Abstract
High levels of bat and bird mortalities have been documented at wind energy facilities; particularly hard-hit among bats are the treeroosting migratory species Lasiurus cinereus, L. borealis, and Lasionycteris noctivagans, which together compose approximately 79% of affected bats. Traditional mark-recapture monitoring methods have proven ineffective for these species due to the fact that these bats roost in small numbers, fly very high, and are difficult to catch. Thus it is hard to tell what effect these deaths at wind energy facilities are having on population numbers. Genetic data may provide a means of monitoring populations when demographic methods are unsuitable. We used coalescent-based simulations to determine the efficacy of genetic data as a monitoring tool for short-term changes in population size. Simulations were run under demographic models parameterized using mitochondrial DNA sequence data and microsatellite genotypes from the eastern red bat, Lasiurus borealis. DNA sequence data and microsatellite genotypes were simulated in both panmictic and structured populations using the computer program, ms, and analyzed using statistical software (microstat) to interpret the results. ms is a coalescent- based program that simulates genetic data under specific population models that are parameterized by initial population size, rate of decline, time since the onset of decline, mutation rate of the chosen molecular marker, and pattern of population structure. Initial estimates of these parameters were taken from previous studies on L. borealis (initial population size = 3.3 million individuals, rate of decline = –1% per year, mitochondrial mutation rate = 10-5 substitutions per gene per generation, no significant population structure). Simulations were allowed to run from 1 to 1000 generations following the initial onset of population decline to determine the timescales necessary to observe significant loss of genetic diversity under biologically realistic conditions. Loss of genetic diversity was assessed using summary statistics including the number of segregating sites, nucleotide diversity, and Tajima’s D for DNA sequence data; analogous measures for microsatellite data included average heterozygosity, θP, and Cox’s Δ. We found that direct measures of diversity (segregating sites and average heterozygosity) are much more informative for detecting population declines than neutrality tests such as Tajima’s D and Cox’s Δ. Between the two types of markers, microsatellites provided more power to detect population declines over shorter timescales (hundreds of generations for microsatellites as opposed to thousands of generations for sequence data). These results demonstrate that even quicklyevolving microsatellite data are unlikely to be useful for the type of year-to-year comparisons needed by monitoring agencies. We conclude that genetic data do not appear to be a useful metric for monitoring red bat population declines due to wind turbine-associated deaths. We emphasize that these conclusions are limited to the population parameters examined in this study, specifically those for eastern red bats facing population declines from wind turbines. Similar questions in other species (e.g., little brown bats facing local extirpation from white-nose syndrome) should be addressed using models appropriately parameterized for those systems.