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
Musings of an Energy Nerd

Ice Dam Basics

How ice dams form, how to prevent ice dams, and what to do if you have an ice dam right now

When melt water reaches the cold eaves of the house, the water freezes.
Image Credit: Image #1: Andrew Kuchling

What do you call the weeks between Valentine’s Day and Easter? It’s ice dam season, of course. Eastern Massachusetts is now the wet-ceiling capital of the world, but this winter, tens of thousands of homeowners from North Dakota to Maine are struggling with ice dams.

I’ve written about ice dams before. This article makes no attempt to be comprehensive; instead, it’s an introductory article to summarize basic information on ice dams. If you still have questions after reading the advice here, you should probably read “Prevent Ice Dams With Air Sealing and Insulation.”

How ice dams form

1. Ice dams form when a home’s escaping heat warms the roof sheathing and melts the underside of the snow layer on the roof. Water trickles down the roof until it reaches the cold roofing over the eaves, where it freezes. After a while, the ice at the eaves gets thicker, forming an ice dam. If the water reservoir behind the ice dam is large enough, water can back up under the roof shingles and damage ceilings. (Image #2, below, depicts the complicated shape of a typical ice dam.)

2. The most common cause of ice dams, by far, is air leakage. Warm interior air leaks through through ceiling defects or recessed can lights, as well as through hidden passages in wall assemblies and roof assemblies, and this warm air contacts the underside of the roof sheathing.

3. The second-most common cause of ice dams is thin or missing insulation.

4. Bad roof design can make ice dams worse. Since ice can form in gutters, roofs with gutters (especially gutters that extend higher than the plane of the roof) are more susceptible to ice dams than roofs without any gutters. Roofs with valleys are more…

GBA Prime

This article is only available to GBA Prime Members

Sign up for a free trial and get instant access to this article as well as GBA’s complete library of premium articles and construction details.

Start Free Trial


  1. user-1135248 | | #1

    when venting isn't
    Okay, I see one major issue with the whole vented thing, whether
    it's an over-roof on top of a robust insulated structure or a
    typical vented-attic setup. When the RIDGE VENT is covered with
    snow, there's no more venting to speak of. There may be a bit
    of thermal break across the vented space, but either configuration
    is probably going to let a bit more heat than expected accumulate
    in what is now basically a dead-end air space that's only open at
    the eaves and wind up warming the roof deck anyway.

    And having R-20+ of snow on top is going to put the deck somewhere
    in the middle of the overall thermal gradient in almost all cases.
    I noticed this when I went to shovel the weight off my shed dormer;
    no ice dams, but still some evidence of melt-n-refreeze going on
    under there.

    Possibly one argument for gable vents, as outdated as they seem
    to be these days... they won't get snowed on.


  2. GBA Editor
    Martin Holladay | | #2

    Response to Hobbit
    The R-value of snow varies depending on its density. Estimates of the R-value of snow vary. According to one source, the R-value of fresh dry snow is R-0.5 per inch. Other sources put the R-value of snow at R-1 per inch, or even at the (unlikely) value of R-2 per inch.

    In most cases, an R-20 layer of snow would have to be at least 20 inches -- and more likely, 40 inches -- deep.

    Snow-covered ridge vents still work, for the simple reason that snow is not an air barrier. Of course, air flow through snow-covered ridge vents may be reduced by snow -- but it won't be stopped.

    Snow doesn't cover most roof ridges for long, because roof ridges are exposed to wind, and in some cases to sun. Snow will cover the ridge of a low-slope roof for longer than it will cover the ridge of a steeply pitched roof. My advice for a good (ice-dam-resistant) roof still holds: no valleys, no gutters, a steep pitch, and standing seam instead of asphalt shingles. These rules can be broken, of course, but breaking the rules moves your roof in the direction of an increasing risk of ice dams.

    "Vented over-roofs" still work, as long as they are part of an overall approach that includes good roof design, attention to airtightness, and adequate R-value. The temperature of the air in the ventilation channels is going to be relatively low, even when the ridge vent has some snow on it, as long as there is an effective layer of R-49 insulation under the ventilation channels, and as long as there is at least a little bit of air flow through the ventilation channels. The air flow rate through these channels will vary, of course; the main driving force will be wind, so more air will move through the channels on windy days than on days when the wind is still.

    The final proof is in the pudding: if you follow my design guidance, you won't have any ice dam problems. People who build roofs according to these guidelines know that these approaches work.

  3. Paul Eldrenkamp | | #3

    ice dam formation

    Your description of an ice dam is a little dated. Check out pages 113-114 of Bill Rose's book "Water in Buildings." Bill refers to Canadian research that indicates that the real problem is not so much the dam itself but the lens of ice that is formed uphill of the dam. "It is this lens that allows the buildup of water, providing a water head sometimes several feet in height."

    Bill's analysis of ice dams in general remains the best I've ever read (no surprise there). I wonder if you could get permission to reprint just that section here on GBA?

