On Global Geodynamics and Earthquake Forecasting: An Interdisciplinary Approach (LDEO News, Winter 08/09)

From the sudden, violent motion of a single earthquake to the more gradual convection of its interior over millions of years, Earth is a dynamic planet.

Geodynamics—the study of the forces and processes driving Earth’s tectonic plates—is an integrated science that draws from geology, geophysics, technophysics, seismology, fluid dynamics, and petrology. It’s natural for Lamont-Doherty scientists—enabled by the tradition of interdisciplinary studies and the breadth of expertise here—to pursue some of the most complex questions in geodynamics. With a combination of observation, theory, computation, and laboratory experimentation, they have begun to illuminate the most difficult problems and to provide the most relevant constraints.

Nowhere does this effort have more impact on society than when significant earthquakes occur. Large earthquakes are among the richest phenomena on our planet, and to study them provides insight not only into their hazard and risk, but also into the deformation processes in Earth’s crust and uppermost mantle. Questions pertaining to mantle convection and plate tectonic movement, formerly considered as two distinct scientific problems, are now being looked at as one coherent system.

It is incredibly productive to view the tectonic movements that triggered last spring’s earthquake in Wenchuan County—the Sichuan province of western China—in a geodynamic framework. Standard plate tectonics, in which rigid plates collide and generate earthquakes at well-defined boundaries, cannot explain entirely the forces that led to this disastrous event. Of course, the underlying cause is India’s collision with Asia, which has given rise to the Tibetan Plateau and the Himalayas. But rigid plate interactions are not sufficient to account for what happened in Wenchaun. Instead, a large body of work suggests that, in China and neighboring countries, the Eurasian plate is deforming internally. Much of this investigation has been an attempt to identify the nature of this deformation: Is it “platelike” (and thus explained by the localized shattering of the Eurasian plate into microplates), or is the deformation spatially continuous (thus behaving more like putty than pottery)? And which forces contribute to this unique style of deformation?

“Earth scientists want to know what happens at all scales—both geospatial and temporal—and how localized systems interact with larger, global ones,” says Ben Holtzman, a Doherty associate research scientist.


The 7.9-magnitude earthquake struck the semirural area outside of Chengdu at 2:28 p.m. on May 12, and aftershocks ranging between 4.0 and 6.9 on the Richter scale occurred over the next several days, weeks, and even months. Results were catastrophic, with more than 70,000 casualties. Most seismologists and engineers had estimated that ground tremors would reach only certain levels, and they built their civic infrastructure accordingly. “But the actual ground shaking recorded was three or four times greater than what had been projected,” says Arthur Lerner-Lam, associate director of Seismology, Geology and Tectonophysics at Lamont-Doherty. “How could we have been that far off?” he asks.

The absence of a comprehensive geodynamic framework (its predictions for brittle deformation and fault failure) contributed to underestimates of the earthquake’s intensity. Transferring basic research to help formulate stronger building codes may have averted the disaster that befell homes, factories, schools, and community buildings.

Inevitably, poor comprehension of the dynamics involved in the movement of continental plates—upon which so many of Earth’s inhabitants reside—hinders our ability to perform accurate risk assessments. “Here’s an earthquake that surprised almost everyone, including the Chinese,” says Lerner-Lam. “Isn’t it important—from a humanitarian perspective—not to be surprised by such events?”

Deep beneath this Chinese territory, complex internal activity complicates prediction efforts on the surface. The deformation of this area is highly diffuse—it cannot be described as a single, large fault (a boundary between two crustal plate fragments). Thus, scientists have begun to ask how one might conceive of deformation when it is distributed along multiple faults. How do these distinct faults interact with one another and how does an earthquake at one fault relieve or enhance the stress at another fault?

In order to delineate the forces driving the plates, scientists are reconceptualizing the problem of deformation. The theory of plate tectonics approximated that crustal deformation occurred at plate boundaries, without recognizing the deformation that occurs within a given plate. “We’re looking at a problem that demands a more regional understanding of earthquakes and brittle deformation,” asserts Lerner-Lam.


For the past 40 years, plate tectonics has been the working theory adopted by researchers. But it is a predominately kinematic theory—it describes the actual motions of the plates, their position, and speed; it depicts how the plates move, but not necessarily why. Although geophysicists have often pondered the reasons for crustal movement, definitive answers have been hard to obtain.

Geodynamics has emerged from the original plate-tectonic paradigm shift, and it is a convergence of knowledge gleaned from different disciplines. It integrates kinematic theories with dynamic modeling of topography, and it surveys the entire earth structure, examining deformation as a result of both surface and internal physics. Researchers now understand that plate motion and deformation are surface expressions of mantle convection. With this synthesis of fields, researchers are starting to comprehend why certain motions occur. And by understanding the reasons why plates shift, scientists can generate better-localized predictions regarding future deformation.

There still exists a gap between the observable, measurable—and experienced—seismic events and the effects of deformation visible in the longer-term record. “We need to link observation to theory—what we can measure now, what we see in the geological record, and what we can model on a computer,” explains Lerner-Lam.

“We’re beginning to understand plate dynamics in an integrated sense—how they move, how they deform, and why,” Lerner-Lam says. “We’re able now to get beyond a purely kinematic view of just the earth’s motions. Now we’re interested in the forces driving that tectonic movement—not just the forces that drive the plates, but also those that contribute to the interactions between these plates.”

