On the performance of automated porpoise-click-detectors in experiments with captive harbor porpoises (<i>Phocoena phocoena</i>)

Recently, automated porpoise-click-detectors (T-PODs, Chelonia-Marine-Research) have been used intensively in monitoring harbor porpoises (Phocoena phocoena) in the wild. However, the automated click-detection-mechanism of the T-POD leads to questions on the characteristics of the detection process. We undertook experiments with six captive harbor porpoises (four subadult males in one pool, two adult males in another) at the Dolfinarium Harderwijk (Netherlands). One T-POD was placed for over a week in each pool, while the behavior of the porpoises was logged by visual observation. Data were analyzed using the T-POD software. A total of 725 431 clicks in 30 090 trains were recorded with 32% of the trains classified as CET HI, 27% as CET LO, and 41% as DOUBTFUL. All three train classes differed significantly in all parameters, except for click duration. We conclude that T-PODs perform generally well in detecting click trains of harbor porpoises but that in any future study trains classified as being of lower probability should be investigated very carefully to avoid the risk of losing valuable information.