Machine Learning Analysis Reveals Environmental and Demographic Patterns of Skin Lesions and Scars in Bottlenose Dolphins

Tamanend’s bottlenose dolphins (Tursiops erebennus) in the Chesapeake Bay are important ecosystem sentinels in a critical but underrecognized habitat, and have been threatened by major disease-related mortality events. Despite this, researchers have yet to conduct a deeper investigation into the correlation among environmental factors (such as temperature and salinity), demographic factors (such as age and sex), and disease susceptibility in these dolphins. Both skin lesions and rake mark scars can be used as potential indicators of disease susceptibility in dolphin populations, as skin lesions often indicate underlying disease, and rake marks are associated with other stressors that may increase a dolphin’s susceptibility to disease. Analysis of these markings requires photographic data, for which manual processing is inefficient, often resulting in long lags between data collection and the complete analysis needed to inform management decisions. This study uses modern machine learning tools to assess indicators such as the prevalence of skin lesions and extent of rake mark coverage across dolphins photographed in the Chesapeake Bay to test hypotheses about the potential patterns between environmental factors, demographic factors, and potential disease susceptibility indicators. I found a significant association between lower temperature and lesion prevalence and between lower salinity and lesion prevalence for one lesion class thought to be most associated with infectious disease. I also discovered that adult male dolphins had higher rake mark coverage compared to adult females or calves, but no significant difference could be found between rake mark coverage for the latter two demographic classes. These findings could suggest a higher susceptibility to disease for adult male dolphins and more generally dolphins in colder, less saline waters. My work provides new understanding into the trends that may drive disease among T. erebennus in the Chesapeake Bay.