Roof Design Basics for Snow Country

Leaks from ponds created by ice dams can frequently be traced to roof design. Roof designs that funnel snow to narrow spaces and narrow eaves are likely to promote the development of ice dams and resulting roof leaks. Roof designs with changing slopes that drain steep roofs to flatter roofs are likely to promote the development of ice in the area where the roof slopes meet. Standing seams of metal roofing can restrict snow runoff and promote the development of ice dams near the lower ends of valleys. Roof designs that include opportunities for warm interior air to reach the underside of the roof are likely to cause roof snow to melt and allow melt-water to run down and refreeze at the eaves, forming ice dams and leak-producing ponds behind the ice dams.

Some design principles to minimize the risk of ice dams and resulting leaks include the following:

1. Keep the roof design simple. Avoid complex roof architecture that requires runoff to change direction or follow circuitous paths to get off the roof. 
2. Avoid or minimize 'waterfall' conditions where runoff dripping from a high eave can freeze on a lower roof.
3. Avoid roof configurations that include a high, steep roof intersecting a lower, flatter roof surface.
4. Make way for snow. Remember that snow cannot get through tight spaces easily passed by water. Roof designs should allow wide paths for snow movement. Avoid tight dormer spacing, tight valleys, and other roof configurations that would restrict snow movement. 
5. Avoid standing seam configurations that restrict snow movement. Snow may move down-slope easily in a direction parallel to standing seams, but standing seams in valleys and roof slope changes can act like brakes, restricting snow movement toward eaves. Snow movement around chimneys and similar items can be restricted by standing seams, so special consideration should be given to a seam layout that will promote effective snow movement.
6. Keep the roof surface cold - especially up-slope from eaves - by keeping warm interior air away from the underside of the roof. Paths for interior air to reach the underside of the roof must be effectively blocked by a complete air barrier on the warm, interior side of the insulation (or by a properly installed, complete air barrier type of insulation).
7. Include support for underlayments, flashing, and roofing membranes at intersections between roof surfaces and related construction. Do not expect watertight integrity where a design calls for a dormer eave to intersect a main roof plan at a point without special construction to support underlayments, flashing, and roofing transitions.
8. Consider roof orientation and exposure when designing a roof. Snow will generally melt sooner on a roof exposed to sunlight than on a more shaded roof. As noted in 2 above, snow melt from an exposed roof can meet colder temperatures and refreeze as ice on the more shaded roof.
9. Minimize use of skylights and roof windows. Even the most energy efficient of these will melt snow that lands on them, and the melt-water is likely to refreeze as ice as it runs down on colder roof surfaces.
10. Do not expect an underlayment product like "ice and water shield" to compensate for design features that promote ice dams.
11. Consider the need for snow removal maintenance. Designs with features that promote ice dams may require frequent snow removal to minimize leaks.

Aesthetics and structural integrity are commonly the first considerations in roof design. Roof designs for snow country should also include basic considerations and accommodation of snow behavior in order to minimize problems caused by ice dams.

Bidders Trust Bid Documents for Take-off

Estimating quantities from a set of plans prepared by another architect reminds me that bidders are likely to rely on the accuracy of the drawings when preparing a take-off for a bid. If the drawings are inconsistent or include discrepancies, those are likely to affect the bids, and they may lead to claims of extra cost during construction, if the successful bidder determines that actual construction of the design requires more material (and related labor) than the drawings clearly indicated. The claims may be disputed as unreasonable based on a documented requirement for the bidder to consider the greatest quantity in the event of a discrepancy, but the limit of practicality may be exceeded where determination of actual quantity for bids would require exhaustive review and computation based on various plans and details. Bid preparation is typically limited to a short period of time due to a combination of the scheduled bid period and bidder attention. This is stated not for the purpose of blaming either the designer or the bidder but instead to suggest that accuracy in bid documents should be optimized in order to obtain accurate bids and to minimize discrepancies and the related disputes. There is an old saying that close bids are an indication of tight documents, meaning that the bidders all saw and bid the same scope. Of course, experience also shows that bids vary for reasons that have nothing to do with the bid documents, but that does not detract from the advantages of well coordinated bid documents.

Building Science and the Risks of Experimentation

Science is experimental; it consists of hypothesis and experiment. The path to success can be littered with experiments that fail. Scientists learn to expect failure along the way and to live with experimental failure as the cost of progress. Scientific design is experimental, and it is accompanied by an expected risk of failure.

The growing popularity of building science today brings increased risks of experimentation to the mainstreams of the building industry and the practice of architecture.

Historically, building design decisions were based on long established and proven practices and material selections. Expectations of reliability rested on proven performance over years or decades or - in some cases - centuries. The practice of experimentation was left mostly to the fringes and outliers. Main-streamers tended to avoid products and systems that lacked a good track record. Established building technology was a focus of learning and skill building; architects and builders could expect to learn from a previous generation and practice for decades with a building technology that would remain essentially the same.

More recently, we have seen and become obsessed with an increasing pace of change. Many equate faster with better, making decisions based on the latest available product or on predictions of the next invention or innovation - perhaps even with a belief that it must be better simply because it is new and not established. However, this kind of experimental approach to building design and construction dramatically increases the risk of building failures, in large part because it discredits time-tested performance and avoids or dismisses time-consuming consideration of the multiple roles played by building materials and the roles played by parties in the construction process.

Valid interest in (and popular incentives for) conservation and quality of resources and processes may have led to a willingness on the part of some to take more risks with experimentation. But questions need to be answered: Who assumes the risk? How much risk? Is there awareness and consent of assumed risk? and, If it fails, who owns the failure? Further, If it fails, how can it be considered a sustainable practice?

New Blog

Many of the posts from this blog have now been exported to a new blog: http://onthejobarchitect.blogspot.com/

Code Talk: Challenges for Architects

Most of the architects I have known are not comfortable with codes. They want to comply with applicable codes, but they find them confusing, tedious, contradictory, or even frightening. Codes are always being revised or superseded. It takes time to determine how a combination of applicable codes can be reasonably applied to a particular building type and scale, and it seems that the codes are changed almost as soon as the architect reaches a level of comfort with the requirements.

One of the reasons that it takes more than a little time to determine the application of codes to a specific project is that codes tend to be dense and voluminous texts that are full of "fine print", numerous exceptions and cross references, and hierarchies that are hard to follow (i.e., "Which article has precedence in this situation?"). Further, the architect is usually charged by statute with the professional responsibility to account for the application of numerous, differing codes - on the same project. In some cases, state or local authorities have adopted parts of different codes that cover similar matters, adding their own hierarchies to interpretation of requirements, and the architects are challenged with having to determine how to resolve gaps and conflicts that have not been addressed by the state or local authorities. Codes that include graphic illustrations of requirements are generally easier for architects to understand, because many - if not most - architects tend to think graphically. In that regard, accessibility guidelines that rely on graphic illustrations have been much easier to follow than text-only codes. Code commentaries or handbooks such as those available from ICC and NFPA can be more useful than the codes themselves due to the use of graphic illustrations. Graphic illustrations can also be enlightening for code writing authorities where the illustrations are intended to cover typical conditions, and the pictures themselves may raise questions that are then considered and addressed by the authorities.

Another challenge for architects is how to satisfy the professional responsibility to apply code requirements in those situations where less than full design services are contracted. If contracted services are limited to preliminary phases of design or other design iterations that exclude detailed drawings and specifications, how should the architect account for code requirements in the preliminary design phases, and how should the architect account for code requirements that would normally be applied to the development of detailed plans and specifications? How should the architect alert the owner (and/or contractor) of the need for the owner or contractor to complete the process of code compliance related to parts of the design that are beyond the architect's contracted scope of services?

I recall the advice or direction of one architect employer to make sure that what you do show on drawings is correct. His comment was not really focused on codes, but it could be applied to the question of code compliance in preliminary design. The code requirements to be considered in preliminary design tend to be large scale matters that would govern detailed development of a design in later phases. For example, a schematic design would consider zoning regulations such as building setbacks and building height and also allowable area and height as established by the applicable building code for the intended building use. That would be an appropriate design phase for consideration and documentation of building code construction type. The preliminary code analysis should reveal any applicable requirement for fire walls to divide the project into 'technically' separate buildings, and appropriate locations for such fire walls could be shown diagrammatically on the preliminary plans together with notation referencing the specific code provisions that would govern development of detailed design in a later phase. If it is not practical to even approximately locate such fire walls diagrammatically, those applicable code requirements should be included in notes that relate to the schematic plans. A similar approach can be followed for other code considerations that relate to the preliminary design, such as wheelchair accessible entrances that will require detailed design (e.g., accessible ramps, railings, door approaches, hardware, etc.) in subsequent design phases. Following this approach, each design phase would include code information appropriate to the phase and an indication of further design that is required in a subsequent phase.

One good reason to develop a comprehensive preliminary design approach to code compliance is to lay the groundwork for subsequent design development and documentation that will be performed by staff in the same office or on the same team. Another, perhaps more significant reason, is that architects are sometimes invited to defend themselves against claims of noncompliance where their services were limited to preliminary design and the code matters in question would customarily be applied to a later design phase (e.g., detailed construction documents). While an argument of exclusion by agreement may be valid, the time and cost to wage the argument after the fact may be a greater problem, especially if the project owner has encountered either an unexpected and costly construction change or post-construction change after the architect was dismissed from the project.