GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted

Community and Q&A

Is Uneven Snow Melt on Roof a Sign of a Problem

MikeCNY | Posted in Energy Efficiency and Durability on

Hello everyone,

I am a long time lurker on these forums. For a very long time I wanted to build a passive house here in upstate NY, but I ended up buying a a brand new construction house in March of 2020. It is supposed to be well insulated, although not anywhere close to passive house levels. I have noticed some uneven snow melt on the roof and it has me concerned something is not right with the insulation details, or the HVAC system. I am attaching 2 pictures from today of this, and I noticed it a couple times in the spring after we moved in. It is always in this area the problem occurs.

In my opinion on a roof that is facing the same direction, with no difference in shading, or slope changes the snow melt should be pretty much uniform if the insulation levels are the same. There is clearly some consistent snow melts in a couple areas of the roof.

What are your thoughts? Do you think there is a problem here?

Thank you,

GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.


  1. ssnellings | | #1

    How is the roof insulated (either "actually" if you can inspect it, or "presumably" if all you can do is refer to the original plans)? If you can't inspect it, what's on the other side of the roof in that location?

  2. MikeCNY | | #2

    It is a insulated sloped roof, so I would have difficulty inspecting it. On the right side of the house looking at it, so from the door and over to the right it a large cathedral vaulted ceiling, that doesn't show any signs of melting problems. On the left side there is a stair case that goes up to a master suite. That seems to be where the issue is.

    On the back side of the house, which faces mostly north there is a dormer for that master suite side, but I don't notice any obvious melting there.

    From my understanding there is sprayfoam, cellulose, and rigid foam used in this house, but I can't tell you for sure when it transitions from one to the other. I am assuming mostly in cellulose in the attic areas, spray foam in the walls.

    I have a insulation certificate from the company that did the work that says R25 in the walls, spray foam, and then attic area is r49 cellulose, with a note about the sloped roof also being r49, but not describing the use of material.

    Thank you

  3. Expert Member
    RICHARD EVANS | | #3

    Hi Mike, does the stairwell have can lights in the ceiling?

  4. MikeCNY | | #4

    The entire cathedral ceiling side of the house has small recessed led lights, they aren't the large can lights, but there is a light fixture that resides in the ceiling drywall. I can post a picture if you would like. This isn't the side of the house that is exhibiting the problem though.

    The master suite does not have recessed lights in the same fashion. It has one lighting fixture on that front side of the roof that isn't recessed. It has one small led recessed light right at the very middle of the house up there. This is the side of the house that does have the melting issues.

    The stair case is just to the left of the front door, and it is open to the rest of the main living area. This has the same led recessed lights as the rest of the main living area.

    Edited, to add in photos of ceiling lights. Again these are mostly on the side of the house that isn't having the issue.

  5. AlexPoi | | #5

    Your trusses are probably acting as thermal bridges. Not the end of the world but it is inefficient and can cause ice dams. This can be fixed by adding some insulation on the oustide when you'll need to reroof your house.

    1. MikeCNY | | #6

      Woohoo in 25 years or more when the brand new roofs life span is up I can fix it!!

      Any ideas why it is only impacting that one side? Maybe the insulation company put more insulation over the cathedral ceiling side, and less on the master bedroom side that is living area.

      1. Expert Member
        MALCOLM TAYLOR | | #7


        There is something more in play than just thermal bridging of the trusses or it would be uniform across the roof.

        1. MikeCNY | | #8

          Thank you Malcolm.

      2. AlexPoi | | #17

        How deep are your trusses/rafters on the side having problem vs the trusses on the cathedral ceiling side? Are both room at the same temperature?

        A 2x10 has a R value of 9-10 vs a r value of 5-6 for a 2x6. This could explain the difference between the two rooms.

  6. Expert Member
    RICHARD EVANS | | #9

    Mike, are you able to reach/ touch the drywall under that section?

    If so, and when it is really cold out, see if it is cold to the touch. If so, it may be that insulation is missing or sprayed too thin in that area.

    If it is warm to the touch, then I would assume it is from a duct or a can light (combined with possibly too little insulation). Can lights can become super hot. Try keeping them off during the next snow and see if it has an impact. Likewise, you can turn your HVAC off and plug in a space heater. See if the spot covers with snow.

