Calibration of sighting rates of harbor porpoise from aerial surveys
Calambokidis, J., J.R. Evenson, J.C. Cubbage, S.D. Osmek, D. Rugh, and J.L. Laake, 1993. Calibration of sighting rates of harbor porpoise from aerial surveys. Report to the National Marine Mammal Laboratory, National Marine Fisheries Service, Seattle, Washington. 41pp.
Harbor porpoise abundance estimates have been made for Washington and Oregon using aerial and vessel surveys (Barlow 1988, Barlow et al. 1988, Calambokidis In Press, Calambokidis et al. 1991, 1992, Green et al. 1992 Turnock et al. In press). A major limitation of these estimates, however, has been that the fraction of animals missed on the survey line was not known. The primary objective of this study was to experimentally determine a correction factor for harbor porpoise missed during aerial surveys.
Two methods were used to evaluate the correction factor for animals missed. The primary method for obtaining a correction factor was to measure the sighting success by aerial observers of specific harbor porpoise groups being tracked from land. Harbor porpoise groups in the study area were tracked from land prior to the aircraft passes. Secondarily, we gathered breath rate data on harbor porpoise groups to determine the proportion of time they are at the surface and likely to be seen by aerial observers. This was similar to the method used by Barlow et al. (1988) in devising a correction factor.
The calibration experiment was conducted off the northwest side of Orcas Island in the San Juan Islands, Washington. Aerial surveys with a twin-engine aircraft were conducted in conjunction with land observations on 7 days between 3 August and 13 August 1992. A total of 547 transect lines were flown in 33 flight hours completing approximately 26 hours of transect surveys. Two teams, each consisting of 4-6 people made observations of harbor porpoise from land. Each team used an electronic theodolite interfaced with a computer data acquisition system. Radio contact between one of the aircraft personnel and the land teams was used to select aircraft transects that would include passing over a harbor porpoise group being tracked from land.
To determine the proportion of harbor porpoise groups seen from the aircraft, several criteria were used to select the harbor porpoise groups that would be used for the calibration and to determine whether a group tracked from land was seen by the aerial observers. Harbor porpoise groups were included as a sample in the calibration only if: 1) they were seen from land within 4 minutes of the passage of the aircraft, and 2) were within 100 m of the aircraft transect line. A harbor porpoise group being tracked from land was considered seen by the aerial observers if: 1) the position of the porpoise determined from land and air was less than 300 m, and 2) the number of porpoise in the group determined from the air was not more than twice or less than half the size determined from land.
Using the parameters described above, 28 of 92 (30.4%) harbor porpoise groups within 100 m of the transect line were seen by the aerial observers and 69.6% were missed. The point estimate for g(0) (the probability of sighting a group that was on the transect line) was 0.324 (SE=0.056). The sample size for this estimate was better than expected and provides a low enough variance that it will be able to be applied without greatly increasing the overall variance of abundance estimates. The correction factor for animals missed on the transect line was almost identical to that determined previously by Barlow et al. (1988) based on breath rate data of harbor porpoise.
Seven different harbor porpoise groups whose surfacing were monitored spent an average of 36 sec (sd= 13.1) at or near the surface and 120 sec (sd=31.0) on longer dives of 30 sec or more. These values likely underestimate the time groups spend at the surface because some surfacings may have been missed. The proportion of harbor porpoise at the surface (potentially visible from the air) based on breath rates was slightly lower than the proportion seen in the calibration experiment. This may have been a result of our underestimating the time porpoise are at the surface (due to missing some surfacings). This also suggests that relatively few harbor porpoise on the transect line that were at the surface when the aircraft flew over were missed during the calibration experiment. The proportion of time we calculated harbor porpoise were at the surface from the breath rate data was similar to that reported by Barlow et al. (1988).