Visibility bias for sage‐grouse lek counts

Abstract

ABSTRACT Declining populations of greater sage‐grouse (Centrocercus urophasianus; hereafter, sage‐grouse) necessitate accurate estimates of population trends. The most widespread approach to monitoring sage‐grouse populations involves counting males attending leks and tracking changes across years. Visibility of male sage‐grouse on leks is likely <100% for numerous reasons, but this potential bias has generally been overlooked. Our objectives were to estimate the visibility bias (failure to enumerate all males on a lek during a count) of sage‐grouse and identify variables that influence visibility of males attending leks by developing a sightability model. We conducted simultaneous counts of male sage‐grouse from a truck using standard lek‐counting procedures and ground blinds during 2007–2009 in south‐central Idaho, USA. We built candidate sightability models using logistic regression with proportion counted from the blind that were also counted from the truck as the response variable. Naïve visibility in our study averaged 79%. Our sightability model suggests visibility bias decreased when sun was shining on the lek and when female birds were present, but increased with time of day. Additionally, cloud cover appeared to negatively affect visibility before sunrise, but improved visibility later in the mornings. Per our model, average visibility of male sage‐grouse on leks decreased 16% from 33 min before sunrise to 90 min past sunrise when both female birds and direct sun were recorded, and decreased 28% over the same time without direct sun or females. Additionally, our model predicted visibility decreased 9% as cloud cover increased from 0% to 100% at 30 min before sunrise, but increased 44% as cloud cover increased from 0% to 100% at 90 min past sunrise. Our study does not suggest the visibility bias we observed necessarily relates to a bias in the lek route index. However, by accounting for visibility bias, detecting population changes should require fewer counts or be evident with fewer years of data. © 2017 The Wildlife Society.

Keywords

• Centrocercus urophasianus
• greater sage‐grouse
• Idaho
• lek counts
• population index
• sightability model