Biomass assessment of a deep-water fish, the orange roughy (Hoplostethus atlanticus), based on an egg survey

Abstract

An egg survey was carried out in 1992 to estimate the biomass of the orange roughy stock that spawns off north-eastern Tasmania. Spawning occurs at depths of 700-1000 m around a single seamount from early July through early August. Orange roughy have determinate fecundity, and their fecundity and the depth distribution and development rate of the eggs are described elsewhere. A random stratified survey was designed for a limited area around the spawning site in order to sample the eggs fully during their first day of development. The eggs were sampled with vertical tows from 1000 m to the surface. There was no evidence of significant mortality or other egg loss during the first day after spawning, when the eggs are mostly below 400 m. Subsequently, egg numbers declined rapidly at an instantaneous loss rate of 0.036 h^-1, primariIy through advection out of the survey area. The estimated biomass of spawning fish was 27 445 t. From the proportion of non-spawning fish in the population (29% of females and 10% of males) and estimated catch of spawners, total mid-season stock biomass in 1992 was estimated to be 34 593 t with a coefficient of variation (CV) of 47%. A bootstrap estimate of the variance indicated a somewhat lower CV of 41%. The point source of spawning activity appears to give rise to a highly patchy egg distribution, which results in the high variance of the biomass estimate. From the catch history of the fishery, virgin biomass of the stock was estimated to be approximately 96 900 t, and the stock was at 28% of virgin biomass at the end of season. An acoustic survey of the spawning orange roughy provided a similar estimate of stock size. The acoustic survey had greatly reduced statistical confidence limits (CV = 6%) but far greater non-statistical sources of uncertainty, i.e. the species composition of acoustic targets and the target strength of these species.