Age classification of bowhead whales using recursive partitioning

Abstract

An algorithm was derived for using morphometric data to classify bowhead whales into three age brackets: over 90 years (‘very old’); 60–90 years (‘old’); and under 60 (‘younger’). Recursive partitioning was applied to a subset of the data from post mortem examinations. This subset consisted of whales with higher quality data scores and with either estimated ages or characteristics of very old animals such as: near-maximum body length and baleen length; heavy scarring; and ancient weapons embedded in them. Statistical analysis suggested that for males, body length and peduncle girth provide the most useful information for this age classification. For females, anterior flipper length and body length were the key variables for classifying age. If anterior flipper length is not available for females, then body length, baleen length and peduncle girth may be used to classify age.