News tips from UC Santa Cruz American Geophysical Union meeting -- fall 1995
For more information, contact Robert Irion at irion@ua.ucsc.edu
(1) Study unveils a way to probe fault zones before a quake hits Session S41A, poster #27, Thursday a.m. 12/14
A new technique may let seismologists estimate which parts of a fault are likely to rupture most severely during an earthquake, even if the fault hasn't broken for a century or more.
(2) A 5000-year "movie" of earthquakes in southern California Session T12B, talk #8, Monday p.m. 12/11
Many a seismologist dreams of having a record of all quakes in southern California for the last 5,000 years. So, Steven Ward made one--using a computer tool called "synthetic seismicity."
(3) A thin, partially molten region in the lower mantle, just above the core? Session S32E, several talks, Wednesday p.m. 12/13
Geophysicists didn't think any component of the lower mantle was molten, not even partially. Now, a peculiar type of seismic wave may be proving them wrong.
(4) Supercomputer models open new window on icy era Session O12A, talk #2, Monday p.m. 12/11
Climate modelers have united to study the Last Glacial Maximum, when ice sheathed much of the globe. One such model offers notable insights into how the climate may have behaved.
(5) Near-surface geology dramatically affects direction of earthquake shaking Session S11C, talk #2, Monday a.m. 12/11
The direction of the most damaging shaking in a quake at any one place often depends more upon geologic structures near the surface than the location or the mechanism of the quake itself.
(1) Study unveils a way to probe fault zones before a quake hits
Thursday, December 14, 8:30 a.m.-noon Session S41A, poster #27 Moscone Center, Hall D Presenter: Justin Revenaugh (408) 459-3055 or jsr@earthsci.ucsc.edu
A new technique may let seismologists estimate which parts of a fault are likely to rupture most severely during an earthquake, even if the fault hasn't broken for a century or more.
The technique relies upon an apparent connection between the pre-earthquake geology of a fault zone and the amount of motion that a quake triggers along different segments of the fault. A similar relationship exists between fault-zone geology and the pattern of aftershocks that strike after the main earthquake.
UC Santa Cruz seismologist Justin Revenaugh, who devised the method, says it cannot predict when an earthquake will happen or how large it will be. However, the method may help researchers refine their maps of seismic hazard by hinting in advance which fault segments will pack the biggest wallop. That information could prove especially useful for the many mysteriously "locked" segments in southern California and elsewhere.
"It's difficult to figure out how much slip might occur on locked faults," says Revenaugh. "This scheme begins to give us a rational way of looking at each fault and dividing it into segments. Those segments are the building blocks that could break in one large earthquake or several smaller ones."
Revenaugh draws his conclusions from an analysis of the magnitude 7.3 Landers temblor and two other sizable quakes that ripped across California's Mojave Desert region in 1992. His study appeared in the November 24 issue of the journal Science. At the AGU meeting, Revenaugh will present further work, based on applying the same technique to the San Andreas, San Jacinto, and other locked faults in southern California.
Revenaugh adapted a technique used by researchers to monitor nuclear tests. Effectively, waves from earthquakes can serve as sonar-like "pings" to probe the earth's crust. Revenaugh's method models the rocks within the crust as a swarm of points that scatter energy, like fish in the water scatter sonar. Quakes from across the globe send waves through the swarm. Seismographs receive the echoes; their intensity and timing tells Revenaugh which points scatter the energy most strongly. Pronounced scattering exposes an abrupt change in the earth's structure, such as a dense network of cracks along a fault.
He notes that the method will not unravel the tectonic complexities of the Los Angeles Basin, because most faults there rupture in a different way and are too close together for the method to resolve. But he feels confident it will shed light on locked segments of major strike-slip faults--such as a long stretch of the San Andreas from Cholame to San Bernardino, which unleashed a huge earthquake (bigger than magnitude 8) in 1857 but hasn't budged since.
(2) A 5000-year "movie" of earthquake activity in southern California
Monday, December 11, 3:15 p.m. Session T12B, talk #8 Moscone Center, room 130 Speaker: Steven Ward (408) 459-2480 or ward@uplift.ucsc.edu
Many a seismologist dreams of having a record of all quakes in southern California for the last 5,000 years. So, Steven Ward made one--using a computer tool called "synthetic seismicity."
With a complete catalog of earthquakes down to magnitude 4 for several thousand years, Ward says, scientists could address some of the most elusive mysteries of earthquake occurrence: Do quakes cluster in location and time? Where are the active blind- thrust faults? Do century-long cycles in seismicity exist? Can past earthquakes predict future ones? Alas, the reliable historical record of southern California quakes is 150 years long at best. Synthetic seismicity, says Ward, can supplement that record--even though the results are not unique.
