Introduction
Changes in elevational distributions of diverse taxonomic groups are occurring globally, with the potential to reshape ecological communities, alter ecosystem function, and affect climate (Pecl et al. 2017). Given that temperature generally decreases by 0.6°C per 100 m increase in elevation (Barry 2008), and the assumption that species distributions are limited by temperature, a dominant paradigm suggests that montane animals and plants will move upslope in response to climate change (Martin 2001, McNab 2003). Theories of upslope elevational shifts in multiple taxonomic groups are often supported by field data (Chen et al. 2011, Freeman et al. 2018). However, in many instances, populations and species are not moving in synchrony with warming temperatures and either have not shifted or shifted downslope (Tingley et al. 2012, Campos-Cerqueira et al. 2017, DeLuca and King 2017, Freeman et al. 2018). The interacting effects of temperature, precipitation, and other climatic variables could lead to large variation in interspecific or intraspecific responses to climate change (Tingley et al. 2012). The direct or indirect effects of land use may further lead to unpredictable shifts in species’ distributions (Fleishman and Murphy 2012).
Understanding mechanisms of elevational range shifts is complicated by the fact that climate variables are not changing uniformly across elevational gradients. For example, in some cases, high elevations are warming more rapidly than low elevations (Pepin et al. 2015). Precipitation rates also typically are greater at high elevations, and climate change is projected to strengthen this relation (Barry 2008, Van Tatenhove et al. 2019). Theoretically, differences in rates of climate change along elevational gradients may result in larger distributional shifts at the upper edges than at the lower edges of a species’ elevational distribution. Additionally, biotic interactions, especially competition, may be stronger at lower distributional limits, stabilizing distributions at the lower edge (Alexander et al. 2015). Conversely, high-elevation taxa often have greater thermal tolerance than low-elevation taxa, proportional to the magnitude of seasonal and diel thermal variation at high elevations (Janzen 1967, Deutsch et al. 2008). High-elevation species’ physiology may therefore result in relatively small elevational range shifts.
Birds, especially long-distance migrants, are highly vagile, and can track microhabitats within and across years (Greenwood and Harvey 1982, Cline et al. 2013, Gow and Stutchbury 2013). Elevational range shifts of birds have primarily been examined in tropical regions, as species richness and endemism of birds often is concentrated in tropical mountains. Tropical species are generally more physiologically sensitive to temperature than temperate species, and the extent of temperature tracking may be stronger in tropical than in temperate bird species (Pollock et al. 2020, Freeman et al. 2021). However, temperate regions are warming at a faster rate than tropical regions (Friedman et al. 2013), and research on the effects of climate change on non-tropical montane species is essential to understanding how elevational shifts may affect global responses to climate change. Recent research on bird species in North America has identified a variety of elevational shifts. The distributions of 84% of avian species in the Sierra Nevada that were documented by Grinnell in the early 1900s shifted over the past 100 years, with 51% of species moving upslope and 49% moving downslope (Tingley et al. 2012). Over 16 years, 9 of 16 low-elevation passerine species in the northern Appalachian Mountains shifted an average of 99 m upslope, whereas 9 of 11 high-elevation species shifted an average of 19 m downslope (DeLuca and King 2017). In the Adirondack Mountains, repeat surveys found that abundance-weighted mean elevational distributions of 42 species shifted upslope by 83 m over 40 years, with shifts observed at the both the upper and lower elevational range edges (Kirchman & Van Keuren 2017). Although elevational shifts in temperate bird communities appear to be more variable than those in tropical communities (Freeman et al. 2021), sizable yet unexplained variation among species is apparent in both temperate and tropical regions.
Mechanisms of elevational range shifts in birds are difficult to investigate, in part due to data or sampling constraints. Particularly among resurveys of historical sampling locations, variation in distribution or abundance between two time points can impede strong inferences (Sparks and Tryjanowski 2005, McCain et al. 2016). In general, upslope shifts are attributed to the direct and indirect effects of climate change, such as changes in the composition of plant species, plant phenology, or primary productivity (Morison and Morecroft 2006, Lenior et al. 2008, Amano et al. 2010). Downslope shifts in bird populations are often attributed to changes in the predator community or other shifts in interspecific competition (Lenoir et al. 2010). Theorized mechanisms of stable population distributions include temporal lags in species’ responses to climate and land-use changes, stochastic fluctuations in population size, and small magnitudes of climate change (Parmesan et al. 2005, Tingley and Bessinger 2009, McCain et al. 2016).
However, if variability in population size is high, upslope or downslope changes in occupancy may reflect stochastic fluctuations. For example, stochastic increases in population size may lead individuals to colonize unoccupied locations, whereas population declines may result in vacant lower-quality habitat (Thomas and Lennon 1999). Annual variability in abundance may be especially high at the edges of species’ elevational distributions (McCain et al. 2016). Therefore, accounting for population fluctuations is necessary to detect a deterministic distributional shift. Population fluctuations can be identified through long-term data, which can capture annual oscillations that often are undetectable in studies comparing two time points, and by inclusion of tests for population variability in statistical analysis.
We examined whether the elevational distributions of two communities of birds in the Great Basin are shifting. The Great Basin is a cold desert characterized by extensive sagebrush shrubsteppe and variable topography. We are aware of few studies that examined shifts in the elevational distributions of birds in arid ecosystems (Iknayan and Bessinger 2020), where species may be at the edges of both their thermal and xeric tolerances. We examined data from long-term, nearly continuous avian point-count surveys in two regions of the Great Basin to explore potential mechanisms of the shifts. These data span a considerably larger area (>100 km) and greater number of elevational transects (35), and characterize annual variability in occupancy more rigorously, than most resurveys of birds and other taxonomic groups (e.g., Moritz et al. 2008, Tingley et al. 2012). We used single-species occupancy models of 32 species to examine elevational movement at three spatial extents: the full elevational gradient and the lowest and highest 25% of the elevational gradient. Additionally, we examined the effects of temperature, precipitation, and primary productivity on single-species occupancy and elevational movement.