Probability of attendance and sightability of greater sage-grouse on leks: Relating lek-based indices to population abundance

Proper management of wildlife requires accurate assessments of populations. Often, actual estimates of abundance are difficult to obtain and information adequate for management decisions can be acquired more efficiently through the use of population indices. Indices, however, rely on the assumption that they are proportional to the population. Recently, populations of greater sage-grouse (Centrocercus urophasianus) have declined, resulting in the designation of the "warranted, but precluded" status under the Endangered Species Act, which necessitates careful assessment at the population level. Breeding populations of sage-grouse are typically indexed from the number of males attending groups of breeding sites; however, assumptions of this index have not been verified.;I investigated visibility bias of sage-grouse on leks to identify variables that affect sightability of grouse during lek counts. To test for inconsistencies in the relationship between the index and population abundance, I modeled the probability of individuals attending leks, including year-specific covariates. By correcting lek counts for sightability and attendance probabilities, I generated abundance estimates that were used to examine the proportionality of the index to the population estimates. To relate estimates of male abundance to that of females, I developed a method for estimating sex ratio from DNA analysis of noninvasive genetic fecal samples. I used periodic detections of radio-marked male and female sage-grouse to estimate monthly survival probabilities and monitored hens during the nesting season to estimate nest survival.;I found that sightability of male sage-grouse on leks was influenced by lighting conditions, lek characteristics, and factors that appeared to influence the level of activity of attending males. The probability of attending leks varied among years and appeared to be tied to winter severity. These results suggest contemporary methods for monitoring sage-grouse populations have limited power to detect modest population changes over short time periods (i.e., < 5 years). Accounting for visibility bias in lek counts did result in improved correlation with the population abundance; however, accounting for variation in lek attendance due to winter severity would strengthen underlying assumptions and improve the reliability of lek-based indices.