Validating a method of predicting timing of hatch of American lobster, Homarus americanus, in nature

Abstract

Models have been developed in the laboratory to predict hatch of American lobsters based on temperature and embryonic eye-size. These models have not, however, been validated in different locations in nature. I compared hatch predictions using two models that differed only in the assumed functional relationship between temperature and embryo development rate to observed hatch in two locations in Atlantic Canada. The best model predicted 91-100% of the range of observed hatch dates and 90-95% predictions fell within the observed hatch periods, although it was not the same model for each location. Although I was able to predict the mean hatch date at the population level, on average only 31% of predictions fell within the observed hatch periods of individual lobsters. Models appear useful to predict hatch, however future research is necessary to determine what influences model success in different locations.