Condor Conservation in Three Dimensions

Can gadgets and processors save endangered species? Advances in biotelemetry combine with the power of supercomputing to keep California condors safe.

A tagged California condor. Image Credit: James Sheppard

A tagged California condor. Image Credit: James Sheppard

The California condor has a leathery pinkish head and a magnificent collar of tufted black feathers. The largest bird in North America, the condor uses an immense 9.5-foot wingspan to soar over the rocky landscape of southern California.

According to James Sheppard, a spatial ecologist at the San Diego Zoo Institute for Conservation Research, the California condor is “a major conservation success story.” These glorious creatures declined quickly over the course of the 20th century, the victims of intentional shooting and unintentional lead poisoning. By 1982, only 22 birds remained.

In 1982 the San Diego Zoo, collaborating with the U.S. Fish and Wildlife Service, California Department of Fish and Game, National Audubon Society, and the Los Angeles Zoo, began an intensive intervention program, rearing condors at the San Diego Zoo and then releasing them, outfitted with tracking devices, into the wild.

There are now roughly 410 condors in the world, half of them in captivity and the other half free-flying, according to Sheppard. Thanks to the hard work of conservationists, the condor population is healthier than it’s been in more than 30 years.

Now California condors are facing a new threat from an unlikely source: renewable energy. Condors soar on the wind. Their current home range, between the Pacific Ocean to the west and the Mojave Desert to the east, comprises the southern Sierra Nevada Mountains and Baja, California, where the winds are strongest and most consistent.

Not surprisingly, this wind profile is equally attractive to energy companies developing large-scale wind projects. This presents a growing danger to condors and other species. According to Robert Fisher, a biologist at the Western Ecological Research Center of the U.S. Geological Survey (USGS), the north-south corridor is a biodiversity hotspot.

Some wind farm developers are working with conservationists to manage the problems. Recently, Sempra Energy, an oil and gas company which has proposed a wind farm in Baja, teamed with Sheppard at the San Diego Zoo to fund a study on the potential collision danger that the wind farm may present for condors.

A map of the condor home range (in red) and the proposed wind farm in Baja, California. Image Credit: Tracey, et. al, PLOS ONE

A map of the condor home range (in red) and the proposed wind farm in Baja, California. Image Credit: Tracey, et. al, PLOS ONE

Sheppard was prepared for the challenge. Using a brand-new spatial ecology lab, Sheppard collects data from telemetry tracking devices on condors and analyzes it on high-spec computer workstations with top-end industry-standard geospatial software.

“We really are in the golden age of animal biotelemetry,” Sheppard says. “… We’ve got a 45-gram, solar-powered GPS transmitter that slips onto the bird’s wing and will last more than a year.” Once a device is deployed on a bird, Sheppard can download the bird’s location data straight off the Internet.

The devices return location data in three spatial dimensions: north-south, east-west, and depth-height. Ecologists use the data to map animal home ranges, because home range mapping gives ecologists information about how animals interact with the landscape, crucial to informed conservation planning.

But despite receiving location information for the x, y, and z axes, ecologists have only mapped home ranges in two dimensions.

Sheppard realized that nobody had yet come up with a workable algorithm for integrating the z data into a home range map, so the study was his opportunity to make use of the vertical location data that was being harvested. To do this, he turned to Jeff Tracey, a computational ecologist at USGS working with Fisher to develop two-dimensional home range maps for bobcats and other southern California fauna.

For the condor study, Tracey developed a three-dimensional, movement-based kernel density estimator (3D MKDE). A kernel density estimator is an algorithm used to compute the missing information between data points. Tracey used a methodology developed in 2007 by John Horn, called a Brownian bridge.

“It’s based on probabilistically using the assumption of Brownian motion between the points,” Tracey says. With this formula, Tracey could map a much more accurate condor home range from the hourly data. Unfortunately, processing the data with an extra dimension thrown in remained a time-consuming computing problem.

“We found that moving into the third dimension really gobbled up a lot of computing capacity,” Sheppard says. Tracey could create the algorithm, but without much greater computing power to process and analyze the data, its potential would remain unrealized.

Enter the San Diego Supercomputing Center — a research unit of UC San Diego and a leader in data-intensive computing which houses massive computers with high performance storage and processing capacity.

Bob Sinkovits, director of the Scientific Applications Group at the Center, has the responsibility of approving scientists’ requests for supercomputing help. When he saw the proposal, he was excited by the chance to help with conservation research.

“Our uses are mostly dominated by the hard sciences,” Sinkovits says, “so it was a very novel domain for us.”

Sinkovits took Tracey’s 3D MKDE software and refined it further.

“(Tracey) had a working version of the software, and my specialty is that I work on other people’s software to make it run faster,” Sinkovits says. “So I was able to speed up the performance by roughly a thousand times.”

3D home range model calculated using GPS tracking data acquired from a wild California condor. Image Credit: James Sheppard

3D home range model calculated using GPS tracking data acquired from a wild California condor. Image Credit: James Sheppard

Previously it had taken half an hour or more to compute the home range for one bird for one month’s worth of data, but Sinkovits got it down to mere seconds, allowing the researchers to process much more data in a much shorter period of time. Calculations that would have taken hundreds or thousands of hours could now be done in about 20 minutes.

“So the problems they’re working on can be solved much faster, but even more important, they can start doing problems they hadn’t really thought of before,” Sinkovits says.

The results of the study were published in 2014, but the plight of the condor is ongoing. According to Fisher, the wind farm in Baja is now in the process of being built.

Officials from Sempra Energy could not be reached for comment.

According to Fisher, the company has developed a “single-species conservation” plan. The company plans to turn off the windmills when its sensors detect a condor in the area. The company can follow each bird and deploy evasive maneuvers to prevent incidents. This solution won’t work for all the affected species in the area, but for California condors, the future looks bright, Sheppard says.

“It’s great to work with the California condor program because it is a very hopeful message,” he says. “Now you can actually go out on a cliff face and see a bird with a nine-and-a-half-foot wingspan soar over you, and that really drives home why we’re actually doing this.”