MAY/JUNE 2000 (Vol. 2, No. 3) pp. 20-21
1521-9615/00/$31.00 © 2000 IEEE
Published by the IEEE Computer Society
Published by the IEEE Computer Society
Guest Editor's Introduction: Computational Earth System Science
|Improving the Process|
|In This Issue|
PDFs Require Adobe Acrobat
Of the many fields of science in which increased numerical simulations and information technology have accelerated recent developments, perhaps none holds more promise than computational earth system science, a field of research within the new science of GeoComplexity. In the past, researchers observed physical and chemical processes to clarify the physics of such natural systems as earthquake faults, the convective engine driving tectonic plates, the evolution of the earth's surface, and the geodynamo deep within the earth's outer core. For example, they devised a variety of observational systems to measure the surface deformation due to earthquakes, to track changes in the earth's magnetic and gravitational fields, to determine hillslope stability, and to quantify the myriad effects of forest fires. They then constructed models to explain the observations and to formulate new hypotheses.
Unfortunately, progress has been slow. The temporal and spatial scales on which many of these processes operate often dwarf human dimensions. For example, large earthquakes occur on faults in the earth's crust over distances up to 1,000 km and involve recurrence times of typically hundreds to thousands of years. The geodynamo is characterized by even longer distances and times, typically several thousand kilometers and hundreds of thousands of years. Longer yet are the scales for the convective processes that drive the tectonic plates, which are on the order of thousands of kilometers and hundreds of millions of years.
In addition, the earth is a closed system, meaning that, on some level, all parts of the solid earth interact with each other over short- or long-range distances. Furthermore, most observations can only be made at the earth's surface—outside the volume of the solid earth—so earth scientists can't determine the values of the physical variables that actually define the nonlinear dynamics.
Improving the Process
Under these conditions, it is easy to see why obtaining definitive results has been difficult. Predicting hazardous events such as earthquakes and landslides is particularly difficult because no laboratory exists that can measure the necessary physical variables, refine the techniques, and apply the results in preliminary ways. The advent of modern, inexpensive, and powerful computational hardware, algorithms, and visualization methods, together with the World Wide Web, is rapidly changing this picture.
Analyzing and visualizing large earth system data sets have recently become routine and are now possible on desktop workstations. Equally exciting is the recent development of numerical simulation technology for complex nonlinear earth systems, along with new analysis techniques (based on statistical physics methods) and powerful visualization tools. More importantly, theoretical advances related to the physics of critical phenomena strongly suggest that the long-wavelength dynamics of model systems are closely related to those of real systems. So, for the first time, numerical laboratories exist that let us use nonlinear simulations to further our physical understanding of highly complex natural earth systems.
In This Issue
The five articles in this issue discuss progress in several of these problem areas. Louis Moresi, Michael Gurnis, and Shijie Zhong, in "Plate Tectonics and Convection in the Earth's Mantle: Toward a Numerical Simulation," formulate a self-consistent model of the formation and driving mechanisms of the earth's tectonic plates. In "Models of Earthquake Faults with Long-Range Stress Transfer," Eric Preston, Jorge S. Sá Martins, John Rundle, Marian Anghel, and William Klein describe recent progress in understanding the physics of simple models of earthquake faults.
"Cellular-Automata Models Applied to Natural Hazards," by Bruce Malamud and Donald Turcotte, describes the common understanding—based on scaling properties of self-organizing complex systems—being developed for various natural hazards. Howard Zebker, in "Studying the Earth with Interferometric Radar," provides an exciting look at a new, satellite-based technique that can obtain incredibly detailed images of the earth's surface deformation caused by earthquakes, glacier flow, and other processes. Finally, "Computational Aspects of Geodynamo Simulations," by Gary Glatzmaier and Thomas Clune, explains how numerical simulations of magnetohydrodynamic processes are revolutionizing our understanding of the earth's magnetic fields.
What is most exciting is that our understanding of the physics of these processes is rapidly improving in ways that, before the information technology revolution, were simply inconceivable.
John B. Rundle is a professor of physics and the director of the Colorado Center for Chaos & Complexity at the University of Colorado at Boulder. His research interests focus on problems in GeoComplexity relating to strongly correlated earth systems that can be attacked using the methods of statistical mechanics. He received his PhD from the University of California, Los Angeles. Contact him at the Dept. of Physics and Colorado Center for Chaos & Complexity, Cooperative Inst. for Research in Environmental Sciences, Univ. of Colorado, Boulder, CO 80309; email@example.com.