People who live in glass houses shouldn't have earthquakes: How well do modern glass-covered buildings hold up?

There’s been a building boom in downtown San Francisco over the past decade, and it can be summed up in one word (several times): glass, glass, glass.  Lots of new towers going up completely covered with shiny reflective glass.  Cool futuristic gleaming towers...or a major disaster waiting to happen in the next big earthquake?  What will conditions on the street downtown among these towers be during a large quake?  I’ve wondered about this and I’m certainly not the only one, and here’s what I’ve managed to learn.  

Figure 1.  Floor-to-ceiling windows in a modern condo, 
built ~2007.  This photo is from my previous home.

I have to admit, I like glass on skyscrapers; I like the way it reflects the sky and blends in without the stark concrete that makes up most of each building built in earlier days.  I like the sleekness.  I even like it in mid-rise buildings like where I used to live, where I enjoyed the fact that there were floor-to-ceiling glass walls so that the little dog could sit on the floor and look down, watching the world go by, rather than staring at a white wall (Figure 1).  However, in a quake, will the windows break and shower the street below with fragments of glass?  Will the panes of glass separate from the structure and fall intact, creating open gaps in the walls that are hazards indoors for people during the shaking and pets afterward who may not realize that they will fall out if they try to lean against what has always been an invisible wall?  What about the people outside, and particularly in downtown areas where anything that falls will fall a long way? 

The reasons to not run out of a building during an earthquake (like things falling right outside the building) are magnified in a downtown area because anything that separates from a building will fall quite a bit farther and will likely be very destructive.  I’ve long had the impression that there would be no safe place to be outside in a downtown city center during a quake, and I’ve wondered if, as people are running out of those buildings (despite it being unwise), I would be fighting against the current to get INTO the building!  To better understand the potential hazards, I've been looking into what happens to high-rise glass during quakes and into how far away from the building one can expect skyscraper rubble to fall.

It has been VERY difficult to get answers from the building engineering/inspection community about this, and a friend of mine who is a building engineer specializing in seismic safety astutely suggested to me that this is probably because the results are so hard to predict.  I can see that; each situation will be different.  Still, to give you an idea of the frustrations encountered along the way, one contact at a major city Department of Building Inspections actually answered my line of questions with the advice to “avoid standing under falling objects at all cost.”  I kid you not.

Another inquiry to an SF Dept of Building Inspections engineer resulted in the advice that if glass DID fall from a skyscraper, it would fall straight down, so we shouldn’t be on the sidewalk but would probably be relatively safe out on the street.  But think about that; ok, even in the absence of wind, there are other influences that will likely be there: (1) If there’s sideways shaking of the whole building, then there’s a sideways force on any glass that separates from the building for whatever reason, probably enough to send it at least a few more feet away from the building.  (2) Even if the base doesn’t move very much, tall buildings sway even in high wind and their highest floors can move by several feet by wind alone (apparently the tall and slender 432 Park Ave building in New York City sways up to 5 feet at the top in the wind).  I’ve heard anecdotal accounts of people watching the tops of tall buildings move in large quakes in places like Japan (there are various videos on the web).  This could amplify the effects of sideways shaking.  (3) It just takes a slight temporary tilt of the ground on which a tower sits to make the top move to the side.  Bottom line: I would not assume that debris from the top of a skyscraper in a major earthquake will fall straight down and land right next to the base of the building. 

In the hopes of finally cornering a building engineer on this, I went to a San Francisco Dept. of Building Inspections workshop on seismic safety in downtown SF a few years ago, and listened to a presentation about tall building safety.  There was a long line of people who wanted to ask questions after the talk, and one woman asked about exactly what I had been wondering; the topic of this article.  The speaker evaded the question similarly to the responses I had gotten before.  I later went to an info table that they had after the formal presentations were finished, and posed the same questions to the people at the table, who said basically the glass won’t shatter and it won’t pop out, and I simply could not pin them down with a “what if” question.

And I’ve lost count of the number of times that building engineers have told me that downtown San Francisco in a major quake would not become a shower of glass because modern buildings don’t do that, to which I’ve replied that many buildings in downtown San Francisco aren’t modern.  That’s typically met with a shrug (figuratively).

So let me try to put this together based on what I’ve learned and seen and read, and remember, I’m not able to officially advise you on where to be in this situation if the experts can’t!  But I will describe my thoughts and you readers can be the judge.  I’ll be curious to see if any comments come in.

