The National Institute for Computational Sciences

UT’s Kraken supercomputer helps track white-nose syndrome

University of Tennessee researchers predict the spread of a lethal disease that’s devastating North American bat populations

Caitlin Elizabeth Rockett

In February 2006, a professional hydrologist and recreational caver documented the North American emergence of what is known as white-nose syndrome (WNS), a fungus-driven disease that is rapidly killing cave-dwelling bats. In six years, WNS has killed at least a million bats and has spread from New York to 15 other U.S. states and four Canadian provinces. The U.S. Fish and Wildlife Service, who officially monitor the disease, have called it “the worst wildlife crisis in memory.”

Above: The remains of bats litter the floor of a cave that has been infected with white-nose syndrome, October 2010. (Image courtesy of Ann Froschauer, United States Fish and Wildlife Service)

As such, ecologists, speleologists and nature-lovers alike are fervently searching for insight into this poorly understood disease, including University of Tennessee (UT) recent professor emeritus Thomas Hallam. Using one of the fastest supercomputers in the world, Hallam and his colleagues have been exploring one of the most fundamental unknowns of WNS.

“The aim of our research was to project, using the little data that we have, how white nose will spread—if it will spread—across the U.S.,” said Hallam, who explained that while the disease exists in Canada as well, available Canadian geographical information is not detailed and has kept the team from examining the spread of the disease there. While there is still much that is unknown about WNS (and even about bats), Hallam’s team was able to accurately simulate how the disease is currently spreading, giving them insight into how it will spread in the future. Their research has also give rise to methods that may help mitigate WNS.

Hallam’s team used the Cray XT5 known as Kraken at the National Institute for Computational Sciences (NICS), which is managed by UT and located on the campus of Oak Ridge National Laboratory (ORNL).

About bats and WNS

Bats—those winged creatures of the night that often (speciously) cameo in horror movies—are far more helpful than harmful. A study published in Science in April 2011 estimated that bats might be worth at least $3 million per year to U.S. agriculture because of the sheer volume of crop-eating insects they consume. The study, led by Justin Boyles, a post-doctoral research fellow in the Department of Ecology and Evolutionary Biology at UT, concluded that if WNS remains unchecked, “we can expect noticeable economic losses to North American agriculture in the next 4 to 5 years.” With all species of bats typically having only one pup each season, it’s very difficult for bat populations to recover from something as devastating as WNS.

Scientists think that WNS was imported to North America from Europe. While the disease has existed for quite some time in Europe, it does not produce the same devastating effects seen across the Atlantic, perhaps because European bat species are heartier, have evolved resistance to the disease, or simply because the European strain of WNS is less aggressive.

Unfortunately, the ultimate outcome of WNS in North American bats is death—wing deterioration and loss of body mass make it impossible for bats to survive. The fungus produces a fuzzy white growth on the bat’s snout, wings, and ears, and sometimes penetrates into the bat’s hair follicles and sebaceous glands (oil producing), generating ulcers on the skin.

Above: A group of infected Myotis lucifugus, commonly known as little brown bats, in a cave in New York. The white “fuzz” on their muzzles is one of the hallmarks of white-nose syndrome. (Photo courtesy of Nancy Heaslip, New York Department of Environmental Conservation)

Of the nine species of bats that have thus far been affected by WNS, all are cave-dwelling, hibernating species. Hallam speculates that bats that live individually under tree bark or rock crevices simply don’t come in contact with each other enough to readily pass on the disease. Hibernating bats huddle en masse during the winter to preserve body heat, easily spreading WNS to one another as the months pass.

WNS is transmitted via direct contact with the fungus that causes it, primarily from one bat to another. In October of 2011, a paper in the journal Nature (also co-authored by Boyles) confirmed that the fungus Geomyces destructans causes WNS.

Bat specialists have noticed unusual behavior in the diseased bats, namely winter flight. Hallam suggests that this odd behavior may have decreased the speed at which the disease spread, particularly into the southern region of America. When Hallam’s team began studying the spread of WNS back in 2008 they had no idea what they were going to find, and their observations have been both promising and disturbing.

