It’s bad enough that fossils, buried deep in layers of rock for thousands or millions of years, may be damaged or missing pieces, but what really challenges paleontologists, according to University at Buffalo researchers, is the amount of deformation that most fossils exhibit. That’s why Tammy Dunlavey, a master’s degree candidate in the Department of Geology in the UB College of Arts and Sciences, and her colleagues are working on a computational method to morph fossils back to their original shapes by calculating and excising the deformation.From the University at Buffalo:Paleontologists Use Computer to “Morph” Deformed Fossils Back to Their Original Shapes
It’s bad enough that fossils, buried deep in layers of rock for thousands or millions of years, may be damaged or missing pieces, but what really challenges paleontologists, according to University at Buffalo researchers, is the amount of deformation that most fossils exhibit.
That’s why Tammy Dunlavey, a master’s degree candidate in the Department of Geology in the UB College of Arts and Sciences, and her colleagues are working on a computational method to morph fossils back to their original shapes by calculating and excising the deformation.
“Our goal is to develop computer programs that can reliably solve the deformation problem,” noted Dunlavey, who on April 1 presented research on a new suite of “retrodeformation” programs at a Geological Society of America meeting (North-Central section) in St. Louis.
The main program is called “MsWellman,” written by the UB researchers in collaboration with H. David Sheets, Ph.D., professor of physics at Canisius College and adjunct associate professor of geology at UB.
MsWellman adapts an approach developed by a structural geologist named Wellman, and works on multiple rock slabs at once.
“Fossils are deformed because they are fossils,” said Dunlavey.
Deformation makes the prospect of gleaning from fossils important data about ancient worlds that much more difficult, according to Charles E. Mitchell, Ph.D., professor and chair of the UB geology department, with whom Dunlavey is collaborating.
While paleontologists traditionally have tried to concentrate on the rare, well-preserved fossils for which deformation is not a significant issue, they increasingly are interested in the many fossils that clearly have been deformed.
“The question our computer program is designed to address isn’t, ‘Are fossils deformed,’ but rather ‘By how much?'” said Dunlavey, noting that millions of years of being buried causes different levels of deformation in fossils.
According to the UB researchers, MsWellman calculates the degree and form of the deformation and then a second program the UB team developed called Retrodef6, uses this understanding to “correct” a representation of the deformed fossil back to its original form.
“We wanted to design a methodology that determines at what point, statistically, fossils can be considered deformed and calculates the amount of deformation based on how much strain they were subject to when embedded in rock, as well as other variables,” she said. “The program then will restore the virtual fossils to their original shape.”
To do that, the UB scientists employed a technique called geometric morphometrics, which documents aspects of shape and size in a specimen based on landmarks, discrete anatomical points that generally are uniform for related specimens.
For example, Dunlavey explained, one might consider the eyes in a human face as a landmark feature, and, since human faces are expected to be bilaterally symmetric, the right eye is expected to be located on the opposite side of the face at the same height as the left one.
In the same way, she said, many fossils are expected to be bilaterally symmetric in their original form, a concept that is a key premise of the UB computer programs.
To gauge the reliability of the new retrodeformation programs, Dunlavey used several fossils of graptolites, which are the remains of an extinct group of marine organisms.
The fossils Dunlavey used lived during the Middle Ordovician Period, some 472 million years ago, before land animals or large land plants had evolved.
Since graptolites were common and evolved rapidly, Mitchell explained, they tend to be useful markers for constructing a timescale in the Ordovician period.
Because their original shape is well-known, he continued, several sets of deformed, slightly deformed and non-deformed graptolites served as an excellent test case for the new computer programs.
So far, the type of deformation the UB team has excised from these specimens is what geologists call structural deformation, changes in the earth’s crust that occur over many millions of years during mountain building.
During their research, the UB researchers discovered that a significant amount of deformation also occurs from the hardening of the soft mud the organisms were buried in, which flattened the fossils, producing asymmetry.
“First these fossils were squashed during this hardening process and then they were smeared during mountain building,” explained Mitchell.
The team plans to apply its computational techniques to both types of deformation to develop methods that will provide the clearest view of what the fossils looked like when they were still living inhabitants of Earth’s ancient oceans.
