Estimation of reference points and a precautionary harvest strategy for the Razor Clam (Siliqua patula) fishery at Haida Gwaii

Abstract

Beaches near Massett, Haida Gwaii have supported commercial fisheries for Razor Clams, Siliqua patula, for over 100 years and have been managed using a catch quota since 2001. Biomass has been surveyed annually since 1994 using hydraulic core sampling and a three-stage sampling design stratified by beach area and beach level. From 2001 to 2008 the fishery was managed using an annual 0.123 harvest rate (instantaneous fishing mortality 0.131 yr¹) that was estimated based on a surplus production model and one year of data. Biomass estimates from 1994 to 2008 ranged from 560 to 1,853 t (mean 966 t). Sustainable yield reference points were estimated using a publicly available program, Yield 1.0, that simulates fishery yields and estimates fishery reference points for user-specified parameters. The input parameters included an estimate of instantaneous M, a Razor Clam growth relationship, and a Ricker stock-recruit relationship based on 15 years of data. The annual natural mortality rate M was estimated to be 0.23 yr¹(coefficient of variation [CV] 0.4 and instantaneous M 0.26 yr¹) based on catch curves constructed from the annual length frequency distribution and ages of clam shell samples. The instantaneous F_msyreference point was estimated to be 0.35 (95% confidence limit [CL] 0.06–0.57), B_msy as 574 t, and unexploited fishery biomass as 1,390 t. A limit reference point at 230 t (0.4 B_msy) and an upper stock reference at 459 t (0.8 B_msy) are recommended. The recommended annual harvest rate is 0.21 based on F_0.2(instantaneous F 0.24, 95% CL 0.12–0.50). The analysis looks at trade-offs among harvest rate at specified reference points, average annual yield, and resulting biomass generated by equilibrium and transient analyses, the latter incorporating interannual recruitment variability. Parameter uncertainty played an important role in the analysis as demonstrated by the wide confidence limits for the various reference points.