About IceDamCheck

Why this exists

Ice-dam advice is usually written for one place and read everywhere. A homeowner in a town that gets two snowy weeks a year hears the same warnings, and the same sales pitch for heat cable, as one in northern Minnesota. IceDamCheck answers the question for the town: in a normal winter, how many days does a roof here have snow on it, freezing air at the eave and enough melt above to build a dam, and which fix is worth paying for. It covers 3,206 US cities with ten thousand or more people.

What goes into the number

Climate is the NOAA NCEI 1991-2020 U.S. Climate Normals at the weather station nearest each city (1,379 stations in all): monthly snowfall, days with at least an inch of snow on the ground, days that stay at or below freezing all day, and days that cross freezing. Where the nearest station does not record snow on the ground, those counts come from the next nearest one that does (292 cities). City locations and populations come from the Census Bureau's Gazetteer and population estimates.

An ice dam needs three things at once: snow on the roof, freezing air at the eave, and melt above it. For each month the model takes the share of days with snow lying, multiplies it by the days that stay below freezing (melt from attic heat) and by the freeze-thaw days (melt from sun, weighted lower because it is weaker), and scales the attic part by how much heat the attic leaks. The sum is "dam days" per winter. Five bands follow from it: under 1 is very unlikely, under 5 low, under 15 moderate, under 35 high, and 35 or more severe.

What the data shows

On a house with a typical attic, 1,697 of the 3,206 cities (53%) get at least one dam day in a normal winter, and 586 (18%) are high or severe. Sealing and insulating the attic moves 797 cities down at least one band, and cuts the high-or-severe group to 91. The worst is Fairbanks, AK, at 80.6 dam days. Snowfall alone misleads: Cedar City, UT gets 60 inches a year but only 5.0 dam days, because its snow melts off between storms.

Known limits

This is a screening estimate for a typical house, not an inspection. Roof pitch, the shape of the roof, skylights, recessed lights, ducts or a furnace in the attic, and how much snow the wind strips off all change what happens on a particular roof. Climate normals describe a typical winter, not a record one. Several cities share a weather station, so their figures match even where the towns differ, and a station at a different elevation from the town is flagged on that city's page. If you already see icicles thicker than a finger along the eave, believe the roof over the number.

Corrections

If a city's figures look wrong, say so through the contact page. Confirmed errors are fixed in the data or the model, which corrects every page that depends on it at once.

Who runs this

IceDamCheck is independently run. It is not affiliated with any roofer, insulation contractor, heat-cable maker, insurer or government agency, and nobody pays to change a city's number.

Written by

Each city page is generated by AI software from the cited NOAA climate normals and Census place data, and no journalist or editor reviews a page before it publishes. The dam days, bands and costs come from that data and the model described above, not from AI; the AI only assembles the sentences around them.