Learning from Finland: The role of geology and Radiogenic Heat Production to predict subglacial heat flow in Greenland

Understanding the spatial variability of geothermal heat flow (GHF) in Greenland is critical for predicting ice sheet dynamics. However, the scarcity of direct observations complicates GHF predictions and most samples of thermal parameters are from its coasts, which requires some form of extrapolation or modelling to predict these under the Greenland ice sheet.

Finland, in turn has a uniquely dense coverage of radiogenic heat production (RHP) measurements. Here, the RHP data reveal a significant spatial variability, which can be correlated with tectonic age units. However, RHP also exhibits variability within the units, which we find to be describable as a random field with two correlation lengths of <10 and c. 50 km as well as substantial white noise variability. If we decimate the Finish data to a strip at the Finish coast and inland border, the statistical distribution (mean and standard deviation) for each unit is almost identical to the whole data set. Hence, using the limited data for Greenland to predict thermal parameters under the ice is justified, if the subglacial geology was known.

Subglacial geology is, however, not well known for Greenland and we test two geological maps. Using conditional simulation, we extrapolate RHP from coastal rock samples, assuming the statistical properties are valid for the whole units. Based on the Finish data, we treat the within-unit variability using a random field with a spatial length scale of 50 km and an amplitude equal to 10% of total standard deviation. We choose to neglect the smaller-scale variations, as they are likely not affecting surface heat flow, due to the smoothing effect of heat diffusion.

Our results show that both geological maps predict significantly different GHF values of up to 20 mW/m² differences within the local spots in the interior of Greenland with a similar unit wide GHF. In northern Greenland and at the transition between southern and central Greenland larger scale differences of about 30 mW/m² can be found, which come from the different structure of the units of both maps and thus significantly different mean RHP values. This underlines the need for reliable geological maps to constrain the distribution of RHP.