Robust assessment of population trends in marine mammals applied to New Caledonian humpback whales
Résumé
Estimating population trends for long-lived, migratory animals is challenging but essential for managing populations. We propose to use a simple but potentially robust method, the direct estimation of population growth rate (PGR) from capture recapture data. As motivating study, we consider an endangered population of humpback whales Megaptera novaeangliae breeding and calving annually in the south lagoon of New Caledonia. Studied since 1996, this population is known to exhibit a strong signal of transience, i.e. the presence of individuals that pass through the sampling area once, inducing detection heterogeneity. Another difficulty is that a second breeding and calving habitat (offshore seamounts to the south) more recently discovered has been surveyed with less intensity. Now, the current direct PGR estimation models cannot deal with spatial sampling heterogeneity. In order to assess the reliability of the proposed method-in general and for our population in particular-we evaluated its robustness using simulations: first, when there are transient individuals; then, when the study area is split into two unequally sampled parts. We found no bias in PGR in presence of transients. The bias with two unequally sampled parts depends on the amount and direction of exchanges, but appears negligible in our case study. The constant yearly PGR of the New-Caledonia humpback whale population at the level of the two habitats was estimated at 1.15 [1.11; 1.20], hinting at outside recruitment. Whenever capture-recapture is feasible, we recommend the Pradel approach to estimating the PGR, validated with appropriate simulations, in order to assess population welfare. (250)
Origine : Fichiers produits par l'(les) auteur(s)