Skip to main content

    A preliminary model-based approach for estimating natural mortality of Antarctic toothfish (Dissostichus mawsoni) in the Ross Sea Region

    Request Meeting Document
    Document Number:
    B. Moore, S. Mormede, S. Parker and A. Dunn
    Submitted By:
    Mr Alistair Dunn (New Zealand)
    Approved By:
    Mr Alistair Dunn (New Zealand)

    Natural mortality (M) is a key parameter in fisheries stock assessments, relating directly to the productivity of a stock, stock status, and estimated precautionary yields. In the Ross Sea region stock assessment for Antarctic toothfish (Dissostichus mawsoni), M is assumed to be 0.13 y-1 and constant across age and sex.

    In 2018, the independent review of the Ross Sea region D. mawsoni stock assessment recommended that CCAMLR consider the robustness and reliability of estimating natural mortality in the stock assessment model. We undertake a preliminary investigation of estimating natural mortality using the assessment model from the Ross Sea region from 2017.

    We estimated natural mortality as a single value for males and females combined, with a uniform prior within the base model used in 2017 for Antarctic toothfish in the Ross Sea region. The model MCMCs estimated M=0.11 y-1 (0.09–0.13 y-1) with B0 estimated to be 92 140 t (70 330–121 950 t); a higher B0 72 620 t (65 040–81 050 t) when compared to the 2017 base case where M=0.13 y-1. The precision of the estimate of B0 when estimating natural mortality was lower and estimated the precautionary yield higher (3419 t compared with 3258 t). MPD simulations suggested that when the rate of natural mortality was assumed in an operating model, it was estimated with relatively high precision and with no bias.

    Further investigations will need to be carried out to consider the effect on model fits and residuals and if robust estimates of the rate of natural mortality can be achieved using MCMC simulations. We recommend that analyses exploring the effect of the model-based estimate of M, including values of natural mortality that vary by age and sex, should be considered in future stock assessment sensitivity analyses