Long-term monitoring of tropical bats for anthropogenic impact assessment: Gauging the statistical power to detect population change
Résumé
Bats are ecologically important mammals in tropical ecosystems; however, their populations face numerous environmental threats related to climate change, habitat loss, fragmentation, hunting, and emerging diseases. Thus, there is a pressing need to develop and implement large-scale networks to monitor trends in bat populations over extended time periods. Using data from a range of Neotropical and Paleotropical bat assemblages, we assessed the ability for long-term monitoring programs to reliably detect temporal trends in species abundance. We explored the magnitude of within-site temporal variation in abundance and evaluated the statistical power of a suite of different sampling designs for several different bat species and ensembles. Despite pronounced temporal variation in abundance of most tropical bat species, power simulations suggest that long-term monitoring programs (>= 20 years) can detect population trends of 5% per year or more with adequate statistical power (>= 0.9). However, shorter monitoring programs (<= 10 years) have insufficient power for trend detection. Overall, our analyses demonstrate that a monitoring program extending over 20 years with four surveys conducted biennially on five plots per monitoring site would have the potential for detecting a 5% annual change in abundance for a suite of bat species from different ensembles. The likelihood of reaching adequate statistical power was sensitive to initial species abundance and the magnitude of count variation, stressing that only the most abundant species in an assemblage and those with generally low variation in abundance should be considered for detailed population monitoring. (C) 2010 Elsevier Ltd. All rights reserved.