Snow Load Code Development and Building Failures and Investigations

Snow Load Code Development and Building Failures and Investigations

“Winter is Coming”   OK, just making sure all you GOT fans are paying attention…

State College hasn’t seen much in the way of snow flakes so far this season but there is no question that winter weather is quickly approaching in many areas of the country in addition to Pennsylvania.  Some locations with higher elevations have already experienced several inches of snow on the ground.  With more snow on the way and the inevitable roof damage and collapses that seem to follow at some point, Paul Rouis, PE,  Principal of Ryan Biggs Clark Davis  in Albany, NY  participated in the Visiting Practitioner Failures Lecture Series to Penn State AE students on the topic of Snow Load Codes and their evolution over the years.  Mr. Rouis also discussed several snow collapse case studies including pointing out lessons learned so that hopefully you won’t make the same mistakes as others.

Structural collapse and roof damage from snow can occur from a variety of sources and often results from a combination of events including snow drifting, roof geometry, storm overloads from ice and snow combinations, design deficiencies  and construction defects just to name a few.   A good background article on this topic can be found in the November 2008 issue of Structure titled “Structural Collapse from Snow Loads” by Michael O’Rourke.  Another good summary of snow collapses can be found from the Building Failures Forum guest post by Alyssa Stangl titled Snow-Induced Roof  Failures and Prevention Methods which also leads to a detailed discussion on the topic on the Failures Wiki site.

On a related topic, the Whole Building Design Guide contains a section titled Considerations for Building Design in Cold Climates: Avoiding Falling, Sliding or Windblown Ice and Snow from Buildings and Structures.  In addition, no snow collapse discussion would be complete without first looking at FEMA P-957 Snow Load Safety Guide.

While snow collapses are nothing to kid about, MKev does occasionally find humor or at least irony in the topic as can be seen in the photo below.

I guess this means I shouldn’t be standing here but how was I supposed to read the sign?

And to start out the commentary, remember the comment by Mr. Rouis about how before the early 1970’s many roofs didn’t have much in the  way of insulation so the escaping heat pretty much took care of long-term snow build up?   See my comments  below from last year about snow melting on an industrial scale.   And, ask me in class about the snow melt idea for one of the campus buildings that fortunately never came to pass.   MKev

Tags: , , , ,

52 Responses to “Snow Load Code Development and Building Failures and Investigations”

  1. WangjaeY
    October 24, 2017 at 8:35 am #

    It was great to have the lecture from the expert in a particular area, snow load. I enjoyed the great overview of the history and development of snow load. It was interested that one of the case study was failed by the improper installation of solar panel. When I was working on the research of solar energy, I always thought the effect of solar panel in the gravity system of the building structure. Even with smaller modification in the snow, this will change story tremendously in the structure when heavy snow. I would like to see in-depth analysis in our class project in ice arena.

    • mkev
      October 24, 2017 at 10:19 am #

      Retrofitting solar on buildings has been the cause of a lot of problems, not just related to snow. Even in new design you still need to look at snow, wind and combination loads.

      We will be discussing the ice arena snow load in great detail as the actual load case vs design is one of the primary items you need to help discover the cause of the collapse.

  2. Tyler J
    October 23, 2017 at 3:07 pm #

    In the world of structural engineering, it is easy to overlook snow loads as simply another load on a building. However, the presentation by Paul Rouis depicted the many scenarios in which snow can drift, melt, or preform in some other manner to cause serious structural problems.

    I particularly enjoyed the scenario and associated images where an entrance had been added to by some sort of roof covering attached to the building. In the presentation, it was shown how snow can drift from a roof above onto an entrance roof causing extreme loading. Often, this load is not planned for as structural engineers are typically not heavily involved in the addition of an entrance roof. This type of scenario is similar to the case study involving the gymnasium that failed due to recently added snow jacks. It seems that often modifications are made to existing structures without considering all the possible affects the additions may have. This is especially true in the case of snow where drift patterns are not necessarily intuitive to the average person.

    Also interesting was the progression of snow codes from the 1950s until now. Though today’s codes sometimes seem excessively detailed, they seem to provide accurate design values. Early codes applied a blanket snow load with little thought to drift, topography, or building use. This makes me wonder how necessary the new codes really are. Certainly, there are advantages to designing with conditions such as drift as we do today, but I wonder if these conditions were thought about prior to code being established? Does anyone know if conditions such as drift were accounted for prior to code requiring it? Or was drift only accounted for as a result of building codes?

    • mkev
      October 23, 2017 at 7:53 pm #

      Excellent question. I can’t answer that one because it is a question that you will need to review as part of your ice arena collapse project!

