Video Transcript: NIST’s Technical Findings in the NCST Investigation of the June 24, 2021, Partial Collapse of Champlain Towers South

NCST Champlain Towers South Investigation | Technical Findings (June 2026)

Judith Mitrani-Reiser: Shortly after 1:00 a.m. on June 24, 2021, the middle and east parts of the Champlain Towers South condominium in Surfside, Florida, collapsed. Only the west part of the tower remained standing. Ninety-eight lives were lost in this tragedy.

The 136-unit condominium building stood 12 stories tall plus a penthouse, and it was built of steel-reinforced concrete. At ground level, a pool deck and a parking deck made up of concrete slabs were supported by columns over the basement of CTS.

The building had stood on this stretch of the Atlantic coast since 1981. There was no obvious cause of the collapse, such as an explosion or a storm. Within days, the National Institute of Standards and Technology launched a national construction safety team investigation to determine the technical cause of the collapse.

I'm Judith Mitrani-Reiser, co-lead of the investigation.

Glenn Bell: And I'm Glenn Bell, co-lead of the investigation. In this presentation, we will provide a summary of our technical findings.

Judith Mitrani-Reiser: Our team of experts has completed the technical portion of its investigation. This includes the analysis of physical evidence and historical records as well as extensive interviews. We conducted materials testing and geotechnical studies. We built and tested building components like those within CTS and ran many computer simulations of the failure. We examined two dozen possible scenarios for where and how the failure started.

Based on this extensive work, we have concluded that the failure most likely began in early June, about three weeks before the collapse, when two connections between garage columns and the pool deck slab failed. Once the first connection failed, other elements of the pool deck were left to carry their loads, but they were not strong enough to handle them due to problems that stemmed from the original design and construction of the building.

When building structures are designed and built to required codes and standards, they have margins against failure, meaning they should be able to support much more load than they are expected to bear. In the case of Champlain Towers South, these margins against failure were too narrow from the start.

NIST investigation revealed that the collapse of Champlain Towers South began with what are called punching shear failures. In this type of failure, the slab breaks around the column, giving the appearance that the column punched through it. It occurs when the load on the slab becomes great enough to cause it to bend and crack around the column, weakening the connection that is holding the slab in place.

After the first two slab-column connections failed in early and in mid-June, respectively, the slab that they supported did not fall immediately. It was held up by the pool deck and street-level parking deck slab surrounding it. When the connections between the pool deck slab and these two columns in the garage below failed, it increased the loads on neighboring connections, eventually causing them to fail as well.

The failures spread to other elements of the pool deck and street-level parking structure, eventually unseating the southern edge of the pool deck slab from a supporting wall. We believe this caused the slab to sag further and eventually break away from the face of the middle section of the tower, seen at the extreme right of this image. This left the slab resting completely on crushed cars and the garage floor below it.

There's no way to know precisely which path these failures took from the initial two punching shear failures. But this scenario we have described here is consistent with principles of structural mechanics, eyewitness accounts and other evidence. Small variations in the sequence could also be consistent with structural mechanics and evidence.

Most importantly, when the pool deck broke away from the middle part of the tower on June 24, 2021, it damaged two connections. The damage weakened these joints, causing the failure to enter and to spread through the middle part of the tower. Once the failure began in the middle part of the tower, it caused what's called a progressive collapse in the tower. The collapse progressed through the middle part of the tower and then to the east part.

Glenn Bell: Low margins against failure existed from the time Champlain Towers South was built, primarily caused by two factors. First, severe and widespread deviations in the building's original structural design from the codes and standards of the day, but also some limitations in those codes and standards. And second, deviation in the building's construction from the design drawings. Loads added to the structure over its life further diminished the margins against failure.

But what caused the collapse in 2021? We found no evidence of any specific initiating event. The final factor that brought the critically low margins of safety to the point of failure was most likely long-term degradation from corrosion.

While we have determined the cause of the collapse, our work is not done. We continue to develop recommendations to improve the structural safety of buildings based on what we've learned. We will publish a report as we have done for our previous investigations, which will include all of the technical analysis and our recommendations. We will then work to ensure that standards, codes and practices are updated as needed.

Throughout this investigation, we have kept top of mind those who were most directly impacted by the collapse of Champlain Towers South. We appreciate everyone who has helped with this work, including the survivors and the families of those who were lost. With their invaluable input, this effort will help make other buildings safer and help prevent tragedies like this from happening again.

Judith Mitrani-Reiser: Now we will walk you through more detailed technical information that summarizes our investigation into the partial collapse of Champlain Towers South.

Whether you're watching your first NIST video on the investigation of Champlain Towers South, or you've been following our updates for the past five years, we would like to provide all our viewers the necessary background information to follow our presentation. Please use a QR code provided on the screen to learn more about the National Construction Safety Team Act, NIST NCST investigation into the partial collapse of Champlain Towers South, and the details of the technical work required to investigate the causes and contributors of the failure.

Champlain Towers South was a 136-unit condominium building in Surfside, Florida, a town in Miami-Dade County. The construction style of CTS is referred to as a reinforced concrete, flat plate structure. That means horizontal concrete slabs, for example, at each floor, sit directly on columns.

CTS had 12 stories plus a penthouse with one parking story below ground. The tower had west, middle and east parts, as shown here. To the south of the middle and east sections of the tower was a pool deck at street level with a hot tub and pool above the basement level. The middle and east parts of the tower join the pool deck at grid line 9.1 shown here.

After nearly five years of extensive interviews, combing through physical and social science evidence, testing evidence taken directly from the collapse site, testing full-scale replicas of parts of the building, and extensive computational and collapse modeling of the building, our team has concluded that the most probable scenario for the onset of the partial collapse of Champlain Towers South on June 24, 2021, is the punching shear failure of slab-column connections in the pool deck shown at 1.

I will now describe this type of failure with an animation. The first image shows a column supporting the slab. As the slab is loaded and the column resists the load, the weight of the slab and other things on it push the slab down. The column resists that load by pushing up. The slab bends down around the column, opening diagonal cracks. As the load increases, one or more of the crack surfaces widens to the point where it can no longer resist the load, and the joint fails; it loses most of its strength.

