SEAwise report on improved predictive models of natural mortality
The SEAwise project works to deliver a fully operational tool that will allow fishers, managers, and policy makers to easily apply Ecosystem Based Fisheries Management. This SEAwise report describes enhanced predictive models of fish survival to produce estimates of natural mortality for use directly in stock assessments. The five major challenges to using natural mortalities that vary over time are: limited knowledge about top predator impacts, in particular marine mammals, on commercial species; limited incorporation of effects of distributional changes; uncertainties around the food selection models used to simulate predator-prey interactions and resulting structural uncertainty in model output and lack of tools to export and review estimates of natural mortality for use in stock assessments.
The impacts of marine mammals on commercial fish was investigated in a study linking the diets of marine mammals to the availability of their prey in a multi-species functional response model estimating marine mammal diets based on estimates of fish abundance are available. There was a strong relationship between prey availability and consumption by seals.
A key factor when using models to predict natural mortality and stock biomass at very high or low fishing mortality is the ability of the model to stabilize the dynamics in these cases. Various feeding responses were tested in a simplified system but none of the current implementations attained the stabilisation effect. Concurrent cases of low prey stock size, low prey fishing pressure and low predator fishing pressure can lead to further instability. This effect may be altered if spatially explicit refuges for the prey species exist. The predictive ability of multispecies models estimating natural mortality was only sometimes enhanced compared to single species assessments. However, the choice of whether to use variable or constant natural mortality greatly impacted the estimated stock recruitment relationship and reference points for stock biomass based on this. Measures of the variability in the estimated natural mortality between years was estimated to provide guidance for Management strategy evaluations in single species environments. The estimation was compared between different areas for the same species, providing a general guidance on the likely level of variability in natural mortality for different species.
Predation mortality is highly sensitive to fishing and increases markedly for all stocks in the absence of fishing as large predatory fish populations rebuild. As fisheries management improves and predatory stock recover, realised biomass gains will be less than those projected today due to higher natural mortality.
Distinct models are currently being developed for the Celtic Sea, Bay of Biscay and Western water and will be integrated for the next SEAwise deliverable on natural mortality.
