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

Potential Ice Damming on New Roof

mpsterner | Posted in Energy Efficiency and Durability on

I’ve worked with an architect to design a pretty good house in Zone 7 (60psf Snow Load Region).

We’re using 16″ i-joists for the roof system with the intent of nailing plywood onto the top flange to create a 1.5″ vent chute. Then, dense packing the vaulted section of the roof with 14″ of dense pack cellulose. Drawing attached. There will actually also be an interior Intello air barrier and a service cavity on the ceiling.

In working with an energy consultant on mechanical design, heat load calcs and “redlining” the building for performance improvement opportunities, they expressed concern over the roof system. The r-value was not high enough (wanted “whole wall” R50 with less thermal bridging) and they wanted to see a 2″ vent chute.

They’re concerned that I will see ice damming on the eaves and that for the level of effort, there are higher performing roof assemblies.

1. Should I be concerned about ice damming with this assembly?
2. Are there other durability and performance concerns with the assembly?
3. Are there other roof assemblies that you would suggest using instead?

The energy consultants would like me to change to either a lumber or i-joist rafter with plywood and 6″ of rigid foam on the roof deck, then doing 2x rainscreen strapping and a second layer of roof sheathing with underlayment and roofing on top of that. I would have to apply the overhangs as outriggers using the rainscreen strapping.

Would the external insulation on the roof be THAT much better that it is worth what seems would be some additional effort and hassle?

CAN I do less external foam on the roof, perhaps 3-4″ to make it easier or do I then introduce concerns around condensation in the roof cavity?

Thank you for your feedback.

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. Jon_R | | #1

    For more R value/less bridging, you can add an inch of rigid foam on the interior side. Or bigger i-joists.

    Consider these or similar to maintain a uniform eave temperature (upper eave/lower roof warming + cold outer eaves creates ice dams).

  2. charlie_sullivan | | #2

    Generally, I think that real problem ice dam cases are a result of egregious air leaks or insulation weaknesses where it's way worse than code, usually combined with inadequate venting. If you execute this design well, you might get a little ice buildup occasionally, but I doubt it will ever cause a problem.

    If you want to make extra sure, you could fur down on the inside, with perhaps 2x4s running perpendicular to the I-joists, which reduces thermal bridging. But the thermal bridging on a 16" I-joist is pretty minimal.

  3. Expert Member
    Michael Maines | | #3

    Bigger vent spaces, more insulation and tighter air-sealing will all reduce the risk of ice dams. Nothing can fully prevent them; in some cases the sun and snow work against the best of envelopes. But I think yours sounds pretty good and should not cause ice dams, as long as you pay attention to air sealing. There is a tiny amount of thermal bridging through I-joist webs but even the strict Passivhaus Institute considers them to be thermal-bridge free because the move so little heat.

    1. mpsterner | | #4

      Thank you all. You're three comments are all consistent, which is great and helpful. I could further eliminate thermal bridging and increase r-value by insulating the ceiling furring/strapping cavity or adding a layer of poly-iso to the interior.

      I could also increase the depth of the vent space, however, wouldn't the actual airflow start to be limited by the area of the soffit venting?

      May I also ask – is relying on i-joists for a roof system a solid choice? I had never encountered this before the architect spec'd this and have some reservations (probably mostly ignorance). Obviously, it helps us achieve the desired r-values. Is an i-joist roof one that will last 100 years and beyond?

      What other "simple roof" systems should I be considering to achieve a minimum r-50 that are easier than 6" of exterior insulation?

      Thank you.

      1. aaronbeckworth | | #5


        Your plan is very much what I'm considering for the roof framing of a small 1 1/2 story home with shed dormer(s). I'm in SW New Mexico, CZ4B and looking for near R49 with 14 inch I-joists.

        I've become a fan of Ben Bogie; he seems to advocate the advantages of I-joist framed vented roofs, as he mentioned in a recent Q&A spotlight
        and in this excellent recent episode of BS* + Beer

        I've been thinking through different options for eave and rake framing. Would you be willing to share a detail showing how your overhangs will be framed and attached to the I-joists? If not, I understand.


        1. mpsterner | | #10

          Hi Aaron, sorry for the delay in response. Thank you for your feedback and links.

          I don't have any detail in addition to what I posted, however, the i-joist manufacturers cover this info pretty well. Download their manuals. Here is a screenshot from one of them.

  4. onslow | | #6


    I would say no, no, no to the three questions. The move to external insulation on the roof would mean a major change to the envelope philosophy. Typically external insulation means the roof becomes unvented and all space under the roof line becomes conditioned. Your inside to outside R ratio needs to keep the sheathing and interior framing elements above the critical cold point for your climate conditions. For guidance I have attached below a pdf that can describe the conditions with better authority than I.