Modeling the disease

According to Hallam, the spread of WNS has slowed considerably in the past two to three years—the disease has been found no further south than Tennessee and has yet to surface farther west than Indiana (although the fungus has been located on single bats in Missouri and Oklahoma, it has not caused mass mortality). Observations and the team’s simulations indicate that temperature has something to do with slowing down the spread of WNS.

Above: A map showing the counties that have confirmed cases of white-nose syndrome. This map is updated as info from a number of sites is provided. (Image courtesy of batcon.org)

“The first time that we ran our code, our models showed more spread in the South than was actually occurring, so something in nature was moderating the fungus,” explained Jeff Nichols, a computational scientist at ORNL and a code developer for Hallam’s team. “That’s when we brought in temperature.” According to Nichols, WNS grows at temperatures between approximately 2 and 20 degrees Celsius (35 to 68 degrees Fahrenheit). This temperature range corresponds closely with the cave environments many bats choose for hibernation and as such, temperature becomes a key factor in how WNS spreads.

Unfortunately, data regarding temperature within caves across the U.S. is not available, so the team used average outside air temperature datasets from the NARR (North American Regional Reanalysis) archives in their code. In addition to temperature, the code took a number of other factors into account; locations of caves in the continental U.S., biology associated with bats, and the pathology of WNS, namely the mass mortality it causes.

The algorithm represents caves from all but four states—some states like Kansas don’t have caves. But because of the staggering amount of caves in the continental U.S. (approximately 55,000), the researchers used the geometric center of a county as a proxy for the entire county, rounding the number of caves in their simulations down to a reasonable 3,300.

The movements of roughly 4.15 million bats were simulated. Over the course of their lives, bats participate in a continuous cycle of three major movements: winter hibernation in caves, a swarming area for breeding in the fall, and a summer roost where females give birth. “This movement forms the timescale and distance scale that we’re interested in,” explained Hallam, adding that their movement can be unpredictable. “Bats do not necessarily hibernate in the same places each time,” further complicating how WNS spreads.

The team used 6,000 processors at a time on Kraken to run simulations predicting how WNS will spread over the next 15 years. While the Southern U.S. still saw moderated spread, nationwide spread is still predicted despite being slowed by the Great Plains and the Rocky Mountains.

Google Earth animation of predicted spread of WNS across the U.S. averaged over 500 runs on Kraken. (Animation generated by Tom Hallam's Research Group, University of Tennessee)

“Our simulations show white-nose reaching the edge of the plains in 2014, then slowing down in 2015 and expanding rapidly once it reaches Colorado,” said Nichols. He and Hallam explain that while biologists know little about bat populations in the East, even less is known about bats in the West. Much more research must be done before scientists can truly understand how WNS will move.

However, there is a more positive result from the simulations—there is a cave-rich corridor across southern Illinois that has not been affected by WNS. Hallam and his team have hopes that the caves of this area can be treated, possible with fungicides. If bats with WNS continue to spread west, this corridor acts as a transitional zone between the East and the Midwest. Once the bats reach these caves, the fungicides would treat the disease. Again, deeper research is necessary; no antifungal treatments have undergone substantial testing in cave environments, and biologists must be sure that any treatments used will not pose a threat to the natural ecosystem of a cave.

Like ecologists across the nation who are studying WNS, Hallam and his colleagues have much more to examine about this perplexing disease. Their future plans involve strengthening their models by incorporating internal cave temperatures, adding more caves, and integrating Canadian data.

About NICS: The National Institute for Computational Sciences (NICS) is a joint effort of the University of Tennessee and Oak Ridge National Laboratory that is funded by the National Science Foundation (NSF). Located on the campus of Oak Ridge National Laboratory, NICS is a major partner in NSF’s Extreme Science and Engineering Discovery Environment (XSEDE).