Metal Roof Ice Dam Help
Hello GBA community,
I recently built a new home following the “pretty good house” principles in CO (zone 5). Much of the build was done with info learned from this site, and everything turned out awesome with one exception – minor ice dams. My roof assembly is a standard vented assembly with dark red metal instead of shingles. I’ve got continuous soffit venting and ridge venting. The attic is a sealed attic with R-70 blown cellulose. I have zero penetrations in my ceiling for can lights and exactly 2 electric boxes for ceiling fans, both of which were air sealed meticulously before drywall, so I’m confident I don’t have interior air leaking into my attic.
The ice damming only seems to happen on days when the temperature is sub-freezing but the sun is shining bright – on cloudy, sub-freezing days I have no melting, thus ice dams never form.
Is the problem likely from not enough air flow exiting the ridge vent, too much cold air entering the soffits, both, neither?….Any advice is appreciated.
Thanks,
Brandon
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Have a look on a few frosty mornings and look to see if you have an even frost pattern on the roof and see if the pattern looks even.
Is this happening on the south side of the house only?
Walta
Walter,
My roof ridge is oriented N/S, so the slopes face west and east. I haven't noticed the frost pattern, but the snow pattern seems consistent across the roof, even over the uninstalled garage portion.
Kammy, I can't find an article but I believe I've read one, and I know I've seen Dr. Joe Lstiburek of Building Science Corp talk about this issue. The sun heats up the wall surface, warm air (even on very cold days) rises into the soffit vent and warms the lower portion of the roof. Snow melts, the sun goes down, the melted snow re-freezes and eventually you get an ice dam. A few ways I recall him recommending to avoid this issue is to paint the wall a light color, locate the soffit vent as far from the wall as possible, and/or insulate the soffit space. Insulating the soffit space makes venting difficult. Or change to an unvented roof assembly--easier said than done.
Figure 5 here:
https://www.buildingscience.com/documents/building-science-insights/bsi-097-de-icing-ice-dams
Brandon,
About one half of my roof area is unvented standing seam metal and ice formation is normal for a metal roof in Colorado sunshine. The other half of the roof area is PVC membrane and the melt patterns differ noticeably between the two materials. The metal melts snow much faster. My whole roof insulation of R50 keeps the BTU loss through the roof well below what is necessary to melt snow. Sunlight on the metal roofing is what is driving your snow melt/icing. Snow melt also occurs on empty metal roofed shed buildings, which in theory should have perfect cooling on the underside of the roof.
It appears to take 144 BTUs to melt one pound of ice. I am guessing a similar BTU requirement to melt snow based on the notion that snow is just fluffy artistic ice. Fresh dry snow fall is supposedly 3-4 lbs a cubic foot, so a 3-4" snow fall should be about 1 lb per sq ft.
Solar energy available at sea level is generally valued at 1000W per sq. meter or roughly 100W per sq. ft. With perfect capture of all energy falling onto a surface, one could expect 3412 BTUs of heat per hour per square meter or 341 BTUs per sq. foot per hour. Metal roofs are not perfect collectors, but pretty good ones just the same.
Even at a 50% collection efficiency, full sun on a square foot of metal roof could provide 170 BTUs of energy in an hour. This is ignoring a number of variables to more simply show that sunlight, falling on a small exposed strip of metal roof can start snow melting. The heat gathering in the metal only needs to melt the adjacent edge of the snow to create melt water. The melt water will then slide down the roof until it gets cold enough to freeze again. The snow covering below the exposed metal roofing will actually act as an insulator that keeps the metal cold and freezes the melt water. Add sun, melt, freeze, repeat.
A gable roof on a 24x36 house might present two surfaces of 15 x38 or 1140 sq ft and in the prior snow fall example carry 1140 pounds of fluffy ice. In a very simplistic calculation framework, it would take 164,160 BTUs to melt off the roof. Imagine that much energy hiding under your eaves. Much easier to believe that much energy falling on top of the roof in the form of sunshine. I believe this is why solar panels are mounted there.
This is why I find it difficult to imagine that warm air rising from under your eaves into your vents could be the main driving force given the relatively low heat capacity of air, especially drier air in winter. Additionally, winds would tend to sweep away any warmed air. The energy falling on a wall surface is more likely to be absorbed into the wall mass much faster than re-radiate back into placidly awaiting air that will sneak into your vents.
Adopting the formula for delivered BTUs in furnace air handling (delivered BTU = Delta T x CFM x 1.08) it appears that it would take an hour of 3.3 CFM airflow with a 40F rise (20F to 60F) to carry 142+ BTU to ONE sq. ft. of roof area. To satisfy the surface area of just one 16" vented bay from eave to ridge (~18s.f.) would appear to need either 3.3 CFM of 40F rise air for 18 hours or 59 CFM for one hour to melt the snow cover over just one joist bay completely. (As I said, this is a simplistic framework I am laying out.)
It is true that you don't need to melt all of the snow to get some icing going. However, one can see it takes a fairly massive amount of warm air over a period of time to achieve melting from underneath. I could believe that warm air entering the soffit vents could create a bit of melting within a foot or two of the entry point, but not more as the air flow rate out your vent ridge length is (hopefully) not close to 180 CFM, the total of 27 16" bays on two sides of roof.
Beside, the traditional diagram of an ice dam has the ice forming just above the top plate because the overhang is colder than the upper part of the roof, which is warmer due to poor insulation above the living space or can light leakage. Your R70 attic insulation means a BTU loss of less than 1 BTU per sq. ft. at a delta T of 50 and exactly 1 BTU if it is -2F outside. Even if a heavy snowfall plugged your ridge vent, the heat loss from below should not be sufficient to cause melting.