Climate‐related range shifts in Arctic‐breeding shorebirds

Abstract

Aim To test whether the occupancy of shorebirds has changed in the eastern Canadian Arctic, and whether these changes could indicate that shorebird distributions are shifting in response to long‐term climate change. Location Foxe Basin and Rasmussen Lowlands, Nunavut, Canada. Methods We used a unique set of observations, made 25 years apart, using general linear models to test if there was a relationship between changes in shorebird species' occupancy and their species temperature Index, a simple version of a species climate envelope. Results Changes in occupancy and density varied widely across species, with some increasing and some decreasing. This is despite that overall population trends are known to be negative for all of these species based on surveys during migration. The changes in occupancy that we observed were positively related to the species temperature index, such that the warmer‐breeding species appear to be moving into these regions, while colder‐breeding species appear to be shifting out of the regions, likely northward. Main Conclusions Our results suggest that we should be concerned about declining breeding habitat availability for bird species whose current breeding ranges are centered on higher and colder latitudes.

Unexpected sources of uncertainty in projecting habitat shifts for Arctic shorebirds under climate change

“…For our habitat suitability models, we used two categories of environmental predictors of shorebird occupancy: climate predictors found in ‘climate‐only models’, and additional predictors in ‘climate+additional models’ (see Anderson, Fahrig, Rausch, & Smith, 2023 for more details). For the climate predictors, we used all 19 bioclimatic variables available from WorldClim 2 (Fick & Hijmans, 2017), which are interpolated from mean historical climate data covering 1970–2000 at a 30 arc‐second resolution (~1 km).…”

“…In our models, there are lags between the observation period for shorebirds, and the observation period for the climate data (1970–2000) and the land cover data (1993–1995). We chose these data sets because they are the only options available that would allow us to include future land cover predictors and the influence of tree dispersal on shorebirds, which we identified as being important variables in Anderson, Fahrig, Rausch, and Smith (2023). We also believe there is likely a true lag in the responses between climate, landcover, and shorebird occupancy, so these time lags are a reasonable modelling assumption.…”

“…All analyses were performed in R version 4.2.3, using packages tidyverse 2.0.0, caret 6.0-94, and terra 1.7-29. See Anderson et al (2024) for the full dataset used in our analysis.…”

“…These Arctic‐breeding shorebirds appear to be shifting their ranges northwards in response to recent climate change (Anderson, Fahrig, Rausch, Martin, et al., 2023). A qualitative risk assessment of the vulnerability of Arctic‐breeding shorebirds to climate change in North America predicted that many of these species will lose more than half of their suitable breeding habitat, at a high level of subjective confidence (>70%, Hector Galbraith et al., 2014).…”

“…Second, we compared the consequences of the pool of candidate variables available for selection. We compare such climate‐only models and models including additional non‐climate variables, which are important predictors of Arctic shorebird distributions (Anderson, Fahrig, Rausch, & Smith, 2023). Third, we tested how assumed dispersal limits in projections of future vegetation affect uncertainty.…”