    • Perry H.
      October 26, 2017 at 12:56 am #


      I did a little research into your question and I found this article. While the article is about drift control for highways and railroads, page 6 shows that the concept of snow drift has been around since at least 1852. I am not fully sure when the first reference to snow drift in relation to buildings first occurred but there is a good chance it was before codes based on this information. More research will be important for our projects.

    • WangjaeY
      October 26, 2017 at 8:40 am #


      I agree that structural engineers often forget about the detailed design on small addition to the existing structure such as the canopy above the entrance. I was also surprised that that canopy was caused the change of the wind and snow behavior and finally failed. It seems that it would have been better not to put the canopy and let the occupants get snow when they enter the building!

  3. CamilleS
    October 20, 2017 at 11:04 am #

    I really enjoyed’s Mr. Rouis’ detailed explanation of the different behaviors of snow loads that he has encountered in his work, that are not addressed by load calculations in present and historic codes. Personally, I feel that my analysis of snow loading is a little too dependent on ASCE 7 dictations of drift and roof loading, because my experience with snow in Texas is about 2 inches of snow every other year, and it melts in about 2 days. (I also thought those things on the roof were decorative and did not know they had a name until this lecture) As pointed out by the snow load drift problem we completed for this class, the code is very complicated to follow even for professional engineers, and as pointed out in Mr. Rouis’ presentation, the code does not even take into account some serious loading conditions that can cause failures. I would argue that based on these cases, snow loading should be based on a consideration of the building construction type and issues that may arise from that, as well as the site specific long term weather patterns, including rain on snow and partially melted snow that can cause the sliding snow condition in Mr. Rouis’ case study. These experience driven considerations seem to be just as important as following ASCE snow loading codes, if not more important than a blind faith in ASCE to cover all loading conditions the building will see in it’s lifetime. Tying back to previous lectures, this is just another reason that engineers should be engaged over the lifetime of a building. In tying back to previous lectures,not just the effects of age or use can have serious impacts on the structural integrity of a building, but also small modifications to the building that require the consultation of engineers, as well as unexpected weather conditions that contribute unexpected loading.

    • mkev
      October 20, 2017 at 4:03 pm #

      Glad to see you brought in the Texas connection. I once investigated a building located in the NE designed by an engineer in the south and they used minimum roof live load only, no snow. I guess that reinforces my point that every engineer should study failures to become better engineers, not necessarily forensic engineers.

      And those “things” on the roof can be decorative! Have you seen the ones that look like eagles? Sometimes called snow birds. That said, they usually are for snow and look good. I personally use them as decoration and have a couple (from failed buildings) on the wall in my home office. Then again, you all know I have no life so you might expect something like that.

    • Geoffrey T.
      October 24, 2017 at 12:04 am #

      Hi Camille,

      It was definitely interesting on how a simple snow jack can cause a collapse within a year of its installation. From Mr. Rouis’s lecture, it is very important to analyze the existing condition before adding any additional attachment. However, I would say that it depends on the building type. If the building is commercial (I would consider the case studies that Mr. Rouis brought to be a commercial building), the drawings need to be reviewed by professional engineer. However, if the building is residential, the owner does not need to consult the professional engineer (even though it is advisable) due to the type of the code used in constructing the building. While commercial building is built under IBC, residential property is built under IRC.

      • mkev
        October 24, 2017 at 10:26 am #

        Both IBC and IRC have structural requirements that must be met. Most code officials will tell you (IBC or IRC codes in force) that a structural modification needs engineer approval. For example, you technically can’t take a bearing wall out of a house without an engineering evaluation. That is an obvious one. Problem with snow jacks is that to most people it is not obvious that they could increase the roof load or snow load distribution.

    • Tyler J
      October 24, 2017 at 8:00 am #

      I agree with a lot of what you said. Code is typically developed as a result of the experiences such as the failures that Mr. Rouis showed us. As you pointed out, basic concepts such as snow sliding down a metal roof, can have complicated formulas to them by the time they make their way into today’s codes. It is a shame there isn’t an easier way to require engineers to design for the various conditions that have been experienced in the industry short of the complicated code. As a structural engineer, code or no code, I want to design buildings that will last and not fail under any foreseeable loading scenarios. To me, it makes sense to go above and beyond the code minimum by taking into account the items you mention such as building construction type, long term weather patterns, rain on snow, and sliding. As engineers, following ASCE and other code requirements only to meet the letter of the code will lead to failures. Code is meant to help us design, but shouldn’t be used as our source of all information and understanding.