Geodynamics looks at how different earth systems interact to produce the deformation observed at the surface. A unified approach relates the forces causing changes in Earth’s soft interior to those causing the deformation of its rigid exterior. Earth’s crust is solid and brittle on the surface, but it gets softer and more ductile with depth. Accordingly, the brittle exterior layers, which break during earthquakes, behave differently from the viscous layers underneath. “Those forces in the earth that move the viscous soup and break the brittle crust are the same underlying forces that drive Earth’s dynamics,” Lerner-Lam states. “We can no longer look at earthquakes as just isolated instances of brittle crust fracture. It’s all related to the stuff churning underneath.”

“If we can begin to understand how they’re all related, then we can begin to approach such problems as earthquake forecasting—one of the most daunting challenges facing global geo–dynamics,” Lerner-Lam says.


The geodynamic story of the Sichuan earthquake began millions of years ago when India collided with Asia. Since then, the crust and lithosphere under Wenchuan County have advanced eastward, effectively squeezed out from between Indian and Asian plates. The crust and upper mantle continue to crash into the fixed Sichuan Basin—a cold continental fragment and low topographic area that impedes the crust’s movement. As they encounter this resistant basin, they have fl owed around it, but not easily.

“When you start squeezing something that is soft and mushy, it deforms all over the place,” explains Geoffrey Abers, a Lamont-Doherty research scientist. “It doesn’t just break and slide on a single fault. Instead, it breaks in lots of little faults over a big area.” In fact, the crust shortens horizontally but also thickens vertically, which works to build topography.

Holtzman likens the process to what happens when a car hits a wall. “It crumples up—it gets thicker. When Earth’s crust thickens, its deeper parts start to heat up. As the crust grows warmer, it weakens and begins to fl ow more easily.”

When India collided with Eurasia, the terrain in the middle was young, hot, and malleable—allowing for greater mobility. But all that squeezing and thickening built up what is now the Tibetan Plateau. The Sichuan Basin—a rigid block—remains in place and undeformed, while the weaker crust continues to thicken and flow. The most dangerous location is at the boundary between these two regions; material in the middle and lower crust can fl ow without breaking, but the brittle uppermost crust continues to amass significant stress. “If you squeeze crème brûlée from the sides, the cream creeps quietly, but the sugar crust on top will bend only a little bit before it cracks,” says Holtzman.

Cracks that pervade Earth’s crust can withstand a certain amount of stress. When their stress threshold is surpassed, the fault “fails,” nucleating an earthquake. “Somehow, the flow of the weaker layers below is coupled to the more brittle upper layers,” Holtzman says. “But how the cracking will be distributed, and in what time frame, we do not really know.”

People who have surveyed fault movement in China have thought that the deformation might have been aseismic—that is, when deformation (such as mountain building) in the crust fails to elicit large-scale earthquakes. “But obviously, in retrospect, that aseismic deformation wasn’t enough to relieve the built-up stress. So what’s the balance between seismic and aseismic deformation?” asks Lerner-Lam. Is there a cyclical pattern involved? “It’s a perfect example of why we need to comprehend the deformation that is distributed along numerous faults. We can no longer think of crustal deformation as occurring along very well-defi ned plate boundaries.”

“Nothing we previously knew about the geology translated into an accurate forecast or model of potential ground motion,” says Lerner-Lam. And thus very little of what was understood about the area’s geology and seismology helped in the design of local buildings.

Lerner-Lam never loses sight of the social good that might result from his colleagues’ fi ndings. He also serves as the director of Columbia’s Center for Hazards and Risk Research, through which Lamont, along with the Earth Institute, draws together the University’s expertise in earth and environmental sciences, engineering, social sciences, public policy, public health, and business. Its goal, ultimately, is to accumulate enough knowledge to inform engineers so that they can build safer structures. “The fact that this earthquake was underestimated, underpredicted, underforecasted, however you want to put it, should be a wake-up call for a lot of people who work in countries where buildings may not be well constructed and where there is a poor understanding of earthquake safety,” says Lerner-Lam.

There are at least two lessons to learn, according to Lerner-Lam. First, earthquakes can happen in more places than scientists previously thought. Until we understand how crust deforms and releases stress, we will never make much progress on the hazards front. Second, we must focus on earthquake engineering and informed construction. “We need to incorporate an understanding of geology and geological inevitability into a policy that makes sense, especially for a developing nation,” asserts Lerner-Lam. “We should be able to use our knowledge from the science side to understand why such seismological events were underestimated, and then the applied side can come up with better tools for making forecasts so that this does not happen again.”


Lerner-Lam hopes that Lamont-Doherty will continue to foster expertise in computational modeling, theory, and observation. “We want to look at observations both in the fi eld and in the lab. We want to build a good computational base to test the theory against our observations,” he says. “Ideally, it would involve a collection of people with each researcher specializing in a particular facet of study such as petrology, seismology, or rock mechanics. We need a coalescence.”

Abers concurs: “The trick is to bring together enough different people with complementary skills and backgrounds.” And Lamont is one of the few places that approaches the task in a multidisciplinary manner.

The collaborative environment at Lamont remains unequaled. “From the cafeteria to seminars—scientists have the opportunity to talk to each other. There are few institutions that have this breadth,” asserts Lerner-Lam. “This is a natural environment for bringing all these pieces of the puzzle together. We have the size and the scope—we can be deep and broad at the same time. That’s part of our history.”
The field has generated a surge of scholarship, and scientists, deciphering how the plates move and deform, are presenting us with an entirely new understanding of the dynamics of the planet.

The study of natural catastrophes is a fundamental concern for earth scientists and something Lamont takes quite seriously. “Basic science is absolutely essential to forecasting geologic hazards,” argues Lerner-Lam. “Then we can begin to talk about what societies can do about them.”

Thanks to Arthur Lerner-Lam, Ben Holtzman, and Geoffrey Abers for their contributions.


~ by Janet Fang on July 3, 2009.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: