Statistics and Observations on the Colorado Blizzard of 2003

This article originally  appeared in the December 2003  issue of Claims Quarterly, published by the CPCU Society, Malvern, PA,,  and is reprinted with permission.

6591305365_66a3dfed06_b-300x225On the morning of March 18, 2003, snow began falling at the rate of one-half inch per hour and at times, over an inch per hour. Overnight, as much  as 18 inches fell in some areas near the foothills.  By Wednesday morning March 19, 2003, most of the roads in the Denver Metro area were no longer  passable, and many businesses began shutting down for lack of employees. Most of the neighborhoods were not plowed because the plows could not get through snow higher than the plow blade. Most residents in the city and outlying neighborhoods could not get out of their streets until Saturday March 22, 2003. In some areas as much  as five-feet of snow accumulated, not including drifting. The official snow level in downtown Denver was 32 inches.

Colorado  sustained the highest insured loss for the first quarter of 2003 in the United States with property and casualty insurers according to Claims Magazine. At $315 million, the majority of these losses were a result of a record-breaking snowstorm that hit the Denver, Colorado metropolitan area beginning  March 18, 2003. This storm caused extensive structural  damage to buildings.

The majority of commercial and residential structural roof collapses observed and analyzed by Professional Investigative  Engineers (PIE) failed  as a result of insufficient  design and/or insufficient construction. Tom Meyers, plans analyst with the City and County of Broomfield and the current president of the Colorado Chapter of the International Code Council, said the failure  mechanisms  they have seen as a result of snow loads include, “missing structural components and damaged structural components.” PIE found that upon structural analysis, many roofs in Denver County, on paper, failed between 10 to 20 pounds per square foot (psf). Today Denver requires between 25 to 30 psf, depending on the category.

A large percentage of the residential failures observed were related to “rafter”- type structures. Many older homes are constructed with rafter-type systems in lieu of pre-engineered wood truss systems or post-and-beam construction. With a rafter-type  system, the rafters rely on the exterior walls to be sufficiently rigid. As the roof structure is loaded, the rafters compress the ridge board at the top and exert outward  pressure on the bearing walls. Based on our observations, these rafter-type roofs typically failed when the bottoms of the rafters were not adequately tied together by means of collar ties, or the bearing walls were not sufficiently rigid. When a rafter-type roof fails, the roof “flattens out” and the bearing walls of the home displace outward causing significant  damage to the roof members and the bearing walls of the home.

In addition to the rafter-type roof failures, some failures were also observed in residential structures with post-and-beam construction. PIE observed failures in both beams and roof joists where either drifting loads exceeded the strength of the members or where the structural members were undersized even for the basic design snow load  as required by most of the building departments in the Denver Metro area.

A large amount of additional claims have occurred due to excessive deflection and cosmetic damage in the structures. Many residential structures were subject to large deflection and “serviceability failures,” including drywall cracking, sagging roof members, bowed bearing walls, and racked doors. In these cases, a serviceability failure has occurred, which is more difficult to assess than a catastrophic collapse.

Jefferson County  requires that building design consider “drifting and unbalanced snow loads.” Jefferson County’s  Basic Snow  Load table  can be accessed through their web site at In addition to Jefferson County,  any building department that has adopted the Uniform Building Code as its governing code requires that drifting and unbalanced snow loads be considered. Many of the structures PIE inspected did not take these conditions  into consideration. During this event we observed drifting snow, typically on north slopes, commonly in the 4 to 6 foot range and reported as high as 8 to 10 feet on one roof collapse we investigated in Windsor, Colorado. According to this snow was “very wet: about 8 inches equaled an inch in precipitation, as compared with the usual 11 inches,” meteorologists said. According to the Rocky Mountain  News, there  was as much  as 1.2 inches  of water in 6 inches of snow. A cubic foot of water is equivalent to 62.4 pounds. Thus, in a foot thickness of snow (2.4 inches of water), the weight was as much  as 12.5 pounds per square foot on the structure. If 3 feet of snow accumulated on the structure, the weight  was nearly 37.5 pounds per square foot.

It is our firm’s opinion there is likely a large number of roofs that have been damaged during this event that are yet undiscovered and will be seen over the next few years, or could fail during a future  less significant event. Many of the rafter-type structures may have flattened out, and some of the structural members may have cracked  as a result  of the snowstorm. Many homeowners or building owners may not have fully inspected their structures for damage, which could cause future  problems.

Tom Meyers said that “snow loading and eccentric drifting” relevant to building codes have been under review prior to this event and are now more likely to be a priority to code officials and the Colorado Chapter ICC. PIE believes there is strong potential for code revisions because of the large number of failures from the March 2003 snowstorm. Several building officials do not believe code revisions are needed, but believe that good plan review, proper field inspection, and code enforcement are the keys to preventing future failures. There are even some code officials that are proponents of a reduction  in the snow loads in the Denver Metro area.

Although there were typical injuries and even deaths as a result  of a storm of this magnitude (i.e. heart attacks and vehicle accidents) what we found amazing was there were no deaths caused by the approximately 100 flat-roof collapses. Our firm investigated commercial buildings that experienced catastrophic roof collapses, and had they been occupied, the fatalities would have been numerous. The lack of fatalities and relatively few injuries can be attributed to businesses closing as the snow accumulated and people taking refuge in their homes. In most  cases, employees just could not get out of their street to get to work. Events of this nature pull communities together. All across the front range and plains of Colorado, people pulled together to unbury cars, dig out driveways and streets, shovel snow off of roofs, and in emergencies providing transportation on snowmobiles.


Tags: , , , , , , , ,