    I am estimating over half of the interior R value needs to be on the outside for your CZ. This will be quite thick and demand the use of long and expensive screws that demand care of placement. These screws do exact a thermal toll of their own. See attached image of my own long screws holding 8" of EPS. And no, you can't short the outside insulation.

    The notion of adding venting above an externally insulated roof strikes me as somewhat silly and also likely to have little to no effect on ice damming, with the distinct possibility of increasing the risk in shoulder seasons. Please note Mr. Lstiburek's observations about warm air gathering under eaves of mountain chalets. Search "Dam Ice Dam" on the Building Science site. Sunlight warming my metal roof is the main driver of melting not heat loss from the living space. Even the banding you will potentially have from the 16" I joists is lower than what it takes to melt snow in a big way. I would worry more about dropping snow and rain further out from your walls.

    You have 11/12 pitch so your venting activity should be strong even for 1 1/2" chutes. The lower pitch lift dormer ( I am guessing the diagram is a cross section mid-line) only has a very short span of insulation to cross. Maybe padding the chutes to 2" there might be helpful, but I doubt it. I trust you are not going to vent at the ridge since you have gables to that appear to tie the whole attic together. In snow country, your ridge vent is likely to remain covered much of the time. Snow may be air permeable when fresh, but the small heat load concentrated at the vent will crust it up. Ask an avalanche victim how much air gets through to judge venting capacity.

    Some ridge vents also allow a bit of snow or hard blown rain past, which would not be good for the cellulose on the ceiling. I would assume you need to meet wildfire safety for vents so whatever vent form you choose, choose ember resistant. Insurance companies are getting very skittish about underwriting these days.

    I would be much more concerned about the length of your overhangs and the lack of clear insulation detailing at the floor bands. It appears that you have two layers of foam continuous insulation wrapping the house walls. The floor banding would be a good place to use foam despite the counter opinions about its use.

    I think the suggestion of solid lumber for rafters is optimistic at best. Even 6 years ago, the 2x12's I got delivered were not very nice. I may be too old and seen too much prime wood, but I think I joists will be more predictable in their quality. That is if you can even get them. Have you done any checking on availability of materials in your area? It may not even matter if where you are building is not a hot spot of activity. Check out the:

    for a quick flavor of the times.

    1. mpsterner | | #7

      Thank you Roger. Really good feedback. This is all sort of what I was thinking, but this consultant seems to believe that this is a somewhat risky roof. All other information I can come across would suggest that it is not risky at all but needs to be detailed properly like anything roof system.

      Valid points regarding venting. I do believe that they were intending that this would be vented at the ridge, though I don't have clear details on that. What other options would I have? Just venting out at the gable end walls?

      Another question I had was even with a 2" venting cavity, wouldn't the limiting factor then be the soffit vent space and how much air would actually come through the strip soffit vent? That is certainly not 24 square inches of air space per foot like the vent cavity is, so I wonder if a 2" space would even help.

      I will look at getting better detailing for the floor banding. As you suggest, perhaps this is an area to use beads of acoustical sealant and spray foam since it is a tricky area.

      As for overhangs–what about the overhangs specifically concerns you? Are you mostly just interested in getting snow and rain dropping off further from the building? This is where aesthetics start to come into play. A 2' eave was about all we could swallow with the cape aesthetic we're going for.

      Thank you for your feedback.

      1. Jon_R | | #11

        > wouldn't the limiting factor then be ...

        This isn't how fluids work. Everything adds resistance, but doesn't create a fixed limit.

        Also note that very short restrictions cause less resistance than longer, larger restrictions.

  5. Expert Member
    Akos | | #8

    Let say your outdoor temperature is around -10F. Assuming no air leaks, with 70F inside, a 15' section of 24OC rafter bay looses 48BTU to the air space. It is not going to take much ventilation to remove that much heat.

    This is not precise but can put a ballpark on flowrate. For the snow to start melting, the underside needs to get up to 32F, so you minimum flow rate through a rafter bay is 48 BTU /(42F deta * 1.08), ~1CFM. This is such a low flow rate, that pretty much any reasonable gap will work, going up to 2" will make no difference.

    Your bigger issue is the sun heating the roof surface causing the snow to melt. No amount of insulation and venting will prevent that.

    I think you have a solid roof, I wouldn't worry about it. The exterior rigid option just costs way more, much more complicated to build and would barely change your overall energy use. Even in zone 7, an R50 roof is pretty good.