      • mkev
        October 24, 2017 at 10:22 am #

        You are correct. Code is the absolute minimum. Engineers are free to establish higher loads or more redundant systems as you see fit (within reasonable budget restraints of course).

  4. benP
    October 19, 2017 at 11:17 am #

    When adding snow birds or bars to a building, it can be pretty easy for the strength of the existing structure to not be taken into consideration. I have seen on county projects were this has been overlooked. Most municipalities in my area use architects to bring projects and see the bidding and construction process thru. On this particular project, the architect hired an engineering firm out of Vermont to design the snow bars, however the calculations were never done to see if the roof was strong enough for the extra weight, so now we have been awarded the job, and we wait for engineering to determine this before the work can be performed.

    Also in the lecture, he presented a case study with solar panels. My senior year of high school we had a job on a restaurant were we installed PV solar panels. The existing roof loading was only calculated to be 7 psf. The insulation was virtually non-existent and we figured the heat loss on the building kept snow from hardly accumulating. Luckily, the town made the solar company hire an engineer to proof the roof was structurally sound and this was disaster was avoided. We ended up getting the job, our solution was to strip the existing roofing material (as per code, abandonment in place is not permitted), build 6 ft knee walls, and put new scissor trusses above the existing roof structure. This was costly to the owner, but it was better then having a collapse. With any hope, the energy savings from having an insulated attic will help offset some of the costs.

    • mkev
      October 19, 2017 at 11:45 am #

      Thanks for bringing in some actual examples to the discussion. There are a lot of examples of added solar panels causing roof strength problems. You discussion also reminded me that what we are not including in our “snow” talks so far is the wind load. Odds are that if the wind is causing downwind drift or scouring on a gable roof for example, then the calculation needs to include the load combination of wind and snow which if stress reversals are already taking place can cause even more problems (depending on the geometry of course).

      • benP
        October 19, 2017 at 2:27 pm #

        I could see this being a major problem with older “kit” buildings. Many older pole buildings in my area were designed by companies from the south where snow was not factored. If someone were to put snow bars on only the lower part of the roof (as is popular in my area) snow could build up on the lower part of the roof similar to what you mentioned. On a weak structure that was designed for a snow load around 12 psf (common for these “kit” structures) with a lopsided loading, the wind can be the force that can make the structure fail.
        Also on the topic of lopsided snow loading due to wind. My experience is the structures without collar ties do not perform very well. I’ve had more then once where I’ve had to jack the ridge back up into place and install collar ties. Depending on how well the rafter was connected to the top plate effects whether more work has to be done to re-plumb the wall. Sometimes we get lucky and when the ridge gets jacked back into place, the walls come back to plumb. If not that turns into another few days on the job. I have also noticed that when the snow is lopsided (before jacking takes place) one wall will stay closer to plumb and one wall will be way out. I’m not sure if this happens every time or just specific to the jobs I’ve been on. I’m curious to hear what your thoughts are? Thanks.

        • EllenW
          October 31, 2017 at 7:32 am #

          I have often wondered how the loads were determined for “kit” buildings. If they were designed by a manufacturer in the south, and were purchased and built in the north, snow loads on the structure would be higher, but wind would be lower. Did this allow the capacity from the designed wind loads to be “traded” to the applied snow loads? I know that in current commercial structures, the gravity and lateral systems are usually separate, but with residential light framing the wood stud walls carry both gravity and lateral, so this may be possible. It’s also possible that the stud spacing in the walls was used for construction convenience and was not governed by strength, so there was enough residual capacity to handle the additional loads present at the building location.

  5. Jared Patota
    October 19, 2017 at 10:09 am #

    How engineers designed for snow loading before the 1950s seemed a little ambiguous compared to what ASCE 7 requires today. Paul Rouis gave a few references to what was commonly used: like the Structural Engineers Handbook and the Steel Joist Institute. For our project with steel roof trusses built in 1948, these references may have been used to design for snow; however, finding the referenced used would not be the best way to find the truss’s design capacity.

    The most practical solution to finding what load the truss can hold would be to analysis the size of the members in the stress. By taking cross section measurement of the trusses significant members and finding the material design stresses, the actual truss capacity can be obtained. We can check the truss allowable capacity against today’s code and see if it passes.

    • mkev
      October 19, 2017 at 10:50 am #

      You make good points here. What if it doesn’t pass today’s codes? Is it unsafe? Can it be occupied? Do we grandfather old structures? Just pointing out that once you have the data, there are still judgement and ethical questions to consider. We also often like to know what the structure was designed for and see if it was appropriate to the time period as that helps give us some base reference point for a larger evaluation. But I agree, you can’t be sure what you have until you do the analysis.

