Evaluation of four methods for estimating parrot population size

We evaluated point transect, Line transect, mark-resighting, and roost survey methods by comparing population estimates to a reference population of Green-rumped Parrotlets (Forpus passerinus) in the Ilanos of Venezuela. All four methods produced similar population estimates, reflected peaks in nesting and fledging, and almost always exceeded the known minimum population size. Congruence among the estimates decreased as the breeding season progressed. The mark-resighting method had the largest confidence intervals, but precision was similar for the other three methods. Point transect surveys appeared to underestimate the population in open habitat. Line transect surveys more accurately estimated the distribution of the population between habitats. Detection probability was more variable for point transects than for line transects. We recommend using line transect surveys rather than point transects when conditions allow. Roost surveys were complicated by the tendency of,parrotlets to change roost sites often, and roost surveys underestimated populations during breeding. Roost surveys may be reliable during nonbreeding, and we recommend further development of roost survey protocol. Behavioral characteristics of Green-rumped Parrotlets allowed us to meet assumptions and requirements of all methods with the exception of mark-resighting. This success may not be replicated with other parrot species.