A comparison of 3 aerial survey techniques to estimate wild horse abundance in the Australian alps

Estimates of the size or density of wildlife populations are essential for effective management. Aerial surveys often are the only practical way of obtaining these estimates for large mammals ranging over a large area. Different aerial survey techniques are available. This paper reports on a study that estimated abundance and population density of wild horses (Equus caballus) in the Australian Alps using helicopter aerial surveys. Three frequently used techniques were compared simultaneously: strip counts, mark-recapture, and line transect. Using data on horse groups over a strip width of 200 m, estimates of horse abundance from mark-recapture analyses were higher (N̂=3,045 modified Petersen and N̂=3,298 Huggins estimators, respectively) than from strip counts (N̂=1,814 and N̂=2,242 for observers 1 and 2, respectively). Use of line-transect analysis yielded the highest estimates of abundance (N̂=5,200-5,830, population density D̂=1.86-2.09 km-2). The coefficient of variation (CV) of estimates varied from 22% for strip counts (0-50 m) to 28% for mark-recapture analysis (0-50 m) and 31-32% for line-transect analysis (0-200 m). Mark-recapture estimates of horse abundance from the modified Petersen estimator were substantially lower when the strip width was 200 m (N̂=2,061) compared with a strip width of 50 m (N̂=4,345). These results suggested that in aerial surveys using strip or mark-recapture analyses, the effect of wide strip width should be examined; otherwise estimates may be strongly negatively biased. Observer 1 estimated mean group size of 4.91±0.61 SE horses and observer 2 estimated mean group size as 3.79±0.57 SE horses. Mean group size observed from the ground was 5.65±0.51 SE. We recommend that aerial surveys be implemented periodically using line-transect methodology to monitor trends in wild horse abundance in the Australian Alps.