How many leks does it take? Minimum samples sizes for measuring local-scale conservation outcomes in Greater Sage-Grouse

Monitoring population response to conservation actions, such as habitat management, is critical to evaluate conservation outcomes. Greater Sage-Grouse (Centrocercus urophasianus) has been the recipient of substantial recent conservation efforts in North America. Sage-Grouse are often surveyed using counts of males displaying on breeding leks, and these lek counts offer a practical method for monitoring Sage-Grouse population trends. Although substantial work has assessed the utility of lek count data for large-scale population monitoring, there has been comparably little effort focused on the use of lek counts to evaluate local-scale management. We used Greater Sage-Grouse lek count data from Oregon, USA, combined with simulation, to evaluate the sample sizes (number of leks, years of monitoring) required to detect a positive outcome of habitat management on population growth. We further assessed assumptions associated with male detection, and compared analyses that both did (N-mixture models) and did not (Poisson regression) account for detection probability. We found that when treatments produced a 5% increase in annual population growth, and leks were monitored for at least 10 years, lek counts produced unbiased and detectable estimates of treatment effects with as few as seven treatment and seven control leks. Using an unbalanced design with a greater number of control leks (n = 16) permitted inference from even fewer treatment leks (n = 4), however, we found no scenarios where use of more control leks permitted detection of smaller treatment effects or allowed shorter duration studies. We found that N-mixture models and Poisson regression of the maximum of three repeated counts produced equivalent results when detection probability was constant, but at the small sample sizes we evaluated, confounding between detection probability and habitat management compromised the accuracy of all analysis methods. Our results show that lek counts hold promise for efficient monitoring of local-scale conservation, but further work is needed to understand the mechanisms that affect male detection during lek surveys