Abstract
BACKGROUND: Commercial viewing and swimming with dusky dolphins (Lagenorhynchus obscurus) near Kaikoura, New Zealand began in the late 1980s and researchers have previously described changes in vocalisation, aerial behaviour, and group spacing in the presence of vessels. This study was conducted to assess the current effects that tourism has on the activity budget of dusky dolphins to provide wildlife managers with information for current decision-making and facilitate development of quantitative criteria for management of this industry in the future. METHODOLOGY/PRINCIPAL FINDINGS: First-order time discrete Markov chain models were used to assess changes in the behavioural state of dusky dolphin pods targeted by tour vessels. Log-linear analysis was conducted on behavioural state transitions to determine whether the likelihood of dolphins moving from one behavioural state to another changed based on natural and anthropogenic factors. The best-fitting model determined by Akaike Information Criteria values included season, time of day, and vessel presence within 300 m. Interactions with vessels reduced the proportion of time dolphins spent resting in spring and summer and increased time spent milling in all seasons except autumn. Dolphins spent more time socialising in spring and summer, when conception occurs and calves are born, and the proportion of time spent resting was highest in summer. Resting decreased and traveling increased in the afternoon. CONCLUSIONS/SIGNIFICANCE: Responses to tour vessel traffic are similar to those described for dusky dolphins elsewhere. Disturbance linked to vessels may interrupt social interactions, carry energetic costs, or otherwise affect individual fitness. Research is needed to determine if increased milling is a result of acoustic masking of communication due to vessel noise, and to establish levels at which changes to behavioural budgets of dusky dolphins are likely to cause long-term harm. Threshold values from these studies would allow managers to set appropriate operational conditions based on quantifiable criteria.