Ornithology has always been a bit like Wikipedia–that is, crowdsourced. Since the early days of naturalism, professional bird researchers have always gotten a hand in their data collecting from amateur bird watchers.
In 2002, there was an attempt to formalize this relationship with the eBird project, a joint venture of the Cornell Laboratory of Ornithology and the National Audubon Society. That project has been so successful, with 48 million observations entered over the last eight years, that researchers have now amassed more data than they can crunch.
Unless, that is, they could score 100,000 hours on the massive National Science Foundation TeraGrid supercomputing network, which is exactly what they’ve just done, reports Nature.
On a single processor, it would take about 10 days just to run a single year’s worth of data on a single species. “But we have approximately 700 species we want to run, and five years’ worth of data,” Steve Kelling of the eBird Project tells Fast Company. “If we’re gonna do this in my lifetime, we need to begin to parallelize the process.”
The eBirders will also be combining their data with satellite data to learn more about how spring blooms correlate to bird migrations. Already the eBird data has revealed new information about the importance of the U.S. Gulf Coast for the migrations of the passerines, a type of songbird.
Though the spirit of citizen science has always been strong in ornithology, says Kelling, “the advent of the Internet has really opened up opportunities for volunteer engagement.” If it takes fleets of volunteer naturalists, the hivemind that is the Internet, plus some of the strongest supercomputers we have to understand their behavior, then maybe bird brains can’t be so dumb after all?