    --Paul Eldrenkamp

  4. GBA Editor
    Martin Holladay | | #4

    Response to Paul Eldrenkamp
    Thanks for your comments. You wrote, "Your description of an ice dam is a little dated. ... I wonder if you could get permission to reprint that section [on ice dams] here on GBA?"

    In fact, the shape of the ice that you call a lens is illustrated in an image (Image #4) posted in my first GBA article on ice dams. (You may have noticed that in this article, I referred readers there: I wrote that this is "an introductory article to summarize basic information on ice dams. If you still have questions after reading the advice here, you should probably read “Prevent Ice Dams With Air Sealing and Insulation.”")

    The caption to that GBA image reads, "This drawing of an ice dam is more accurate than the other (more simplified) drawings. Based on careful observations by researchers from Canada, where residents know a thing or two about ice dams, the drawing comes from a CMHC publication, "Attic Venting, Attic Moisture and Ice Dams"."

    Moreover, that earlier article on ice dams benefits from a posted comment by Bill Rose, who wrote, "Don Fugler and other Canadians have refined the picture [of the shape of the ice that forms on a roof]. It includes formation of an ice layer at the bottom of the snow extending up the roof, and melted water gets beneath this ice."

    While the refined analysis by Don Fugler and Bill Rose, explaining the shape of the ice lens that forms on a typical roof, is helpful -- I always wondered how water could back up as far as it does to cause so many leaks -- a full explanation of the shape of the ice lens is (fortunately) unnecessary for those seeking to implement ice dam remedies.

    I am reproducing Image #4 from my earlier article below. (I have also added this drawing to the article on this page, as Image #2, to clarify for readers what a typical ice dam looks like.)


  5. BrucePPH | | #5

    Ice Dams and HVAC
    Attic duct work also needs checked for leaks. Its not just a blower door. Many supplies run to the edge of the exterior wall and are located close to windows. All connections sheet metal and flex duct need sealed.

  6. GBA Editor
    Martin Holladay | | #6

    Response to Bruce Chyka
    You're right, of course. Fortunately, most HVAC contractors in cold climates know better than to install ducts or HVAC equipment in a vented unconditioned attic. (That problem is much more common in Texas and Florida than in Vermont.)

    If you are a cold-climate homeowner with an ice dam problem, and you are unlucky enough to have been bamboozled by an HVAC contractor who was stupid enough to put ducts in your attic, you have my sympathy.

    The best solutions are either (a) to move the ducts and HVAC equipment to a new location inside the home's thermal envelope, or (b) to transform your vented unconditioned attic into an unvented conditioned attic. (For more information on this option, see Creating a Conditioned Attic.)

    Trying to address the issue by leaving the ducts in place, and merely sealing the duct seams and improving the duct insulation, usually results in disappointment.

    Thanks for your comments, Bruce. I have edited my article to reflect your points.

  7. qofmiwok | | #7

    How do you feel about insulating eaves in snow country like Joe Lstiburek suggests?

    1. GBA Editor
      Martin Holladay | | #8

      I assume that you are talking about the situation that Lstiburek discussed in his article called "De-Icing Ice Dams." (The illustration shown below comes from that article.)

      This type of problem is relatively rare, but can occur in very snowy climates, especially on the south side of buildings with dark siding. In my opinion, insulating roof overhangs isn't usually necessary -- but in rare cases, insulating a roof overhang may fix a stubborn problem.

  8. guillow | | #9

    Wouldn't a PV system always solve ice damming? As long as it provides sufficient coverage of roof? Essentially creating and "cold"/"vented" overroof?

  9. montlamothe | | #10

    Hi GBA folks, this is an older thread but my question is in line with the conversation.

    I'm building a house in southern Washington in a town called Trout Lake, it’s just south of Mt. Adams, lots of winter snow, fairly hot summers. The ground snow load is 134psf, which apparently is quite high. Ground snow load is very locally defined here, due to lots of elevation changes in the area. We're building a 1600 sqft rectangular, 1-story house, with a simple shed roof with generous 3-4ft eaves/overhangs. House is located in the middle of a field, with a covered connection to similar garage. The roof is held up by large exposed glulam beams. My architect and structural engineer have designed an unvented roof assembly using 9” of polyiso (roughly R-60) with a standing seam metal roof.

    I’m worried about ice dams. Assuming proper air tightness and vapor permeability, should I over-vent the insulation by elevating the metal roof 2”, making it a “cold roof" as Dr. Joe advises in heavy snow areas? Or does this seem like overkill for my region of the country? The snow loads are high, but winters are typically short, from December to February.

    1. Expert Member
      MALCOLM TAYLOR | | #11


      Posts here in the Recent Article Comments disappear very quickly. You will have better luck putting this in the Q&A.

      Sorry, I don't know enough about ice damming to offer any help.

      1. montlamothe | | #12

        Thanks Malcolm, I'll migrate it over.

Log in or become a member to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |