Quantifying the Dive Behavior of Coastal Sharks
Mentor:Nick Whitney, Staff Scientist, Mote Marine Laboratory
Developed in the early 2000s, high resolution data loggers provide the opportunity for a much more precise measure of tracking marine animals than was previously possible. However, with this increased detail, new methods of analyzing behavior must be developed. This is especially true for the diving behavior of coastal shark species, behavior that could not be accurately viewed using prior technologies. In this study, we demonstrate the significant increase in accuracy and descriptive ability
of hi-res data loggers over previous technologies as we compare two methods of classifying the dive behavior of sharks, a visual identification schema and a numerical method performed by wavelet analysis and K-means clustering. In this study, K-means clustering was shown to be much more time efficient than visual analysis. Additionally, it provides a consistent, objective measure while visual observation was more subjective. Furthermore, four of the six clusters were primarily (≥40%) associated with one of the visual dive types. Because of this, we suggest using numerical methods in future studies. In particular, work should focus on determining what behaviors are explicitly linked with each cluster through velocity, acceleration, and video feeds. These dive types can also be used to address questions of post-release recovery and metabolic activity.