Roof Collapse Due to Water Ponding

Here are a few photos on a current structural engineering project I am working on.

It’s a 1950s commercial block building 100’x50’ with a flat roof. It has a block wall down the center creating a series of 20’ x 25’ bays. Every 20’ has a W16x36 steel beam with 20’ 2x10 at 16" centers. The aerial photo from Google clearly shows the bays with water stains in each bay due to previous ponding. The last bay in the top right corner has standing water when the photo was taken. The roof rafters recently broke and caved in about 18 inches. For now it was temporarily shored up until repairs can be made. The roof failure was due to several factors:

  1. Being flat the roof has no positive drainage other than running off the back wall.
  2. The natural sag (deflection) of the 2x10 joist due to the dead weight of the roof creates a a low area for water to pond. This was a EPDM roof with 1" ballast rock on top.
  3. Wood will experience additional long term deflection 1.5 to 2 times the initial deflection called creep. Creep is basically the stretching and/or compression of the wood fibers over time, which is permanent.
  4. The roof joists were #2 grade lumber.

Over time the initial deflection allowed water to pond creating more sag until one or more rafters cracked at large knots in the joists. When a joist breaks its load is transferred to the adjacent joist, which fails and you see the picture.

Turns out the #2 2x10 joist at 16" centers were doomed to fail from day one. The weight of the dead load combined with the snow load would exceed the allowable bending stress of the joists by 21 percent. The repair is to install new 2x10 Select Structural Grade joists at 12" centers as a temporary repair for the damage. The remaining bays constructed the same are now in question. Modern building codes don’t allow flat roofs. Two percent is now the minimum for a low-slope roof. I am meeting with the owner and the city building code official to see where we go from here.

As commercial inspectors don’t assume a flat roof with evidence of ponding that hasn’t failed in the last 70 years can’t fail tomorrow. Make note of any signs of ponding, including clogged drains on sloped roofs that also create potential ponding failure.


Once the new framing is installed, I would be recommending they strip the whole roof of that ballast and install a new tapered insulation with adhered EPDM along with roof drains as required.

1 Like


Have you installed the tapered insulation.

Yes, on some of my past jobs, of course a sub-contractor. I was the project supt. at the time.

1 Like

Were all the layers in the built-up wedge glued together and glued to the roof?

Yes, and here are a few pictures of roofers working around a drain area. It was on metal deck, but no different on a wood deck. Blocking height around the drain is to compensate for whatever thickness insulation you want to add on the roof and then the tapered insulation goes on top according to the drawing which is usually drawn up by the tapered insulation manufacturer to fit the roof design.


Interesting read, many thanks to both of you. Always an education.

1 Like

I agree with Marcel that is best to remove the ballasted roofing system as the rock ballast adds significant weight to the structure, usually about 10+ psf. An adhered PVC, TPO or EPDM single-ply roof can be installed over a tapered insulation system for positive drainage if there is not slope in the structure now. The insulation is normally mechanically attached with long screws and plate washers to the roof decking. I worked for SBC telephone company (now AT&T) many years ago in facility engineering and we did many of these reroofs. Most of the central office equipment buildings had large reinforced concrete flat slab decks. Tapered insulation with single-ply roofs worked well. Suggest you talk with a commercial roofer in your area for guidance.