To study dolphin populations and better understand how they’re affected by climate change and natural resource development, a group of scientists has created an algorithm that can recognize distinct dolphin click patterns among millions of clicks in underwater recordings.
Kaitlin Frasier of Scripps Institution of Oceanography, California, who led the study, makes autonomous underwater acoustic sensors with her colleagues to record dolphins’ echolocation clicks in the wild. These sensors are non-invasive tools that help researchers study various aspects of dolphin populations and how they’re affected by their ever-changing environment.
Though very useful for recording millions of dolphins’ clicks, these instruments don’t make it easy for humans to recognize any species-specific patterns in the recordings. That’s why the researchers decided to make use of machine learning to create a smart, automated algorithm that can uncover consistent click patterns in giant datasets.
In the study, the algorithm was able to recognize consistent click patterns in a huge dataset of over 50 million echolocation clicks that were recorded over a two-year period in the Gulf of Mexico. The team believes that the new algorithm could also help them distinguish between dolphin species in the wild.