Modelling environmental and survey influences on lek attendance using long-term lek survey data

摘要

Targeted lek surveys, in which historically active lekking areas are surveyed over multiple years, are commonly used to monitor lek-mating birds. Lek surveys are usually used in habitats and species in which individuals are not easily detected when they are not lekking. Unfortunately, lek surveys suffer from a number of shortcomings, most notably the failure to observe lekking individuals when they are near but not actively attending a lek. Like other sources of imperfect detection, inconsistency in lek attendance may bias estimates of annual population sizes or the proportion of lekking arenas active in a given year. In turn, this may result in unreliable estimates of trends, distributions or other parameters of interest. An improved understanding of the factors affecting lek attendance is crucial to improving the efficacy of many lek monitoring programmes and the utility of lek surveys in estimating population and distributional dynamics. We assess variation in lek attendance in the Lesser Prairie Chicken Tympanuchus pallidicinctus using a subset of data from a long-term monitoring programme in eastern New Mexico, USA. Focusing on lek-years in which at least one male was observed, we used generalized linear models to assess the influence of lek size and survey and environmental characteristics on the probability that at least one male was observed in attendance. Our best-supported model provided evidence for influences of each of these sources of variation. Models suggested that surveyors may improve the likelihood of detecting lek attendance by making small changes to survey duration, targeted survey dates and by considering environmental factors such as weather, habitat and human disturbance. In addition to improving detection, consideration of lek attendance patterns and associated behaviours may improve estimation for more effective conservation in a variety of lekking bird species.