This web application uses cookies and other tracking technologies to ensure you get the best experience.
SC/68C/ASI/02
Download this file922 KB |
Download |
View directly in browser922 KB |
View in browser |
Resource ID
19255
Access
Open
Document Number
SC/68C/ASI/02
Full Title
Humpback whale Breeding Stock G: Updated population estimate based on photo-id matches between breeding and feeding areas
Author
Fernando F?lix, Jorge Acevedo, Anelio Aguayo-Lobo, Isabel C. ?vila, Natalia Botero-Acosta, Andrea Calder?n, Benjam?n C?ceres, Juan Capella, Romina Carnero, Cristina Castro, Ted Cheeseman, Luciano Dalla Rosa et al.
Publisher
IWC
Publication Year
2021
Abstract
We report a new mark-recapture-based population estimate for the humpback whale Breeding Stock G (BSG), defined by breeding grounds on the northwestern coast of South America and southwestern Central America and feeding grounds around the Antarctic Peninsula and southern Chile. Photographic fluke catalogs from 23 research groups working in both breeding and feeding areas were compiled in the largest photo-ID matching effort ever made for this stock. A total of 6,354 unique individuals including 1,698 (26.7%) from feeding areas and 4,656 (73.3%) from breeding areas covering the period 1991-2018 were used for this purpose. The dataset was fitted to closed population models to estimate population size and Jolly-Seber models to estimate apparent survival, both implemented in the software Mark. Mixture models with two different data types, full likelihood and conditional likelihood, produced similar results of 11,784 and 11,786 (SE = 266 for both estimates) whales, respectively. In both cases, a model with two mixtures {Mth2} provided the best fit. Two Cormack-Jolly-Seber with Pledger mixtures models produced apparent survival estimates for the two mixtures (0.924 and 0.959, SE = 0.003 and 0.008; respectively). The new population estimate is 181% higher than a previously obtained in 2006. The annual rate of increase in the 27-year study period was 5.07%. Sources of bias were associated with effort heterogeneity, population stratification and the time scale. These and other sources of bias should be considered in future modeling estimates.