Life-history traits inform population trends when assessing the conservation status of a declining tiger shark population

Abstract

The assessment of the conservation status of wide ranging species depends on estimates of the magnitude of their population trends. The accuracy of trend estimates will depend on where and how many locations within a species' range are sampled. We ask how the spatial extent of sampling interacts with non-linear patterns in long-term trends to affect estimates of decline in standardised catch of tiger sharks (Galeocerdo cuvier) on the east coast of Australia. We apply a Bayesian trend model that uses prior information on life-history traits to estimate trends where we use data from all regions versus spatial subsets of the data. As more regions were included in the model the trend estimates converged toward an overall decline of 71% over three generations. Trends estimated from data only from northern regions or southern regions underestimated and overestimated the regional decline, respectively. When a subset of regions was modelled, rather than the full data-set, the prior informed by life-history traits performed well, as did a weakly informed prior that allowed for high variation. The rate of decline in tiger sharks is consistent with a listing East Coast Australia tiger sharks as endangered under local legislation. Monitoring programs that aim to estimate population trends should attempt to cover the extremes and mid-points of a population's range. Life-history information can be used to inform priors for population variation and may give more accurate estimates of trends that can be justified in debates about the status of threatened species, particularly when sampling is limited.