    • Richard T.
      October 23, 2017 at 7:59 pm #

      Jared I like your approach to finding the capacity of a truss. However, for this to work, the truss would nearly need to be perfect. Assuming that all the members are made of the same material and that all have the same properties. In manufacturing, there is tolerance and I think it’s very difficult to judge the variances between each member. Another thing is how do you accurately model the connections between each of the members? There’s further variance because now you introduce human error and each person who puts together their truss might tighten the bolt slightly different.

      Maybe in the future when robots build our buildings this might be more attainable.

      • David K.
        October 25, 2017 at 7:57 pm #


        one could argue that this is the case for new construction as well. I believe at some point we have to make reasonable conservative modeling assumptions on the behavior, get as close as we can to estimating the material capacities, and accept the fact that we are not going to achieve the perfect model. Taking steel coupons and idealizing connections can lead to an acceptable model in many cases.

  6. HarryB
    October 19, 2017 at 8:22 am #

    One take away from the Snow Load Code Development and Building Failures and Investigations lecture was that Mr. Rouis suggested to read the mill building hand book to find the snow load used to design the historic building in Centre Hall. This was interesting because it meant that older building codes were not determined by region but by the industry that they were being used in. This could be because buildings were more likely to be built by the companies within the industry themselves instead of general contractors.

    My other takeaways from the lecture had to do with the ice rink collapse project. Mr. Rouis mention that reversal of loading was often a concern when dealing with unbalanced loading of snow drift. He also mentioned that this was mostly a concern with wood connections. Moving forward in the project, a deeper analysis of the connections of the glu lam arches should occur. Also, in the pre-engineered gymnasium case study the collapse happened because of two changes, snow jax and a removal of bracing. Therefore a list of renovations or additions to the Ice rink roof and structure would be helpful.

    • mkev
      October 19, 2017 at 10:46 am #

      By coincidence I have a copy of the 1932 edition of the 1903 Milo S. Ketchum Steel Mill Buildings on my desk. I will copy some appropriate pages for you. Your building was 1948 so we will need to get some “updated” resources. That said, resources were sparse “in the old days”. The AISC manuals going back to the early days also listed some information on snow.

      On the arch connections, see my response to Jeremy below.

  7. Shangmi X.
    October 19, 2017 at 7:09 am #

    From the lecture provided by Mr. Rouis, I learned about the development of snow load codes and snow failure cases. Many factors should be taken into account when considering the snow load such as building location and roof condition. As the development of the snow load code, many factors were added into the code for calculating snow load. In the snow load calculation we did in homework, the factors considered are snow exposure, thermal condition building occupancy and wind effects.
    From the snow failure cases, I learned that providing theoretical analysis to designing is not enough. Although we took into account all the known factors, there is still some unpredictable situation created by nature. By looking into the codes which is based on historical research, investigating the site condition of the building is still important during calculating and designing. The snow load design for a building must be unique according to the specific building condition.

    • mkev
      October 19, 2017 at 10:37 am #

      Good points Shangmi. You also need to keep things in perspective of cost when the building is originally designed. Will it really cost that much more to design a roof for 30 psf in lieu of 21 psf for example? It is cheep insurance to do that for unusual cases or complex roofs and the owner is often more than willing to pay a small additional cost when you explain the reasons to them.

  8. Perry H.
    October 19, 2017 at 2:49 am #

    While considering the problems associated will snow drift and overload conditions I thought why not just prevent the snow from accumulating in a known problem area in the first place. After doing a little research I found products such as heat cables and mats that can be installed and operated during a snow storm and will melt the snow before it begins a accumulate. I wonder if this could be a viable solution to snow drift considering operation cost won’t be too high due to the fact that the system only needs to run during a snow storm.

    • mkev
      October 19, 2017 at 10:35 am #

      The Dome at Syracuse U. uses the proactive approach and just turns up the heat when it is expecting a large snow fall. It is of course a fabric structure so that works. Heat trace / tape has the problem of being easily damaged, is often unsightly, and sometimes can’t work fast enough to keep up with snowfall. It is used strategically sometimes and we can discuss that more or perhaps others will find examples to add to these comments.

    • Pete Pitilis Jr.
      October 23, 2017 at 1:49 pm #


      Products such as heat cables and mats are definitely a viable option but I’m unsure of their limitations. My thesis building has a 9-story high roof and 6-story low roof which tower over and existing 3-story historical building. The new snow load imposed on the existing building was very large and exceeded any snow load design criteria designed for initially. Therefore, the existing building needed a solution to carry the new snow loads. The owner did not want spend the money on improving the structural system to carry the new loads. So the owner chose a system which utilized heat cables to prevent any snow accumulation from occurring.

