Browsing by Subject "humpback whale"

Now showing 1 - 7 of 7
  • Thumbnail Image
    Item
    Open Access
    Bayesian assessments of the southern hemisphere humpback whale breeding stock B using three different models for stock-structure
    (2010) Müller, Andrea; Butterworth, Doug S; Johnston, Susan J
    Three different models are considered for Southern Hemisphere humpback breeding stock B: Model 1, a single fully-mixed stock; Model 2, reproductively independent stocks B1 off Gabon and B2 of the southern African coastline; and Model 3 with two stocks as in Model 2 but with B1 consisting of two sub-stocks, one of which migrates to Gabon along the southern African coastline. The models are fitted to mark-recapture data (both photo-ID and genetic) available for the Gabon and South African regions. The purpose of the paper is one of illustrating the stock-structure models and associated methods of analysis, anticipating that further discussion and selection of specific model input assumptions will take place during the Scientific Committee meeting.
  • Thumbnail Image
    Item
    Open Access
    Bayesian assessments using models which allow for interchange on the breeding grounds of Southern Hemisphere humpback whale breeding Sub stocks C1 and C3
    (2009) Johnston, Susan J; Butterworth, Doug S
    Bayesian stock assessment results for breeding sub-stocks C1 and C3 using models which allow for interchange on the breeding grounds as well as mixing on the feeding grounds are illustrated for four models – the sabbatical, tourist and migrant models and the resident model (for which interchange is set to zero.) Results are also presented for a range of sensitivity tests. The availability of photo-id data allows the estimation of interchange rates. The estimates of these interchange rates are generally low with posterior median estimates all below 6% p.a., and estimated trajectories are fairly similar for all models considered. With single exceptions, current (2006) posterior median population sizes relative to pre-exploitation levels are all estimated to exceed 80% for C1 and 90% for C3.
  • Thumbnail Image
    Item
    Open Access
    Capture-recapture analyses of humpback whale population sizes and increase rates: breeding sub-stock B1
    (2008) Johnston, Susan J; Butterworth, Doug S
    Maximum likelihood and Bayesian approaches are applied to estimate parameters of a simple experimental growth model for the humpback whales of breeding sub-stock B1 from photo-ID and genotypic capture-recapture data for all observation sites combined and for Iguela only. The maximum likelihood estimates are generally rather imprecise, and some estimates for the annual growth rate r are demographically unrealistic. Bayesian results, for which a prior restricts its r to a demographically plausible range of [0; 0.106], are more reliable and suggest a population of size 7600 [95% PI: 6600-8800] in 2003, with r at 8.5% [95% PI: 3.7%-10.4%]. These results, are however, intended primarily as illustrative of the methodologies, and to facilitate further discussion and model refinement. This will need, inter alia, to consider the appropriateness of aggregating data from different sites.
  • Thumbnail Image
    Item
    Open Access
    Is the length distribution of C1 and C3 humpback whale catches consistent with an age-structured version of the resident model
    (2009) Müller, Andrea; Butterworth, Doug S; Johnston, Susan J
    A sex- and age-structured BALEEN II population model is fitted to population abundance and trend as well as photo-id capture-recapture data for the C1 and C3 humpback whale breeding sub-stocks. The model is of the “Resident” type, i.e. no interchange between breeding grounds, though the whales do mix on the feeding grounds. Uniform selectivity on the 1+ population is assumed for both regions. A particular aim is to address the question of whether length distribution differences between the two regions are a reflection of different levels of past exploitation. Comparison with length distribution data for both regions does however indicate a greater proportion of larger males than anticipated in the C3 catches, and the reverse effect for both males and females in the C1 catches.
  • Thumbnail Image
    Item
    Open Access
    Is the length distribution of C1 and C3 humpback whale catches consistent with an age-structured version of the resident model
    (2009) Müller, Andrea; Butterworth, Doug S; Johnston, Susan J
    A sex- and age-structured BALEEN II population model is fitted to population abundance and trend as well as photo-id capture-recapture data for the C1 and C3 humpback whale breeding sub-stocks. The model is of the “Resident” type, i.e. no interchange between breeding grounds, though the whales do mix on the feeding grounds. Uniform selectivity on the 1+ population is assumed for both regions. A particular aim is to address the question of whether length distribution differences between the two regions are a reflection of different levels of past exploitation. Comparison with length distribution data for both regions does however indicate a greater proportion of larger males than anticipated in the C3 catches, and the reverse effect for both males and females in the C1 catches.
  • Thumbnail Image
    Item
    Open Access
    Preliminary results for a combined assessment of all seven southern hemisphere humpback whale breeding stocks
    (2010) Müller, Andrea; Butterworth, Doug S; Johnston, Susan J
    This paper reports preliminary results from the development of an assessment process that aims to include all seven Southern Hemisphere humpback whale breeding stocks in a single joint assessment, with the purpose of allowing highlatitude historic catches (i.e. catches taken south of 40˚S, where mixing amongst the populations occurs), to be allocated to breeding stocks in proportion to abundance, rather than on set ratios. The approach can be broadened to allow for uncertainties in the placement of the boundaries assumed to link high latitude catches to breeding stocks. Because of the interaction between populations arising from the procedure to allocate high latitude catches amongst breeding stocks, the conventional SIR-based Bayesian approach proved impractical to expand. Instead uniform priors on the various preexploitation level (K) parameters were assumed with the intent to later iteratively adjust these to account for their being informative about the values of the intrinsic growth rate (r) parameters. Initial results (which will need later refinement) are presented purely for the purposes of illustrating the application of the approach.
  • Thumbnail Image
    Item
    Open Access
    Prior incoherence within a Bayesian assessment of the Southern Hemisphere humpback whale breeding stock B population
    (2010) Müller, Andrea; Butterworth, Doug S
    In very simple terms, a Bayesian analysis involves drawing estimatable parameter values from some prior distribution, computing population dynamics and assigning a likelihood value to each combination based on comparisons to data containing information on population size and/or trend. A posterior distribution may then be constructed and conclusions drawn about the parameter estimates. In Model Ia (see Appendix) r B1 , r B2 , ( ) 1 arg ~ ln B Nt , ( ) 2 arg ~ ln B Nt are the parameter values drawn from priors for the intrinsic growth rate and the log of the recent abundance for the two populations under consideration.