How Supercomputers Are Helping Save Endangered Species

Ecologists get a leg-up from Big Data.

Habitat models produced with the help of a supercomputer. Image: PLOS One

It's always exciting when ecological conservation gets a techie boost, and that's exactly what is happening at the San Diego Supercomputer Center (SDSC). The SDSC has partnered with the US Geological Survey and the San Diego Zoo to track the range of three endangered species in 3D using the center's supercomputers. The team published their results in PLOS One this month.

The result is a dramatically improved method for tracking threatened wildlife. One immediate benefit of using such a powerful setup is the reduced processing time. “We were able to speed up their software by several orders of magnitude,” said Robert Sinkovits, director of SDSC’s Scientific Applications Group, in a release. “In this case, calculations that had formerly taken four days to complete were finished in less than half an hour.”

But beyond the quick turnaround, the partnership's 3D models provide researchers with a much more detailed picture of the migratory patterns of terrestrial, avian, and aquatic animals. The team selected three species representing land (giant pandas), air (California condors), and ocean terrain (dugongs, a close relative of the manatee) to track with with movement-based kernel density estimators (MKDEs), which collectively are one method of modeling how species move throughout habitats, which provides crucial data for conservation efforts.

This figure from the paper shows how the team builds range and habitat models in three dimensions. Image: PLOS One

“Someone read a draft of our paper and said, wow, this list reads like ‘The Justice League’ of endangered species,” wrote lead author and computational ecologist Jeff Tracey in a USGS blog post. “They’re definitely all very charismatic species of conservation concern. I suppose the California condor would be Superman and the dugong would be Aquaman, but I wonder who the giant panda would be.”

Maybe some kind of anti-Flash? Regardless, instead of vigilante-style masks and capes, individuals from each species sported collars with MKDEs and GPS units. These instruments recorded the animals' movements in vertical space as well as depth and width. The team put one of their 3D maps, which represents the movements of a breeding pair of condors, onto YouTube, and the video is worth watching on its aesthetic merits alone.

Supercomputer map. Video: YouTube/SDSC. 

Of course, there are also many practical applications for the maps. “If you wanted to find out where condor home ranges intersect with planned turbine sites using traditional 2D analysis, you just stack the two maps and pick out areas that overlap,” said co-author James Sheppard of the San Diego Zoo in the USGS post

“But both condors and turbines occupy 3D space: condors fly at certain altitudes, while turbines protrude into the airspace to a certain height and volume," he continued. "A simple 2D analysis can’t take this into account, and might inaccurately estimate the risk of bird-turbine conflicts, depending on the geography at hand.”

The same logic applies to the vertical movements of dugongs, pandas, and whatever other species get hooked up with MKDEs. Accordingly, the same preventative measures can be taken by industries and activities that encroach on their habitat. It's a fascinating result, and should the technology get deployed further, data-driven conservation could have a big impact.