A model-based approach to standardizing American lobster (Homarus americanus) ventless trap abundance indices

Abstract

Fishery-independent ventless trap surveys are an integral component to assessing American lobster (Homarus americanus) population trends, as they can sample complex, heavily fished habitats where most survey gear has difficulty accessing. U.S. American lobster stocks have been assessed within state waters using a standardized ventless trap survey since 2006. However, confounding survey attributes that may contribute to catch variability have not been investigated and some discontinuity in sampling has resulted in missing estimates of abundance. We constructed sex- and stock-specific generalized linear mixed models to discern the dynamics between lobster catch and individual survey factors and removed these sources of variability when producing continuous abundance indices. Soak time, day of year, and unique site had measurable contributions to the variability in lobster catch per ventless trap. Generally, sex- and stock-specific abundance indices from this model-based approach and a traditional design-based approach exhibited similar trends. The two approaches’ magnitudes and trends for the Gulf of Maine were nearly identical. For Southern New England, model-based index trends were smoother and lower in magnitude than the design-based estimates. The greatest difference in the two approaches’ trends for the Southern New England indices were in the early and terminal years. This work serves as an example of how variability associated with fixed and random effects of a survey can be accounted for when producing abundance indices used in stock assessments.