Evaluation of an Unmanned Airborne System for Monitoring Marine Mammals

Tests of an unmanned airborne system (UAS) for surveys of marine mammals were conducted near Port Townsend, Washington. Sixteen surveys were conducted over a 10-d period to find 128 simulated whale targets (4 to 9 per survey). Various weather conditions were encountered, and searchwidths and altitudes were varied to establish optimal search parameters for future surveys. Logistic regression models were applied to estimate how detection rates were influenced by target color, degree of target inflation, shutter speed, searchwidth, and Beaufort wind force. Beaufort wind force was the strongest predictor of detection rates with color and degree of target inflation also included in the model that best fit these data. Overall detection rates of simulated large whale profiles using UASs were similar to published estimates of detection rates during manned aerial surveys for marine mammals, except the search area was much smaller (narrow strip width) when using the UAS. The best detection rates were obtained when Beaufort wind force was lowest (~ 2). The UAS tested showed promise for replacing manned aerial surveys for monitoring distribution and abundance of large marine mammals; however, improvements are required before the UAS would be an efficient tool for detection of all species. Side-by-side comparisons are needed between the UAS and manned aircraft to evaluate any differences in detection rates from the two platforms.