First, glass and building design have both evolved considerably over the last century.  The glass itself, in very old buildings, breaks into shards.  You can see in photos of older glass that you get jagged edges and what are essentially glass daggers; if these fall, you don’t want to be in their way (Figure 2).  If larger panes of glass dislodge, I have heard that they can sail quite far if they are falling from a tall enough building in the wind, but I can’t verify this.

Figure 2. Some older windows break as shards (click images to zoom in).  Counterclockwise from top: A building in downtown San Francisco after the 1989 Loma Prieta quake (KTVU Fox 2). Broken windows in a building in Fukuoka, Japan after the 2005 quake (By Oarih (Own work), CC BY-SA 3.0, via Wikimedia Commons).  Two buildings in Nepal after the 2015 quake, one with no visible window damage and the other with extreme damage and windows broken into shards (photo taken by my friend Laura Whitehurst, used with permission; she points out that we have very different building construction in California!).

In contrast, modern window glass is tempered glass, and frequently laminated.  This is extremely strong glass that fractures without falling apart unless there is an extreme impact, so you get that fracture spiderweb pattern so often seen in car windshields after something hits them, or a familiar pebbly texture.  It’s likely that if an interior piece of furniture fell against a glass window in a modern building, even if the glass cracked, it would still be there but with the fracture pattern.  If the quake was strong enough to truly break the glass, then this glass breaks into pellets rather than sharp jagged shards, as you know if you have seen the aftermath of a smash-and-grab auto burglary (Figure 3).  Still, pellets falling from high up are probably going to be like bullets when they hit the ground.  

Figure 3.  Tempered glass fractures into pellets (click on the image to zoom in).  All photos taken by me, showing shattered tempered glass (probably laminated) in a bus stop ad window and a grocery store display case (I guess people get upset about late buses and expensive dairy products?), a familiar sight of smashed car window glass on the ground, individual pellets, and a smashed window that I photographed (no, I didn’t do it) shortly before finalizing this article.  The lamination membrane of the smashed car window is visible around the top of the hole.  (For readers born in the 21st Century, that round copper thing for size comparison is a penny.  You know, a 1-cent coin that will buy you approximately one pellet of smashed tempered glass.  Um, a coin; you know, cash?  The thing you don’t need a cell phone to—oh never mind…)

There’s an interesting website about laminated and tempered glass that was supposed to also have a video at that site showing how tempered glass breaks, but it’s not there.  Instead, I found this YouTube video of two guys who clearly have the best job in the world, not to mention the best accents in the world, showing the breaking process in slow motion and I highly recommend you take a few minutes out to watch this whole thing because it just keeps getting better and better; best on a full screen of a computer monitor to be effective.  (They say in the video that the glass was laminated, but I confirmed with one of them via e-mail that it was not laminated, and that it was indeed tempered.)

Figure 4.  Breakage patterns of different kinds of
glass, from a  Peerless Products blog.

To my surprise, basic skyscraper glass is not even tempered nor laminated; rather it is heat-strengthened, according to several trade websites including this one from Thermomax.  Figure 4 is a nice graphic comparing how different kinds of glass breaks.  Somehow the breakage pattern of heat-strengthened glass is not comforting to me.  Apparently, heat-strengthened glass is preferred over tempered or laminated glass due to better optical qualities (giving new meaning to the phrase “killer views”?).  However, I did get confirmation from the SF Dept of Building Inspections and online sources like that Thermomax website that floor-to-ceiling windows are required to be tempered and/or laminated glass; in fact, I believe that applies to any glass panel with a bottom edge less than 18 inches from the floor.  So it sounds to me like, ironically, the all-glass towers about which the typical person worries the most are probably tempered/laminated and will remain intact, whereas the older style towers in which windows are interspersed within a concrete façade are likely to have the more dangerous consequences of glass breakage.

However, advances in how windows sit in their frames make modern buildings safer.  The main expected challenges to window glass in a quake would be twisting and torquing of the window frame, rather than impact, unless something heavy flies toward the window and hits it.  If the glass was securely attached to the frame, a little twisting could stress the pane, but glass panes that make up the envelope of modern glass towers are in flexible gasket mounts that allow them to float a bit in the frame, so the building can undergo a bit of flexing and twisting without forcing the glass panes to break. 

It sounds to me that in my old mid-rise condo in SF’s Mission Bay neighborhood with floor-to-ceiling windows (Figure 1), even if furniture flew right toward the window, the glass would be unlikely to shatter or fall out; this is reassuring with regard to safety for the two-legged and four-legged occupants of such modern condos.  Still, I tried to have relatively light furniture by those windows.

