Analysis of iPod-recorded audio can track the struggles of endangered species

Using iPods, microphones, and a computer relay system, researchers from the University of Puerto Rico have built a network that can track the fluctuations of animal and insect populations throughout the tropics — even down to the hour. The tracking technology is important for determining the vitality of species' populations in the long-term, especially as climate change alters their habitats. But it's been difficult to do this with existing methods, which often required careful human measurement.
 
AN IPOD, A SOLAR CHARGER, AND A COUPLE COMPUTERS:
Instead, the researchers built a system that's almost completely automated: Audio is collected by iPod touches that are continually powered by a solar charging station. An app records a single minute of audio every 10 minutes and then forwards it along through a nearby computer system and onto an online database. From there, an audio analysis system determines what species all of the recorded chirps, shrills, and shrieks come from — and it can reportedly analyze more than 100,000 of these clips every hour.
The only human input comes in training the computer to properly analyze the audio. In a series of sample clips, a researcher must first isolate the different sounds made by any species that could be present in the field recordings. Once it learns, the system is fairly accurate — though it isn't perfect. The researchers noted that it sometimes misidentified environmental noises like thunder and wind as apes and frogs. But as the study's lead author, Mitchell Aide, says, it might be the best alternative: "Sending biologists to the field is expensive ... and it is impossible to clone expert field biologists so that they can monitor various sites simultaneously."
While the system provides nearly immediate access to population levels, it'll primarily be used as a way to track animals and insects over years — not hours or days. The system is being detailed in a paper published today on PeerJ, but the research team has already tested it for five years while monitoring an endangered frog. The frog's population declined over the first four years, but shot up during the fifth. It's reportedly a common pattern, but the continued long-term measurements enabled the team to identity it. Now the team hopes that its persistent analysis system can be used elsewhere, providing researchers with an incredibly detailed picture of the environment — and a massive catalog of audio recordings to back it up.