Improved Estimation of Bird Abundance and Energy Consumption by Combining Point and Transect Observations

Point counts and transect surveys are both common methods for estimating bird abundance (by species, genus, or family) in a given environment. Sometimes both methods are used in the same study. Point counts reflect abundance directly, while transect observations of bird feeding instances reflect both the bird abundance and the mean energy consumption rate per bird. In this research we develop an improved estimator of bird abundance based on both sets of measurements, and derive error bars for these abundance estimates. The improved estimator is a linear combination of the two individual abundance estimates obtained from the two data types. The combination weights are obtained through a comparison of two separate energy estimates. The mathematical methods developed are applied to a dataset consisting of point counts and transect surveys of Ecuadorian frugivore bird species in six different types of environment (three elevations and two degrees of forestation). The improvement in accuracy achieved by this combined estimator is verified by simulations applied to bird family data. We also obtain 90% confidence intervals for the bird family relative abundance estimates, and verify the confidence intervals using simulations.