Locating the Source of Water Intrusion Can Be Tricky

As architects/roof consultants, there is nothing we hate more than to get a call from a client who says, “My new roof is leaking.” Yet, that is exactly what happened to us not long ago. My firm had put a new thermoplastic PVC roof system on a high-profile government building in central New Jersey. The owner was my long-time client, and I ran the project, so I was intimately familiar with it and utterly shocked to get this call about six months after the project was completed. We had just experienced a three-day nor’easter that began on Thursday night and ran straight through to Monday morning when the client arrived at the building to find numerous leaking areas.

I responded by immediately going to the building. I was accompanied by the roofing system manufacturer. As the client led us around the building, water was dripping through suspended ceilings all over, which gave us the sinking (almost apocalyptic) feeling you hope to never know. However, when we went up to examine the roof, much to our surprise, there was no blow off; no seams torn; in fact, no apparent defects at all. Our thermoplastic cap sheet looked perfect on the surface.

What we did find, however, was large amounts of water trapped between this cap sheet and the 90-mil bituminous base sheet underneath. This was creating large water-filled blisters on the roof that looked like an old waterbed as you walked up to and around them. No matter how hard we looked we just couldn’t find defects in the membrane surface or at any of the flashing connections or terminations that could be causing this. There was, however, a likely suspect looming adjacent to and above our roofs. The building experiencing the roof leaks has one whole face that is an aluminum-framed glass curtainwall. It extends up above the roof lines, slopes up (from the vertical) forming a peaked skylight, which then slopes back toward these roofs that were holding water. On the upper roof, sawtoothed skylights of the same construction were dripping water when we first arrived. This gave the only clue to where the “smoking gun” may lie.

METHODOLOGY

Water was dripping from the saw- toothed skylights into a planter in the 4-story atrium. The client said that was typical with all hard rains. Armed with this clue, and no other apparent explanation for such a large amount of water intrusion, the owner engaged us to find out what indeed was the root cause of this problem.

In a couple days, the dripping subsided and most of the water blisters had dissipated or at least were reduced and stabilized. In the interim, I assembled a team consisting of a roofing restoration contractor (this is not a rip and tear production contractor but one especially geared to finding problems and making associated repairs), skylight restoration contractor and testing agency capable of building spray racks onsite to deliver water wherever it’s needed. With this team, I embarked on a systematic investigation that would make any “detective” proud.

First, we plugged the roof drains and let water pool on the roof until the en- tire surface was wet. Meanwhile, “spot-ters” inside the building were looking for any sign of water intrusion using lights above the dropped ceilings. When this showed nothing, we began constructing spray racks and running water for set intervals on every adjacent surface rising above and surrounding the lowest roof in question. We first sprayed the exposed base flashings, then rose up to the counterflashing, then further up the wall, then to the sill of the windows above, etc. Then we would move laterally to a new position and start again.

This proved painstakingly tedious, but we knew that making the building leak was not enough; we had to move slowly and systematically to be able to isolate the location to determine what exactly was leaking and why. It is important when applying water this way to start low and only after a set period move upward, so when water does evidence itself as a leak, you know from what elevation it came.

After an entire day of spraying the rising walls surrounding the first (low) roof area, we could not replicate a leak. Somewhat frustrated—and rapidly burning the testing budget—we began the second day focusing on the adjacent peaked skylight, which is more than 75- feet long.

Again, we started low, where our base flashing tied into the knee-wall at the base of the skylight, below the aluminum-framed sill. Still no leaks. Late in the day, when we were finally up to the glass level, we sprayed water from the ridge and let it run right down the glass onto our roof below. Finally, we found some leaking occurring at a skylight flashing to wall connection. OK, that was reasonable to anticipate and easy to correct.