      • mkev
        October 23, 2017 at 3:13 pm #

        I would be interested to know if the heat cables are connected to the emergency power since big snow and ice storms are one source of power outages? Also, how fast are they designed to eliminate the snow / drifts etc.

    • Shubham K.
      October 25, 2017 at 10:52 pm #

      You bring up a very good point. Heat cables and mats integrated into a roof design could help melt snow before it accumulates. Roof tested and designed with these systems could perfect the approach rather than an easily available add-on system sold on amazon.

    • Shangmi X.
      October 26, 2017 at 8:02 am #

      I think heat cables and mat might do something to help reduce the snow load, but it will be not efficient unless for the small house. The melting speed might not enough for melting the snow from snow storm. If In order to make huge surface area of snow to melt, it will increase the heat equipment which might cause additional load. However, I think making change of the roof surface will be a efficient way to reduce the snow load and snow drift, such as changing the roof material or surface pattern to increase the friction to keep the snow from drifting.

  9. Keunhyoung Park
    October 18, 2017 at 11:19 pm #

    The lecture by Mr.Paul Louis devoted to the historical development of snow load and case studies. By following snow load in from early to current code philosophy of snow loading was introduced. In these days snow load for a building is calculated in not only accounting local environment and roof shape but also critical snow loading scenario. I think this development in snow loading calculation is a continuum of trend in structural engineering such as performance based design.
    In addition, drastic changes in snow load is result of reflecting of newly discovered element affecting to snow loading’s characteristic. For example, third generation code in 1970 adopted thermal condition to consider how thermal effect change snow on the roof. And wind effect on the roof also considered. Lastly, in recent code, moving snows by wind and roof’s geometry participate to figure out most critical snow loading scenario. Sliding and drifting with various roof shapes are assessed.
    Although with these efforts to consider snow loading in diverse aspect, I learned from the case studies in today’s lecture that structural engineer always should keep in mind that there is always probability of unexpected things. Even structure was designed appropriately, it can be failed due to poor workmanship in construction. And we saw that theory can not fully express snow’s behavior from the many photos in the slides.
    Frankly, I was so impressed when steel beams and plates in the photos were buckled and bent, because I did not see real cases of these failures due to unexpected loading behavior and improper retrofitting as much as I learned as theories.

    • mkev
      October 19, 2017 at 10:31 am #

      I think “expect the unexpected” as you note is a good motto for snow related design at least until you have considerable experience in the industry and have “seen it all” so to speak. There are still going to be some one-off weird or unusual failures. O’Rourke discusses some of those in his article and we will get to a few more in our class discussions.

  10. Richard T.
    October 18, 2017 at 7:24 pm #

    I want to thank Paul for coming and teaching us about the history of snow loads in the United States. What I found very interesting was how basic it was to determine snow loading back then. In the past, one can simply follow a chart and if they knew the location or the latitude line that the building fell on it would provide them with a snow load.

    Even though the code wasn’t as developed it was somewhat reliable. If the code didn’t know what the snow load was in a certain area, it would just tell the engineer to go out and measure it themselves. In a way this puts more responsibility to the designer and it teaches them that these numbers that they’re using are not arbitrary and that they are derived from things that occurred in real life.

    One case study that I found most intriguing was the one where the owner decided to put snow jacks on their roof. True, an engineer can probably foresee there will be a build up of snow to the point where maybe the roof can’t handle it, but it surprises me that no one foresaw this problem during the change. I think that if owners want to change how they’re buildings work they need to have it pass through the eyes of an engineer so that things like this don’t t happen in the future. Maybe that’s how it is now.

    • mkev
      October 18, 2017 at 7:29 pm #

      If you have never been exposed to structural engineering, you just don’t think about things like accumulated snow. Most owners think their buildings are fine for just about anything. See my note to Nick S. below about Holuba Hall. It almost happened here on campus!

    • Nick S.
      October 19, 2017 at 10:08 am #


      I completely agree with your statement that the owner should of had an engineer look into if the addition of the Snow-jaks, but unfortunately this is not always the case. A majority of owners are not going to go through the hassle of bringing an engineer in for something that seems so simple to install. By bringing in an engineer you now have to pay them for their time and expertise. Many owners lack the construction knowledge necessary to know that when you are making any changes or additions to a roof structure you should consult an engineer. Instead they will listen to the contractor who is telling them that this product will improve their situation, not understanding the repercussions that it could have. Overall, in any situation where there is a change or addition to the roof structure the expertise of an engineer should be conducted. This will ensure that a failure will not occur that could potentially cost the owner money or a safety hazard to people.

