Summary: | We developed a simple, precise and relatively inexpensive technique to estimate moose (Alces alces) populations in small census zones (< 1500-2000 km2) such as parks and wildlife reserves. A two-phase sampling plan was adopted. In phase 1, we used fixed-wing aircraft to survey the entire area using flight lines spaced 500 m apart in order to locate track networks, draw them on 1:50,000 topographic maps, and determine their area. In phase 2, a sample of the identified track networks was intensively searched by helicopter to count moose. total population was estimated by extrapolating the mean number of moose per track network to the total number of networks counted in phase 1. We compared three approaches to estimate the mean number of moose per track network and its variance: a direct estimation using the arithmetic mean per track network, and two methods that accounted for the size of track networks (quotient and regression estimations). The regression estimation model provided the most precise estimates. A confidence interval of 20% (α = 0.10) can be achieved by counting moose on 30 to 50 % of the track networks. This two-phase approach can reduce survey costs by 25-35 percent as compared to the usual total count preciously used in small census zones.
|