La Habra and The Importance of Being Fast
Yesterday evening around 9pm local time, a Magnitude 5.1 earthquake struck the LA area near La Habra, east of Downtown. The quake was strongly felt throughout the LA Basin, causing minor damage in the vicinity of the epicenter. Several people were injured but thankfully there have been no reported deaths from this event.
This earthquake appears to have occurred on the Puente Hills Fault. This is a “blind” fault, so called because the fault itself does not reach the surface of the earth so we can’t see it in the geologic features of the LA Basin. Instead, we rely on earthquakes that occur on the fault to tell us where it lies and what hazard it represents. The Puente Hills Fault is a thrust fault, meaning it accommodates convergent or “squeezing” motions within the LA Basin. It dives shallowly into the crust beneath LA and presents a very geographically distributed seismic hazard, though the actual risk of a major earthquake on the fault is very poorly constrained. The trouble is, the fault underlies the most densely populated parts of the LA basin so it has the potential to generate an earthquake that, even if only moderate by global standards, could harm many people. And a large M 7 earthquake, such as the one shown in the scenario to the left, could devastate the entire LA Basin with almost no warning at all.
It is exactly this sort of fault hazard that motivated us to build the system we have. I have seen in my work in the earthquake warning community a focus on very large, very rare earthquakes as the worst-case model for earthquake warning system design. The trouble is, these earthquakes are expected to nucleate on faults at some distance from urban areas: think of the “Shakeout” scenario, in which a massive M 7.8 earthquake ruptures the southern San Andreas Fault, but has an epicenter at the Salton Sea, far from Downtown LA. Similarly, think also of a 1906-repeat earthquake on the northern San Andreas, which in the “worst case” will nucleate off the Mendocino coast. It’s true, this would be a much more damaging earthquake than one with an epicenter near San Francisco (remember directivity), but from an earthquake warning design perspective, this is not the worst-case scenario! These events are so far away from urban centers that it takes over a minute for the damaging waves to arrive from the epicenter. This is plenty of time for any earthquake warning system to figure out what’s going on and issue ample warning to these population centers.
The true worst-case scenario for any earthquake warning system is exactly the sort of earthquake that occurred on Friday: a moderate or larger earthquake that occurs directly beneath a population center. First of all, these earthquakes are much more common than the M 7-8 “big ones” which may occur only once every several decades. And because they occur much closer to people and buildings, they can cause as much damage to a local area as a much larger, though much more distant earthquake. But most importantly from an earthquake warning perspective, these earthquakes provide scant seconds between the nucleation at the epicenter and the arrival of damaging waves in vulnerable areas. In our view, this is the earthquake we need to design for, not the large and distant event. If we were to build a system dedicated to the large and distant event, we would be consigning everyone near the epicenter of any earthquake to their fates with no benefit whatsoever from the warning. If, on the other hand, we can build a system that can act in the fraction of a second between the detection of the local earthquake and the onset of damage, then we can give some small benefit to everyone near the epicenter, and that much more benefit to people more distant from the epicenter of a larger quake. What’s more, because of the hard work of speeding up the solution for the local earthquake, the problem of the large and distant event becomes that much easier to solve.
The benefits of this design philosophy are plain to see in the map to the right, which shows the warning times for Friday’s earthquake from the proposed SWS earthquake warning system for the LA Basin. Right at the epicenter we provide a scant tenth of a second warning, but this is better than zero (or less than zero!) warning, and the rapid processing pays dividends in places like Downtown LA, where this approach yields over 7 seconds of warning. In our view, this is the only way to provide earthquake warning: protect everyone you possibly can, with as much warning (and therefore value) as you possibly can. Anything less is a dereliction of our duty to the people of California.