If other slab-column connections adjacent to the failed connections have not failed, the adjacent columns may keep the slab from falling, as shown here. This was the case in the first connection failures I've described so far. But if there isn't support for the slab through adjacent connections, the slab will drop. It is as though the column punches through the slab. The slab slides down the column, and as the slab slides down the column, the top layer of the slab reinforcement pulls down on the failure cone over the column, breaking it, and the concrete breaks apart as the slab continues to drop until finally the slab drops to its final resting place. As the slab drops, the top reinforcement breaks the failure cone over the column, leaving the column’s top and the hooked-shape reinforcement as shown here.

When structural elements and connections are designed and built according to applicable standards, they are expected to have a large margin against failure, meaning that the expected failure load and deformation is well above the loads and deformations that are expected to occur in service. But in the pool deck slab-column connections of CTS, those margins against failure were zero at the time of failure.

After these two-slab column connections failed in early June and by mid-June, the slab around these connections was able to remain standing for some time through a combination of bending resistance in the slab and flat arch action, technically called compressive membrane action. This behavior prevented immediate collapse. However, when the first two slab-column connections failed, the loads on neighboring slab-column connections increased, causing them also to fail on June 24, 2021. The failure spread through a large portion of the pool deck and adjacent street-level parking deck.

I will now present a likely scenario of how failure could have initiated and progressed at CTS. The animation opens with this image of the collapse site, overlaid with an outline of the pre-collapsed building shown in black. We highlight in red a slice of the pool deck and garage that we'll use as an example to demonstrate a scenario of the initiation and progression of failure.

This scenario is consistent with principles of structural mechanics, eyewitness accounts, and other evidence, but it is not the only such scenario. There are other progression scenarios that are consistent with structural mechanics and the evidence.

We then zoom into the elevation view of the slice, showing an initial punching shear failure at grid line 13.1 and the sagging of the slab that follows. The failure spreads to neighboring slab-column connections shown here occurring at grid line 15. In this scenario, the sagging grows, eventually unseating the southern edge of the pool deck slab from the wall, and then causing punching shear failures at slab-column connections at grid line 11.1, which are strengthened by beams.

The failure continues to spread north until it reaches the southern face of the tower, where sections of the concrete are pulled from the connections between the pool deck slab and the column towers. In this scenario, the slab ultimately rests completely on crushed cars and the garage floor below it. When the pool deck slab and beam fully disconnect at its northern edge, where it meets the middle part of the tower shown here in green, it severely damages two critical structural connections, where the pool deck's slabs and beams join the columns and slabs of the tower. The damage weakens these joints, causing the failure to progress into the tower, shown with the two yellow arrows, and then spreading through the middle part of the tower in the area marked 2.

The downward movement of the collapsing floor slabs of the middle section of the tower pulls the east section toward the middle, causing the east section to sway back and forth. This swaying is initially restrained by a portion of the middle section of the tower. However, continued collapsing of the middle section allows the east section of the tower to sway westward until the capacity of the lower-story columns is exceeded, and this section of the tower begins to fall downward.

Our team's meticulous studies of the physical evidence we collected, the observations shared by eyewitnesses and those familiar with the building, and the testing and the results of computational structural analysis has led us to determine the following most likely failure scenario for Champlain Towers South. First, the building suffered punching shear failures of slab-column connections of the pool deck approximately three weeks, and then again, one week before the tower collapsed. The initial slab-column connection failures eventually cause additional slab-column connections to fail, causing a progression of collapse of parts of the pool deck and the street-level parking, eventually unseating the pool deck slab at its southern edge. And second, this failure damaged two critical joints where pool deck slabs and beams joined columns and slabs of the tower at the northern edge of the pool deck. The damaged connections crushed due to the loads on the columns, causing two tower columns to drop and the failure to spread throughout the middle part of the tower. And finally, those failures progressed through the east part of the tower.

The next few chapters of our presentation will show how our team determined this most likely failure scenario and why Champlain Towers South tragically collapsed on June 24, 2021.

We'll describe the initiation and progression of failure of Champlain Towers South in three distinct chapters shown here. Chapter 2 is focused on the initiation of failure of slab-column connections in the pool deck and the progression of this failure through a large part of the pool deck slab and the street-level parking slab. Chapters 3 and 4 are focused on progression of the failure into and through the middle and east parts of the tower and arresting at the west part of the tower.

Chapter 2 will now address the failure initiation. Our analysis shows that slab-column connections at the six locations in the pool deck shown here in yellow were candidates for initiation of the failure. The structural analysis together with photographic evidence and key observations made by eyewitnesses indicates that the connections at K-13.1 and L-13.1 outlined in red are the slab-column connections that likely failed first.

The next few slides will show a timeline of structural distress starting approximately three weeks before the collapse. In early June 2021, just three weeks before the failure of Champlain Towers South, eyewitnesses saw and captured in photos severe distress in a masonry planter wall and a planter box sitting on the pool deck. The location of this distress is shown on the left with a blue dot and arrow.

The photo in the upper right shows distress three weeks before the building collapse. This includes a long horizontal crack in the planter wall identified here with a blue arrow. An enlargement below shows severe distress and displacement in the corner where the planter wall meets the planter box. Another photo of the same planter taken 14 months before this photo shows no distress. Therefore, significant movement occurred during this time period.

One year ago, we shared that the slab-column connection at the intersection of the two dash lines in the upper right-hand image or at grid point K-13.1 has one of the lowest calculated margins against failure in our analyses to that date. An eyewitness described to our team that approximately three weeks before collapse, the gate between the pool deck and street-level parking had become jammed and unusable. The location of this distress is shown with a blue dot and arrow in the left-hand image. The photo in the upper right shows the gate between the pool deck and street-level parking taken several years before the collapse. It shows no damage to the gate.

The image in the lower right is an artistic rendering of what the eyewitness described to us. They reported that the northern portion of the gate shifted down vertically relative to the southern portion by less than half an inch. This view and subsequent similar views show parts of the pool deck structure, its slab-column connections, and the basement columns. At A, we see the slab-column connection at grid point K-13.1 near where that planter and gate were damaged.

Our analysis of the building's design and its condition at the time of collapse, our testing and computational modeling of the structure, and the damage observed by eyewitnesses and documented in photographs indicates a punching shear failure at point A. The video clip in the lower left side is the same video shown earlier and demonstrates a likely scenario where the slab sags after the slab-column connection failure.