    I-joists are pretty easy to build with, they are much lighter thus easier to handle and straight. About the only drawback is that you can't drill or notch the flanges, so some of the details are different than a stick built roof.

    P.S. With a structural ridge, I would be tempted to skip the mini attic and cathedral the whole space. You do have to detail your air barrier continuity across the ridge and at the dormer but not much extra work.

    1. Jon_R | | #12

      Note that ice dams typically form above 20F. There is nothing wrong with the sun melting snow on the roof - ice dams occur if the eaves (or portions of) aren't equally warmed and it refreezes there.

      "In ice dam regions, where ground snow loads are greater than 30 lb/ft2 (146 kg/m2) you need 2 in." - Lstiburek

      1. mpsterner | | #13

        Hi Jon_R and Akos,
        What are your thoughts on the fact that even if I provide a 2" vent space by furring down the top flange of the I-joist, the ventilation is going to be limited by the volume of air that can travel through the soffit vent, which is less than the 2"x24" vent cavity. Strip vents are only 2” wide and perforated or louvered so you are not actually getting 24 free square inches per linear foot, rather only ~9 SQ IN NFA (net free area)/LF.

        Furthermore, the vent space air is provided from the soffit through "notching" in the LSL rimboard on the face of the I-joist, which would also have less net free area than even the soffit. Image attached.

        So, do these bottlenecks in net free area for airspace ultimately render my ventilation strategy ineffective even if I do have a larger vent space.

        Please advise on what you all would suggest.

        1. aaronbeckworth | | #14


          I was really hoping you’d get answers to the question of notching the rim board to connect the I-joist vent channels with the enclosed soffit. I asked about this in a recent Q&A post.

          In comment #15 I asked,
          “The question I have is how best to cut vent openings in the rim board. I picture the I-joists fastened to a rim board at the top plate. For the sake of the air barrier formed by the CDX vent baffles it would be best to limit the vent openings to no greater than the depth of the top flange. Would it be easiest and/or best practice to bore holes through the rim board or some other method? Would it be best to do this before the rim board is installed or after?”

          Malcolm Taylor replied,
          “It depends on how the overhangs are supported. Usually the I-joists continue out, so then rather than a rim-board there will be blocking between the rafters that can be kept down several inches. If there is a continuous rim-board it gets a bit dicey as the attachment is through the top flange of the I-joists, so you don't want to be removing too much material in-between.”

          I wonder if your architect had intended to terminate the I-joists with a continuous rim board, as shown in your comment above, or use blocking between each I-joist. It would be great to get some responses focused on the transition of vent channel to soffit assuming typical eave soffit framing. Seems like an important detail that gets little discussion.

          1. Jon_R | | #15

            > notching the rim board

            Not my area, but in some cases, the vent is brought through the sheathing to exit from underneath the shingles (or I assume metal). See "Shingle Over Intake Vents". This provides some ice dam advantage in that it never sucks in solar/wall heated air (as soffit vents can). Like a ridge vent, there may be issues with snow restricting airflow.

          2. mpsterner | | #16

            Hi Aaron, thank you for the feedback.

            I asked the same question to the architect and he provided me with the attached drawing. I am not sure that I love this notching in the LSL, but it will likely work.

            Jon R. had pointed out earlier that air doesn't work exactly like liquid where there is an fixed volume that can fit through a pinch point (like the notching). I am not sure I understand the science behind this though.

            On a different note, I just heard from our local lumberyard and regional truss manufacturer and they're VERY concerned about being able to even get the i-joists I've spec'd here and especially the large spanning ridge beam, which is essential. They said that it could be September before they're able to get that.

            So, with that, I am looking at what other options I may have for getting a high r-value cathedral with materials that I can source.

          3. Jon_R | | #17

            Maybe a steel truss for the ridge beam?

          4. Expert Member
            Akos | | #18

            If you look at the TJI application note on roofs, you can cut a pretty large hole through the roof rim board (P17):


            This opening is a couple of times larger than the free open area typically needed for a pitched roof, I doubt it will add a lot of restriction.

            As for vent opening and air sealing, this should be done at the ceiling level, sealing the vent baffle doesn't do all that much besides adding a lot of extra labor.

            Supplies issues are definitely real and annoying, I've had to use I-joists two steps up on the flange width because it was all that was in stock. Try to design an assembly where there is a bit of wiggle room on size and material. Usually you can go from I-joists to LVL / PSL.

            I've also had good luck with built up CFS beams. Availability is not an issue in these and can be ordered in any length. Since it is steel, it needs to be fully inside the thermal envelope, so it does need a bit of design work. Since they are site built, they are pretty light and much easier to place than a wide flange iron beam.

