Data from: Influence of topographic features and vegetation on climate refugia in the boreal plains

Climate-change refugia can be defined by the strength of influence of physiographic and biotic properties on local atmospheric conditions relative to regional averages. In flatter landscapes, thermal heterogeneity and decoupling may be limited. We sought to quantify the magnitude of climate decoupling produced by topographic variation and vegetation in North America's boreal-parkland-grassland transition zone. We deployed iButton data loggers at four hill and 12 river valley systems along a latitudinal gradient in Alberta, Canada (N = 240) between 2014 and 2018. We modelled monthly temperature metrics by season as a function of topography and percent vegetation cover using general linear mixed effect models. Summer Tmax varied ~6 oC across the entire elevation gradient sampled, and ~3.5 oC within the range of topographic conditions sampled, after accounting for latitude. We found that canopy cover type had a negative effect on most temperature metrics, with coniferous forests having the strongest effect on Tmin and Tave. Spatial predictions derived from models based on iButton sensors resulted in similar regional patterns compared to a gridded climate product (ClimateNA), but the difference between them increased with potential solar radiation, vegetation cover and terrain roughness. We observed that local Tmax on steep, north-facing slopes with relatively low levels of potential solar radiation were 0.5 – 1.5 oC cooler than the surrounding landscape on average, while terrain roughness may further contribute to 0.5 – 0.7 oC. When combined, these local landscape characteristics may compensate for expected warming by the end of the century. Topographic and vegetation characteristics that are related to colder local climates should be considered in the evaluation of future climate-change impacts and to identify potential refugia from climate-change.