A land-based infrared (IR) camera is used to detect endangered Southern Resident killer whales in Puget Sound,
Washington, USA. The observations are motivated by a proposed tidal energy pilot project, which will be required to
monitor for environmental effects. Potential monitoring methods also include visual observation, passive acoustics, and
active acoustics. The effectiveness of observations in the infrared spectrum is compared to observations in the visible
spectrum to assess the viability of infrared imagery for cetacean detection and classification. Imagery was obtained at
Lime Kiln Park, Washington from 7/6/10-7/9/10 using a FLIR Thermovision A40M infrared camera (7.5-14μm,
37°HFOV, 320x240 pixels) under ideal atmospheric conditions (clear skies, calm seas, and wind speed 0-4 m/s). Whales
were detected during both day (9 detections) and night (75 detections) at distances ranging from 42 to 162 m. The
temperature contrast between dorsal fins and the sea surface ranged from 0.5 to 4.6 °C. Differences in emissivity from
sea surface to dorsal fin are shown to aid detection at high incidence angles (near grazing). A comparison to theory is
presented, and observed deviations from theory are investigated. A guide for infrared camera selection based on site
geometry and desired target size is presented, with specific considerations regarding marine mammal detection.
Atmospheric conditions required to use visible and infrared cameras for marine mammal detection are established and
compared with 2008 meteorological data for the proposed tidal energy site. Using conservative assumptions, infrared
observations are predicted to provide a 74% increase in hours of possible detection, compared with visual observations.