Assessing the factors influencing discard mortality of demersal fishes using a semi-quantitative indicator of survival potential

Understanding the factors affecting the likelihood that discarded fish will die can contribute to better management of resources by enhancing the potential for successful live release and by improving the estimation of otherwise unaccounted fishing mortality. Semi-quantitative measures of individual fish vitality or physical condition, obtained by at-sea observers aboard commercial fishing vessels, are often used as an indicator of survival potential for discarded fish. The present study and previous ones have shown that these measures relate well to eventual survival. However, observer subjectivity in fish vitality scoring can affect the precision and accuracy of inferences drawn from an analysis of the observations. Here we propose the use of a mixed-effects multinomial proportional-odds model, which is appropriate for modelling ordinal vitality data and is a useful approach for addressing observer scoring subjectivity. This model was used to analyse data collected for eleven fish taxa captured by four gear types. The effect of eight factors previously shown to affect discard survival was evaluated. The gear type used and amount of time that fish spent on deck prior to discarding most strongly and consistently affected the distribution of fish among vitality levels. Sea surface and air temperatures, and fish body size, were also important factors for a number of taxa, while other factors such as the depth fished, catch size and fishing activity duration were important only for certain taxa. A random effect in the model, used to account for observer subjectivity, was significant for most taxa and fisheries. Failure to account for this effect could affect both the precision and accuracy of inferences on the survival potential of discarded fish.