Scientists studying the calls made by killer whales and pilot whales have a big problem: these whales talk too much.
Because they make so many different sounds it is very hard to work out what these noises might mean. A first step would be to understand the typical sounds these animals make, and that’s where volunteers visiting Whale.FM can help.
Robert Simpson of Oxford University, one of the researchers behind the project, told Scientific American ’s Mariette DiChristina:
‘When you visit Whale.FM, you are presented with a sound clip of a recording of a whale. The idea is to match the big sound that you see/hear with one of the smaller ones underneath.
‘All the pairings go into a database and we use that to find the best pairs of sounds and build up our understanding of what the whales are saying to each other. Basically: we need help decoding the language of whales.’
Whale.FM is the latest in a series of ‘citizen science’ projects led by or involving Oxford University scientists (others include Old Weather , Ancient Lives , and Galaxy Zoo ) and is a collaborative effort of Scientific American , Zooniverse and the research institutions WHOI , TNO , the University of Oxford, and SMRU.
One of the questions always raised by citizen science projects is whether volunteers can perform tasks as well as professional scientists. To test this the team took a selection of calls where they already knew the call category and tested them against the groupings given by people visiting the site.
‘We found that the Whale.FM volunteers grouped up the sounds in the same way that professionals would,’ Robert comments. ‘It agrees very well. We found approximately 90 percent agreement in our preliminary test. Our volunteers are amazing!’
Perhaps the biggest surprise, he says, is that their results could be used to help improve automated algorithms for decoding whale sounds:
‘There are tens of thousands of whale calls out here. It would seem that Whale.FM can help narrow the big problem into a smaller, more manageable one.’