Abstract
Project Objectives
This project was intended to evaluate the use of automated echolocation detectors as a means of estimating bat activity levels during the pre-construction phases of wind energy development. An intensive array of detectors was deployed at a wind energy development and the resulting data was used to demonstrate the analytical methods used to recommend the number and configuration of detectors necessary to produce precise estimates of bat activity during a given time period. Additionally the project demonstrated a model selection procedure to determine which design variables (e.g., meteorological, detector location) were most important for explaining observed patterns of bat activity.
Methods
An array of 26 ANABAT II echolocation detectors was established at 12 locations to measure bat activity at the Dillon Wind Energy project site near North Palm Springs, California. From 25 October 25 - 5 December, 2007 detectors were deployed on 4 meteorological towers at heights of 2, 22, and 52m above ground. Detectors were configured to automatically record echolocations from before dusk to after dawn each day. Eight additional bat monitoring towers were added to the site on Dec 13 - 15 2007 with detector microphones attached at 2 and 22m above ground. The number of files that contained bat echolocations was tabulated for each detector and each night from those mounted on meteorological towers.
Meteorological conditions were measured at the towers and linked to echolocation data from the same date for use in analyses. Generalized mixed models were used to establish relationship between number of bat files recorded, tower identity, height and meteorological variables. The model that best fit the data was selected and used to estimate the precision of estimates of bat activity that could be established with more of fewer towers.