How a collapsed pool deck could have caused a Florida condo building to fall
After the deadly catastrophe at a high-rise beachfront condo building in Surfside, Fla., in June, suspicion quickly focused on a pool deck outside, in part because years of deterioration had left it badly damaged. But a mystery remained: How could the partial collapse of a patio have brought down much of a 12-story structure that had stood for 40 years?
A Washington Post examination — based on witness accounts, photography, construction plans, interviews with engineers and a computer simulation using a 3-D model of the complex — explores that question. The review identifies possible sequences of events where the collapse of the pool deck into an underground parking garage at Champlain Towers South could have weakened key columns, triggering the wider disaster.
Using the simulation, a team led by Khalid M. Mosalam, an engineering professor at the University of California at Berkeley, found that an initial deck collapse that extended beyond those key columns, where the deck joined the building’s facade, could have overloaded the columns and caused them to buckle. Mosalam’s team created its 3-D model and simulation for The Post.
Other engineers outlined a second possible scenario involving a less extensive collapse of the pool deck. In this sequence, the deck remained attached to the columns even as it collapsed, tugging or twisting them and the surrounding beams until the columns fell.
Both theories are consistent with a growing number of witness accounts that the pool deck collapsed first. The cause of that failure remains unknown. Some engineers pointed to water corrosion-related issues identified there in 2018, but others said such damage alone was unlikely to have caused the collapse.
Some suggested that a yet-to-be-discovered problem in the parking garage — or even deeper underground — could have brought down the pool deck and possibly the building. Investigations of the disaster, which killed 98 people, are advancing on multiple fronts and are expected to take many months.
Allyn E. Kilsheimer, a longtime engineer leading an investigation of the collapse for Surfside, told The Post that he is looking at several potential collapse sequences resulting from the pool deck failure, including the ones engineers proposed to The Post.
“We’re looking at every possibility,” said Kilsheimer, who added that investigators are “researching 20 or 30 things” that could have caused the initial deck collapse.
The Commerce Department’s National Institute of Standards and Technology and the Miami-Dade Police Department are also investigating the disaster.
[The shattered lives of Champlain Towers South]
Joe DiPompeo, the president of the Structural Engineering Institute of the American Society of Civil Engineers, said that both theories put forward by the engineers are “plausible” accounts of how a failure in the deck could have led to the building collapse.
DiPompeo and other engineers cautioned that nothing will be conclusive until investigators finish analyzing the wreckage, which has been removed for examination.
Mosalam, the director of Berkeley’s Pacific Earthquake Engineering Research Center and a fellow of the American Society of Civil Engineers, stressed that his simulation depicted a “possible sequence of events” and that it does not provide “the definite answer we all are looking for.”
The president of the condo association board and the board’s attorney did not respond to messages seeking comment for this article.
The troubled slab
Champlain Towers South, which contained 136 apartments, was built on a concrete slab about 9½-inches thick that spanned about 55,000 square feet. A grid of steel reinforcing bars, known as rebar, ran through the slab to boost its strength and stability.
The pool deck was laid on the southeastern quadrant of that base slab, with a pool and Jacuzzi in one corner. Almost five inches of additional layers including concrete, waterproofing and pavings sat between the slab and the surface that residents walked on, according to a 2020 engineer’s report.
Raised concrete planter boxes were installed in various locations around the complex, including at the northern edge of the pool deck, near to where it joined the building.
To the west of the pool deck, on the same concrete slab, was a driveway and a parking area. Separately, a single-level basement parking garage — accessed by a ramp on the building’s north side — was built into a 10½-foot-deep hole beneath the slab, so the slab was the garage’s ceiling. Rows of concrete columns ran vertically to the roof from the foundations beneath the garage. The columns held up more steel-reinforced concrete slabs, which made up the building’s floors.
Concrete in the pool deck was repaired at least four times between 1996 and 2013, according to town records. Multiple palm trees and planters not in the 1979 plans were added to the slab, archived photographs show. By late 2017, most of the palm trees were replaced with smaller bushes.
In October 2018, engineer Frank P. Morabito reported to residents that he had found “major structural damage” to the pool deck, adjacent driveway and planters, where failed waterproofing had allowed corrosion.
Water damage was visible on the garage ceiling close to where the concrete planters sat on the deck above, video recorded last summer shows. The condo association’s board warned residents in October that a lack of waterproofing on large sections of the slab had “exposed the garage to water intrusion for 40 years,” and said significant work was needed. “Where there is waterproofing, it has failed,” residents were told. “Water has gotten underneath and caused additional damage to the concrete.”
As part of a $15 million repair package for a locally mandated 40-year recertification, Morabito planned to tear up the pool deck and driveway, down to the base slab, and dig up and replace the planters. Contractors were to install new drains, waterproofing and materials to block water and roots from escaping the planters and damaging the concrete.
Nothing in Morabito’s assessments warned that the pool deck was in danger of imminent collapse.
The deck collapses
About 1:14 a.m. on June 24, ground-floor resident Sara Nir heard a noise so loud that she thought it was construction-related. Nir has said she went to report it to a security guard in the lobby, where, about a minute later, she noticed that the pool deck and adjacent surface-level parking area outside had partly collapsed.
Around that time, Adriana Sarmiento and her husband heard a noise while swimming in a pool at a resort directly across from the entrance to the Champlain Towers South underground garage. She said that at about 1:18 a.m., she pointed her cellphone toward the entrance ramp and began recording what appeared to be water running down from the ceiling of the garage and a small pile of debris on the floor.
Based on the location of a numbered column visible in the video, the water and fallen debris appear to have been roughly below the planter beds on the pool deck’s northern edge.
“The ceiling had collapsed in that area at that point. We know that,” said Kilsheimer, the investigator.
About 1:24 a.m., someone who had called 911 to report that the roof of the parking garage had collapsed said they saw the tower come down. “Oh my god,” the caller said repeatedly, according to recordings released by police. “The whole f------ building collapsed.”
The precise shape and size of the section around the pool deck that collapsed cannot be determined from photographs. Part of the deck ended up buried under rubble, obscuring the edges of the collapsed section. Engineers said some of the damage to the deck that is visible in photographs could have been caused by fallen debris.
The concrete slabs that made up the floors of Champlain Towers South helped stabilize its columns by providing horizontal support that prevented bending. The Mosalam team’s simulation, developed using the plans from 1979, examined what would have happened if a collapse extending out from the pool deck stripped away that stabilizing force for key columns where the deck joined the building.
In that scenario, the deck’s collapse would have left these 16-inch-thick columns “unbraced” for about 24 feet, from the floor of the parking garage up to the ceiling of the first floor. The columns would have effectively become twice their intended length — and as a result would have had only a quarter of their strength.
The calculations by Mosalam’s team, which included his former student Carlos Arteta of the Universidad del Norte in Colombia, showed that before the deck collapse, the “demand” on the columns in this area was about 60 percent of their capacity. Once the deck was removed, they found, the demand probably would have exceeded the columns’ capacity, causing them to buckle.
According to the team’s model, the buckling would have drastically increased the stress on the columns deeper into the building, which the team found were among the most heavily stressed in the building before the collapse. These columns probably would have then failed as well, setting off a cascade as the floors of the building crashed down, Mosalam said.
“Columns doubled in height, loads increased on the columns because of the loss of this support, columns failed, and then the rest of the structure failed,” he said, summarizing his team’s hypothesis.
Video shows that the central section of the building, above those key columns, collapsed first. The eastern portion remained standing for several seconds before it teetered and fell.
Vertical rebar reinforced all the tower’s columns, according to the plans. A limit on how much rebar should be in a column is contained in an American Concrete Institute code. Two engineers said the limit, which is unchanged since 1979, is intended to leave enough space around the steel rods for poured concrete to set properly and bond to the rods.
Some columns toward the bottom of Champlain Towers South — including the key columns in Mosalam’s model — would have contained excess rebar if contractors followed a drawing in the plans that showed how vertical rebar typically should have been arranged where the columns joined each floor. It is unclear whether builders followed that drawing in all cases. Another part of the plans specified that rebar in the areas where the columns joined the floors should be made to comply with the code. The potential code violation was first reported by the Miami Herald.
R. Shankar Nair, a longtime Chicago-based engineer, told The Post that exceeding the rebar limit should not by itself have caused the columns to fail. “But it may be indicative of a general carelessness in the design,” he said.
William M. Friedman, the architect whose firm designed the building, died in 2018. Sergio Breiterman, the consulting engineer who certified to Surfside that the building was up to code, died in 1990. Breiterman’s business partner, Manuel Jurado, could not be reached to comment. Jurado, 92, previously told The Post that he did not remember the firm’s work on Champlain Towers South.
A second scenario
Other engineers said the way the pool deck was joined to the building could have helped to bring down columns in that same area, even if the pool deck collapse was less extensive than Mosalam’s model assumed.
The tower was built on an elevated base, with the first floor set at least four inches higher than the exterior slab. A vertical concrete “drop beam” provided the elevation to that first floor.
Because of how these structures were joined, even as the pool deck collapsed it may have remained connected to a critical column or columns, said Albert Bleakley, an associate professor of engineering at the Florida Institute of Technology.
A drawing of the first floor in the building’s plans called for the rebar running through the pool deck to end in an L-shape that connected it to the drop beam.
If these connections held, the collapsing deck may have exerted a destructive tugging and twisting force.
The precise mechanics of that force could have taken several forms, engineers said. It could have been exerted directly on the columns or through beams running from the columns out into the deck.
Or the collapsing deck could have acted like a wrench and twisted the drop beam to the point of “mutual destruction of both the beams and columns at those connection points,” said Joshua Porter of Consult Engineering in Punta Gorda, Fla.
