Temporal evolution of the Mediterranean fin whale song

Area of Science:

• Marine Biology
• Bioacoustics
• Animal Vocalizations

Background:

• Fin whale songs are crucial for understanding their behavior and population dynamics.
• Passive acoustic monitoring is a key tool for studying marine mammals in their natural habitat.
• Previous studies have identified distinct fin whale pulse types and inter-pulse intervals (IPI).

Purpose of the Study:

• To analyze fin whale songs using passive acoustic recordings from the Pelagos Sanctuary.
• To develop and apply a convolutional neural network (CNN) for automatic detection of 20 Hz fin whale pulses.
• To investigate temporal trends in fin whale song characteristics, including IPI and pulse center frequency.

Main Methods:

• Utilized passive acoustic recordings from 2008-2018 in the Pelagos Sanctuary.
• Employed a low complexity convolutional neural network (CNN) for automated detection of 20 Hz fin whale pulses.
• Classified detected pulses and measured inter-pulse intervals (IPI), analyzing temporal variations.

Main Results:

• Successfully automated the detection of 20 Hz fin whale pulses using a CNN, robust to data variability.
• Confirmed the local relationship between pulse type and IPI with a significantly larger dataset.
• Identified inter-annual shifts in IPI and an intra-annual trend in pulse center frequency.

Conclusions:

• The developed CNN provides an effective tool for fin whale pulse detection in diverse acoustic environments.
• The study confirms and expands upon previous findings regarding fin whale song structure and variability.
• Results offer valuable long-term data for understanding global fin whale song trends and behavioral ecology.