Relation between Occupancy and Abundance for a Territorial Species, the California Spotted Owl

Abstract

Land and resource managers often use detection-nondetection surveys to monitor the populations of species that may be affected by factors such as habitat alteration, climate change, and biological invasions. Relative to mark-recapture studies, using detection-nondetection surveys is more cost-effective, and recent advances in statistical analyses allow the incorporation of detection probability, covariates, and multiple seasons. We examined the efficacy of using detection-nondetection data (relative to mark-recapture data) for monitoring population trends of a territorial species, the California Spotted Owl (Strix occidentalis occidentalis). We estimated and compared the finite annual rates of population change (λt ) and the resulting realized population change (Δt ) from both occupancy and mark-recapture data collected over 18 years (1993-2010). We used multiseason, robust-design occupancy models to estimate that territory occupancy declined during our study (Δt = 0.702, 95% CI 0.552-0.852) due to increasing territory extinction rates (ε(1993) = 0.019 [SE 0.012]; ε(2009) = 0.134 [SE 0.043]) and decreasing colonization rates (γ(1993) = 0.323 [SE 0.124]; γ(2009) = 0.242 [SE 0.058]). We used Pradel's temporal-symmetry model for mark-recapture data to estimate that the population trajectory closely matched the trends in territory occupancy (Δt = 0.725, 95% CI 0.445-1.004). Individual survival was constant during our study (φ(1993) = 0.816 [SE 0.020]; φ(2009) = 0.815 [SE 0.019]), whereas recruitment declined slightly (f(1993) = 0.195 [SE 0.032]; f(2009) = 0.160 [SE 0.023]). Thus, we concluded that detection-nondetection data can provide reliable inferences on population trends, especially when funds preclude more intensive mark-recapture studies.