Research Feature
May 11, 2004
Odessa, Texas, received its ironic name in 1881 from Russian railroad laborers, who must have longed for the temperate Black Sea resort in their hot, prairie exile. The Texan Odessa lies in the Permian Basin, site of some of the country's richest oilfields. About eight miles southwest of the town (which is now, thanks to oil, a metropolis of nearly 100,000), the barren pastures boast the second-largest meteor crater in the United States.
); "A geological cross section through the Odessa meteor crater.")
see the full-size image.
Chances are that your personal catalogue of meteor craters begins and ends with the largest meteor crater, the Great Crater near Winslow, Arizona, also called the Barringer Crater or the Canyon Diablo crater, which is three quarters of a mile across and 600 feet deep. The site has been open to the public since the 1920s, astronauts trained to walk on the Moon there, and its picture has become the canonical crater photo.
The Odessa Crater is far smaller, only 160 m (175 yards) in diameter, and the original depth of 30 m (~100 feet) is now only 2 m owing to the gradual accumulation of sediments since the meteor struck an estimated 25,000 years ago. With its rocky rim rising another 2 to 5 m above the plain, it is far from unimpressive--but still, few have heard of it. Until five years ago, the site was basically uncared-for.
What hit Odessa? How big was the meteor and what was the impact like? Answers have been a long time in coming. The details are only now emerging, thanks to new computer simulations, performed at the Texas Advanced Computing Center (TACC), based on early data from the site. "This crater is finally getting some respect," says Dr. David L. Littlefield of the Institute for Computational Engineering and Sciences at The University of Texas at Austin (UT Austin), who carried out the calculations together with graduate student Paul T. Bauman.
The simulations of the impact, done on TACC's 224-processor IBM Power4 cluster, called Longhorn, showed that early estimates of the size and trajectory of the Odessa meteor were wrong and that a much larger object coming in at a nearly grazing angle had made the crater. "Our work has both validated the code we used and enabled us to reconstruct the actual event," Littlefield says, "and it is also an interesting story about the progress of science."
Drilling to investigate the crater and four surrounding smaller craters, begun in 1939 by geologist Glen Evans, was completed in the 1940s by E. G. Sellards of UT Austin's Bureau of Economic Geology. A shaft was sunk in the center, in hopes of finding a large core remnant of the nickel-iron meteorite, but only fragments were discovered, and the diggers reluctantly concluded that the collision had caused it to disintegrate. Core samples from the excavations helped delineate the stratigraphy of the impact area, and the data found their way to the Texas Memorial Museum on the UT Austin campus. An artist's rendition of a cross section through the center of the crater (Fig. 1) was exhibited by the museum, but it was only in 2000 that Evans and Charles Mear published a detailed paper on the geology--in a museum bulletin issued by Baylor University.
Littlefield and Bauman used the original stratigraphic data, which were supplied by Dr. Ann Molineux, curator of non-vertebrate paleontology at the Texas Memorial Museum. Molineux believes that Littlefield's subsequent calculations at TACC validate the fundamental museological idea. "The work David and Paul did is a wonderful example," she said, "because it shows how old data are never exhausted--data that have been saved can be re-used in new contexts at later dates, when more is known or better methods of analysis are developed."
As Littlefield points out, it was not only new methods that were developed, but also a revolution in scientific thinking about the Earth's encounters with meteors. Because other geologic processes tend to erase the evidence, meteor craters are not the obvious suspects in the constitution of the Earth's crust. Yet there was an epoch when the Earth must have been bombarded by meteors as often as--or more often than, as it presents a larger cross section to meteors--the Moon. Tectonic processes have disposed of those craters, and since the current geological epoch began, the Earth has suffered only occasional hits. Most have gone into the oceans, which cover 6/7 of the Earth's surface. Those that hit in the tropics were quickly overgrown.
Very few meteor falls have occurred throughout human history. When Yale professor Benjamin Silliman and a colleague authenticated a fall in Weston, Connecticut, in 1807, President Thomas Jefferson--an Enlightenment scientist himself--was extremely skeptical. "I would rather believe two Yankee professors would lie," he reportedly said, "than that stones have fallen from the heavens." Arguments among geologists over whether crustal features were volcanic or meteoric in origin were often heated, and astronomers seemed to have little interest in meteors once they had landed. By the 1950s, the world list of craters contained only about 20 sites.
