Julian Lozos came to talk to us on July 17th, 2014 on measuring earthquakes – how they are taken and how they are used with an important distinction between measured and calculated data.
Corona Heights where our Randall sits has very visible evidence of an earthquake. Julian referred to the fault on the side of the hill as a “slip & slide”. It is very rare to be able to see one so exposed — you can see the polish from the pressure and the striations showing which way the surfaces slipped relative to one another.
After a quake, we want to know where was it? How big? Will there be aftershocks? The audience was invited to weigh in on what questions come to mind immediately after feeling a moderately small shake. It turned into a long, interesting discussion. Tsunami turn out to be a low risk in the Bay Area due to distance from faults that cause them. One at north end of the San Andreas (“triple junction” zone) can produce tsunamis but they come at an angle that diminishes them. Northern CA and Oregon north would be hit hard. In the Bay Area, only a landslide displacing lots of water would cause a tsunami. The 1906 SF quake caused a one inch tsunami.
Another way water can be a problem is seiching–water in a basin sloshing around to the point of generating dangerous waves. This would not happen on the Bay, since it’s shallow, but it might be a problem in Lake Tahoe.
Bigger quakes get named, even though the names are often not very directly related to the quake location, in order to give them media handles for discussion and easy identity.
Sections of the Hayward fault are slipping in “seismic creep” that means you can see the gradual movement (about as fast as a fingernail grows) by breaks in walls, buildings, roads. Other parts of the fault, and most of the San Andreas, lock together, building up even more stress as the creeping parts move. The locked parts only move in a rupture, which is another term for an earthquake. Hayward’s old city hall was abandoned because the fault goes right through it. Now that fault is a couple blocks away.
There are various ways earthquakes are measured: Acceleration: which is compared to the acceleration of gravity. Sometimes an earthquake can be more than or even twice the force of gravity. The biggest quakes (by energy released) don’t necessarily produce more shaking movement; it depends on surface composition and all sorts of other factors. Saturated and soft soils shake more.
During a quake, we measure shaking, but some techniques are improving that may give the ability to measure displacement as it happens. Various things are measured: P waves (Pressure; Primary) are like a sound wave in rock. They are the fastest thing emitting from the rupture and arrive first as a result moving at the speed of 6km/second. They feel like the come from below. S waves (Shear; Secondary) are the side to side waves. Love waves (named after a person) come third, more slowly, and are like a tail wagging side to side. Rayleigh waves are most destructive and shake in elliptical motion up and back then down.
P & S waves are generated mostly by the fault itself. Love & Rayleigh are mostly lensing and interference patterns caused by “sloshing” within the bedrock and the soils, waves reflecting off the earth’s mantle, other rock formations, fault surfaces, etc. The bay doesn’t effect the waves’ movement because it’s shallow. If you actually see the ground move in waves that’s probably a big quake and a Rayleigh wave.
Shaking measurement can be a crowd-sourced thing, too. Modified Mercalli Intensity is calculated from “measurements” or qualitative descriptions of many observers reporting how much they felt or damage they saw. Anyone can report in and should report — even if you don’t feel a quake that you heard about — this helps keeps the data accurate. Accelerometers in laptops and other devices (which exist to protect hard drive when dropped) can be networked to be seismographic info sources — all you have to do is download a free small application.
Since the biggest and worst earthquake damage comes last, early warning systems analyze the relationship between the P wave and S wave then try to get things in order before the later waves. This can provide a minute or two warning, depending on the distance to the epicenter — the farther, the more warning. An early warning might allow enough time to turning off gas and machinery, for example. Early warning systems in place since 1986 in Mexico (due to huge Mex City 85 quake) and Japan since 2008. This kind of technology is available now, but currently not for the general public, due to budget shortages and the complexity of psychology of how to get people to react appropriately. Currently PG&E, Google, BART and some others have access to a beta version only.
Earthquakes are located by triangulation using three or more stations. The origin is complex, because multiple places on fault can rupture at once. It might take years of computer models and analysis to determine all the sources, interactions and effects.
