Simulation exercise to ascertain the relative utility of collection of various data types for informing future Southern Hemisphere humpback whale assessments and Breeding Stocks D, E1 and the Oceania stocks

• Main Authors: Ross-Gillespie, Andrea, Butterworth, Doug S
• Format: Report
• Language: English
• Published: University of Cape Town 2015
• Subjects: Southern Hemisphere; humpback; Breeding Stocks D; E1; Oceania stocks; Humpback Whale
• Online Access: http://hdl.handle.net/11427/17633; https://open.uct.ac.za/bitstream/11427/17633/1/Ross_Gillespie_Simulation_exercise_to_.pdf

The existing three-stock model for Breeding Stock D (West Australia or BSD), Breeding Stock E1 (East Australia or BSE1) and the collection of Oceania breeding sub-stocks (referred to as Breeding Stock O or BSO for convenience) is used to simulate future data which might be collected for these stocks, to ascertain which have the best potential to improve estimates of precision of key quantities associated with the population dynamics. For BSD, a future estimate of absolute abundance in 2017 and a series of relative abundance estimates each year from 2016 to 2020, where all estimates have CVs of 0.25, are considered. For BSE1, a similar further relative abundance series is considered, as well as new mark-recapture data collected over 2016 to 2020 where sighting probabilities are taken to be half the average achieved previously. Sighting probabilities are dealt with in a similar manner for future mark-recapture data considered for BSO. Both the further absolute abundance estimate for BSD and especially more mark-recapture data for BSO show some potential for providing improved precision in parameter estimates. However, there seems to be little increase in precision to be gained from further relative abundance data, bearing in mind that these results presume the current three-stock model to be correct. The three-stock model is further used to estimate the range of future observations that would remain compatible with that model’s assumptions. Actual data collected in future could be compared to such ranges to check on the model’s ability to continue to reflect reality.