Thanks to efforts made by some astronomers, however, the list was gradually lengthened. The pioneers included Robert Dietz in the United States and Carlyle Beals in Canada, and a second generation led by Robert Grieve in Canada and the late and very well known Eugene Shoemaker in the United States. By 1989, then, meteoritics expert H. Jay Melosh of the University of Arizona could look at a list of about 150 known craters and write, "As recently as 1950 most astronomers believed that the lunar craters were giant volcanoes and all but a few geologists derided the idea that the earth's surface had been scarred by impact structures kilometers in diameter. . . . geochemists are just now beginning to realize that nearly all of the material now residing in planets has been processed through high-velocity impacts." (Melosh, 1989: Impact cratering: A geologic process, Oxford Monographs on Geology and Geophysics, p. v.)
The plate tectonics revolution of the 1960s and the trips to the Moon helped validate this view, Littlefield notes, as did a new means of distinguishing between volcanoes and impact craters: shocked quartz crystals are found in the ejecta and in the breccia under meteor craters, but not around volcanoes. The idea that a meteor had been responsible for ending the dominance of dinosaurs some 65 million years ago, as it gained acceptance, completed the transformation of our thinking. The Earth was not a protected corner of the Universe, but subject from time to time to the most catastrophic events.
Littlefield has been at UT Austin since 1996, using computational methods for predicting the effects of high-velocity impacts in solid and fluid media. He has computed at TACC for many years. "We deal with problems that have high strain rates, large deformations, and strong shocks--mostly highly transient events," he explains. Much of his funding has come from the Department of Defense, for investigations of the effects of explosions or of the penetration of armor by projectiles. "Planetary impacts are a new departure, but they are also of greater interest in the context of the new 'catastrophism' of astronomy and geology," he says.
When he learned that detailed geological data were available for the Odessa crater, including measurements of the crater and rim and a stratigraphy of the layers of soil and rock that the meteor hit, Littlefield saw an opportunity to test his computational codes and validate them in a new regime, with a broader range of input parameters.
Littlefield and Bauman used CTH, which he describes as a code embodying "Eulerian finite-volume continuum mechanics," in which calculations are made on a mesh of points representing the physical domain. CTH is widely used in the academic, government, and industry communities for high-velocity impact analysis. "What we needed for input was material models of the constitution of the meteor and of the surface it impacted," Littlefield says, "plus some guesses about the size and speed of the meteor and the angle of attack." The output was a redistribution of the material in the model, showing the crater and the material ejected from the site, with gaseous material removed.
Littlefield and Bauman's preliminary calculations used impact conditions suggested by the geological studies: an iron meteor 4 m in diameter coming in at an angle of 60 degrees from the vertical at a velocity of 12 km/s. These could not produce a crater volume great enough, or one of the right dimensions, so further runs were made. "We used a lower bound of 12 km/s for the velocity and an upper bound of 26-27 km/s," Littlefield says, "with the original meteor diameter of 4 m and a larger diameter of 6.87 m, coming in at a lower angle, 77 degrees from vertical."
These calculations also failed to reproduce the Odessa crater morphology, and in particular the evidence that the meteor dug up and turned over about 7 m of the subsurface in creating the crater rim. Finally, the calculations were repeated with much larger meteor diameters (9 and 15 m) and a near-grazing angle of 84 degrees from the vertical. These inputs produced a model crater of the right shape, although still smaller than the Odessa crater (Figs. 2 and 3).
"We're homing in on the answers," Littlefield says, as the calculations continue. "What we have shown is the great sensitivity of the code to the variation of inputs and thus the need for accurate material models to take best advantage of the code's flexibility. Because we have an answer to shoot for in the case of Odessa, at least in terms of the size and shape of the crater and the amount of material upended and ejected, we can come a lot closer to a better estimate of the velocity and trajectory of the meteor at impact."
Littlefield and Bauman have already shown that the meteorite impact energy was more than 50 times larger than originally believed. "As we worry about the near-Earth and Earth-crossing asteroid orbits," he says, "we need to know what effects to expect from a hit the size of Odessa--which probably occurs with much greater frequency than the dinosaur-killing impacts.
"We needed the speed of a machine like Longhorn at TACC to do these calculations at all," Littlefield notes, "and many of our runs used 64 processors at a time. I think we could easily scale the calculations to more processors on faster machinery."
The mythic giant jackrabbit of Odessa may have been the only witness 25,000 years ago (although the dig did turn up the bones of a Pleistocene mammoth), but Littlefield's simulations can inform our present wonder--and our hopes never to be near such an event!
As for the crater itself, interest is also increasing. Dallas attorney Tom Rodman took its part and shepherded it into the modern era: a museum was opened at the site in 2000, and the land is now in the care of the Park Service. Molineux of the Texas Memorial Museum will incorporate a movie of the new computer simulations into an educational CD; contact her for more information annm@mail.utexas.edu.
