? Like the opposite of a super heated liquid?

  4. The frost distribution is basically the same on the two cars, with significantly more frost on the south side. There’s a single maple tree on the north side of the driveway, but other than that, they’re out in the open.

    They also tend to have more frost on the windshield than either of the side windows, but I’ve always attributed that to an angle effect, the windshield being more gently sloped than the side windows.

  5. The south side gets more ‘sun’ … are the cars parked before the sun goes down? Not that that explains anything (and the west side would get more sun in the PM anyway, so the back should have the same pattern as the south, and the front windshield the same pattern as the north side).

    You need to do two things on two different days: 1) Take one car to the car wash and park it as usual (after drying); and 2) park the cars in the opposite direction for a couple of nites.

    Get more data!

  6. My first inclination is to say that past the Autumnal equinox, the sun rises South of due East, so the southern side of your car should see slightly more sunlight than the northern side from dawn onward. This would suggest that the southern side should have _less_ frost, not more, so I’m stymied.

  7. Stephen, the sun is always south of the zenith if you are north of 23.5 degrees latitude. I think any north-south temperature difference due to direct solar heating would go away within an hour after sunset anyway (and if it’s cloudy there wouldn’t be any anyway).

    JohnV, yes there is supercooled liquid, from 0 to -40C water needs a condensation nuclueus, so you can have clouds of supercooled water. I didn’t think it could exist in contact with a solid, but apparently it can if the water is pure enough.

  8. Clay B, he’s scraping in the morning, presumably after sunrise. I was talking about direct solar heating melting frost in the morning, not latent heat from the night before. Of course that would go away.
    Nor am I talking about the sun being anywhere near zenith. In the winter, the sun RISES south of east, and in the summer it RISES north of east.

  9. Adhesion is only a monolayer deep. Make it your monolayer.

    Rain-X your windshields and frost won’t grab. Raise wipers. Wash and dry glass, wipe with rubbing alcohol, dry with paper towel. Wear rubber gloves and have ventilation. Spray on Rain-X, wipe with Rain-X wetted folded paper towel. Let sit 10 minutes. Wipe off with folded paper towel a couple of passes, repeat. Lightly buff a couple of passes. Your glass surfaces are silanized. (Outside only – inside is counterproductive.) Lower wipers.

    Do the headlights, too. No need to buff. Clean, spray, light wipe. Crud won’t stick.

    Lasts a month or three. The silanization cleaves off when hit with an ammonia window cleaner.

  10. My theory (or at least the beginnings of a theory):

    The sun melts a little bit of the first layer of frost. When it refreezes, it’s more of a continuous sheet of ice rather than a layer of ‘individual’ frost particles.


  11. This effect/explanation only works if the ambient temperature is near the freezing point: it’s radiative heat transfer. On a clear night, your clear-sky radiation reservoir is at 3K, but your tree is somewhere around 275K. The north side “sees” the tree and tends not to form frost. Even on a cloudy night, the sky temperature will be lower than freezing. You can get frost formation on a clear night even if the ambient temperature never reaches the freezing point.

    But if this is happening on even colder nights, then I have no clue.

  12. I always assumed that frost appeared most heavily on the front windshield, and less preferentially on the rear window and hardly at all on the side windows because it needed some sort of dust grain to form water droplets which freeze into frost. Meteoric dust and atmospheric dust in general would preferentially deposit on a more horizontal surface (like the front windshield), and on a more vertical side window facing the prevailing wind (seems weird that that would be a south facing window). Where more dust is deposited, you get more frost.

    This is a wild guess from a non-scientist. Think it has any validity?

  13. Differences in relative humidity (and therefore the dew point) could make a difference. (If the relative humidity is higher, dew or frost will form at a higher temperature.) Since both sides of the car have frost on them, and you’re looking at a difference in amount of frost rather than frost vs dew, that might explain the difference.

  14. I vote for partial obscuration of the open sky by the maple tree. Frost forms frequently on still nights if the sky is clear and the temperature is close to freezing because surfaces radiate quite well to the cold night sky. Daytime heating has nothing to do with it. Night breezes tend to suppress the formation of frost because they mix the air and tend to prevent the formation of a subfreezing layer next to surfaces, so the direction of any wind is unlikely to have a significant effect. I suppose I could conceive of a light wind tending to suppress the very cold, frost-forming layer on the windward side of a car and not suppressing it on the leeward side, but would that happen most nights?

    Clay B, if you are north of the arctic circle, in which direction would you see the sun at noon on the first day of summer? Maybe you meant during winter the sun is always south of the zenith if you are north of 23.5?

  15. Stephen, the sun is always south of the zenith if you are north of 23.5 degrees latitude.

    The sun at local solar noon is always south of the zenith, but as Stephen points out in his subsequent comment, it rises and sets north of the east-west line in the summer, and south of that line in the winter. Extreme case: In places like Tromso where the sun never sets near the summer solstice, the sun passes north of zenith at local solar midnight.

  16. My husband and I have noticed the same effect with our car windows. We’d tentatively concluded that it had to do with air-flow, as affected by the car’s position relative to the house and garage; that the greater exposure to the wind on the north side of the car made it harder for the frosting over to get started.

  17. Eric: woops. You are, of course, almost precisely right. In the northern hemisphere summer north of the tropic of cancer, the sun is always south of the zenith at solar noon. Now, I think that just about covers it. Of course it has nothing to do with the frost issue.

  18. The sky exposure thing might be it. It’s not anything to do with the sun, which is still behind the house at the time when I’m scraping the windows.

    It’s kind of hard to test, though, because the tree is big enough to block both spots in the driveway, so there isn’t a spot to park facing east-west where the tree doesn’t block anything. I suppose we could leave one of the cars on the street overnight, but the street gets sun sooner then the driveway.

    I guess I could check the neighbors’ car. Their driveway is parallel to ours, about ten feet to the north.

  19. BTW, “frost” is deposition, the opposite of sublimation, and – for reasons I would love to understand – forms when the air temperature is above freezing but below the dewpoint. Must be some combination of radiational cooling (open to the sky) pulling out the latent heat.

    I have a related observation: I regularly get dew on the side windows on one side of the car and not the other. This is clearly due to the crown of the road where I park my car. One side window is nearly vertical (and stays clear), while the other is at a steep angle to the sky (and gets dew on it). As with frost, the windshield gets heavier dew and heavier frost. Clearly the condensation is not simply onto the cold surface from the air, but has some effect from exposure to the sky (via radiational cooling?). Could it be the same effect? In my case, the dew forms on the north side. I should park the other way around and see what happens.

    So … Are your cars on a level surface? Do you see the same thing with dew as you do with frost? Have you considered putting sophisticated instrumentation out there?

  20. Tom @13 : On a clear night, your clear-sky radiation reservoir is at 3K

    Um, no. The background radiation from the night sky does have a lower temperature than that coming from the tree, but it’s coming from regions in the atmosphere with temperatures ~ 250 K or so.

    (A cloud-free night sky is pretty clear in the optical, but not in the IR; more to the point, the air itself is glowing in the IR.)

  21. CCPhysicist, it’s the same effect for dew and frost. The surface facing the sky cools to a temperature below that of the air. For dew, it’s below the dewpoint and for frost, it’s below freezing. The angle of the surface is very important. The closer it is to horizontal, the better the view of the sky, assuming no obscuration by maple trees.

    Dew and frost formation (especially at above freezing temperatures) depend strongly on calm air. If there is wind, the radiational cooling doesn’t work very well because of the constant replacement of cooler air next to the surface by air with a higher temperature. The air cools the surface very efficiently to its own temperature but tends to keep it there instead of letting it continue to cool.

    Clear, dry air is pretty transparent in certain IR wavelengths, but it doesn’t really matter because the effective temperature the night sky is low enough to let it cool to below freezing in the right conditions.

  22. I’d be seeing if it was the cars themselves.
    Get a couple of panes of glass and prop them on the lawn, some north/south, some east/west, some a little in between and see what effect that has.

    I was going to say that north-facing slopes heat up faster due to more sunlight, but that’s both here on the other side of the planet *and* directly contradictory to observed evidence in this case.

  23. clearly you have something breaking the symmetry of the car. remove the tree and see what happens!
    well i guess the other thing breaking the symmetry would be that the sun doesn’t rise in exactly the east and set exactly in the west, but we seem to have already ruled the sun out as a factor(unless we’re wrong).
    so what else breaks the symmetry… clearly not anything within the car or the car itself (unless one window is preferentially cleaner than the other for whatever reason. rolling down one window preferentially over the other etc)…

  24. We see the same effect on our car here in England.

    However, we are much further North here (51 degrees)and our road is narrow and lined with tall houses. As a result, the low winter sun only shines on the car for a few hours around midday — and the car sits in the shade for a few hours before the frost forms in the evening.

    So I’m with those who rule out the sun.

    We park on the road, so the car can point either way — I’ll try to keep track of the frost this winter and report back!


  25. Hmm. Your car faces east, and you get more frost on the south window. That means that you get more frost on the window closer to the building, and I thought it would be the opposite side… (or did I misunderstand the geometry?) When I lived in Canada someone told me that it was common knowledge that the side of the cars facing buildings get less frost. Or so I seem to remember. Maybe ask a Canadian?

  26. In my experience (based on various hotel car parks etc.) it’s likely to be the tree, as it’s warmer than the sky. I once parked overnight facing the brow of a hill, and there was thick frost on the windscreen despite the air temperature (so far as I could tell) never going below zero.

    (the cars are symmetric under reflection about the axis, after all)

    That must be very useful when travelling to countires where they drive on the left…

Comments are closed.