    • Jared Patota
      October 19, 2017 at 10:44 am #

      I agree Richard, engineers should always be apart of any alteration to an existing structure, especially if the addition will contribute more load to the structure. Insurance agencies are also in agreement of this statement. You are liable for your own damages if you change or add to the structure’s use if it is not stamped and approved by a structural engineer. This could make the engineers job a lot more tedious and less fruitful business wise but engineers are never about the money or the fame anyway.

  11. David K
    October 18, 2017 at 6:51 pm #

    One of the issues that was discussed during Mr. Rouis’s lecture involved the validity of structural standards, such as ASCE7, predictions’ for design snow loads. One of the case studies that was presented involved a roof failure of a pre-engineered metal building due to drifting snow. The point of interest that arose from this case study involved the sliding of a snow drift due to roof surface conditions. The roof of a pre-engineered metal building is generally rather “slick” sheet metal supported by cold-formed steel framing. In this case, the snow drift was able to move down away from the drift zone, as predicted by ASCE7, and overload a portion of the roof that was not designed for such increased snow loads.

    A second major part of the discussion was that of improper building modifications. Mr. Rouis presented several cases in which the building owner decided to modify the existing building in some way that either increased snow loads (adding sno-jax to the roof) or decreased structural member capacity (removing LTB bracing).

    This lecture provided the class with an appreciation of the complications of snow loading and an increased awareness of the time-dependent nature of snow loads due to building modifications.

    • mkev
      October 18, 2017 at 7:23 pm #

      You make a good point and I believe Mr. Rouis confirmed it. The codes can not cover every single case you will encounter or they would be even more complicated and voluminous than they are already. Another example of “not in the code” is the really long building I mentioned in class that drained to one end. The roof is so big you almost think it had its own climate zone. Some of these cases are known to those who study the details but not necessarily to many engineers. Developing a good sense of engineering judgement and making sure your structures have redundancy goes a long way to keeping those “weird cases” away from your buildings.

  12. EllenW
    October 18, 2017 at 3:13 pm #

    Mr. Rouis lectured to our Building Failures class on the topic of snow loads and failures. I found the information on snow loads from historical codes very interesting. The earliest codes had very few influencing factors on snow loads and resulted in simple uniform loads applied to the building. Due to the low number of historical building snow failures, these numbers were apparently sufficient to cover all of the drifting, sliding, and rain-on-snow conditions that are designed for now. As the codes became more sophisticated and complex, factors for building location, topography, exposure/wind effects, and insulation were added. These factors allowed a lower but more precise snow load to be designed for, but it also increased the difficulty of calculating the design load and required detailed measurements of local snow conditions. The current codes have many snow load conditions that must each be considered separately in the design. Drifting snow in both the leeward and windward directions must be computed and compared as one load condition. Sliding snow and rain-on-snow are two other conditions that must also be considered. Complex roof geometry adds another level to the difficulty in calculation the correct loads. I would be very interested to see if all of the loads from these conditions could be condensed into one envelope load condition. If so, it may be possible to apply one higher load to the entire roof in the manner of the earlier codes. By simplifying the snow calculations, it would ensure that all possible conditions are designed for, which would reduce the number of failures caused by design mistakes.
    Thank you to Mr. Rouis for taking the time to share with the class.

    • mkev
      October 18, 2017 at 3:22 pm #

      An interesting thought that would certainly simplify the calculations and perhaps prevent errors in dealing with complicated and complex formulas and assumptions. I think it might work under certain framing types and not so well under others. My thought is that the relative amount of the drift load assigned to each member based on tributary area (width) would be the key. Individual members spanning in two directions directly under a drift would have to be designed for the worst case, not an average or smeared case to be safe. A two way slab or reinforced concrete shell structure would have more redundancy and be more capable of spreading the load. Still, I would be interested in any comments on how to simplify the code without compromising safety.

  13. Geoffrey T.
    October 18, 2017 at 11:43 am #

    On September 17, 2017, Paul Rouis of Ryan Biggs|Clark Davis gave us presentation on snow collapse investigation. He started with the history of snow code that goes all the way back to 1916. The thing that I found interesting from his presentation was how snow load provision had changed since then. Back then, the snow load was only based on your latitude. If the building was located south 35 degree, the snow load should be 20 psf. Anything above certain latitude, just add 10psf to the maximum value. Compared to current snow load provision that includes map with different isobars across the map. Even though the number is still rather arbitrary, the standard is a lot more specific compared to the code back in 1916.