Assessing whether these connections would have held is difficult because the building’s plans are contradictory in places. A close-up drawing of that joint on the page showing the design of the first floor omitted the L-shaped bar connecting the pool deck to the drop beam. Without these bars, the pool deck may have more easily separated from the drop beam instead of twisting it and the columns.
Mosalam’s model did not incorporate the drop beam. He said it would not have affected the sequence in his simulation because if the deck collapsed as he presumed, the drop beam would have collapsed with it. Other engineers agreed that this was reasonable.
Why did the deck collapse?
The cause of the pool deck collapse remains a mystery for investigators. Some engineers said prior reports of water damage in the area, and photographs of the wreckage, may provide important clues.
Several engineers said that, based on photographs of the site, the slab that made up the pool deck and surface-level parking area appeared to have experienced a “punching shear” failure, in which a slab separates from its supporting columns and falls — as though the columns had punched through it.
Common reasons for punching-shear failure include an increase in the weight bearing down on a slab or a weakening in the slab — caused by severely damaged concrete, for example — that leaves it unable to hold its load, engineers said.
In photographs of the site, parts of the steel rebar that reinforced the Champlain Towers South slab can be seen sticking out of the columns where the slab once sat. These bars were laid in two grids, one closer to the top of the slab and one closer to the bottom. The building’s plans called for eight top bars and four bottom bars to run through every column.
In such a design, the bottom grid of rebar is more important for limiting the impact of punching-shear failure, experts said. When a column punches through a slab, strong bottom reinforcement can slow or even prevent the slab’s total collapse.
Kilsheimer said investigators were eager to check whether the concrete and steel used in Champlain Towers South, and the arrangement of the rebar, matched the building’s plans. Mosalam and three other experts said they saw no clear signs in photographs that the rebar arrangement deviated from the plans.
According to the plans, two of the bottom bars were spliced together at each column rather than running continuously.
Splicing rebar together as it passes through columns does not violate building codes in effect now or at the time of the construction, though today’s code requires four bars rather than the two in the plans as well as greater overlap where those bars are spliced together, according to engineers and a copy of the current guidelines. But Mosalam said that rebar spliced close to columns is more likely to slip out of slabs, allowing them to collapse.
A photograph of a core sample engineers took from the pool deck last year appears to show that the base slab’s concrete was cracked and crumbling, said Porter, who added that this probably was caused by corrosion. He said this process may have split the base slab into separate layers, weakening it and making it easier for the columns to punch through, causing the deck to collapse. Some engineers cautioned, however, that damage in the core sample could have resulted from the drilling required to take the sample.
Daniel Linzell, an engineering professor at the University of Nebraska at Lincoln, said that the damage flagged in Morabito’s 2018 report and later by the condo board may explain the pool deck failure. “Those things could certainly have led to this collapse,” he said.
Jason Borden, an engineer who examined Champlain Towers South last year while preparing a bid for work, said possible leaks from the planters near the critical columns could have corroded the concrete.
“If the deterioration of the concrete is concentrated at the planters, and the planters are adjacent to the building, we can surmise that there might have been some concrete deterioration at that joint between the floors in the building where they could have impacted the overall integrity of the building,” said Borden, a regional director at O&S Associates.
It is also possible that the deck collapse occurred in stages, Kilsheimer said.
“I think you’re going to find out that it probably collapsed twice. And so that’s what we’re working through now,” Kilsheimer said. He declined to elaborate, but previously noted to The Post that residents had reported hearing several loud noises in the minutes before the building came down.
Why did the west side survive?
The western portion of the building remained standing despite facing some of the same dynamics that engineers said may have brought down the rest of the building.
“From an engineering perspective, the interesting question is why the columns under the building that remained standing did not buckle,” Bleakley said.
Some columns beneath the western section were left at double height when the deck collapsed, photographs show. Kilsheimer said the fact that this did not cause the west side to collapse is a reason to question Mosalam’s hypothesis.
Other engineers, however, said the part of the structure that didn’t fall may have been protected by larger 24-by-24-inch columns in the portion of the parking area near the pool deck, as well as horizontal concrete beams between the columns at the second-story level. The collapsed portion of the building lacked those reinforcing beams and had mostly 16-by-16-inch columns in the area where the pool deck joined the building, the plans show.
The Mosalam team’s calculations showed that the columns in the western section were shouldering less demand than those that triggered the collapse in his simulation, and that as a result the western columns remained well within their capacity even at double height.
In addition, the western portion of the building stood on columns, as if on stilts, elevated above the surface-level parking lot. The same slab spanned the pool deck and the parking lot, so there was no need for a drop beam to join the two areas. This means the deck collapse may not have acted on the columns there in the same way that it did on the columns in the central portion of the building, engineers said.
Several engineers also noted that the western part of the building featured shear walls — continuous vertical plate-like walls designed to resist sideways forces — around the elevator shaft. This was probably the strongest part of the building, the engineers said.
“I think the shear wall protected this part and carried a big part of the load,” Mosalam said.
Dalton Bennett and Beth Reinhard contributed to this report.