Palaeontology - Palaeontologists develop new way to find new dig sites

National science journals are putting the spotlight on two Western Michigan University professors for discovering a new way to predict where fossils are hidden. 

University of Michigan graduate student in Anthropology Craig Wuthrich at the 2009 fossil site in the Great Divide Basin in Wyoming [Credit: Robert Anemone]

Traditionally, finding fossils is a "serendipitous and intuitive" event that comes with lots of reading and luck. 

Paleontologist and WMU anthropology professor Robert Anemone and WMU geography associate professor Charles Emerson said technologies are available and should be used to develop clues about where fossil sites are located. 

Their researched method - using a neural network, infrared electromagnetic radiation and satellite imagery - is being recognized as a possible option to help paleontologists prioritize where to spend time and resources out in the field with better results. 

Glenn Conroy, an anatomy and anthropology professor at Washington University, was a partner in developing this method. 

The professors were invited to present the neural network approach at the Society for Vertebrate Paleontology meeting in Las Vegas last week and were featured in national scientific journals Nature and New Scientist. 

Finding fossils 

During a field excursion, paleontologists camp out in a region and use their eyeballs and other tools to hunt for fossil treasure. 

"We understand the geology of the region and we don't just wander around but in the field it's a lot of intuition," Anemone said. "We want to add a more rigorous, predictive tool and we're trying to pioneer the use of new tools from geographic sciences in the search for fossils." 

Anemone has been leading field crews of students and other professionals to the Great Divide Basin in southwestern Wyoming since 1993 to collect mammal bones and fossils from the Paleocene and Eocene eras, 55 to 48 million years ago. 

In 2009, Anemone's team took a wrong turn and found themselves in an area of land with recognizable traits or hints of holding fossils. They "crawled around" for an hour and there it was: at least 100 partial mammal jaws with teeth. 

It was the greatest find he ever "stumbled upon," he said, with the discovery of one new rodent species and two new species of early primates that roamed the earth 50 million years ago. 

He said if the rodent were around today, it'd be some sort of "desert squirrel." 

"It's not just history, it's pre-history," he said. "It's the deep past. Fifty million years ago the only way we know what the earth was like and the living inhabitants that existed is by people going out and collecting fossils and studying the geology of the things that were alive." 

A better way 

After the big find, Anemone said he knew there had to be a better way. 

He asked Emerson, who has an extensive knowledge of satellite imagery, to get involved in 2010. They partnered to develop the neural network approach and have been conducting research for the past year. 

"We suggest that the geospatial sciences have earned a place in the paleoanthropological tool kit, and that 21st century research must increasingly rely on the kinds of sophisticated spatial analyses that can only come from collaborations with our colleagues in the geographical and geospatial sciences," it says in their study. 

While the use of GPS and satellite imagery is not entirely new in the profession, this method goes further by training a neural network - or the software brain - to recognize the characteristics and electromagnetic radiation data of a landscape to project the probability of finding fossils. 

By training the network to recognize "the fingerprint" of fruitful fossil sites, they hope it can find more. 

New approach 

The results are promising, with 85 percent accuracy in the testing stage. Next summer their model will be used to identify where to conduct field research. The results will show if the approach actually increases the number of fossil finds. 

Gerald Smith, the curator emeritus in the University of Michigan Museum of Paleontology, attended the Las Vegas meeting and said the neural network approach has the potential to change the field. 

"Some of us spent the day after the meetings looking for new fossil sites without any success, so in the future it's possible that their approach will be an important tool," he said. 

For Anemone, more fossils means more research. 

"In the time we are working with - 50 million years ago - there was a major event of global warming," he said. "The earth's climate was warmer than what it had ever been, so we are interested in the effects of climate change in the past on living things so we better prepare for climate change today by seeing the past events." 

But, they have to find the fossils first. 

The Kalamazoo Gazette 

Source: The Archaeology News Network