Temperature data
We combined the Rallarvägen 2021 dataset with soil temperature estimates from the ‘Global maps of soil temperature’, which provides global modeled soil temperatures averaged for the period 1979-2013 at a spatial resolution of 1 km² for 0-5 cm depth (SoilTemp; Lembrechts et al. 2021). These maps were constructed by calculating differences (i.e., offsets) between in situ soil temperature time series from over 1,200 1-km² pixels, summarized from 8,500 unique temperature sensors across all the world’s major terrestrial biomes, and coarse-grained air temperature estimates from ERA5-Land (Lembrechts et al. 2020). ERA5-Land provides a reanalysis dataset with monthly free-air temperatures at a spatial resolution of 0.1 x 0.1 degrees (Copernicus Climate Change Service (C3S), 2019). We used the mean annual soil temperature (SBIO1), mean soil temperature of the warmest quarter (SBIO10; hereafter named Mean Summer Soil Temperature (MSST)), and mean soil temperature of the coldest quarter (SBIO11). We extracted temperatures for each transect using the extract function from the raster package in R (Hijmans, 2021).