UC Berkeley

Climate change is widely considered to be one of the most important and omnipresent threats to global environmental health and biodiversity. Responding to changing climates, species are expected to shift their geographic distributions in order to remain in physiologically and ecologically favorable climates. These shifts may be species-specific, and different responses of species to a rapidly changing climate have unknown consequences for biotic communities. Despite recent evidence of shifts in mammals and plants, evidence for major changes in the distributions of birds in response to climate change is sorely lacking. Additionally, our understanding of the ecological patterns in which range shifts occur is rudimentary. This is particularly true of montane regions, where there is no clear picture yet as to how animals respond to warming along elevational gradients.

To investigate this subject, I, along with the Grinnell Resurvey Project at the Museum of Vertebrate Zoology (MVZ), resurveyed birds at 95 sites in the Sierra Nevada mountains of California. These sites were originally visited 80-100 years ago by Joseph Grinnell and colleagues from MVZ. The sites were distributed among three primary elevational transects, each extending from near sea level in the Central Valley to the highest peaks of the Sierra Nevada. Comparisons of bird species assemblages at sites across time allowed unique inference on the long-term changes to species' ranges and community composition in California's montane regions.

Our knowledge of current and recent climate-induced impacts have lagged behind predictive work due in part to the difficulties inherent in studying changes over time, requiring both reliable historical data and a robust method to compare these data to contemporary observations. Resurvey studies, where sites with historical species lists are revisited, often after many decades, are critical tools in understanding distributional changes over time. As we show in my first chapter, resurvey studies are increasingly used in ecology to infer extinction, colonization, and range shifts, yet most authors struggle with or completely ignore the problems arising from comparing occurrence data collected by different observers using different methods, sometimes even in different locations. A modeling technique known as occupancy modeling provides a flexible framework by which range shifts can be estimated while accounting for the diverse problems associated with historical data, particularly detectability. In subsequent chapters, occupancy modeling is the key analytical tool I employ permitting robust comparisons of occurrence data across time.

Given climate change, any particular species is expected to shift in geographical space in order to track, or remain within, its favorable niche in climate space. Using occurrence data for 53 western US birds from all resurveyed sites, my second chapter investigates whether documented occurrence changes over the 20th century provide evidence in support of niche tracking. Based on movement directions of occurrence in climate space, we found evidence of climatic niche tracking for 91% of species, with some species tracking only temperature and some tracking only precipitation. Additionally, there was a strong relationship between the environmental factors limiting species' ranges on a continental scale, and the factors tracked over time. Two-season occupancy models further demonstrated that extinction and colonization probabilities for a species were most strongly related to the climatic relationship between a site and the species' niche centroid.

Evidence that species are tracking their climatic niche does not, per se, describe how species are moving in geographic space. Using a larger sample of 99 bird species, my third chapter catalogues the elevational movements of species over time and seeks to test the naïve hypothesis that all species will shift upward in elevation in order to track a warming climate. While species did, on the whole, shift up more than they shifted down, this naïve hypothesis only described 56% of measured range shifts. Alternative hypotheses providing species specific predictions of upward or downward shifts based on site-specific climate change almost universally outperformed the naïve hypothesis. Many species did not shift in parts or all of their range, despite climatic expectations to do so, and traits defining these species and differentiating them from moving species are explored.

The cumulative effect of hundreds of species shifting geographic ranges in an individualistic manner has an unknown effect on local species diversity. My fourth chapter uses a hierarchical multi-species occupancy model to estimate richness change and turnover at sites based on all 210 observed breeding bird species in the Sierra Nevada. The results illustrate that richness has broadly declined across all elevations, but that turnover has been greatest at the lowest and the highest elevations. The results also demonstrate the importance of accounting for detectability using methods such as occupancy modeling, as analyses of community change using naïve detections of species show opposite trends than those inferred when accounting for false absences.

Overall, my dissertation provides a detailed picture, over a uniquely long time span and broad geographic area, of how bird species have responded spatially to changing climates over a century. These movements have been shown to be more diverse than previously described and are not likely to be predicted by simple ecological relationships. As we prepare for greater climate shifts during the 21st century, recovering and utilizing our knowledge of the past will be critical to anticipating and understanding the impacts of the future.