The snow load duration factor is 1.15 times the standard 10-year sustained maximum load rating as long as the accumulated time under this load is not more than 2 months. Does this mean that

  1. a. the snow load being at the duration factor of 1.15 for 2 months continuously (experimental case) is at the same level of safety as

    b. permanently being at the standard load rating (control case)

when assuming that the experimental case is well under the under standard load rating for the rest of the year, or

  1. does the experimental case impart additional risk compared to the control case from the long-term perspective?

Basically, this is asking whether the standard practice allows the control case to use up the safety margin (factor of safety minus one) of the standard load rating only because its short term means that the probability of other loads happening simultaneously will be negligibly small, or whether the safety margin of the standard load still applies on top of the maximum load at full duration allowed by the duration factor.

For example, if the snow season is especially long one year but not especially intense at any given moment, meaning that the roof at any given time is at the forseeable maximum snow load accounted by the duration factor but that it is at that load for a whopping 4 months rather than the expected maximum of 2 months, does that mean the roof will cave in, probably causing the pancake effect to collapse the entire building too?

As for another example under main question, will having the load at any given time within the timeframe moderately exceeding the snow load duration factor, but not having the duration exceed that, make the roof or even building collapse?

Under the hypothetical case that load duration factors in standard practice already use up all of the safety margins for the standard load ratings, this means that exceeding the load duration factor also exceeds the safety margin. Since this exceeds the safety margin, does this mean that exceeding the load duration factor will guarantee a collapse?

In civil engineering in general, does being at the rated limit mean that there is a negligible chance of failure if everything is done perfectly (too small to be measured, such as 1ppb) and does being at the upper limit of the safety margin practically guarantee a failure if everything else is done perfectly (too large to measure a successful case, such as 99.9999999% failure rate)?

  • $\begingroup$ Would only looking at 10 years of data be the issue or should 100 years be examined? ie has snowfall amounts been reducing over time? $\endgroup$
    – Solar Mike
    Jul 1, 2022 at 5:31
  • $\begingroup$ Yes, 100 years and even 1000 years should be examined. Also, this is only a hypothetical question for logical purposes, so the conditions would be perfectly controlled. As a result, average annual snowfall over any 10-year interval remains constant, just that once every so many (arbritrary amount) years, the snowfall wildly exceed expectations so it overloads the roof to a twice the difference between the load duration factor (with either one at the limit and the other at twice the amount over (how much the load duration factor is over the standard load limit allowed). $\endgroup$ Jul 1, 2022 at 6:00
  • $\begingroup$ Also, all other factors, such as humidity, live loads on the rest of the structure besides the roof, and foundation settlement, remain the same between the different scenarios at any given time, so they are control variables even though they will change somewhat within each scenario over time. $\endgroup$ Jul 1, 2022 at 6:14
  • $\begingroup$ Snow (over time from wind and sun) will compact and wet snow (from rain or melting) is heavier and you must factor in slope, roof shape, drifting, sun exposure. So exceeding safety margin does not mean the roof will collapse, but does increase the chance of collapse. $\endgroup$ Jul 2, 2022 at 17:05
  • $\begingroup$ @StainlessSteelRat Okay, I see. I always thought thought the rated amount is practically 0% chance of failure (for example, lies outside 7 sigma confidence interval), the interior of the safety margin is just a sliding scale (transition zone) between totally safe and totally dangerous, and the outside edge of the safety margin is practically 100% chance of failure (for example, lies within 7 sigma CI). Apparently, I was wrong. So what is the typical/standard percentage chance (or confidence interval number of standard deviations) of failure rate at the upper limit of the safety margin? $\endgroup$ Jul 3, 2022 at 3:30

2 Answers 2


The problem is you deciding to push the limit in one area and the owners will have to know that and maintain all the supports for the roof that entire 100 or 1000 years. Not likey, and it isn't collapse that is the outcome, it's FAILURE. One small part failing that isn't visible starts the process and then a portion of the system fails or "collapses", the failure may be at the foundation because of the roof load. I had a house that a stack of timber holding a portion of an upper floor shear a concrete foundation. That ended with removing 13 cubic yards of soil from below grade and replacing 20ft of foundation. It was a concrete foundation over a 100 years old and that why real professionals say "They don't build them like they used to." Pause ......."Thank God". It's a line used on remodeling or updating old homes. You pull something open and say that. Everyone on site knows we have a problem. Everywhere in the world you're only seeing the 1 out of a 100 buildings that didn't fall down or burn.👻


Timber structures are somewhat unique in that their strength and serviceability characteristics are adversely affected by the duration of time that the structure will carry a specific load. Normal load duration is ten years. This doesn’t mean that the specific load will be applied for ten consecutive years. Load duration is the cumulative amount of time the full design load is expected to act over the life of the structure. Snow loads for instance are based on a 50 year mean recurrent snowfall with a low probability of being exceeded in a 50-year period. In most years the specified snow load will not be seen. In other years it may occur, but not continuously. If the economic life of a structure is 30-years, it becomes more likely that the design snow load will occur for 12 of every 36 months. But this too is somewhat unlikely since it only snows in winter months. The safety factors are generally larger than load duration factors (LDF). So, use of the 2-month LDF for snow in heavy snowfall regions should not compromise the structure to the point of imminent failure. However, use of the 10-year value is probably not a bad idea for snow loads even though it may be slightly conservative. If sound information regarding load duration is available, a higher value of LDF is justifiable. The designer should not use LDF’s above the values prescribed in the design standard. Use of a LDF large enough to degrade safety factors is definitely not a good idea. Correct application of LDF's should not degrade the safety factor. It can be difficult to obtain specific load duration information. A typical consulting firm will often have standard practices regarding application of LDF’s. We learn much more from practice than we do in the classroom. A bit of advice; If a structure experiences a failure of any kind for any reason, you do not want to be the engineer that misapplied the LDF, even if it had nothing to do with the problem.

  • $\begingroup$ This wall of text is hard to read, there are formatting tools that can help. Also, some woods get stronger over time... $\endgroup$
    – Solar Mike
    Dec 22, 2022 at 7:07
  • $\begingroup$ I'm sorry you didn't like my answer. I write the answers in a Word document and paste into the answer box. Some time after that it always gets run together. I have to remember to fix it. I didn't know about the formatting tools. I apologize again but I thought you were referring to the load duration factors in the standard. They all apply to woods for which certain strength values are adversely effected by long term loads. $\endgroup$ Dec 23, 2022 at 2:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.