Grove, Thomas; King, Ruth; Henry, Lea-Anne; Stevenson, Andrew (2022): Modelled annual abundance of humpback whales Megaptera novaeangliae around Bermuda, 2011-2020 [dataset]. PANGAEA, https://doi.org/10.1594/PANGAEA.945442

Modelled demographic parameters of North Atlantic humpback whales (Megaptera novaeangliae) visiting Bermuda are provided for every year from 2011 to 2010. The data set was constructed to determine abundance and abundance trends of humpback whales visiting Bermuda, a migratory stopover.

Photographs of the ventral side of humpback whale tail flukes were taken during dedicated vessel-based surveys between December and May each year between 2010 and 2020, and were daily whenever possible (weather-permitting). Surveys followed a haphazard regime to maximize encounters with whales and focused spatially on the southwestern Bermuda Platform and the Challenger Bank seamount*. Surveys were conducted in closing mode; whales were approached and photographed on detection, and a focal follow was conducted at each encounter. Images of suitable quality were used for individual photo-identification via scarring patterns and coloration (Calambokidis et al., 2001). The resulting catalogue was used to construct annual sighting histories for each identifiable animal.

Annual demographic parameters, including abundance, were reconstructed from sighting histories using capture-recapture (CR) methods. A Cormack-Jolly-Seber (CJS) model (Lebreton et al., 1992) was fit through maximum likelihood estimation to estimate annual apparent survival (Φ) and detection probability (p). The final model specification was informed by goodness-of-fit tests and AICc. To account for variable annual effort, p was linked to modified annual survey effort (based on number of survey days) via logistic regression (resulting in a different p value for each year). To account for transient animals that only visit the study area once, Φ was stratified into survival following the first sighting (Φ0) and survival following subsequent sightings (Φ1+). From the output of this CJS model, the following parameters were derived: annual rate of transience (T), number of transient animals (NT), number of non-transient animals (NR) and total abundance (Ntot). Ntot was calculated using a modified Horvitz-Thompson estimator (Horvitz and Thompson, 1952). T was calculated based on survival estimates and the observed number of newly sighted animals and re-sighted animals each year. NT and NR were calculated using T and Ntot. Confidence intervals (95%) for all parameters were derived using a stratified bootstrap approach (1000 replicates).

The downloadable .xlsx file contains two tables in one work sheet:

1. PARAMETER VALUES. This provides the maximum likelihood estimate, lower and upper 95% confidence intervals for each estimated parameter from the modified Cormack-Jolly-Seber model and modified Horvitz-Thompson abundance estimates. Each parameter either has a constant value over time (year = constant), or a separate value for each estimable year. Note: it was not possible to estimate the value of each time-varying parameter for all years.

2. PARAMETER DEFINITIONS. Information is provided about each parameter, including mathematical notation, description, limits, units and extra notes.