"Insights into these questions are so profound that the dream is worth pursuing, at least as a thought experiment realized through computer simulations," Ward says. "Inferences on long-term behaviors of earthquakes drawn from plausible computer simulations are no more suspect than those drawn from conventional tactics that stitch together woefully short earthquake histories, or which boldly extrapolate laboratory results over orders of time and length."
Ward, a geophysicist at UC Santa Cruz, has developed and refined his synthetic seismicity model over the last several years. In general, the model uses the laws of how stress accumulates and releases in the earth. Ward constrains the model with physical data where they are available, such as the locations of fault segments, the characteristic magnitudes of quakes there, and geodetic data that map strain throughout a region. Ward's past studies have shown that his simulated quakes happen far less regularly than researchers have assumed.
At the AGU meeting, Ward will discuss his latest work--a maplike 5,000-year catalog of synthetic seismicity for 14 major faults in southern California. The model accounts both for bends in the faults and for stresses that build up across faults, not just along them. Highlights of the study, which has been submitted for publication, include the following:
* The largest earthquakes, magnitude 7.5 and above, happen with some degree of regularity at any given location. "They really do control everything," Ward says, "but they only happen every 200 to 300 years."
* Moderate quakes (magnitude 6 to 7) occur essentially randomly, while smaller quakes (magnitude 5 to 6) show some clustering patterns.
* Some cycles appear in the numbers of magnitude 5 and 6 quakes. Variations in the timing of small earthquakes can foretell cycles of larger events, but only when averaged over periods of at least 100 years.
"This is a computer model, but it satisfies all of the observed data we know," Ward adds. He says the model can contribute to an overall map of seismic hazard for the region.
(3) A thin, partially molten region in the lower mantle, just above the core?
Wednesday, December 13, 4-5 p.m. Session S32E, talks #10, 11, and 13 Moscone Center, room 304 Speakers: Quentin Williams (408/459-3132 or quentw@rupture.ucsc.edu), Thorne Lay (408/459-3164 or thorne@rupture.ucsc.edu), and Ed Garnero (408/459-5139 or eddie@rupture.ucsc.edu)
Geophysicists didn't think any component of the lower mantle was molten, not even partially. Now, a peculiar type of seismic wave may be proving them wrong.
Careful analysis of recordings of this seismic wave, still in progress, seems to point to an unusual material within a thin layer at the base of the mantle, just above the core-mantle boundary. The wave travels so sluggishly through this layer--at least 10 percent more slowly than expected--that researchers suggest some portion of the rock there may be molten.
To make matters more enticing, the layer does not span the globe. Rather, seismologists have spotted it in only one place: a zone 1,800 miles beneath the south-central Pacific Ocean. Other workers already had identified a particularly warm patch of mantle in a broad swath above this area.
"The seismic discovery of this very slow layer is exciting," says seismologist Ed Garnero of UC Santa Cruz, "but the physical interpretation is a totally open frontier." Garnero and two other UCSC researchers, mineral physicist Quentin Williams and geophysicist Thorne Lay, will discuss the unusual layer or refer to it during parts of their talks at the AGU meeting.
The seismic wave illuminating the layer is a "diffractive" wave. The wave skims for some distance along the mantle side of the boundary between the molten outer core and the solid lower mantle, a dynamic part of the inner earth. As the wave ripples, some property of the material in this zone slows it down. That shows up as an extra bump in recordings of earthquakes. Garnero, who is scrutinizing the recordings with John Vidale of UCLA and others, believes the layer is no more than 40 kilometers thick, and perhaps is as thin as 5 kilometers in spots.
Enter Williams, who studies the properties of inner-earth minerals under high pressures and temperatures. "If the velocity anomaly is 10 percent and the layer is only about 20 kilometers thick, I don't see how you can do it other than with a partial melt," he says. The logical candidate for such a melt would be magnesiowustite, a mineral in which an atom of either magnesium or iron links up with an oxygen atom. Even so, the fraction of melt would be low, since magnesiowustite probably composes only 20 to 40 percent of the lower mantle. "This region has not lost its rigidity," Williams says. "It's still behaving like a solid."
Williams notes his interpretation depends critically on the seismic data. If the layer is thicker than suspected, or if the velocity changes are not as great, other notions would suffice. For instance, the region may feature strong chemical interactions between the outer core and the lower mantle--an explanation many researchers find more palatable.
If the partial-melt hypothesis bears up, it could provide a new gauge of the temperature of that zone. Such fixed points of reference for the inner earth, Williams says, are hard to come by.
(4) Supercomputer models open new window on icy era
Monday, December 11, 2 p.m. Session O12A, talk #2 Moscone Center, room 310 Speaker: Lisa Sloan (408) 459-3693 or lcsloan@earthsci.ucsc.edu
Climate modelers have united to study the Last Glacial Maximum, when ice sheathed much of the globe. One such model offers notable insights into how the climate may have behaved.
