Structural Considerations for Roof Restoration

fall 2016 issue

Structural Considerations for Roof Restoration

By Phil Brearton, P.Eng., LEED-AP, Associate, Stephenson Engineering Ltd. &

Peter McAteer, MASc., P.Eng., Associate, Stephenson Engineering Ltd.

This article was originally published in the Fall 2016 edition of Pushing the Envelope Canada, which is published twice per year by Matrix Group Publishing Inc. for the Ontario Building Envelope Council. It has been reproduced with permission.

When a roof replacement project is being considered, it is very important to consider whether the roof structure needs to be reviewed as part of the project. Although we should expect the roof structure was designed and built by competent professionals two, three or perhaps many decades ago, it behooves the designer to consider whether the pro-posed new roof assembly will affect the future performance of the roof structure.

One must also consider whether changes have already been made to the building or immediate neighbourhood that effect the original roof structure. Overall, this review should be conducted with the awareness that the roof structure’s past performance is not a reliable predictor of future performance. The opposite is the case, as roof structures can be expected to be exposed to more extreme weather and greater snow and rain loads.

The following provides a review of issues that should be considered before replacing a roof membrane system.


When designing a new roof assembly, snow loads need to be considered from to-day’s perspective. We have a better under-standing of the effects of drifting and snow density and experts are expecting that snow loads will continue to increase in Ontario, particularly in the north, due to climate change. It is not a good assumption that past roof performance is indicative of future performance.

It should already be standard practice for roof designers to determine whether screens or rooftop equipment have been added, parapets have been raised or near-by higher roofs have been added since the previous structural review of the roof. All of these changes will affect the calculation of the roof’s snow load. The designer must also consider whether the code, at the time of construction, allowed for a lower snow load.

If the building was constructed prior to 1995, it is possible the roof was under-de-signed in comparison to today’s code. In the 1950s, snow loads were calculated as the same value as the weight of the snow on the ground, with zero allowance for drifting snow at low to high roof junctions.

During the early 1960s, the typical roof snow loads were reduced to 80 per cent of the ground snow loading. In 1965, consideration was given for additional accumulation due to shape variations or the influence from adjacent buildings. In 1985, the Ontario Building Code (OBC) in-creased the design unit weight of snow to 15 pounds per square foot (PSF), and in 1990, the code further increased the design unit weight to 18.75 PSF. In 1995, further refinement was given to accumulation factors for the calculation of snow loads on large flat upper and lower roofs. As mentioned later in this article, further changes can be expected, as snow density assumptions are being revised and historical climate data is being found to be an unreliable predictor of future snow loads.

Future snow loads can also be expected to increase as thermal insulation is added to the roof assembly because the building heat can no longer melt the snow. Increases in thermal insulation have been required by the OBC. For instance, a 1950s building may have one inch of fibreboard insulation, which would have properties of less than R3, whereas today’s code requires R40.

This effect of increased insulation was noted many years ago in the Canadian Building Digest 193: “Heat loss through the roof may also cause a significant reduction in the load, especially where the maximum load results from snow accumulations over a relatively long period. Many older roofs have been saved from collapse as a result of reduction of load due to melting.” A roof’s past structural performance is not a reliable indication of future performance, even if past weather is indicative of future weather patterns.

Unfortunately, future weather pat-terns are expected to be worse than historic weather patterns. In 2013, the Inter-government Panel on Climate Change (IPCC) concluded that climate change has already warmed Canada 1.5°C since 1950 and 3.0°C in northern regions. Canada is also trending to greater precipitation and “heavier events.” In northern Canada, the change is particularly worrisome, prompting the issue of snow load risk management guides from both CSA (i.e. CAN/CSA-S502-14) and FEMA (Snow Load Safety Guide FEMA P-957.

To factor for climate change, adjustments to snow loads are being made to Canadian codes. The 2015 National Building Code of Canada (NBCC) increased snow load values for many northern com-munities by an average of 15 per cent in the two-decade period following, with in-creases ranging from three to 35 per cent. It is likely that snow loads will increase further in many communities once changes to snow densities and rain on snow are factored into future updates.

Nevertheless, Engineers Canada has stated, “As the climatic design values in codes and standards become more out-dated relative to a changing climate, de-signers face even greater challenges and professional risks.”

“It is becoming clear that the historic-al climatic design data in the NBCC will become less representative of the future climate and that many future climate risks will be significantly under-estimated,” re-ports the IPCC. “A one-in-20 year annual maximum daily precipitation amount is likely to become a one-in-five to one-in-15 year event by the end of the 21st century in many regions.”


The need to retain water on roofs is generally understood by roof professionals in Ontario. Rain water must be slowly dis-charged from roofs to the municipal storm water systems to help prevent flooding. Significant floods over the past few years have driven this point home. By design, flow controls or weirs at roof drains en-sure the flow rate is managed and over-flow scuppers allow discharge of rain water if the water accumulates to over six inches above the drain elevation.

The scupper requirement was added to the OBC in the 2006 version so the roof designer needs to confirm whether scup-pers need to be cut into the building as part of their project. If the scuppers are installed, six inches of water, or 30 pounds per square foot, will be the rain water maximum load on the roof. If scuppers are missing, the maximum rain water load will become a function of the severity of the storm and the height of the parapets.



The topic of wind uplift is typically concentrated on the concern that the roof assembly not be damaged by high winds. The effects of increasing roof assembly weight are obvious. The potential negative consequences of decreasing the weight of the roof assembly are less obvious.

If a roof assembly is to be converted to a ballasted protected membrane roof (PMR), then the overall increase in weight can be significant. A lightweight system could be less than two pounds per square foot, whereas the Dow 508.2 de-sign guideline for ballast on inverted roof membranes recommends ballast weight up to 22 pounds per square foot at corners and perimeters, depending on the use, configuration, height and exposure of the building. If the new roof is intended to be an inverted, or PMR, assembly, the roof structure could be subject to significantly higher loads. It is also important to note that this guideline is proposing ballast weights to prevent the underlying assembly from being disturbed by high winds. If it is intended to prevent the insulation from floating, ballast loads will need to be higher than the Dow 508.2 guideline.

Applying more weight to the roof is an obvious flag for the designer. The potential negative consequences of decreasing the roof membrane assembly weight are less obvious. A common example is a corrugated steel deck on open web steel joists with a ballasted roof. If the intention is to replace the membrane system with a conventional unballasted system, the structural framing originally designed to receive a roof membrane assembly dead load of 0.6 to 0.8 kilopascals (12-15 pounds per square foot) will now only sustain approximately 0.10 kilopascals (two pounds per square foot). While reducing the dead load clearly offers increased capacity against the snow loads noted earlier, the net wind uplift force acting on the roof may now be higher than that for which the roof structure was designed. This possibility should be investigated and the ability of the open-web joists to resist any increased uplift should be confirmed by a qualified structural engineer.



Special attention is also warranted when altering roof loads to buildings that are framed with timber trusses. While the truss strength can be approximated by the analysis of the various member sizes, the connections can be a limiting factor. It should also be noted that the bottom chord of pre-1970s construction wood trusses may be considerably overstressed by todays’ standards. This is due to changes made in the late-1960s to the test methodology for determining wood’s tensile strength. As before, a structural engineer should visit the site to evaluate the existing dead loads and the general condition of the wood, and to measure the truss member sizes and layout.


It is not uncommon to be given the opinion that a building structure that has stood for the past 100 years is, therefore, “time proven.” Although old buildings are a testament to the skill and care of our predecessors, giving an old building a “free pass for structural integrity” for perpetuity assumes that the building will be exposed to the same weather we experienced in the past, and we know this is not a good assumption. This belief also assumes the building has undergone no deterioration, no renovations and no changes in use.

During the life of a building, it is not uncommon for building alterations to be made by non-professionals without due consideration to the effects on the roof structure—parapet heights could have been raised, rooftop equipment, sign-age or screens could have been added, a low canopy might have been constructed over an entrance, a neighbouring tall structure could have been erected close to the building, or perhaps a roof’s use may have been converted into a roof-top patio as “outdoor living space” with the requisite planters and patio stones. There are many possible building alterations that, if not designed by a professional, could have compromised the roof structure. When assessing a roof, the designer must rely on drawings, building maintenance staff interviews, and the designer’s personal knowledge of typical past building construction materials and methods before deciding that a structural review is appropriate.

Assessing the condition of structural components is often difficult because the structural components are not accessible for review. Nevertheless, the roof designer should be on the lookout for a sign or report of structural deterioration. Chronic water leaks in buildings, for instance, can lead to serious damage to almost all structural materials including steel, wood, concrete and masonry.


With the replacement of a roof membrane assembly comes an expectation the building will be reliably protected for the next 20 years or more, over which time, weather conditions are expected to worsen. In order to achieve this, during the design process, the designer must assess whether the roof structure has been affected by past or proposed roof renovations.

If in doubt about the roof structure, it is recommended that a structural engineer be retained to assess the load capacity of the roof to perform for the foreseeable future. Where the original design drawings and design loads are known, the review process is relatively straightforward. Where no original de-sign information is available, evaluating the capacity of the structural framing is somewhat more difficult and is beyond the scope of this article. In either event, the structural engineer should visit the site so that all the existing dead loads acting on the roof structure (heating, ventilation, and air conditioning, lighting, suspended ceilings, etc.) can be evaluated and so that the structural framing and its condition can also be carefully evaluated.

Peter McAteerPeter McAteer, MASc., P.Eng., is an associate with Stephenson Engineering Ltd. He specializes in the review, alteration and adaptive re-use of existing structures, focusing on renovations, additions and building retrofits. He is currently assigned as the small projects team lead.

Philip is a senior professional engineer with more than 26 years of experience in building science, codes, standards, LEED and design of new and retrofit of building envelopes for all types of building components. Phil develops excellent client relationships and provides LEED sustainability and Building Science Professional Engineering Services for Stephenson Engineering.Phil Brearton, P.Eng., LEED-AP is an associate with Stephenson Engineering Ltd. He has an extensive background in building envelope failure investigation and repair as well as new construction de-sign consulting and commissioning.