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

Ice Dams: Totally preventible? Or Not?

tundracycle | Posted in General Questions on

At a meeting this past week with our builder and his insulation contractor the topic of ice dams came up.  They both said that they can be minimized but are not totally preventible (in MN). 

Their primary bit was that there is only so much that can be done in roof valleys and some other places.  True? Or with enough effort on design and insulation can they be totally prevented almost regardless of roof design?

My thought is that certainly with a PERSIST build and simple roof they’d be totally preventible. And that similarly a simple roof and PGH could totally eliminate them?

FWIW, the roof plan for our new house:


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. this_page_left_blank | | #1

    "with enough effort on design and insulation can they be totally prevented almost regardless of roof design?"

    You've got a contradiction there. I'm sure it's possible to design a variety of roofs to never have ice dams. But to design one to do so regardless of design... what does that even mean? If you mean start out with a design prone to ice dams and then tweak the design to prevent all ice dams, I'd say no, that's not necessarily going to be true.

  2. tundracycle | | #2

    Thanks Trevor. So, with my roof (posted above), standard stick frame construction, and all HVAC within the envelope as an example. Should it be reasonably possible to insulate that roof in a way that no ice dams will form? Or will they be inevitable with that roof design?

    To complicate it a bit. In this case the attic is fully finished so the insulation (builder's plan is 7" closed cell) is in the roof, not the ceiling. Some is vented with Duro-vent (another post coming with questions on this), some cannot be vented due to valleys or other bits.

    So perhaps my question is... Is a very simple roof required to prevent ice dams or can they be prevented in more complicated roofs such as ours?

    1. GBA Editor
      Martin Holladay | | #3

      W. Ramsay,
      North-facing roof valleys in Minnesota are an accident waiting to happen. It's best not to have such valleys, if possible.

      Preventing ice dams means many things -- to some builders, it means installing lots of electric heating cables in the valleys, and plugging the cables in all winter long.

      A solid copper roof (with soldered seams) is invulnerable to ice dam damage, of course -- but not immune from ice dams.

    2. Expert Member
      Dana Dorsett | | #4

      >"To complicate it a bit. In this case the attic is fully finished so the insulation (builder's plan is 7" closed cell) is in the roof, not the ceiling. "

      That's a lousy plan from an ice damming perspective, due to the thermal bridging through a mere 7" of thermally unbroken rafter, with R8.5 stripes between the R49 cavity fill. Ice dams get started by melting at the lower-R sections/features of the roof, refreezing as the flow reached the higher R sections. The melt-out at the rafter stripes is an ice dam starter, primarily during warmer but still sub-freezing weather.

      As little as 2" of EPS (R8.5- R9 ish @ 40F) above the roof deck would cut the melt-out at the stripes in half. A 4" EPS nailbase panel has ~3.5" of EPS, and would run about R15 @ 40F, a 6" EPS nailbase panel would be north of R21. A 6" nailbase panel combined with 4" of HFO blown closed cell on the underside of the roof deck would deliver a code-min R49+ at center cavity, with the framing fraction at the rafters about R26 or more during the snowy months. That cuts the heat flux through the thermally bridging rafters by 2/3 compared to a 7" HFO blown closed cell foam solution.

      Every inch of thickness reduction on the closed cell foam adds about 1.7" of thickness to the EPS to hit the same R-value at center cavity. But it adds about R6 to the framing fraction, reducing the ice damming potential.

      With 4" of closed cell foam on the underside of the roof deck it's safe to add fiber insulation to the interior(up to the combined R value of the exterior panels + closed cell spray foam), further reducing the thermal bridge, and increasing the center cavity R at low cost. If they are 2x8 rafters stuffing some R13s in there for a full fill of the remaining 3.24" would add about R4 to the thermally bridging rafter, and about R12 to the center cavity R. That would allow you to get away with a 2.5" nailbase panel and still hit code min, but the thermal bridge would be "only" R17-R18 (still less than half the heat flux through 7" of rafter.

  3. Expert Member
    Peter Engle | | #5

    The dormers are also going to make life difficult. It's tough to wrap a foot of insulation on a dormer only a few feet wide. Dormers, cathedral ceilings, north-facing valleys are all design features that bleed tons of heat and beg for ice dams to happen.

