Predicting delayed post-harvest mortality in American lobster (Homarus americanus) using reflex impairment

Post-harvest mortality in the American lobster (Homarus americanus) fishery can result in a significant loss in revenue for the largest single species fishery in North America. At least 2% of American lobster landed in Maine, USA die before they reach consumers. The landed value of the Maine lobster fishery was $730 million USD in 2021. Yet, the industry loses roughly 952 metric tons annually due to post-harvest mortality each year. The lobster supply chain is a network of harvesters, dealers, and distributors that facilitates the transport of live product domestically and internationally. The majority of product loss comes in the form of delayed mortality as a result of stressors within the supply chain. Because of the high volume of lobster transported through the supply chain and its many links, a standard protocol is needed to quickly diagnose whether a high-value live lobster will survive the trip to the consumer. Based on empirical observations, we developed a reflex action mortality predictor (RAMP) model to reliably predict subsequent mortality, days after exposure to the supply chain. A suite of physical characteristics, injuries, and discrete reflex actions of 975 lobsters that were monitored at two standard lobster dealer facilities were recorded to build logistic RAMP models. Results suggest that body size along with four specific reflex actions and five types of injury are significant predictors of mortality five days later. The reflex actions were eye motion, pereiopod motion, 3rd maxilliped retraction, and 2nd maxilliped motion. The five injuries were missing chelae, damaged chela, damaged antenna, damaged carapace, and damaged uropod. The RAMP model can be an important tool in identifying supply chain stressors that impact lobster quality and inform efforts to improve the efficiency and resiliency of the industry. The highly predictive, non-invasive, quick, and cost-effective nature of this method has potential to become a versatile tool for both industry and scientific applications.