"Right now the code says a structure must be engineered to have a 90 percent chance of avoiding total collapse. But many experts believe that is not enough.
“Ten percent of buildings will collapse,” said Lucy Jones, the former leader of natural hazards research at the United States Geological Survey who is leading a campaign to make building codes in California stronger. “I don’t understand why that’s acceptable.”"
One way that reads to me is that an event over some threshold has a high probability to cause the complete collapse all of those buildings. 90% doesn't seem like a very high threshold either, if it corresponds straight out to a 1-in-10 event that seems quite bad. If it's a 90% chance of survival in a 1-in-100 event, that's better, but not great.
Current building codes are intended to design to a 10% probability of collapse in the "Maximum Considered Earthquake", which is roughly a 2475-year event. The 2475-year event has a 2% chance of occurring during the (assumed 50 year) lifetime of a building.
Nah, the building is (mostly) just as good at the end of that time. It's just a convenience to make the probabilities easier to conceptualize. If we designed buildings such that the design earthquake had a 5e-6 daily chance of occurring, we'd just be confused.
If there is a 10% chance of collapse every 2475 years, that is ~1.1/10million per day chance, or 0.11 micromorts. That assumes 100% death during a collapse, and that the person spends all day in one of these buildings.
0.11 micromorts/day is very low, which makes the design criteria reasonable. Of course, I suspect our ability to predict infrequent events with any degree of accuracy is unlikely.
I suspect it would be the latter, similar to other disaster planning.
This gets you into that domain where most people really don't like to go, but... you don't get the choice of 100% perfectly likely to stand up, so where do you draw the line? You don't have an alternative. A line will be drawn. If you don't draw one deliberately, you'll draw one accidentally.
But if you want to have some fun with ways the probability estimates could be badly broken, correlated failures can be the way to go. Suppose for the sake of argument that there is an earthquake in the future that if each building was subjected to it individually would cause 10% of the buildings to come down. That doesn't mean that 10% of the buildings will fall down in real life, though, because the failures could be correlated and coupled, if one building falling down makes another building falling down more likely.
If I'm running the physics and numbers in my head correctly [1], even in an extremely large earthquake the Hollywood image of buildings toppling each other like dominos is still exceedingly unlikely. They'll go down, not sideways. But the additional shocks they create in both the air and the ground in the process could tip some other buildings over the edge. We know from 9/11 that buildings going down can seriously damage other nearby buildings.
So, densely packing buildings with a 90% survival rate can result in a net survival rate less than 90%.
And of course, if it's 90% at Richter x, there's no law of the universe that says x is the limit. x+1.5 is still perfectly possible.
[1]: It's hard to google up the amplitude of an earthquake in a distance measure, rather than Richter scale or something. But it looks like "several centimeters" of transverse motion would be a very large earthquake, which isn't going to put the center of mass of a building out of the area described by its foundations. Which isn't anywhere near a complete description of what it would take to tip a building like a domino, but is enough to say it's a long way from happening.
(My guess is that it's actually impossible to tip a tall building like a domino as any possible attempt to exert the forces necessary to do so would itself be enough to destroy the building. See also "why Superman must also be telekinetic", because it doesn't matter how strong you make an 747, there is no place on that plane it can be picked up by any amount of pressure exerted by two human-sized hands without the hands tearing right through the plane, rather than nicely carrying it. (A handful of non-telekinetic Supermans could do it if the plane lowers the landing gear and Supermans support the tires with their backs instead, but that'll be tricky to "fly" in other ways. Especially since the best way to support the plane will be with the Supermans being sideways, to maximize area of contact with the tire.) Similarly, some sort of Superman-esque force that applies equally to all atoms would be required, rather than the real forces we can apply via external pressures.)
x+1.5 is possible, but super unlikely at the upper end of seismic events. Given the way the Richter scale works though, even x+0.1 could be devastating.
It's the latter, 10% collapse in a 1 in 100 event. I used to do similar work in floodplain management. I don't even know what a 1 in 500 event is in earthquake terms, is it even possible to build something that withstands that level of quake?
“Ten percent of buildings will collapse,” said Lucy Jones, the former leader of natural hazards research at the United States Geological Survey who is leading a campaign to make building codes in California stronger. “I don’t understand why that’s acceptable.”"
One way that reads to me is that an event over some threshold has a high probability to cause the complete collapse all of those buildings. 90% doesn't seem like a very high threshold either, if it corresponds straight out to a 1-in-10 event that seems quite bad. If it's a 90% chance of survival in a 1-in-100 event, that's better, but not great.