However, we are more concerned here with what is happening out on the street, and wondering if we can be safe in the middle of the street around tall buildings during a major quake in the typical downtown area, whether in the US or elsewhere in the world.  I’ve found very few images of windows in tall buildings after major quakes, intact or otherwise, but the ones that I have found suggest that even if your downtown has a lot of modern buildings, the presence of older ones in the mix could make the ground below dangerous.

Here are some photos of post-quake glass that I think are notable.  In Figure 5, the top shows sharp glass shards on the ground after the 2011 Christchurch, New Zealand quake. The lower left shows the area outside of a store in Napa, California after the 2014 quake; you can see tempered glass pellets but also an intact window pane that fell out (notably, ground floor commercial windows use glass that doesn’t have the same standards as tall building windows since it won’t fall very far, but I wanted to include it as an example of what windows can do).  The lower right is probably the most compelling to me; it’s an actual sharp large glass fragment after the 1994 Northridge, California quake that fell from higher up and landed quite a few feet away from the building; apparently not from the broken window that is shown (note the shapes don’t match).  The information accompanying the photo online says “Shard of broken untempered glass that fell several stories from a multistory building in the 1994 Northridge Earthquake” and this is exactly the sort of thing that concerns me in areas with a lot of older tall towers.  

Figure 5.  More distressing street-level glass after earthquakes (click images to zoom in).  Attributions counterclockwise from top: Glass shards on the ground, by Andy Miah [CC BY-SA 2.0], via Wikimedia Commons.  Shattered and intact glass in Napa, https://twitter.com/WayneFreedman/status/503552790334283776 used with permission, @WayneFreedman and Abc7news.com.  Glass shard on Northridge lawn, from a FEMA report, Figure 6.3.1.4-1; photo courtesy of Wiss, Jenney, Elstner Associates.

I have one more photo by itself as Figure 6 so that you can more practically zoom in, as there is a lot of interesting detail when you do that; from the 2010 Chile quake, described at the source as “Glazing damage, due in part to pounding with the structure at right during the 2010 Chile Earthquake.”  Notice all the glass out in the street, even from this relatively short building, and some of it is sharp shards.  I’m not sure which structure on the right did the pounding, and you might suspect that the red ledge caused some of the glass to divert further out in to the street, but extenuating circumstances and exceptions tend to be, well, the rule.

Figure 6.  Chile 2010 glass shards in street from building (click image to zoom in), from FEMA report, 
Figure 6.3.1.4-4; photo courtesy of Antonio Iruretagoyena, Rubn Boroscheck & Associates. 

So what are we to conclude from all this?  I will say that I’ve come out of this with a much better appreciation of the strength of glass than before.  If people are looking up at glassy towers and picturing the kind of glass in pitchers and single-family home plate glass windows that get shattered by a wayward baseball, that’s not what’s up there; it’s MUCH more resilient and the panels can float a bit within their frames, as I mentioned earlier.  Many of these modern all-glass towers might get fractures in the glass but it will still stay intact.  A retired building inspector pointed out to me that while the lowest floors of skyscrapers might experience flexing and torquing, the upper parts of the towers will be moving as a whole unit, and any damage to the windows would probably be mostly limited to that caused by impact from within, rather than the shaking.  In fact, to be fair, all of the examples that I showed of smashed tempered glass in Figure 3 are from intentional impact, not shaking.  

Building construction standards have varied greatly around the world and over time, and older buildings without strengthened/tempered/laminated glass probably lead to riskier street-level situations in major earthquakes than modern buildings do.  My examples in Figures 2, 4, and 5 show that window glass CAN break.  However, my investigation suggests to me that this breakage is relatively unlikely to occur high up in glassy towers, where panes will move within their gaskets and even any glass that fractures will be resistant to shattering, suggesting that the risks of being outside around skyscrapers are not worse in downtown tower areas than in regular commercial areas.  And practically, trying to run into a building while other people are (erroneously) trying to run out isn’t going to work.  I suppose getting as far from the buildings as possible out in the street (assuming traffic has stopped) is probably still the best thing to do, getting down on the ground and curling up as small as possible while covering your head, trying to avoid downed electrical wires.

Even still, I wonder if I can make a few bucks selling bullet-proof umbrellas in downtown San Francisco on the anniversary of the 1906 quake.  Just a thought.

>>back to blog