Shaking is a direct measurement and is not logarithmic; magnitude is a calculation and is logarithmic. A magnitude 7 has 30 times the energy released in a magnitude 6. Intensity maps created by Mod. Mercalli system have a downfall: they depend on how many people in a region report and how the regions are defined. Historical reconstructions of past quakes can use this method to calculate approximate magnitude of long ago quakes from diaries and news reports. Richter scale is used sparingly, it was designed for the LA Basin specifically and sometimes used for the quick and dirty first approximation. The numbers can change later after humans take a careful look after the original machine analysis. How long an earthquake lasts is one of the main factors in determining total magnitude (energy) released.
An aftershock is another quake, if the aftershock is bigger, than the earlier quake is re-termed a “foreshock”. Aftershocks can help determine length of rupture and depth. If they are getting father apart in time, they are aftershocks. They may or may not get smaller, though.
We can also take measurements after quakes: the offset if the quake rips the surface. A fault may be single line deep down but at surface it splays and splits, so there can be lots of offset measurements. 1906 is the first time offsets were systematically measured and the book of the report is huge. Nowadays LIDAR (portmonteau of ‘light’ and ‘radar’) is done from helicopter lasers. This sends signals to measure things on the ground and change is measure by comparing measurements to old data. InSAR measures deformation from space satellites. GPS is used for ground deformation measurements. They can track locations of fixed points and see that they are really not so fixed. There are so many new sensors that the split itself can now be measured as it happens in some places.
Contrary to popular opinion, small quakes don’t release enough energy to help prevent big ones later. You’d have to have magnitude 4 quakes every few minutes to keep the big one from occurring later. Stresses present when a fault ruptures don’t disappear, they just go and “bother” other faults bringing them closer to failure. Along the ruptured fault, the stress is reduced. At ends and bends the fault increased stress shoots out. This explains sequences of quakes. We also can’t release stress by bombs but we can measure fault locations by echoes from explosions or ambient vibrations such as traffic. To see what’s happening inside Parkside section of San Andreas, they dug 3km through it and are now getting lots of data from deep inside. It seems magnitude ones hit every 32 hours. Even though it seems like clockwork in some ways, the overall system and larger ruptures are impossibly hard to predict.
Pushing and pulling of the crust is always going to happen on our planet. Faults don’t move except by the end extending and they don’t go away ever, unless that piece of the crust is subducted. Once it’s there, it’s a weak place in the crust and it can break more easily than surrounding areas.
Seismometers have only existed since the 1880s. Layers of soil accumulation show different offsets, and break the old surface at different depths from which years can be calculated. On Hayward fault they’ve done trenching and seven or eight prehistoric quakes show up so they know an average of one big quake every 120 years. On the Red Sea, trenching shows thousands of years of non-frequent quakes with very little deposition due to desert conditions. We can find shaking by precariously balance rocks. How long has it been there without being knocked down? We can tell by sun exposure changes — how long since it eroded into it’s current balanced shape.
Anatolian fault: calculations show that the next part set to go off (if pattern continues) is right by Istanbul. The San Andreas part that is considered highest risk is furthest south in Palm Springs. San Gregorio is “decently high risk” but in water so it’s harder to tell what it’s doing.
Dynamic modeling: more efficient than waiting to see what behaviors a fault shows. Modeling lets you get at the physics of the observations, and allows picking the problem apart into smaller manageable chunks. They can compare the model to past measurements and tell what can we possibly can expect from possible future quakes. This kind of modeling was hard to do until strong recently with ever increasing computing power. Multi-cycle simulators over time use a stress and re-stress model.
All measurements come together to create rupture maps. UCERF maps (# 3 just out last week) includes all the models and measurements. UCERF4 will include current modeling.
Hazard maps are made from the UCERF map and includes the likelihood of a rupture based on the underlying geology.
Fracking only produces small “induced” quakes and only if it’s done in an area where other faults exist and can be triggered. Southern CA for example would be a bad place to do fracking. They won’t make a fault, but any old place might have some old faults — like Oklahoma. The cause of these fracking quakes isn’t the frack but the reinjection of the water into the ground.
Quake predictions are not possible. The debate now is not: can we predict quakes with what we know; it’s, ‘Will we ever be able to know?’ These are inherently chaotic systems and may be too hard to ever predict.
Comparisons:
- Most powerful quake measured: Chile 1960 was magnitude 9
- 911 was about magnitude 3
- Hiroshima was magnitude 6
- A 50 megaton bomb would be magnitude 7
- The space rock that killed the dinos was mag 13!
SF Natural History Series 