  4. user-723121 | | #6

    W Ramsey; Roofs without ice dams are insulated to code and beyond and vented from eave to ridge. The rafters have an energy heel to allow for full insulation over the exterior wall top plates. The house is very airtight, 1.5 ACH50 or less. Like all who said above, avoid potential problem areas and intersections (no saddles, valleys meeting valleys, etc.).

    I see problems ahead for your roof design, insulation and venting plans. This winter in Minneapolis has been very demanding of building systems, ice dams showing major deficiencies. It is a good time to learn from this experience, especially when building new. As I like to say to anyone considering a building project, "spend a lot of time planning".

  5. 730d | | #7

    My experience in Minnesota is, A Code minimum insulated attic or rafter with chutes/ baffles particularly on vaulted rafters will 99% of the time mean no ice dams. 50% percent ventilation at bottom and 50% at top.

    Snow is a good insulator and if it's deep enough and with the sun shining it will melt and dam
    even over an unheated space. Generally not a problem . 5/12 roof pitches facing south west with a valley have been the only ones that have given me trouble. The 5/12 seems to be the wurst pitch for causing snow to accumulate at greater depth. With some bad luck , enuff snow and sun at the wrong time of year and I wurry. It does not matter how you insulate or ventilate as it all happens above the roofing. Use any kind of ice and water you like and it still might leak once every ten years.

    1. 730d | | #8

      What I am trying to say is the snow and sun are creating the heat and dam. Not the heat from your building.

      1. 730d | | #9

        I was was on the south shore of lake Superior a few years ago and observed dozens of old er homes, 100 years old or so. Really large elaborate homes. There solution seemed to be the bottom four to six feet was metal roofing with asphalt the rest of the way up. Probably the snow was more likely to slide off and or it was easier to us a snow rake.
        It was kind of nice looking also.

  6. Expert Member
    Peter Yost | | #10

    HI WR -

    Martin wrote a great article that should be read before discussing ice dams in detail: "Ice Dam Basics" - In this article, Martin references several great resources for exploring ice dams more deeply.

    Just two additional quick comments:

    Simply-designed roofs with continuous air control layer and soffit-to-ridge ventilation in every roof framing bay don't result in ice dams OR ice dam damage.

    Many builders and roof contractors believe that they can vent their way out of any ice dam problem; that is the problem: most ice dams result from combined air and thermal bypasses that no amount of ventilation can overcome.


  7. Jon_R | | #11

    I'm going to suggest a radical approach to ice dam prevention. The problem is cold eaves and a warm upper roof (a thermal balance issue). Apparently all recommendations revolve around making the upper roof somewhat less warm (agreed, this is often sufficient). But consider making the eaves less cold. Insulate the soffit (reducing downward heat flow). Make sure eaves have good air circulation to the slightly warmer attic air. Most importantly, don't vent the soffits with outside air (use one of several alternatives for intake vents). Pulling outside air through eaves just makes them colder (exact opposite of what is needed).

    1. Expert Member
      BILL WICHERS | | #12

      Soooo, how is this different from just using heat cables on the eaves? Making anything warmer implies more heat loss from the occupied spaces, which is inefficient from an energy use standpoint.


      1. Jon_R | | #13

        > Making anything warmer implies more heat loss from the occupied spaces

        No. A warmer attic achieved with less heat loss out attic vents causes the opposite effect. And it's an option - once you do a better job with the eave/attic thermal balance, you need less ventilation of the entire attic. Less attic ventilation -> warmer attic -> a little less occupied space heat loss.

        Heat cables (or better, mats under the shingles) are the 100% solution (eaves can be kept reliably warmer than the upper roof), but people don't realize how little energy is involved in the ice dam issue and seem to prefer passive attic solutions (despite active systems being used everywhere else in the house).

  8. burninate | | #14

    When we talk about a simple roof, we're trying to talk about one without any valleys, minimizing the number of planes, with no chimney, and no dormers. Gable, gambrel, or hip, but not a combination of cross gables and different shapes. These are the most vulnerable spots on your roof, and if you give them half a chance they will try to leak on you.

  9. Colin63 | | #15

    The main cause of ice dams are freezing and thawing cycles, snow melts freezes melts freezes and climbs uphill lifting shingles, j-channel etc... Reduce your shingle reveal from 5 5/8" to 5 1/4" in known problem areas, including a quality ice and water shield. Use a two part low slope roll roof if necessary around tight roof transition to dormer areas. GAF makes just about every matching color. Insulate the house, not the roof! If the roof stays cold less chance of ice dams.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |