This wildfire ADU burn test from the Insurance Institute for Business and Home Safety (IBHS) shows exactly how fire spreads from a burning accessory dwelling unit to a neighboring structure, even when that ADU is built to California Building Code Chapter 7 fire-resistance standards. Researchers at the IBHS research center in Richburg, South Carolina, vary wind speed, building spacing, orientation, and wall and window construction to measure the radiant, convective, and direct-flame exposure a target building experiences during a live burn.

TRANSCRIPT:

Introduction to the IBHS research center (0:00–0:56)

Today we're headed to Richburg, South Carolina, to the Insurance Institute for Business and Home Safety's research center. I've come here in the past to witness a manufactured hurricane using their giant wall of wind, and we even featured some of that footage during the inaugural season of Building Resilience.

They also have a roof farm where they run longevity testing on various shingle brands, taking sections of exposed roof into the hail lab and shooting them with hailstones.A computerized micro scanner measures the effect of the hail on the shingle, and the results get published. The hail itself is homemade for consistency across different densities.

Today, the wildfire team is looking at accessory dwelling units—ADUs, like a granny flat in the backyard—but this one comes with test panels and a team of firefighters standing by.

Setting up the ADU burn test and target structure (0:56–2:33)

In response to the Palisades conflagration, the team wants data on how close is too close, and whether beefing up fire-resistant materials changes that distance. Ultimately, they're looking at what a burning structure exposes other buildings and materials to. There's an ADU that will be lit, and a target building—or downstream building—nearby.

As the ADU ignites and the fire progresses, the team studies what impact that has on the downstream structure, since understanding what buildings are exposed to in suburban conflagration events is central to the research. When fires move into a community and homes start to ignite, the question is what the neighbors experience, and ultimately what has to be designed against and resisted in order to better protect structures and reduce wildfire losses and destruction.

The U.S. has a significant housing shortage, so many people, particularly in California, are adding ADUs wherever they can. The Palisades disaster showed clearly how wildfire can overwhelm a neighborhood, so the question becomes how to add density in a way that reduces the potential for property loss and fire spread.

Test variables: wind speed, construction type, code compliance (2:33–3:36)

The test looks at a range of variables: wind speeds between 30 and 50 mph, different types of ADU construction, different window types, and different wall sidings—whether or not they're compliant with California Building Code Chapter 7—and how that affects the exposure the target building experiences. They're also testing the spacing between buildings and even orientation, comparing a building set straight-on versus rotated 45 degrees, to understand what homes need to be designed to resist.

Fire can jump from one building to the next essentially the same way heat transfers without fire—through conduction, convection, and radiation—being close enough to radiant heat that a structure can spontaneously combust.

What makes the ADU code-compliant (3:36–4:48)

The ADU being set on fire is compliant with California Building Code Chapter 7, which covers fire-resistance-related construction. What makes it compliant comes down to the wall: a layer of gypsum sits between the siding and the sheathing, which allows combustible siding to still meet code. The second element is dual-pane tempered glass windows. The third is the open eave design—while it looks like an open eave, hidden behind the wood is another layer of gypsum board between the roof sheathing and the visible wood, which is what allows it to comply.

Despite meeting the code's fire-resistance standard, the team is going to set it on fire anyway and test how the fire spreads to other buildings. The gypsum sheathing is expected to be the last thing to burn, but the siding and studs create what amounts to a fire sandwich, and the roof overhang is expected to catch quickly—making the gypsum sheathing something of an afterthought and illustrating that partial fire-resistance measures don't always add up to meaningful protection.

Instrumenting the target structure to measure heat exposure (4:48–7:37)

From a scientific standpoint, the goal is a consistent ignition of the structures across all the variables being changed—wind speed, distance, orientation—so the research question becomes: once ignition happens, what impact does it have on surrounding structures?

The surrounding structure in this case is a small building packed with roughly half a million dollars of test equipment and data-logging boxes, so the goal is to harden that structure enough that it doesn't ignite at all—it isn't built to mimic a real-world structure, since the priority is protecting the instrumentation and being able to measure everything.

There are three types of instrumentation on that building: heat flux gauges, which measure how much heat is being applied to the wall; thermocouples, which measure the actual wall material temperature; and radiometers, which isolate just the radiative heat—the kind you'd feel from the sun. Several wall panels, some with windows and some without, are placed at different distances around the sacrificial ADU to observe how heat exchange can cause combustion.

The instrument shack itself is essentially fireproof, but it has sacrificial panels attached to it made of combustible wall materials that could ignite during the test. If one starts to ignite, a rip cord lets the team pull the panel down to the ground so it doesn't threaten the shack—allowing researchers to measure everything the building experiences and test different materials without putting the instrumentation at risk.

Inside the ADU, the setup mimics a real home, right down to a small laundry hamper and potato chips on the counter, so the combustion materials are as representative as possible of a real household. Sacrificial cameras and sensors also beam data and photos to the cloud until they're engulfed in flames, letting the team monitor from a command center in real time when the fire finally penetrates the wall and ignites inside the structure. A post inside has thermocouples spaced every couple of feet from floor to ceiling, measuring the temperature profile at different heights as part of the data set.

Monitoring ignition and fire spread in real time (7:37–9:12)

Eventually, the fire inside will grow large enough to break the window glass, and since the wind is blowing toward the target structure, the fire will move into the roof and the house will become fully engulfed. At that point, the research focuses on whether flames reach the target structure for direct flame contact, the radiation coming off the burning structure, and convective heating—air pushed toward the target getting heated by the flames and, in turn, heating that structure.

Interview: IBHS chief product officer on windows as a weak point (9:12–10:14)

IBHS Chief Product Officer Alistister Watt explained that his teams take the research and turn it into action through codes, mitigation programs, and education, working to influence people toward more resilient homes across America. He pointed to windows as one of the weakest links: wildfire behaves a lot like a hurricane in that keeping it out of the structure limits the damage, but once it gets inside, the outcome changes dramatically.

Windows are one of the easiest paths for fire, wind-driven embers, and water to get in. Intuitively, the glass seems like the vulnerable part, but it's usually the window frames that melt first—prior research and observation show the frames tend to fail well before the tempered glass does.

The controlled burn and test results (10:14–11:28)

The first step in a controlled burn like this is igniting the cribbing, which catches quickly with the fans left on idle before being turned up to bring the flames across, and things get significantly hotter once the roof comes in. The wall panels sit 20 and 30 feet from the ADU, with the tech shack 30 feet away, directly downwind of a wall of wind that runs from a low rumble of 12 mph up to about 55 mph.

The whole ADU was fully engulfed within 10 minutes. The 15-foot siding sample caught fire, but the other window system, with two panes, was unharmed—a notably strong result for glass and frame surviving at that distance. Watching embers attack combustible material like mulch and vegetation makes clear this was a realistic test of what happens in the real world, and it's no surprise that when one house catches fire, unprepared neighboring houses catch too, leading to a broader conflagration.

Key takeaways on fire-resistant design (11:28–12:49)

Tempered windows, gypsum sheathing, metal doors, and other fire-resistant features don't stop fire on their own. With this Chapter 7 construction—gypsum behind combustible siding, 2x4 framing, and gypsum on the inside—the siding failed fairly quickly as the gypsum, coated in combustible paper, began to disintegrate on the outside. The windows gave way next, letting flames in and attacking the gypsum from both sides, so the walls failed before long. It took a bit longer than expected for the fire to reach the eaves and attic, but that failed fairly quickly too, with complete failure around the 14-minute mark.

Fire is relentless—like water, wind, and gas combined, it probes for the weakest opening the way water does, while traveling and behaving like wind. A systemic, whole-neighborhood approach to fire-resistant design is the best defense, and it starts at the property line in suburban neighborhoods.