Abstract:
|
Collisions between large ships and whales have been documented in every ocean where they are co-located. Although these collisions can have significant population, management, and public-relation consequences, quantitative estimates of collisions are unavailable or are biased due to non-reporting of collisions or detection of carcasses. We used Extreme Value Theory (EVT) to estimate the probability of a collision using distances of ship-whale encounters collected by observers on the bow of cruise ships in Glacier Bay National Park, Alaska, USA (GLBA). Using 1412 unique ship-whale encounters, we parameterized a Generalized Pareto Distribution (GPD) to infer the probability of events, including the distance of 0 m, i.e., a collision. Using the fitted GPD, we determined the probability of a collision to be 4.1e-4 (95% bootstrap CI: 0, 3.3e-3) per surfacing event within 300 m of the bulbous bow. This equated to an average rate of 1 collision per 4858 vessels (95%CI: 600, Inf) transiting GLBA and the surrounding waters, or based on current levels of ship traffic, one collision every 26 years (95%CI: 3.25, Inf).
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.