    2. Expert Member


      Good suggestions here! I was about to suggest a similar idea on catheralizing(?) the entire ceiling.

  6. onslow | | #9


    I have attached a picture to support my concern about the eave's length. I live in a semi arid area of SW Colorado so our snow fall might average about 48" total over the whole winter. We have had a few big dumps of 20 plus inches in one go in past years, but typically it arrives in 3-9" batches. Our snow tends to be the nice powder skiers come for which is quick to condense downward or seemingly evaporate. Just the same this pile (shown as of a few weeks ago) will be with me until late April, as it is on the north side of the house. It was 60F today, but it just laughs. Your snow load parameter suggest similar conditions, unlike the lady in Wyoming that is expecting 7-8 feet each winter.

    My stucco belt line is about 8" above the grade on this side, yet the snow pile is up onto my stucco about 8". So far, as each season ends, the stucco has come through fine. Depending on your own cladding choices, results may vary. I will note that I have a disguised surface trench with perf pipe running the length parallel to the house to direct melt water off to the trees beyond. The section of roof feeding the pile is only 14'. The roof plane on the second floor above is 1 in 12 pitch, so the snow just sits and melts off, not slides off. It is not contributing to the pile. Your much more aggressive pitch may well shoot the snow out further than the lazy snow curls I get.

    I am glad to not see gutters on your plan view. I was warned not to go there, as the melt water would freeze in them and eventually tear them off. I also avoided snow brakes on my metal roof, which has proven a good choice. Most of the time snow slides off at a glacial pace thanks to the ice slick under the snow. The ice slick forms due to sun heating the metal roofing enough to send trickles down under the snow cover. It is not from heat losses, the roof is R58 through center and best calcs show over R50 whole roof. As noted I did not go with a vented roof. The choice was forced by my electing to go with low pitches and what I call Faux-dobe styling. Parapets and low pitch roofs just don't work with a vented approach.

    I think the mini attic would be a bit easier than cathedral-izing the space. Plus the collar ties may be helping fight wind racking more than you think despite the large ridge beam. Fitting in lighting will be a bit more straightforward as well. Assuming winds are part of your local conditions, the negative pressures developed from blowing around and over the house should help to pull air out of the attic volume via gable vents. I will leave calculating the square footage needed to be effective to others. Just be sure to use 1/8" mesh behind any grilles on gables or soffits. As Akos has shown, the heat losses being vented are quite small, so the gap size should be sufficient. You are correct to note that many soffit venting choices don't add up to the vent channel cross section.

    My snow curls and snow pile....

  7. parker_h | | #19

    Hi Mike,

    Great questions and discussion you began here. Would you mind sharing what roof details you ended up with? I'm in the middle of designing a roof system with the same concerns.

    I recently asked a truss manufacturer about making a 16" deep parallel chord mono truss that rests on either side of a ridge beam to replace a typical i-joist. It would use a 2x4 top and bottom chord - did you ever look into that? For my design, it allows for a top chord tail extending out to provide the roof overhang structure without sistering on additional lumber. It also provides openings in the web with no additional labor of cutting holes, and I'm assuming no need for stiffeners at the bottom or top bearing. It would also seem to greatly reduce thermal bridging to the roof sheathing.

    I would welcome any feedback on the pros and cons of this idea.

    1. mpsterner | | #23

      Hi parker_h,
      I was building right in the midst of really bad supply chain issues, so I actually had to adapt my roof system anyway. I did very close to what you're describing but I had some specific aesthetic things I was also trying to accomplish. I didn't want to move to a deeper parallel chord truss because it started to mess with the new england cape look we were going for and we already had deep chords on the roof and 1st floor system.

      I've attached some pictures of what we ended up doing, with a structural ridge assembly as well to keep the height as low as possible but still get 18" of insulation. I also had 4" of external rock wool overlap where the double top plates were, so that helps too.

      I've never seen any ice dams, icicles or anything. Not even from the sun actually, though I am sure we'll see some from the sun at some point.

      Let me know if you have any other questions. Good luck!

  8. user-723121 | | #20

    A deeper parallel chord truss would not need a center beam. I have used a 24" deep truss with a dedicated air space. With this you get full insulation at the exterior wall, snow will lay uniformly on this roof in a cold climate winter.

    1. parker_h | | #22

      user-723121, thanks, I originally looked into that. It makes sense to me, but it does require at least 8" additional joist depth, and that depth would only become greater with increasing spans and snow-loads because of the unique need for the tall moment connection at the peak. The mono trusses supported by a ridge beam also provides greater flexibility at shed dormers and asymmetrical gables.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |