The concept of sidewall runoff is that wind driven rain can blow against the side of a building and increase the amount of storm water that falls on a lower roof below. Each of the two model codes addresses this subject differently.

The IPC is quite brief on this subject and states as follows:

1106.4 Vertical walls. In sizing roof drain and storm drain piping, one half of the area of any vertical wall that diverts rainwater to the roof shall be added to the projected roof area for inclusion in calculating the required size of vertical conductors, leaders, and horizontal storm drainage piping.

In its brevity, this section generates the potential for dramatically oversized storm piping. Consider a building with a tower above a low roof offset as in Figure 1. Taken literally, this section requires that 50 percent of each of the four facades be added to the building storm system. But, since rain cannot hit all for sides of the building at the same time the four facades should not be additive. This section should state that only the largest façade needs to be considered for sizing piping that serves multiple facades.

The UPC in contrast is quite verbose and confusing on this subject. It reads:

1106.4 Side Walls Draining onto a Roof. Where vertical walls project above a roof so as to permit storm water to drain to the roof area below, the adjacent roof area shall be permitted to be computed from Table 11-1 as follows:

(1) For one wall – add 50 percent of the wall area to the roof area figures.
(2) For two adjacent walls of equal height – add 35 percent of the total wall areas.
(3) For two adjacent walls of unequal height – add 35 percent of the total common height and add 50 percent of the remaining height of the highest wall.
(4) Two opposite walls of same height – add no additional area.
(5) Two opposite walls of differing height – add 50 percent of the wall area above the top of the lower wall.
(6) Walls on three sides – add 50 percent of the area of the inner wall below the top of the lowest wall, plus allowance for the area above the top of the lowest wall, per (3) and (5) above.
(7) Walls on four sides – no allowance for wall areas below the top of the lowest wall – add for areas above the top of the lowest wall per (1), (3), (5) and (6) above.

What the strange verbiage of this code section struggles to express is that it is addressing the interior of a structure protruding above a roof – a courtyard of sorts. This becomes evident in scenario (6) where it speaks to the area of the inner wall. The described geometry is pictured graphically in the UPC Illustrated Training Manual (see Figure 2). It is strange that such detail would be expressed for such an uncommon condition, while the more common condition of Figure 1 is neglected. Here again, as with the IPC, the treatment of the combined sidewall effect for a four-sided structure is completely ignored.

So in my opinion, sidewall runoff should be treated with 50 percent for a single wall, and 35 percent for two adjacent walls with common piping. If all four walls have common drainage piping, then the larger two adjacent walls would prevail with an allowance of 35 percent for sizing the common piping.

Let’s look at an example. Let’s assume this is New York City where the rainfall rate is 3 inches per hour (or 6 inches per hour if you have combined overflow drainage). NYC Code is based on the IPC. In Figure 1 the storm drain system is dramatically simplified for discussion purposes. As noted, the area for the main roof (drains 1 and 2) is 10,000 square feet. Since each drain carries 5,000 square feet and assuming the slope is 1 percent, each storm lateral will be 6 inches and the combined vertical leader will remain 6 inches. Drain 3 has a combined area of 22,500 square feet including the sidewall, so that lateral will be 10 inches. Same for drain 5. Since drains 4 and 6 each carry 15,000 square feet, the lateral will be 8 inches. For sizing the vertical pipe, the way the code is written one could argue that the total is 85,000 if you include all four sidewalls, but that wouldn’t make sense. That area would require a 15-inch pipe sloped at ¼ inch per foot.

Another approach would be to use just the largest façade for sidewall, which would be the one draining to either drains 3 or 5. In that case the combined total would be 50,000 square feet, nearly half the total above.

A compromise solution to these two approaches would be to take the worst “real world” scenario, which assumes that the rain is being driven by wind at a 45-degree angle to the building such that it is hitting two facades. In that case a reasonable design would be to include 35 percent of the two adjacent facades (as is mentioned in the UPC) draining to drains 3 and 4 (or 5 and 6). In such a case, the total drainage area would be 52,500, which still requires a 15-inch drain. But, it could slope at 1/8 inch per foot.

The dramatic nature of this example doesn’t entirely capture the effect that sidewall runoff can have in storm pipe sizing, but the square footage numbers paint the picture clearly. Neglecting sidewall runoff, the roof area is 35,000 square feet. Taken to the extreme, the sidewall runoff can increase the drainage area to 85,000 square feet.

However, having discussed the code perspective on this subject, it is my contention that sidewall runoff is a bunch of bunk. In the real world, it just doesn’t happen to any significant extent, no matter how much sense it seems to make on paper. How can I say that? Well, for 15 years I worked in Manhattan and was fascinated by this issue. It rains a lot in NYC and there are many wind driven rain storms. Many a time, I stood outside under an awning watching the façade of a glass high-rise, and never did I see any water sheet flowing down the face of a building. Many a time, I stood inside a glass high-rise during a wind driven storm, and never did I see water sheet flowing down the window I was watching.

Looking at another example, for a 600-foot tall, 200-foot wide glass tower, the numbers suggest that the water flowing down the face of such a building would be:

600’ x 200’x 0.5 x 3” x 448 / 43,560

This equals 1,851 gpm. That’s an enormous amount of water, as it would surely be visible to the eye. It would be like your car’s windshield in a brushless carwash during the power wash cycle. It just doesn’t happen.

Why doesn’t it happen, even though it seems that it should? I’m not entirely sure. There are a lot of variables. But, I’m sure it has something to do with the surface tension of water, the atomizing of raindrops, and a host of other parameters. If anyone reading this disagrees with me and has proof to the contrary, I would surely like to hear about it.

Timothy Allinson is a senior professional engineer with Murray Co., Mechanical Contractors, in Long Beach, Calif. He holds a bsme from Tufts University and an mba from New York University. He is a professional engineer licensed in both mechanical and fire protection engineering in various states, and is a leed accredited professional. Allinson is a past-president of aspe, both the New York and Orange County Chapters. He can be reached at laguna_tim@yahoo.com.