Approximately one week before the collapse, an eyewitness reported a recurrence of the problem with the gate between the pool deck and street-level parking, which had again become jammed and unusable. The location of this distress is shown with the blue dot and arrow in the left-hand image. The eyewitness reported that this time, the gate suffered a more significant vertical displacement downward. The northern portion of the gate displaced vertically down by approximately 1 inch relative to the southern portion of the gate. This observation is significant because this gate is adjacent to the planter box that in the same period of time showed severe distress and displacement prior to the failure on June 24, 2021.

Approximately one week before the tower collapsed, an eyewitness described the persistent water leaks in the ceiling of the garage drying up, and a sudden concentration of water began to flow down on all four faces of the column located due east of the planter that suffered damage just two weeks before. The location of this column at grid point L-13.1 is shown with a blue dot and arrow in the left-hand image. The eyewitness shared that the flow of water increased daily and that there was a change in the height of the ceiling of the parking garage or the pool deck slab around the column. The eyewitness shared that the leaking was as if the water from the pool deck above was being funneled down this one column. It is apparent from the discoloration and the paint streaks in the photo of the column taken in November of 2020 that this column had a history of water damage.

Based on our analysis of the building's design and of its condition at the time of collapse, our computational modeling of the structure, and eyewitness accounts, this damage is, again, indicative of punching shear failure at grid point L-13.1. The video clip in the lower left demonstrates the slab deforming following the punching shear failure. After the two slab-column connections punched three weeks and one week before the tower collapsed, the slab near the punched slab-column connections did not immediately fall into the basement. But instead, they moved downward, likely 1 or 2 inches at the two points labeled A. The slab area shaded in gray was able to transfer the load borne by the punched connections labeled A to adjacent columns, shown here in orange and circled in the figure. This prevented an immediate collapse of the pool deck slab. We’re certain that all slab-column connections around these two would have been much more vulnerable, and that any of them could have been the next to fail.

Less than 24 hours prior to the collapse, one eyewitness described the planter shown here detached from the pool deck, creating a gap in between the planter and the concrete deck. The location of this distress is shown with the blue dot and shaded oval in the left-hand image. This observation was made in the early morning of June 23rd, 2021, and persisted the entire day with the same extent of separation between the planter and the pool deck surface. The last observation was 11 hours before the tower collapsed with no subsequently reported changes to the gap between the planter box and the pool deck by the eyewitness. The lower image is a drawing, not to scale, from the interview with the eyewitness, where the approximate location of the planter’s separation from the pool deck slab is shown at the pink line and where the handwritten notes summarize the observation.

Approximately nine hours before the tower collapsed, one eyewitness described a leak coming through a crack in the garage ceiling that got progressively worse. The eyewitness described the water leak had started as a drip and that the bottom of the slab of the east side of the crack was lower than the west side. The location of this distress is shown with a blue dot and shaded ovals in two upper images. The image on the upper right depicts the garage ceiling cracks annotated in turquoise; construction joints in the slab annotated with pink lines; the key pool deck and street-level parking deck observations annotated with blue dots; and the location of this garage ceiling leak highlighted with an oval. The area where this leak formed is in a part of the garage that historically leaked and was repaired over the years.

The lower image is a drawing, not to scale, from the interview with the eyewitness, where the approximate location of the crack and leak are shown with a red line and black circle. The eyewitness account is captured with a yellow sticky note. The eyewitness described the water drip turning into a water faucet just a few hours before the collapse of the building.

In summary, eyewitnesses observed damage to the planter box and the gate near grid point K-13.1 approximately three weeks before the collapse. Two weeks later, more severe damage is again observed in the gate separating the street-level parking and the pool deck, and water is observed being funneled down the column at grid point L-13.1. Then 17 hours before and again 11 hours before the collapse, one eyewitness notices a separation between the planter and the pool deck. Then nine hours before the collapse, a water leak is observed in the garage. This leak worsened and was described as a faucet pouring water into the garage just three hours before the tower collapsed.

At approximately nine minutes before the tower collapsed, the last vehicle entered the Champlain Towers South garage. Several other vehicles entered and were parked in the garage earlier that evening. An example of the area visible to eyewitnesses in the garage is shown in this video, with columns where the first punching shear failures occurred along grid line 13.1 shown in yellow; those where failures next occurred along grid line 15 shown here in orange; and those where failures likely occurred last along grid line 11.1 shown here in red. The columns at the southernmost face of the tower where the pool deck met the building along grid line 9.1 are shown in green.

Any fallen slabs in the garage would have been observed by any of the eyewitnesses who walked in this part of the garage. However, none of the eyewitnesses in the garage and other parts of the building observed fallen pool deck or street-level parking deck slabs nine minutes before the tower collapsed. Instead, several eyewitnesses reported hearing concerning sounds approximately eight to nine minutes prior to the tower collapsing. These sounds heard from various locations in the building are described as crackling sounds or a sound associated with building renovations.

Approximately seven minutes before the tower collapsed, the building's own fire alarm logged a trouble signal with no further description or details in the log. At approximately six minutes prior to the tower collapsing, several eyewitnesses, some who were already awake and some awakened by the activity in the pool deck area, saw the street-level parking collapsing or fully collapsed. One eyewitness described seeing the cars at the street-level parking deck in a strange position, almost as if they were tilted down. Prior to observing the cars, one eyewitness heard a metallic screeching sound possibly associated with the cars shifting. Another individual observed the street level parking collapsed into the garage. Several individuals recalled hearing car alarms going off around this time.

The aggregated areas visible to all individuals who reported witnessing the collapsed street-level parking is shown in green. It is possible that the slab was progressively collapsing east of the street-level parking deck, but no eyewitnesses had lines of sight that far east during this time. At around this time, the first 911 call is made from inside the building at 1:16 and 27 seconds a.m., alerting authorities to what the caller perceived as an earthquake at their building.

Based on our analysis of the building's design and condition at the time of collapse, and our tests and computational modeling of the structure, the damage observed in the garage and in other areas of the building suggest that the slab-column connections punched first at A, likely followed by the connections at B, and then followed by punches at C, and then the unseating of the slab along the southern wall at C.

There is no way to know precisely which path the failures took from the initial two punching shear failures. But the video clip in the lower left is consistent with structural mechanics, eyewitness accounts, and other evidence. Small variations in the sequence could also be consistent with structural mechanics and evidence.

At approximately five minutes prior to the tower collapsing, several eyewitnesses begin to sense that something was terribly wrong in their building. One eyewitness described seeing the pool deck collapsing in the north-northeast direction. Other eyewitnesses observe the progression of collapse in the pool deck progressing from south to north, one bay at a time, as if dominoes were falling in a sequential chain reaction. Our line-of-sight analysis shows that a portion of the pool deck is obscured for some of these individuals. The visible area for these eyewitnesses is shown in green. The area not shaded, meaning not visible by these eyewitnesses, explains how they observed a south to north-northeast progression while others observed a south to north progression.

At the same time, a few eyewitnesses felt a sudden wind in the lobby of CTS and nearby. Several eyewitnesses heard sounds they compared to a jet engine or described as a loud whoosh. Shortly after the eyewitnesses observe these events, a second 911 call is made from CTS at 1:17 and 49 seconds a.m. saying that an earthquake has occurred and that everything exploded down into the garage. The fire alarm central station made a phone call to CTS a few seconds later at 1:17 and 55 seconds a.m.

Based on our analyses, the damage observed in the pool deck by eyewitnesses, a scenario for progression of collapse is depicted in this 3D rendering of the building and in the video below. In this scenario, the slab-column connection punches first at A, likely followed by the connections at B, followed by punches at C and the unseating of the slab along the southern wall at C, and then followed by the punches at D.

At approximately five and four minutes prior to the building collapsing, several eyewitnesses with limited line of sight, shown here in green, observed the pool deck collapse just south of the southern edge of the tower. Based on our structural tests and analyses and the damage observed in the pool deck by eyewitnesses, a scenario for progression of collapse is depicted in this 3D rendering of the building and the video below it. The video shows the slab and beams tugging at the joints of the slabs, beams and columns at E, and eventually the slab and beams breaking off these joints and pulling some of the joint concrete away with it. The video also shows how this scenario is consistent with multiple eyewitness accounts, indicating the pool deck collapse generally proceeded from south to north.

There is no way to know precisely which path these failures took from those initial two punching shear failures. But the scenario we have described here is consistent with the principles of structural mechanics, eyewitness accounts, and other evidence. Small variations in the sequence could also be consistent with structural mechanics and evidence. However, all variations involve the spread of the failure from the two initial slab-column locations to other slab-column locations, unseating of the slab from the south wall, and the northward collapse advancement of the pool deck and street-level parking slabs, which then initiated the collapse of the middle part of the tower.

At this time, several building residents hurry through the lobby, briefly exchanging their experiences with one another and trying to understand what is happening in the building. A few individuals walk out of the lobby and towards the street-level parking area to get a closer look at the partially collapsed parking deck and the sunken cars. At this time, several eyewitnesses also report hearing car alarms. One eyewitness standing north of Champlain Towers South captured the partial collapse of the pool deck in a video. A single frame of this video at 1:18 and 18 seconds a.m., approximately four minutes before the tower collapse, is shown here. Our team used what is known as super resolution analysis to clarify and sharpen objects. Contrast and brightness enhancements were also made. Debris on the parking garage floor is observed in this enhanced video and demonstrated in this image. Flashing lights from a car alarm triggered by the impact can also be seen reflected on the eastern wall of the garage ramp in this video.

We've just described the how of the collapse of the pool deck and street-level parking deck. What about the why? What caused slab-column connections in the pool deck and street level parking to fail? The causes are multiple, and the bulk of them happened at the time of original design and construction. As we reported in previous updates, NIST contracted with an experienced structural design firm to prepare an independent check of the original structural design of CTS for conformance with both codes and standards that applied at the time of original design and construction and against codes and standards that would apply today. That design check found pervasive instances where the flexural or bending design strength and the strength of slab-to-column connections in the pool deck and part of the street-level parking slab failed to meet both the original and current codes and standards.

This graphic summarizes the check against original codes and standards. The yellow and red icons indicate areas of moderate and severe understrength respectively. Please note that there are very small differences between this final version and the preliminary version we showed in our June 2025 update. But the overall conclusion is the same. In some locations, the design provided less than half of the code-required strength. We found that the top layers of reinforcement in the pool deck and street-level parking slabs were not placed at the elevation specified on the structural engineering drawings. Our measurements of the pool deck specimens removed from the collapse site showed that the cover was generally about 2 inches rather than 3/4 of an inch shown on the drawings. This deviation, while seemingly minor, significantly diminishes the strength of the slabs and the slab-column connections.

As illustrated on the left, the design drawings required that at least 25% of the top of slab reinforcing for the area near columns, known as column strips, was to pass over the column. We found that typically less than 25% of the reinforcement passed over the columns. An example is shown on the left. The highlighted red rectangles in the image from the design drawings show that there were to be 16 bars in the column strips in each direction. Therefore, four bars were to pass over the column in each direction. The bottom left photo shows an example where only two bars in each direction passed over the columns, half the number required.

On the right, we found that the spacing of the top reinforcing bars in the areas of the pool deck slab near columns range from about 20% to 40% wider than required by the structural design drawings, resulting in less reinforcing in the area of the slab near columns than required by design. Both these deviations significantly diminish the strength of the pool deck slab and slab-to-column connections. Records of whether these changes were approved by the design engineer or observed by inspectors in the field are not available.

The final construction of landscaping features on the exterior slabs of CTS varied from the design drawings. One of these variations is a large boxlike planters on the pool deck shown with the red dash rectangle in the aerial view on the right. The drawings on the left showed no such planters. Those planters held palm trees for much of the life of the building, although they were removed following a hurricane in 2017.

Cracking, leakage and evidence of reinforcement corrosion caused the condominium association to undertake a major rehabilitation of the pool deck 15 years after the building was constructed. The rehab involved removing most of the original quarry tile, although some quarry tile was left in place as shown in the core sample schematically illustrated here. A membrane was placed over the original sloping concrete topping. The new pavers on sand setting bed were added. The new pavers and sand added load to a system that was already structurally inadequate.

Glenn Bell: We found that some of the reinforcing steel in the pool deck and street-level parking slabs was corroded. The degree of that corrosion was quite variable. These are images of reinforcement from pool deck specimens retrieved from the building collapse site. The top image is an example of little to no corrosion. The middle image shows that corrosion sometimes varied significantly along the length of the bar. And the image on the bottom shows a high level of corrosion. Reinforcement with moderate to high corrosion tended to be more common in areas with a higher corrosion-inducing environment, such as under planters or locations of high exposure to chloride-laden water.

Evaluating the actual loads and deformations under which the slab-column connections would have been expected to fail has been extremely challenging. This is an exercise in establishing the loads and other conditions that existed at the time of failure. We call this the demand at the time of failure. And it's an exercise in establishing the expected failure strength and deformation of the connections at the time of failure. We call this the capacity. When the demand exceeds the capacity, failure is expected.

We described in earlier updates the meticulous work of our team in evidence collection, testing and analysis that has informed this determination of demand and capacity, such as examination and measurements made of the hundreds of physical specimens retrieved from the collapse pile, material testing, geotechnical tests and analysis, and study of the 40-year history of CTS. All of this informs the analysis described in the next several slides.

I will describe how we evaluated the demands on and capacity of the slab-column connections with three tools. One, finite element modeling; two, laboratory tests of slab-column connections; and three, critical shear crack theory, which will be explained later. We used high-fidelity finite element modeling to determine the loads on the slab-column connections and the deformations of the slab at each of the connections. These models capture geometric and material nonlinearity and include effects of shrinkage, creep, corrosion of reinforcement, and in-plane stresses from the restraint of dilation of the slab.

As part of the investigation’s structural laboratory test program, NIST contracted with a U.S. university with special expertise in load tests of concrete structures to build and load-test to failure eight full-scale replicas of typical CTS pool deck slab-column connections. The specimens included replicas of a portion of a column supporting the pool deck and a section of the slab 10 foot, 6 inches square. The specimens were built as faithfully as possible to replicate the conditions existing in CTS at the time of failure. This included reproducing concrete mixture ingredients and proportions with aggregates used to make the concrete from the same source in the Miami area believed to have supplied aggregates for the original construction of CTS.

The specimens were loaded by eight hydraulic jacks arranged around the slab perimeter, simulating the loads on the connection. The left-hand image shows the start of a test. The right-hand image taken from the end of a test shows the extent of the failure cone at the top surface of the slab. To study the effects of long-term corrosion of the top mats of reinforcement in the slab, we intentionally corroded the top reinforcement of some of the specimens to a target degree. The left-hand image shows a saltwater bath and electrodes at a patch of slab over the column of the test specimen used to induce corrosion in the reinforcement. The right-hand image shows corrosion of the reinforcement at the end of a test.

We used critical shear crack theory or CSCT, the basis for the punching shear provisions in European codes, to determine where demands on the slab-column connections may have exceeded their capacities. In U.S. codes, CSCT is also used to determine the minimum area of reinforcement over slab-column connections. In CSCT, a diagonal shear crack that develops as shown in the upper right reaches a critical width at which the concrete can no longer bear the shearing forces from the applied load, and the connection fails. The width of the critical shear crack depends on the depth of the slab and the rotation of the slab in the vicinity of the column. The demand on the connection is represented by the blue load rotation curve shown on the graph. This is determined from our finite element analysis. The capacity is represented by the failure criterion shown in red, which is determined by the CSCT, our laboratory test results, and the degree of any corrosion in the steel reinforcement. If the calculated blue load rotation demand curve intersects the red failure capacity criterion, considering uncertainty in both, then failure is predicted.

Our analysis shows that multiple compounding causes and contributors led to failures in slab-column connections in the pool deck and street-level parking slab. The most significant of these factors stemmed from the time of original design and construction and therefore were baked in before the building was even occupied. These were design understrength caused by severe and widespread deviations in the building's original structural design from the codes and standards of the day, but also some limitations in those codes and standards and deviations in the building’s as-built construction from the design drawings, specifications and industry standards.

Other factors that reduced the margins of safety to the point of failure include planters placed on the pool deck that were heavier and more extensive than shown on the design drawings; fill and paving added to the pool deck during rehabilitation efforts; and degradation over time, the most significant factor for which was likely corrosion of the reinforcement, exacerbated by porous concrete, concrete cracks that leaked, and ineffective waterproofing.

But it's important to recognize that the first two of these, design understrength and deviations in the as-built construction, caused the bulk of the critically low margins against failure from the time construction was complete. We'll now move to Chapter 3, progression of the failure into the tower. Chapters 3 and 4 contain images, video footage and other content that some may find disturbing. If you desire to skip this content, please proceed to Chapter 5.

So how and why did the collapse spread into the tower? This is a very important question for the investigation because if the pool deck and street-level parking slab collapse had not spread into the tower, the disastrous part of this failure would not have occurred. As we've described, the pool deck and street-level parking collapse advanced to the north until it reached the boundary with the building tower at gridline 9.1. In that collapse, sections of slab and two embedded beams, known as beams A at grid lines K and L and highlighted in yellow, separated from the tower. Here's another view of the beams A and how they joined the basement and first-story columns of the tower at gridlines K and L. This is a view from near the end of the pool deck collapse sequence that we showed. It illustrates the conditions at the two joints K and L.

We see a collapsing beam A and slab between grid lines 11.1 and 9.1. At grid line 9.1, the beam A and slab section separates from the slab-beam-column joint, pulling the beam and slab reinforcing and sections of concrete out of the joint and damaging it. This substantially weakened the joint at grid line 9.1, which was bearing the load of 13 stories of structure from the column above. The two slab-beam-column joints had two vulnerabilities that made them susceptible to failure when the pool deck slab beams broke away from the joints at grid points K-9.1 and L-9.1 in the pool deck collapse. The concrete of the slabs and beams of the floor system, shaded in gray, had a lower design strength than the columns above and below it. This made the floor concrete susceptible to crushing. There were no column ties in most of the floor concrete region of the joint to balance the pulling forces of the collapsing pool deck and prevent buckling of the longitudinal column reinforcement.

The specimens we built for the laboratory testing of these joints included both of these vulnerabilities. Through our university laboratory test program, we built and tested two full-scale replicas of these slab-beam column joints. On the right is a photo of the test specimen in the lab. The middle image is a schematic drawing of the test setup. We first applied a vertical load of 650,000 pounds to the laboratory test replica specimens shown at the red arrow in the middle image, representing the load in the column at the time of failure. We then broke away this section of the specimen representing the pool deck slab and beam by applying loads to the yellow arrows 1 and 2, simulating the pool deck collapse.

On the left, the beam A with the pool deck slab above breaks away from the slab-beam-column joint, damaging it. The right shows a view from the top side of the laboratory test specimen showing damage to the joint. After the beam and slab broke away, the slab-beam-column joint lost strength, and the column started to drop. Here is the laboratory specimen after the test. The steel reinforcement in the column had buckled. For safety reasons, the test was stopped after the column dropped a small amount but had lost load-carrying capacity. This illustrates the joint before the pool deck collapsed. As the slab beam of the pool deck breaks away, the joint is damaged. We think the joint initially held the weight of the tower, but after several minutes, crack growth further damaged the joint, leading to the final failure with buckling of the column longitudinal bars.

As the slab-beam-column joints crushed, the columns on grid lines K and L began to drop. Footage from a security camera looking onto the south face of CTS, as described in detail in our previous updates, provides important evidence about the progression of the failure into the tower. This frame of the video footage was taken about one and a half hours before the collapse. Note, please, the column grid lines I, K, L and M as references in the following sequence. This is the first frame of the video that shows the collapse. The roofline near K and L has dropped about 100 inches or approximately one story height. We can see that the 12th floor line at K and L has dropped about the same amount with K perhaps a bit ahead of L. The 12th floor line at I has dropped less than K and L, and the 12th floor line at M has not yet started to move.

Now back to that initial video footage frame taken one and a half hours before the collapse. We see a unique shape of the building's facade between grid lines K and L. Notice how its width varies over the height. This unique shape was created by the introduction of hurricane shutters to some but not all windows. Note that the landscaping blocks the view of stories one and two. Going back to that first frame of the tower collapse, we see again that the penthouse roof parapet has dropped about 100 inches. The unique facade shape has dropped vertically the same amount, and importantly, it has dropped uniformly, meaning its shape has not changed. As the progression proceeds, the unique facade shape drops but still does not change shape. This tells us that the column failures at K and L were low in the building below the third floor.

To summarize the progression of the collapse into the middle part of the tower: One, the pool deck collapsed up to the edge of the south face of the tower several minutes before the tower collapsed, damaging the first-level slab-beam-column joints at grid points K-9.1 and L-9.1. Two, the first-level slab-beam-column joints had inherent vulnerabilities stemming from the time of original design and construction. Three, the damage caused by the pool deck collapse, exacerbated by the inherent vulnerabilities of the first-level slab-beam-column joints, caused the joints to crush several minutes after the pool deck collapse reached the tower. Four, then the columns themselves at grid points K-9.1 and/or L-9.1 started dropping.

Chapter 4 will now address the spread of the failure through the tower, addressing first how the collapse spread through the middle part of the tower and then the east part of the tower. In previous updates, we described our analysis of footage from a security camera inside one of the units in the 11 stack of Champlain Towers South. As shown here, the 11 stack units are a vertical stack of residential units between grid lines L and M, whose unit numbers ended in 11. Please note that the dropping column on grid line L, indicated by the yellow highlighted red arrow is at the west or left-hand side of the view of the security camera inside. Our team employed computer vision techniques to track and measure movements in this unit. The top right image is a view inside that unit. It shows the progression of movements, starting with a reference frame taken during the day before the collapse. The colored lines show the movement from that reference frame to a time interval between 1:22 a.m. and 4 seconds to 1:22 a.m. and 15 seconds; about two seconds before the precipitous drop of the tower. We see the left-hand side of the room moving down relative to the right. Note in the lower left-hand image, the locations of the columns at grid points L-9.1, L-8, M-9.1 and M-8. The plot on the lower right shows the vertical drop of the slab near the columns L-9.1 and L-8 relative to the slab locations near the columns on grid line M, which appear to have been stationary during this footage. The drop shown is from 1:22:04 a.m. to 1:22:15 a.m. and represents the total drop from a video frame taken the day before the collapse. Note that some of the drop already occurred between the day frame and the start of the measured drop shown in this graph. The slab near column L-9.1 dropped more than the slab at L-8, and the rate of drop increased dramatically at about 1:22:14. The total drop near grid point L-9.1 from the day before the collapse to 1:22:14 a.m. was about 18-27 inches.

This is from video footage from another source, a security camera in an upper-story corridor of CTS. We've shown the analysis of this footage in previous updates. The camera is looking from west to east towards grid lines I, K and L. This footage overlaps in time with the 11-stack footage I just showed. On this frame, from 1:21:55 a.m., we've placed a white reference line at the intersection of the floor and the south wall of the corridor. Here, we see the floor beginning to sag around grid lines K and L. This frame is at 1:22:15, which overlaps with the movement we see in the 11-stack unit below. And here's that sagging once again.

This shows the relative locations in a plan or bird's eye view of the movements we can see in the 11 stack footage shown in the blue circles and the upper-story corridor footage seen in green circles. The locations marked X are not moving. While the corridor footage is taken at a different floor than the 11 stack footage, the movements are in similar floor areas when viewed from above. They tell consistent stories about where in the building movements were occurring.

How did the failure progress horizontally? It's not clear whether the columns at K-9.1 and L-9.1 started to drop simultaneously, or whether K dropped before L, or whether L dropped before K. Regardless, the collapse progressed from the south to the north in the middle part of the tower by progressive punching shear failures at the columns. Further evidence of the progression of the collapse in the middle part of the tower is shown here. In the lower left-hand image, we see the top of the penthouse roof exposed as the structure collapses from south to north. In the lower middle image, we see the tops of columns at grid points K-4 and L-4 poking above the top of the roof as the roof structure moved down around them. And in the lower right-hand image, we see a portion of the main roof to the extreme north of the middle part of the tower is exposed as the penthouse roof collapses. This is all consistent with the collapse progressing from the south to the north.

The collapse of the middle part of the tower arrested where the middle part met the west part. Why did it stop there? The dotted line at the left-hand side of the blue shaded middle part of the tower shows Zones A and B, which denote two different conditions where the collapse arrested. Zone A is an area where the floor and roof slabs connected to a concrete wall. You can see the concrete wall in the photo on the right. Zone B is a region where the slab and column system of the middle part of the tower extended into the west part of the tower. There's no concrete wall in Zone B. The slab collapse arrested at Zone A for the simple reason that the strength of the concrete wall prevented the failure from spreading beyond it. The photograph shows that the floor slabs tore away from the concrete wall as the concrete fractured, the reinforcement fractured, and/or the reinforcement connecting the slabs to the concrete wall tore out of the slabs.

The condition at Zone B is more complicated. The yellow arrows highlight stripes in the bottoms of the slabs. These are areas where the steel reinforcement bars tore out of the bottom of the slabs. The next few slides show how this happened. Our video and structural analysis show that in the region of the middle part of the tower at Zone B, indicated by the gold arrows, the collapse proceeded to the columns on grid lines I and H before stopping short of E. Following the failure at grid line H, the collapse proceeded towards grid line E. Whether the collapse would continue into the west part depended on whether the slab would fail at 1 before a punching shear failure occurred at 2, or whether a punching shear failure would occur at 2 before any slab failure at 1. Our structural analysis shows that failure would occur at 1 before 2, and so the failure did not advance into the west part of the tower.

This slide shows the state of the columns and floor slabs at the boundary between the west and middle parts of the tower before the collapse. The slab first failed in the section of the slab that has no top reinforcement near the end of the reinforcement and where it was weak and bending. Following this failure, a strong downward force was exerted on the bottom slab reinforcement on the left-hand side of the break. And that downward force caused the bottom reinforcement to tear out of the bottom of the slab. The slab to the right of the break continued to drop. On continued dropping, the reinforcement completely separated from the bottom of the slab near the column at E. The stripes visible in the bottom of the slabs are where the bottom steel had torn out.

To summarize the collapse of the middle part of the tower: One, as columns at grid points L-9.1 and K-9.1 started to drop from damage to the first-floor joints caused by the pool deck collapse, some of their loads were transferred to other parts of the structure. Two, this initiated punching shear failures at the slab-column connections as the collapse progressed to the north. Three, the collapse arrested where the middle part of the tower joined the west part of the tower because the resistance of the slab support at the concrete wall and the punching shear strength at the columns at grid line E were stronger than the resistance of the floor slabs themselves. Four, the middle part of the tower was vulnerable to progressive collapse.

So how and why did the collapse then spread from the middle to the east part of the tower? Here's a view of the middle and east parts of the tower from the south face video footage. The perspective of this view and those immediately following have been altered so that the south facade of the building is approximately square, level and proportional, allowing us to more easily visualize movements. This frame was captured before the precipitous drop of the tower. I will step through many of the footage frames to summarize what the footage shows us about how and why the east part of the tower collapsed.

This is the first frame of the precipitous drop of the tower. We can see that parts of the structure west or to the left of gridline M have dropped about one story height. The structure to the right or east of gridline M does not appear to have moved down at this point. And we see no evidence of slab-column-connection failures at gridline M at this time. The east part of the tower is moving west or left in the image. This is likely caused by the westward pull of the collapsing middle part of the tower. Though not annotated in this frame, at this time, the structure at gridline P near the roof also moves south as it moves west.

Now, the part of the structure between gridlines N and O has started to drop, but the westward movement of the east part of the tower has paused. Here we see that several floors of the upper part of the tower are hinging or folding down. The east-west movement of the east part of the tower is still paused. Here, the collapse of the middle part of the tower has advanced, and the westward pull of the tower has diminished. The initial westward pull and release on the structure has set up an oscillation, and the east part of the tower is now moving east. As the collapse of the middle part of the tower progresses to the north or towards the back in this image, its stabilizing effect on the east part diminishes. The oscillation of the east part of the tower reverses, and it again moves west. But because of the diminished influence of the middle part of the tower, it does not also move south, and it continues to move west.

At this point in time, most of the middle part of the tower has collapsed. The approximate extent of the remaining east part of the structure at the roof and 12th floors is shown by the magenta outlines at the top two stories. The now collapsed middle part no longer provides resistance to the western movement of the east part. With no resistance to further side sway from the middle part of the tower, the east part continues to sway to the west. At this point, the 12th floor has moved west about 21 inches. The lateral sway, known as drift, exceeds the capacities of the lower-story columns, and the columns start to fail. The column failures are evidenced by the top parapet moving down, where shown by the arrow at the upper roof line. This is the beginning of the final fall of the east part of the tower. The structure drops and falls to the ground.

Post-collapse, we found several columns, most likely from the east part of the tower, that evidenced punching shear failures in that part of the structure. The view on the left illustrates that when punching shear failures at slab-column connections occur over multiple stories of a single column, the slabs drop around the column, leaving in the collapse pile, shown on the right, column sections that are multiple stories high. The yellow patches in the illustration at the left, and in the photo with multistory column sections on the right, show the punching shear locations where there were floor slabs. The post-collapse appearance of multistory columns is evidence of punching shear failures in the east part of the tower.

In summary of the collapse of the east part of the tower, one, the collapsing middle part of the tower created and then released a westward pull on the east part of the tower, resulting in oscillation of the east part and a large side-sway known as drift to the west. This large drift caused columns in the lower stories to fail. Two, multistory columns, most likely from the east part of the tower, in the post collapse pile are evidence of punching shear failures in some of the slab-column connections. Three, the east part of the tower was vulnerable to progressive collapse.

In Chapter 5, we'll discuss some additional points of inquiry of our investigation, including the questions of why the collapse occurred when it did, and what factors did not contribute significantly to the collapse of Champlain Towers South. An important question of this investigation is why the structure collapsed on June 24th, 2021, 40 years after construction was complete. While problems stemming from the time of design and construction and loads added to the structure created critically low margins against failure in parts of the pool deck, these factors alone do not fully explain a failure in 2021. We found no evidence of any specific discrete initiating event such as sudden overloading or trauma to the structure on or just before June 24th, 2021.

We studied the potential sources of long-term degradation that reduced the already thin margins against failure to zero. We believe the first punching shear failure occurred in early June 2021, when significant distress in the planter walls and planter box was noted. That distress is consistent with failure of the slab-column connection at grid point K-13.1, and we believe that corrosion was the primary degradation mechanism. Expanding on the assessment that corrosion of steel reinforcement of the pool deck slabs is the most likely scenario for long-term degradation of the pool deck structure, we know that corrosion of reinforcing steel was evident more than 25 years before the collapse and had progressed to the point that structural repairs and waterproofing were conducted in 1996 and '97.

Problems persisted after those repairs. There was continued leakage and rust staining. In some locations, the concentration of chloride was high enough to sustain corrosion of the reinforcing steel. We have assessed either by direct measurement or visual grading the degree of corrosion of hundreds of samples of the steel reinforcement from the pool deck and street-level parking slabs of CTS. The degree of corrosion of the reinforcement in those samples was highly variable. While the physical specimens we retrieved from the collapsed pile included few examples of the steel reinforcement at some of the most critical slab-column connections in the pool deck, moderate to high degrees of corrosion were measured and observed in steel reinforcing bars extracted from nearby specimens in the pool deck.

Both our structural laboratory tests and published reports of tests by others show that moderate corrosion can significantly reduce the capacity of connections. It is well known that when concrete is loaded to high levels of stress near its failure level, it may not fail immediately, but may fail after a period of time, even if the loads are unchanged. This is called delayed failure under sustained loads. This phenomenon does not explain the very long-term delay in failures between the time of original design, construction and the subsequent application of loads to the pool deck to the time of the partial collapse. However, once the slab-column connections failed at grid point K-13.1 in early June and grid point L-13.1 in mid-June, the stresses in the system were changed, and sustained load effects may have been a factor in the specific day the collapse occurred, that is, June 24th, 2021. Sustained load effects also could explain the several-minute delay between the end of the pool deck collapse and the collapse of the tower.

Things that most probably did not contribute significantly to the collapse include vibrations from the construction of 87 Park. 87 Park is an 18-story luxury condominium built just south of CTS and completed in 2019. The northern boundary of the soil excavation for the construction of 87 Park was close to the southern boundary of CTS. To support the basement excavation for 87 Park, steel sheet piles were driven into the ground using a vibratory hammer rig approximately 9 feet away from the south basement wall of CTS. The vibrations from this pile driving were felt by people inside CTS, and there was concern these vibrations were damaging CTS. The image on the left shows a plan of the 87 Park excavation to the south of the CTS property with the CTS property to the north. The image on the right is looking west, illustrating from left to right, the sheet pile driving, the south wall of CTS, and the CTS pool deck and tower.

To evaluate the potential structural damage to CTS from vibrations, we investigated the conditions of the soil, rock and foundations and superstructure of CTS, studied the history of the construction of 87 Park, obtained records from accelerometers of ground vibrations near the south wall of CTS during sheet pile driving, conducted nondestructive testing at the building to establish the damping and attenuation properties of the subsurface conditions, and created and ran numerical simulation models for the soil, rock, foundations and superstructure to establish levels of vibration in various parts of the structure. Our analysis shows that the vibrations from sheet pile driving would have been greatly reduced by the south basement wall of CTS and the CTS sheet pile wall and would further attenuate, that is reduce, rapidly before reaching critical parts of the pool deck structure. Vibrations at the critical slab-column connections were too small to have caused structural damage. Our analysis also corroborates, however, reports that people inside the tower would have perceived these vibrations. Humans are very sensitive to vibrations and can sense levels of vibration well below those that would damage even a fragile structure.

Other things that did not contribute significantly to the collapse include foundation failure, sinkholes or differential settlement, hurricanes and storm surge effects, impulsive loads such as vehicle impact, explosion or items dropped from a crane, and accidental loads or overloads caused by the roof repair and roof anchor project ongoing at the time of the collapse.

Judith Mitrani-Reiser: The collapse of Champlain Towers South was one of the most tragic building failures in U.S. history. We have therefore worked diligently to fulfill the charge of our investigation to determine the most likely technical cause or causes of the failure, to recommend specific improvements to building standards, codes and practices based on the findings made, and to recommend any research and other appropriate actions needed to improve the structural safety of buildings.

We have so far determined the most likely technical causes of the failure, and we can now focus on finalizing our recommendations. Building codes require that building structures be designed and built to have large safety margins to prevent failure. A structure's margin against failure is defined as the extra capacity the structure and its elements possess beyond the loads that are expected to be placed on it. Champlain Towers South did not have those large margins against failure. Problems in its pool deck structure stemming from the time of original design and construction meant that parts of the pool deck had very low margins against failure. Champlain Towers South was, therefore, a very vulnerable building from its earliest days.

Loads added to the structure over its 40-year life and lack of durability of the structure and waterproofing system eroded the very low margins of safety to the point of failure by early June 2021. This led to the collapse of the pool deck, and then the middle and east parts of the tower on June 24, 2021. While it was the low margins of safety and degradation that started the failure in the pool deck, it was insufficient strength and reinforcement detailing at the connection between the pool deck and the tower that caused the deadly partial collapse of the tower.

In our final report, in addition to our principal findings on the technical causes of the partial collapse, we will provide all of the analysis of evidence, test results, and computational modeling that supports the team's findings. Our final report will also include recommendations for changes to codes, standards and practices, further research, and other actions to improve building safety based on the findings of our investigation. While we are confident in the technical findings presented today, details may be subject to modification should new evidence come to light before the publication of our final report.

While this summary has been highly technical, what has driven the demanding and meticulous work of our investigative team is the human element, the loss of 98 lives. No one in the U.S. should ever have to go to bed wondering if their building will collapse during the night. We are confident the recommendations from this investigation will move us closer to fulfilling that expectation.

NCST recommendations have widespread impact on codes, standards and practices throughout the United States and the world, particularly for how we ensure the ongoing structural safety of existing buildings. The lessons from Champlain Towers South will be incorporated into engineering guidelines and public policy to help prevent similar failures in the future. We look forward to collaborating with all interested public and private agencies and organizations for improved building structural safety. By doing so, we can honor the lives of the Champlain Towers South victims.