    Another thing that I found interesting from his talk was the fact that all the code in Vermont State was reinforced state wide, creating standardization for one whole state. Even though Mr. Rouis did say that this was possible because Vermont State was small, I still think that it is better and more efficient than going to each municipality and ask for which version of the code that they are using right now. I am curious to hear everybody’s opinion on why we do not reinforced state-wide code.

    • mkev
      October 18, 2017 at 1:44 pm #

      One of the papers you will be reviewing for our discussions on site specific snow load (CS) is about New Hampshire. I don’t believe they have a single source code enforcement agency but they did remap the entire state so that there is a map similar to ASCE 7 but that eliminates CS areas. They used the CS technique developed by the Cold Regions Lab to make it work. Paper is on Canvas or see the following two references: and

    • CamilleS
      October 24, 2017 at 9:36 pm #

      I think an interesting argument for state-wide code is the fact that professional liscensure is controlled on a state by state basis. By implementing a snow code that covers the snow loading for the entire state, professional engineers that are lisenced in that state will be able to understand the snow loading in any part of the state they work in, instead of using site specific snow loading. Perhaps there is an argument to develop state snow codes.

      Then again, it takes states long enough to adopt updated IBC codes that I wonder how effective it would be to encourage states to develop these snow codes. Pennsylvania still has not adopted an updated IBC code since 2009 state wide. The political work that it would require may make that proposal improbable in some states.

  14. Megan F.
    October 17, 2017 at 5:06 pm #

    Paul Rouis gave a very informative and helpful presentation on snow codes and snow loading. It was interesting to learn about different unique snow conditions that can exist and cause structural elements to fail. Learning about snow density and how that can affect the loading conditions and drift is an important thing to think about. The more dense the snow, the heavier it is and the less likely it is to drift/blow off. The one case study presented show the effects of dense snow build up acting on a metal roof. When the snow melts on a metal roof, there is ultimately no friction to keep it from sliding, and when the snow is dense and wet, the mound of snow can move onto an area of the roof that was not originally designed for the extra load. For this case, the building met all code requirements however, the code did not address drifted load moving as a solid mass. I think that snow load is something that requires a lot more research because there are so many different scenarios that can exist and it is almost impossible to predict what will occur.

    • mkev
      October 19, 2017 at 10:53 am #

      Good summary. Let me point out another example for discussion. Think about a commercial greenhouse in say Pennsylvania. They are often built tight against each other with connected valleys at gable roofs. What happens in those cases. Look no further than some of the garden areas in the big box building supply stores. Hint: We had one of those collapse in SC in the 1990’s.

      • Megan F.
        October 23, 2017 at 8:02 pm #

        In the cases of the connected greenhouses, the snow can build up in the valleys between the units causing a great increase in load on the exterior walls. Additional bracing should be placed on those bearing walls in between each greenhouse to allow the extra load to build up. The greenhouses should also begin to be built with more space between each unit to allow the snow to accumulate. There could also be a temporary heating system that can turn on in the case of extreme snow and cause it to melt quicker. I think for many of the greenhouses that are already built with a gable roof and tight against each other, the owner needs to brace the walls and prepare a plan to get the snow off as soon as possible after a heavy snow storm.

        • mkev
          October 23, 2017 at 8:12 pm #

          Exactly right Megan. And, if you think metal roofs get slippery when they heat up, think about glass. So one failure mode of green houses is sliding snow into the valley.

          • HarryB
            October 23, 2017 at 8:31 pm #

            Dr Parfitt,
            Should an engineer design for the lateral pressure due to snow drift as well? I am visualizing two parallel green houses with a small gap of separation between them. The snow drift might lead to snow filling in the gap and creating pressure on the glass walls of the green houses.

  15. Nick S.
    October 17, 2017 at 4:53 pm #

    On Tuesday, October 17th, the AE 537 Building Failures class had the pleasure to hear a presentation by Mr. Paul Rouis. Mr. Rouis is a 1985 Penn State graduate from the Architectural Engineering (AE) program. Currently, Mr. Rouis is working for Ryan Biggs|Clark Davis Engineering & Surveying, D.P.C. where he is Principal-in-charge for engineering design and renovation of culture, education, medical, industrial, and parking structure projects. A fun fact about Mr. Rouis is that he is the first AE to be hired by Ryan Biggs|Clark Davis. To begin the lecture Mr. Rouis gave a brief introduction on Ryan Biggs|Clark Davis and explained that they are one of the largest womon-owned engineering firms in Upstate New York. They specialize in both structural and civil engineering, repair and restoration, construction support, inspection, testing, and surveying.

