Updated COVID-19 GuidelinesLearn More

By Edward W. Pon, AIA

As a Building Envelope consultant, we will be called upon to address and diagnose water infiltration issues or leaks at fenestration assemblies: windows, doors, skylights, curtain walls, storefronts, etc., or at other building assemblies. As there are many industry-developed standardized ASTM or AAMA tests for fenestration, we will not discuss them here. Instead, this is a discussion about how we conduct a water test at those “non-fenestration assemblies”. We will explore the methodology and logic in determining the sources of leaks through other construction assemblies. 


In water testing “other building assemblies” requires an understanding of all types of construction, and it requires knowledge of building science and the nature of materials. The water of course comes in many forms, and it can move in sometimes mysterious ways. Water can come from plumbing leaks, or it can come in the form of groundwater, rainwater or condensation, or water vapor. 

First off it is helpful to know whether we are dealing with a plumbing leak or a weather-related leak. Does it leak only after a weather event or is it constant in which case it is likely a plumbing leak? Sometimes it’s both as in the instance of a storm drain. 

If it is a plumbing leak, it can be a supply line, waste line, storm drain, steam or hot water heating, or condensate line. If it is coming from ductwork, is it condensation or is it stormwater coming into the ductwork? 

If water is coming thru the foundation wall it could be groundwater or it may be a broken pipe or an underground stream. Sometimes testing the water can help identify the source. The presence of chlorine will generally indicate a municipal water main break, E. coli will indicate a sewer main break, and the absence of either can indicate groundwater or an underground stream. Some bacteria but no E.coli would generally indicate surface stormwater. 

We are often first looking at where the water has entered the interior. We need an understanding of general construction. If water is appearing on the ceiling, what is the construction of the ceiling? Is it plaster adhered directly to a concrete slab? Concrete encased steel beam, terracotta flat arch or is it a suspended gypsum board or suspended ceiling tile system? We need to understand what these materials are, their density of these materials, how thick they are, and how they are assembled. Are we near plumbing lines, storm drain lines, domestic water lines, steam lines, and mechanical ductwork? Are there any layers of membranes or impermeable coatings? Are there chemical water repellents? 

We need to investigate the ceiling cavity if any and we need to look at the floor or roof above. We need to perform destructive probes if necessary. When performing probes, we need to carefully look at each layer of material that we encounter almost in a way that an archeologist carefully removes layer after layer of material. A coating can sometimes be nothing more than a thin discolored layer that can easily be overlooked. 

Leak from suspended ceiling. Measure location relative to exterior walls and windows in two directions so the leak can be located from the roof.

Lead from suspended Ceiling Sullivan Engineering

The suspended ceiling space directly above the underside of the roof slab.

Discolored Wet Spots Sullivan Engineering

Look for discolored or wet spots, water, or rust stains from pipe sleeve

Pair of pipe sleeves located on the roof and measured from exterior walls or parapets corresponding to the pipe sleeves observed from below.

Pipe Sleeve partial application Sullivan Engineering

Close examination of pipe sleeves reveals the sloppy or partial application of silicone sealant where absorptive and still-wet fiberglass insulation is partially exposed to the elements.
In this instance, the source of the leak was clear, and no water test was needed.























We need to get a 360° view around the entire area. We need to see if the water is traveling horizontally. A roof leak location inside often does not always correspond directly to a leak on the roof immediately above the leak location inside. For instance, water could drip down a sloped rafter or framing member and travel along that framing member before it drips down to the ceiling where it becomes visible. The surface tension of water can surprisingly change the direction of a leak a far distance. 

Do not underestimate the devastating effect of surface tension and the lack of a simple drip edge can have on a leaking situation. In the photo below, the lack of a proper drip edge from the balcony slab above was one of several factors that caused the water infiltration and catastrophic failure and collapse of the gypsum-based fascia and soffit.  

Surface Tension example Sullivan Engineering



In another example, a recently installed roof drain and drain flashing were suspected to be leaking after a heavy downpour. Below is the suspected drain and flashing. 

Raintrol Drain Sulliavn EngineeringRaintrol Drain Sulliavn Engineering

The drain is a JR Smith 1083 Raintrol drain that allows for stormwater flow rate adjustment. These drains were installed throughout the roof to slow the flow of stormwater which was often overwhelming the building’s stormwater system. In a recent heavy downpour, leaks occurred two floors below the roof. The drain and flashing were suspected. We performed a water test to locate the source of the leak by using a process of elimination. 

Wall opening Drain Sullivan Engineering

Wall opening

Cast Iron Storm Drain Sullivan Engineering

3” cast iron storm drain line 








  1. An opening in the storm drain plumbing chase wall had been made where leaks occurred.
  2. We tested the drain and drain flashing at the roof first by removing the strainers and plugging the drain with an expanding drain plug. Approximately 4” of water and an area of about 5’- 6’ in diameter was flood tested for about 30 mins. No water was observed in the chase or on the pipe.
  3. The drain plug was then removed and within 30 seconds, water was observed streaming down the side of the drainpipe. See the link to the video here: IMG_0894.MOV (egnyte.com)



By first testing the roof drain and flashing we were able via the process of elimination conclude that the breach in the system was not in the roof membrane or drain body but in the drain line somewhere between the drain body and the location of the opening in the plumbing chase wall. 

The important concept to remember is that you must use the process of elimination and isolate the area being wetted so that when it does leak you can accurately attribute the leak to a specific location or component of construction. It is also vital to allow for sufficient time for water to migrate or work its way through the material or construction assembly. If the process is rushed it is easy to mistakenly attribute the leak location at a later stage of the test when the leaking may have occurred at an earlier stage in the testing. And finally, common sense tells you that when spray testing with either a garden hose or spray bar, remember because of gravity and the nature of water flow, you need to “start low” and work your way up higher in elevation along the slope or up the wall or up all the key construction assemblies to the highest suspected breach point. 



For example: in the photo below this is a 2nd-floor terrace where leaks were reported below the terrace floor under the brick walls. 

Water Testing Sequence Sullivan Engineering

The water testing sequence was as follows using the “start low” rule

Flood Test Sullivan Engineering roof

  1.  Flood test the roof surface and drain, 2” of water for about 1 hr. 
  2. Flood test the base flashing, dam the area with sandbags, or use 2×6 boards to raise the water level up 5” against the base flashing for about 1 hr. 
  3. Spray just above the metal counter flashing for about 1 hour. 
  4. Spray the low aluminum windowsill on the left side for about ½ hr. 
  5. Spray the brick wall on the right side for about 1 hr. 
  6. Spray the high precast concrete windowsill on the right side for about 1 hr. 


A wet spot was observed after about 25 min. after stage 2 of testing


Breach found Sullivan Engineering

Breach found in EPDM base flashing








The roof slab construction was concrete, so we need to allow for at least 1 hour for the water to work its way thru the concrete. Windowsills were tested for ½ hour because leaks usually become evident more quickly thru sheet metal. 

In this leak investigation, we only had to go to step 2 before we found the breach in the EPDM base flashing. The roof was covered with rigid insulation board and pavers which were removed before water testing. 

Sometimes it is a good idea to reconfirm your findings especially if water is slow to migrate through dense construction materials. You can stop your test or lower the water level to the prior test stage in a flood test or stop spraying, observe the reduced flow, (sometimes by counting the rate of droplets per minute ), and then retest the area and reobserving the flow rate change. This will take some additional time but this way you can make 100% certain you have found the source of the leak. 

In the roof drain example above, we did not “start low” but then we were only testing two possible assemblies, 1) The roof and drain body and 2) the drain line below the roof. There we used the water collected on the roof to test the drain line. This best simulated torrential rain and provided a sufficient volume of water to adequately test the drain line. A garden hose suspended in the drain line may not have provided enough water volume to adequately test the drain line. 

In conclusion, no two leak investigations are exactly alike. Some leak investigations are resolved in 20 min. and others can take several years to fully resolve mainly because sometimes leaks can have several sources especially when the construction is complex. It can take a long time to cycle thru several investigations, tests, and repair cycles before all sources are discovered and repaired. One such investigation involved a complex assembly of the roof terrace, two sections of the parapet wall, skylight, skylight gutter, and scupper drain from the gutter to the roof terrace. Each component or intersection needed to be isolated and tested and repaired. Eventually, it was discovered that there were leaks coming from the cap flashing of the parapet wall, skylight, skylight gutter, and scupper drain. This investigation took almost 3 years to complete. 

It is important to just follow a few simple principles and be patient. It is also important to be aware of where water is flowing and be wary of letting the test go on for too long. Do not test overnight unless it is fully monitored or there is no chance of potential water damage.  All tests need to closely be monitored with a second person with two-way communication from the inside so the water test can be stopped immediately once water infiltration is observed. Take note of each test stage’s start times, dwell times, and end times. There is such a thing as being too patient. Letting a water test run too long can sometimes result in unintended leaks and costly water damage to adjacent areas and yes … lawsuits. Be sure you inform the client of potential damage, obtain written owner consent, and our standard liability waivers before testing is begun. 


Recent Posts

About Sullivan Engineering, A Rimkus Company

Sullivan Engineering provides high-quality building envelope restoration and compliance solutions.

We partner with facilities managers and account executives to provide technical expertise and project management for building envelope restoration, compliance, and maintenance.

Our solutions reduce the overall building life cycle maintenance costs by creating long-lasting, high-quality work for years to come.