Abstract
Turbines at wind projects pose a threat to birds and bats flying at altitudes within the rotor swept area. These animals die from colliding with the turbine blades. Estimating mortality, or the total number of bird or bat fatalities at a wind project, is critical to understanding environmental impacts of wind energy, and potential population effects to at-risk species. Mortality is estimated by conducting carcass surveys, and then adjusting observed counts for carcasses that land outside the searched area, carcasses removed by scavengers, and imperfect detection. The goal of this research is to assess if and when mortality estimation can be achieved by only searching the roads that lead up to turbines and pads beneath turbines (R&P). Accounting for unsearched area, or equivalently estimating the proportion of carcasses that land on R&P, is pivotal to the success of mortality estimation from R&P searches. Methods based on previously proposed ideas to account for unsearched area are tested and compared using simulations to determine what situations and methods are the most appropriate for mortality estimation based on R&P carcass surveys under different spatial carcass distributions. Results of the simulations are utilized to develop an estimator of mortality based on R&P searches under isotropic carcass distributions. The theoretical properties of the estimator are derived, and utilized for standard error calculation and confidence interval creation. Because optimal methods differ for isotropic and anisotropic spatial carcass distributions, a hypothesis test is needed to determine if it is valid to assume an isotropic distribution.