    Following the introduction to who Ryan Biggs|Clark Davis is the presentation focused on discussing snow load, how code came about, and case studies. Mr. Rouis began this part of the presentation by talking about some of the earliest snow code. From these earliest codes it treated snow load as a uniform live load with some considerations for geographic location and roof slope. Mr. Rouis went on to state that as time went on a more involved consideration began to be placed on snow load as a better understanding of its affects on a structure were seen. Following this discussion on code Mr. Rouis went on to talk about the three (3) types of wind affects that play a role in how snow is loaded on a structure. These are unbalanced snow, drift, and sliding. All three need to be investigated to fully ensure that a structure is designed properly.

    After providing the class with an understanding of snow load and how code came to be, three case studies were investigated. These case studies were completely different snow load scenarios, but showed how important understanding snow load is to ensuring that a structure will not fail. The two case studies that stuck out to me were case study one and three. Case study one dealt with a salt storage dome collapse. The dome was 158 feet in diameter and 63 feet to the crown and was only 5 years old at the time of the collapse. Ryan Biggs|Clark Davis investigated this collapse and found that there were flaws in the original design and quality control problems to just name a few. Case study three was about a pre-engineered metal building (PEMB) frame failure and when Ryan Biggs|Clark Davis investigated this issue it was found that PEMB do not have significant load capacity and that proper bracing and bridging is essential for structures to function as intended by the designer. Reason that these two stuck out to me was because of the individuals negligence in checking to make sure the building could with stand the snow load that would be applied on it. Case study one showed that not following the specification of the materials being used greatly affected the structure causing a failure to occur when a snow load was applied to it. Case study three showed that the addition of a material (Sno-jax) allowing for additional load (snow) to remain on the structure that would usually not be there to cause a failure. After hearing these two scenarios it got me thinking why no one thought to further investigate whether any of these changes could have affect on the structure, specifically altering the load (snow) handled? For myself, as an engineer, I would be expected that I investigated that these changes would not affect the overall safety and quality of the product being delivered. Unfortunately, in these situations this is not the case.

    Overall, the presentation was very informative and opened my eyes to how critical snow loading is on a structure. These two case studies show that further investigation and understanding needs be conducted when designing, renovating, or adding anything to the roof structure.

    • mkev
      October 17, 2017 at 5:06 pm #

      Much closer to home, think about the roof on Holuba Hall. There was a proposal at one point to add Snow Guards to retain snow and prevent it from sliding on students and/or the outdoor practice field. MKev asked if they knew what parameters had been used for the roof snow load calculation. Metal Bldg manufacturer had taken a roof load reduction for sliding snow….so what happens if we then prevent the snow from sliding?

  16. JeremyS
    October 17, 2017 at 2:34 pm #

    The lecture earlier today from Paul Rouis of Ryan Biggs│Clark Davis provided an insightful discussion on the history of snow loads and case studies of snow load failures. Snow load codes in building go back as far as 1916. Back then snow loads were simply determined from a graph which plotted the latitude versus the snow load in psf. While this is much easier than the strategy we have now it is also less accurate. In today’s code the snow load is based on a number of factors such as location, elevation, building occupancy, building thermal status and roof slope. Furthermore the code requires that drift be analyzed under certain conditions such as a step up or down in building profile or a gabled roof. These two scenarios are the most common cause of snow load failures as shown in table 1 of the structural collapse from snow loads article by Michael O’Rourke. The third most common snow load failure is caused by an open air and freezer building where the snow on the roof can accumulate due to the thermal properties of the roof which prevent the snow from melting off. This is a very possible scenario that caused the collapse of our ice arena project. Mr. Rouis described the condition of unbalanced loading which causes load reversal. From what I remember of the lecture Mr. Rouis said that this load reversal causes compression members to act in tension and vice versa. He also went into more depth to explain how destructive this can be for Glu Lam compression members that switch into tension members. The result is that the tension members are torn out of their connections. This sounds extremely similar to what happened at the ice arena where the supports were left intact however the Glu Lam arch was torn out of the support. I think that this scenario of unbalanced snow loading could be a serious factor in our project.

    • mkev
      October 17, 2017 at 3:28 pm #

      Stress reversal is certainly a possibility for the glu lam arch structure. To be sure you need to look at the loading case or cases and the specific geometry. In general it is more likely to show up in steel or timber trusses where you have a lot of connections of individual members. I believe Mr. Rouis was describing a truss when he was discussing connections but as noted that is not exclusive (the wood salt dome was another possibility). Definitely gives all of you something to think about in your deliberations.

Leave a Reply