[1] Borehole-based reconstructions of ground surface temperature (GST) have been widely used as indicators of paleoclimate. These reconstructions assume that heat transport within the subsurface is conductive. Climatic interpretations of GST reconstructions also assume that GST is strongly coupled to surface air temperature (SAT) on timescales of decades and longer. We examine these two assumptions using records of SAT and subsurface temperature time series from Fargo, North Dakota; Prague, Czech Republic; Cape Henlopen State Park, Delaware; and Cape Hatteras National Seashore, North Carolina. The characteristics of downward propagating annual temperature signals at each site clearly indicate that heat transport can be described as one-dimensional conduction in a homogeneous medium. Extrapolations of subsurface observations to the ground surface yield estimates of annual GST signals and allow comparisons to annual SAT signals. All annual GST signals are modestly attenuated and negligibly phase shifted relative to SAT. The four sites collectively demonstrate that differences between annual GST and SAT signals arise in both summer and winter seasons, in amounts dependent on the climatic setting of each site.

No abstract

[1] Inversions of borehole temperature profiles that reconstruct past ground surface temperature (GST) changes have been used to estimate historical changes in surface air temperature (SAT). Paleoclimatic interpretations of GST reconstructions are based on the assumption that GST and SAT changes are closely coupled over decades, centuries, and longer. This assumption has been the subject of some debate because of known differences between GST and SAT at timescales of hours, days, seasons, and years. We investigate GST and SAT relationships on daily, seasonal, and annual timescales to identify and characterize the principal meteorological changes that lead to short-term differences between GST and SAT and consider the effects of those differences on coupling between the two temperatures over much longer time periods. We use observational SAT and subsurface data from Fargo, North Dakota; Prague, Czech Republic; Cape Henlopen State Park, Delaware; and Cape Hatteras National Seashore, North Carolina. These records comprise intradaily observations that span parts of one or two decades. We compare subsurface temperature observations to calculations from a conductive subsurface model driven with daily SAT as the surface boundary condition and show that daily differences exist between observed and modeled subsurface temperatures. We also analyze year-to-year spectral decompositions of daily SAT and subsurface temperature time series and show that dissimilarities between mean annual GST and SAT are attributable to differences in annual amplitudes of the two temperature signals. The seasonal partitioning of these amplitude differences varies from year to year and from site to site, responding to variable evapotranspiration and cryogenic effects. Variable yearto-year differences between mean annual GST and SAT are closely estimated using results from a multivariate regression model that associates the partial influences of seasonal meteorological conditions with the attenuation of annual GST amplitudes.

No abstract

No abstract