    You might be able to find the issue by isolating these potential causes.

    Keep us posted!!

    1. MikeCNY | | #10

      Thank you Rick, in the master bedroom upstairs I can touch the drywall easily. I couldn't really tell the difference with the back of my hand. It may have been slightly cooler. I will try to test that some more. Maybe get a IR gun to get temperature readings.

      I have also contacted the company that did the work when the house was built. Try to get them out here with a FLIR camera to see what's going on.

      I don't think it is the recessed lights. Based only on that fact that the vast majority of them are on the side of the house that doesn't have any melt right above them. But I also don't have them on all the time anyways, so it is hard to judge. I also got on a ladder to touch one of the recessed lights in the stairwell, and after an hour of being on it was warm to the touch, but not hot by any means. The drywall wasn't that noticeably different temperature either. But it could be I guess.

  7. Expert Member
    BILL WICHERS | | #11

    Those look like air leaks to me. I'm a big fan of going out on a winter mornings to look at frost patterns on the roof as a sort of nature-provided thermal image of the roof. I have almost the same pattern on my own roof above two can lights in a cathedral ceiling (put in by the previous owners, and an oddball size I haven't been able to retrofit yet).

    Thermal bridging of a truss or beam will usually result in a more even pattern in the frost. With thicker snow it's harder to tell though. Ideally you want ONLY a morning frost on your roof when doing this type of investigative work. Morning frost is a thin, relatively even coating over the entire roof surface that better shows thermal issues compared to thicker snow. Thicker snow can mask smaller issues, and can be sneaky in other ways since it has some thermal mass that combines with the warming morning air as the sun comes up to mask other things.

    Without knowing the interior layout, it's hard to be sure. I'd guess leaky recessed cans near the bottom of those melt lines. You could potentially also have some separating insulation in those areas, but my money is on an air leak for each of those melt lines.

    You could potentially "cap" a can light as a test. Remove the bulb first, then use some painter's tape to tape a piece of paper over the can. Make sure the paper is taped to the drywall and not the perimeter of the can light itself. This only really works if you have some reasonably predictable weather though, since you need a fresh frost or light snowfall to see if anything changes. Smoke sticks are a more reliable test -- see if the smoke is "sucked out" through the can. With as much of a leak as you could potentially have, you can do this test without a blower door running so it's a cheap test to run.


    1. MikeCNY | | #12

      Thank you Bill. I think I agree there might be a massive air leak somewhere.

      I think I noticed some condensation streaking in these same areas over the summer. The behavior was harder to see in the summer.

      I don't necessarily think it is the recessed LED lighting, but it certainly could be since they are known to sometimes be the cause. Like I mentioned to Rick Evans in a previous post almost all of the recessed LED units are on the right side of the house looking at it, and this side doesn't show the same melting. This right side of the house on both the front and the back is a cathedral celling full of LED lights.

      I am going to attach the type of recessed light bulb used. They aren't the old school can lights, with a light bulb screwed into it. These are all individual LED units that seal relatively tight to the ceiling with a pair of spring loaded arms, and have a cable that runs up to connect to a power supply.

      The front left side of the house showing the melt is where the master bedroom is upstairs. There are almost no lights up here. There is one fixture on the ceiling that is not recessed, and I will attach a picture of that as well.

      I am almost wondering if it is a HVAC duct issue. There is a duct that runs right up around the area of the worst melting, there is a vent that is in the master bedroom on that front facing wall.

      Type of LED lights used

      Thank you,

      1. Expert Member
        BILL WICHERS | | #13

        Those lights should leak much less than the old style can lights.

        A leaky duct could be the cause, or a leaky hole around the duct. Another possibility is that the duct itself is displacing enough insulation to cause the melt pattern you’re seeing.

        Isn’t investigative work fun? ;-) an IR camera might be able to narrow things down a bit before you go opening things up. You could also try waiting for a frosty morning, then crank your furnace up to 80F for a little while and see if the melt lines rapidly expand. If they do, then you know the ductwork is somehow involved in the problem.


