Estimation of annual, time-varying natural mortality and survival for Eastern Bering Sea snow crab (Chionoecetes opilio) with state-space population models

Abstract

Sex-specific, state-space population models with size and stage structure and fit to survey and fishery data were developed for the eastern Bering Sea (EBS) snow crab for the years 1982–2017. Motivation for model development was to estimate population dynamics by shell condition for female crabs; estimate sex-specific annual, time-varying natural morality rates; and explore in general the feasibility of state space modeling for EBS snow crab. Model fits were very good and mean natural mortality estimates agreed with previous studies but several quite high and/or quite low annual natural mortality values were estimated for both sexes. Estimated mean female natural mortality was 0.49 yr−1 (s.d. = 0.01) with annual values from 0.04 yr−1 to 3.34 yr−1. Estimated mean male natural mortality was 0.36 yr−1 (s.d. = 0.07) with annual values from 0.03 yr−1 to 0.91 yr−1. Episodic high and low annual natural mortality estimates indicate potential model overfitting which an autoregressive or random walk estimation framework for natural mortality may address. We consider the models as proof-of-concept estimation frameworks and their results preliminary. After further refinement and testing, they could be a complementary approach to ongoing stock assessment modeling or prototypes for state-space assessment models. Additional process variability, such as for growth and maturation, can be incorporated in future work. Abundance estimates by mature female shell condition makes feasible estimation of annual and biennial spawner abundances, necessary for accurate egg production estimation. State-space population dynamics models of other Chionoecetes populations with both size and stage-based data (e.g., maturity status or shell condition) can be developed utilizing this study’s modeling framework.