The modelers convened under the auspices of the Paleoclimate Modeling Intercomparison Project (PMIP), an international effort to compare the results of different approaches to simulating past climates on the earth. One of the first phases of PMIP focuses on the height of the last ice age, about 21,000 years ago. The researchers will present their early results publicly for the first time at the AGU meeting.
Debate simmers over a particular aspect of the last ice age: How much did the tropics cool? According to CLIMAP, a 1970s study that analyzed fossilized marine organisms, tropical oceans were no more than 2 degrees C cooler than today, while other parts of the globe were at least 5 degrees C colder. But in recent years, scientists tracking clues such as groundwater chemistry, corals, and snow lines on mountains have claimed evidence of much chillier tropical conditions during the ice age. The argument is not merely academic, for it may shed light on how circulation in the ocean--one of the planet's main conveyor belts of heat--reacts to climatic changes.
UC Santa Cruz paleoclimatologist Lisa Sloan used a powerful model to probe the ice age: GENESIS, which runs on a supercomputer at the National Center for Atmospheric Research. PMIP organizers supplied a set of initial conditions used by all of the models, including levels of solar radiation, concentration of carbon dioxide in the atmosphere, sea level, and the extent of ice sheets. For the first time, GENESIS and two other models used that information to compute sea-surface temperatures. Sloan's results tend to side with CLIMAP: a modest, but not extreme, cooling of the tropical ocean.
Sloan also had GENESIS consider a scenario where more heat flowed through the ocean, from the tropics toward the poles. This substantially cooled the tropical seas, but at the expense of heating up polar regions too much to agree with data from the ice age. "Ocean circulation patterns probably were different during the ice age than they are today," says Sloan, "but not necessarily in a way that increased heat transport toward the poles."
GENESIS also concluded that the massive ice sheet over North America did not force the jet stream to divide, contrary to what researchers had assumed. PMIP scientists developed a new model of the planet's icy blanket at the Last Glacial Maximum, with a somewhat thinner layer of ice overlying North America. The jet stream, so critical to weather patterns, then whizzes intact across the continent instead of splitting in two and rejoining in the North Atlantic. All PMIP models thus far agree on this result, Sloan notes.
(5) Near-surface geology dramatically affects direction of earthquake shaking
Monday, December 11, 9:15 a.m. Session S11C, talk #2 Moscone Center, room 124 Speaker: Ornella Bonamassa (408) 459-4426 or ornella@earthsci.ucsc.edu
The direction of the most damaging shaking in a quake at any one place often depends more upon geologic structures near the surface than the location or the mechanism of the quake itself.
UC Santa Cruz seismologists Ornella Bonamassa, John Vidale (now at UCLA), and others first reached that conclusion five years ago by studying records from the 1987 Whittier Narrows earthquake and aftershocks of the 1989 Loma Prieta quake. Now, a detailed field study in the Santa Cruz Mountains has confirmed and refined the results, giving the team enough data to construct 2-D and 3-D models of the curious phenomenon.
The experiment used an array of 24 seismometers spaced 10 meters apart on a remote slope near the Zayante Fault. The team used different tools to bombard the site with two types of seismic energy: compressional "P" waves and shear "S" waves. Directions of the strongest ground motion swiveled by as much as 45 to 90 angular degrees among the sites, in apparent response to variations in soil and rock conditions within the top 20 meters of the terrain.
Next, the researchers used mathematical methods to invert the data and create a 3-D map of the speeds of seismic waves through the study area. They found that S waves traveled 400 meters per second in some spots but as quickly as 700 meters per second in others--extreme variations for such a small study area. Because the stations were so close together, the velocity map was accurate to a scale of just 1 meter. "This is the first time such powerful tomographic and waveform studies have been applied to such fine-scale P-wave and S-wave structure," Bonamassa says.
The map revealed the surface of an old landslide under the study area. The scientists suspect that jumbled and loosely packed material above the landslide amplified ground motions in preferred directions. At sites where the near-surface geology is more uniform, they say, the directions of the most severe motions would depend not on local site conditions but on the location of the quake and its focal mechanism--that is, the way in which the fault plane ruptures.
The team made a successful 2-D model of these "directional site resonances" using three inputs: the velocity map, the topography of the site, and the scattering of seismic energy. At the AGU meeting, Bonamassa will discuss early work on the 3-D model, which also is promising but has not yet reproduced the largest variations in motion from the study.
As the technique for analyzing these "directional site resonances" evolves, Bonamassa says, it should help structural engineers. In particular, long structures such as dams or large buildings may experience unforeseen stresses if the ground beneath them amplifies earthquake waves in a few different directions. Engineers may be able to design safer structures by ascertaining the probable directions of strong shaking at each site.
Press Releases Home | Search Press Releases | Press Release Archive | Services for Journalists
Maintained by:pioweb@cats.ucsc.edu