Abstract

Weight estimates are critical to monitoring the health of captive animals as well as understanding the ecology of wild ones. However, for killer whales, measuring live body weight can be challenging, and is not available under all conditions. Historical measures of length, axillary girth, dorsal girth and body mass measurements were obtained for 23 captive killer whales, (Orcinus orca). Animals were categorized by sex (males n=9, females n=14) and ecotype (North Atlantic Ecotype #9, n=15; Resident Ecotype #6, minimum 25% Resident genetics, n=6; and Subantarctic Ecotype #5, minimum 50% Subantarctic genetics, n=2). Measurements were averaged by month. The relationship between the linear measurements (dorsal girth, axillary girth, body length) and animal weight was estimated by fitting a variety of models including linear, logarithmic, and polynomial. The optimal equation for estimating the weight of a whale in this dataset is a polynomial equation that estimated the volume of the whale as an ellipsoid sphere, where the volume is determined based on the circumference at the center (dorsal girth) and the length. The obtained equation predicted values that were within 10% of the animals’ measured weight 82% of the time. The second best model is a linear model also incorporating length and dorsal girth that predicted values that were within 10% of the animal’s measured weight 74% of the time. Although the second model is slightly less accurate, it is likely the more convenient for everyday use. Differences in predictive accuracy were not seen by sex or ecotype. However, there was a sex by ecotype interaction in which male Northern Atlantic whales had higher residuals (differences between measured and predicted weights) than the other sex/ecotype combinations.

Models that fit the data closely can be used to predict body weight from linear measurements. Models using a different set of measurements that fit the data less closely show promise for differentiating animals at different body condition scores. A preliminary body condition scoring system was able to rank the animals and obtain results comparable to other methods of estimating body condition.

Acknowledgements

Many thanks to the zoological teams at SeaWorld Orlando, San Antonio and San Diego for collecting extensive morphometric data over the years.

* Presenting author