Annual male lek attendance influences count-based population indices of greater sage-grouse

Abstract

Populations of greater sage-grouse (Centrocercus urophasianus) are monitored using counts of males displaying on breeding leks (lek counts). When count-based indices are used to assess population growth (λ), an implicit assumption is that detection is constant through time and space. If detection depends on breeding behavior, annual variation in the proportion of individuals that attend a breeding site may lead to a violation of this assumption. We used 8 years of banding data from male sage-grouse in eastern Nevada and capture–mark–recapture analyses to evaluate how temporary absence of males from lek sites during a given year may influence estimates of population trends derived from lek counts. We estimated the proportion of variance in annual lek count trends that corresponded with an independent estimate of λ, versus variance associated with temporary absence. The probability a male sage-grouse attended 1 of our study leks at least once in a given year ranged from a low of 0.56 (±0.22 SE) to a high of 0.87 (±0.11 SE). Variance in annual lek count trends was associated with both realized λ (semipartial R2 = 0.57), and sampling error associated with temporary absence (semipartial R2 = 0.40). We found discrepancies between lek count and realized λ in 3 out of 7 intervals, whereas estimates of λ for the entire study interval were extremely similar between count-based and capture–mark–recapture methods (λ = 0.90 ± 0.05 SE and λ = 0.91 ± 0.05 SE, respectively). Temporary absence was influenced by male density during the previous year and associated with exotic grasslands surrounding leks, although some uncertainty was associated with this latter effect. Lek counts are well suited for estimating λ across multi-year intervals, whereas annual estimates of λ should be viewed cautiously if variation in annual lek attendance is not directly incorporated. © 2013 The Wildlife Society.