A sensitivity analysis of a simple krill individual-based model designed to investigate length-based recruitment indices

As part of the CCAMLR Scheme of International Scientific Observation, observers on-board Krill (Euphausia superba) fishing vessels measure the length of individual Krill sampled from commercial catches. These extremely valuable data may be used to monitor the state of harvested Krill populations. For instance, investigating length frequency distributions (LFDs) and their change over time provides insight into the underlying population dynamics. Three main processes drive temporal changes in LFDs: growth, mortality and recruitment. Understanding how these processes interact is a prerequisite to the interpretation of any observed trend. To illustrate these interactions a simple mechanistic individual-based model of krill population dynamics was subjected to a sensitivity analysis. Starting from a base scenario of cyclical annual recruitment, plausible ranges of growth, mortality and recruitment rates were tested and their effects on two length-based recruitment indices (the monthly median length and proportion of individuals smaller than 40mm, F40) were investigated. Results indicate that the annual span (maximum minus minimum) of F40 would be the best index of annual recruitment among those tested, and, that length-based indices may be used to determine the timing of recruitment events. However, since the population size structure at a given time is the result of a mixture of several annual cohorts, using such indices to quantify the intensity of a given recruitment event would need to take into account the relative magnitude of previous recruitment events.