Refining mortality estimates in shark demographic analyses: a Bayesian inverse matrix approach

Leslie matrix models are an important analysis tool in conservation biology that are applied to a diversity of taxa. The standard approach estimates the finite rate of population growth (lambda) from a set of vital rates. In some instances, an estimate of lambda is available, but the vital rates are poorly understood and can be solved for using an inverse matrix approach. However, these approaches are rarely attempted due to prerequisites of information on the structure of age or stage classes. This study addressed this issue by using a combination of Monte Carlo simulations and the sample-importance-resampling ( SIR) algorithm to solve the inverse matrix problem without data on population structure. This approach was applied to the grey reef shark (Carcharhinus amblyrhynchos) from the Great Barrier Reef (GBR) in Australia to determine the demography of this population. Additionally, these outputs were applied to another heavily fished population from Papua New Guinea (PNG) that requires estimates of lambda for fisheries management. The SIR analysis determined that natural mortality (M) and total mortality (Z) based on indirect methods have previously been overestimated for C. amblyrhynchos, leading to an underestimated lambda. Updated distributions of Z and lambda were produced for the GBR population and corrected obvious error in the demographic parameters for the PNG population. This approach provides opportunity for the inverse matrix approach to be applied more broadly to situations where information on population structure is lacking.