Modeling the response of oak woodland birds to changing climate using 27 years of monitoring data

Abstract

Using 27 years of point count data collected at the San Joaquin Experimental Range, Madera County, California, we modeled the importance of weather and climate variables on annual variability in the abundance of 35 bird species, to assess the potential responses of birds to changing environmental conditions. We used semi-parametric Poisson generalized additive models (GAM) with smoothing functions to explore relationships between bird abundance and the independent climatic and weather variables and to extract the best parametric polynomial model. We then selected among three parametric generalized linear mixed models (Poisson, quasi-Poisson, and negative binomial) on the basis of how well they explained the overdispersion of the counts. This approach also allowed us to estimate and control for variability among observers. Temperature variables were important in explaining variation in bird abundance. More species responded to cold than to hot temperatures: 18 were sensitive to cold, or were more abundant following warm winters or springs, whereas only 4 species were sensitive to heat. Because minimum temperatures are increasing faster than maximum temperatures, species sensitive to cold may benefit from warming. Abundance of Anna’s Hummingbirds (Calypte anna), California Scrub-Jays (Aphelocoma californica), Oak Titmice (Baeolophus inornatus), and Bewick’s Wrens (Thryomanes bewickii) decreased with warm temperatures, so an increase in mean temperatures is likely to affect these species negatively. One species, the Turkey Vulture (Cathartes aura), was more abundant after hot summers. Response to precipitation varied, with the number of species that increased in abundance with increasing precipitation being roughly equal to the number that decreased. El Niño Southern Oscillation (ENSO) events are predicted to increase in frequency and severity with climate change. The effects of ENSO vary geographically; for example, species that occur at lower elevations appear to respond differently from those that breed at higher elevations, where winter precipitation is dominated by snow rather than rain. In our low-elevation study area 12 species responded positively to warm, wet El Niño conditions; only 4 responded negatively. We found no clear pattern of response to long-term variation related to the Pacific Decadal Oscillation (PDO). We identified four species that decreased following drought years, but one of these, the Red-tailed Hawk (Buteo jamaicensis), was primarily responding (positively) to wet years. Abundance of Nuttall’s Woodpeckers (Picoides nuttallii), Canyon Wrens (Catherpes mexicanus), and Lesser Goldfinches (Spinus psaltria) decreased with drought, so these species should be negatively affected by the increased frequency and severity of droughts projected with climate change. Oak woodland birds are closely tied to oaks and the acorns they produce, but questions about the ability of oaks to shift upward in elevation and latitude raise doubts about the ability of these bird species to adapt to climate change. Long-term monitoring data are useful for investigating and predicting avian responses to climate change. Understanding potential responses will help inform conservation decisions and planning for an uncertain future. We hope our predictions for how California oak woodland birds may respond to the changing climate will help prioritize future research and facilitate conservation.