        1. MikeCNY | | #14

          Thank you Bill. If the insulation company doesn't come out soon with a thermal camera I may try some tests with the furnace to see if I can see anything.

    2. woobagoobaa | | #15

      " I'm a big fan of going out on a winter mornings to look at frost patterns on the roof as a sort of nature-provided thermal image of the roof." I believe this roof provides a textbook example of Zephyr's point. R38 (6") closed cell spray foam on the underside of the board sheathing, interior rafter edges not encapsulated.

      1. MikeCNY | | #16

        Now that is definitely some thermal bridging!

      2. Expert Member
        BILL WICHERS | | #18

        That's exactly what I mean! :-)

        From that pic, you can see the rafters clearly, and you can see very even frost between the rafters. I would guess that there are vent chutes between the rafters here, and that the insulation doesn't cover the rafters at all. It's also possible that the insulation is directly against the roof sheathing inside.

        The other thing you can see is no "splotches" -- you have a pretty consistent pattern over the entire roof. If there were splotches, each splotch would represent either an air leak or place where the insulation had pulled away from the roof. It's common to see "plumes" around things like vent penetrations too, which often indicate an air leak around the vertical rise of the vent pipe.

        If you get a few good pics of your roof after a frost like this, you can see all the things you need to seal up and fix inside. Go up in the attic with some caulk and canned foam, seal everything you find that looks suspicious in the areas of the melt patterns you saw on the frosty roof. After doing that work, wait for a similar frosty morning, go outside, and marvel at the more-even frost pattern that shows that you got a lot of leaks sealed up.

        Note that extra melt around things like ridge vents means you have a small amount of leakage over a large internal area, and that warm air gets concentrated around the vent where it's leaving the attic. Random splotches on the roof surface are "point leaks" -- single leaks of relatively high leakiness. I'd go after those splotches first since those are the low-hanging fruit where you can get the most benefit from the least work.


  8. MikeCNY | | #19

    In addition to my initial post and pictures here is the roof on a cold morning when there is some frost in everything. You can really clearly see there is something going on in this area, which is a sloped insulated roof surrounding a "attic" master suite.

    I still haven't gotten the insulation company that did the work when the house was built out here, but it is the holidays. I will contact them again at some point to come out here with a FLIR camera to get some thermal images.

    1. Expert Member
      BILL WICHERS | | #20

      That’s definitely not thermal bridging. Those are air leaks! Classic melt patterns for air leaks — see how they start low and solid, then taper off as they get near the ridge? That’s because the warm air from the leak is cooling off as it rises.

      Thermal bridging would show as a smaller (usually) line, and would be more solidly defined over its length since the conducted heat would warm the roof more evenly because it wouldn’t be “cooling off as it rises” due to thermal conduction not involving air movement.

      It will be interesting to see what the source of these leaks is. Please be sure to post back when you find out!


      1. MikeCNY | | #21

        Thank you Bill, my guess is definitely the hvac ducts, which reside around that area. Sort of a knee wall area at the front of the house there.

  9. user-5946022 | | #22

    Can't post in the Q&A spotlight as I'm not a member, but with the recently added insulation and associated photos, it appears the owner is trying to prevent warm air from seeping between the triple rafters and up to the roof. The owner's idea of applying tape to the underside of the rafters to air seal that gap was confirmed.

    It looks like the tails (low ends) of the rafters are completely enclosed in foam on the underside. Thus that end of the rafter will take on the outside temperature. Is there any reason to not enclose the balance of the entire portion of the accessible rafters with open cell foam? That way the entire rafter would be the closer to the same temp as the upper edge and the sheathing temp, instead of being a thermal bridge from accessible area to the roof. The open cell foam is vapor permeable, so it would also allow the rafter to dry out if it gets condensation on it. This seems like a way to solve this thermal bridging issue, but there is probably a gotcha in my logic. Just want to learn what it might be.

    1. MikeCNY | | #23

      Thank you, I think I might try the tape approach to reduce the air leakage through that group of rafters.

      The other problem with that group of three rafters is that they go into a sloped ceiling in the master bedroom that is enclosed in drywall. So I can't get to their entire length easily. Within that pocket I can try my best to reduce the air leakage.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |