Ground temperatures in the top few centimetres of the soil profile are key in many biological processes yet remain very poorly documented, especially in the polar regions or over longer timescales. They can vary greatly seasonally and at various spatial scales across the often highly complex and heterogeneous polar landscapes. It is challenging and often impossible to extrapolate soil profile temperatures from meteorological air temperature records. Furthermore, despite the justifiably considerable profile given to contemporary large-scale climate change trends, with the exception of some sites on Greenland, few biological microclimate datasets exist that are of sufficient duration to allow robust linkage and comparison with these large-scale trends. However, it is also clear that the responses of the soil-associated biota of the polar regions to projected climate change cannot be adequately understood without improved knowledge of how landscape heterogeneity affects ground and sub-surface biological microclimates, and of descriptions of these microclimates and their patterns and trends at biologically relevant physical and temporal scales. To stimulate research and discussion in this field, we provide an overview of multi-annual temperature records from 20 High Arctic (Svalbard) and maritime Antarctic (Antarctic Peninsula and Scotia Arc) sites. We highlight important features in the datasets that are likely to have influence on biology in polar terrestrial ecosystems, including (a) summer ground and sub-surface temperatures vary much more than air temperatures; (b) winter ground temperatures are generally uncoupled from air temperatures; (c) the ground thawing period may be considerably shorter than that of positive air temperatures; (d) ground and air freeze–thaw patterns differ seasonally between Arctic and Antarctic; (e) rates of ground temperature change are generally low; (f) accumulated thermal sum in the ground usually greatly exceeds air cumulative degree days. The primary purpose of this article is to highlight the utility and biological relevance of such data, and to this end the full datasets are provided here to enable further analyses by the research community, and incorporation in future wider comparative studies.