[S25-P] Poster

[S25-P-02] Evaluating methods and uncertainties in the inversion of downhole temperature data for palaeoclimate studies, Australian examples

^○Sandra McLaren, Roger Powell (University of Melbourne, Melbourne, Australia)

Downhole temperature data have been used for many years to try and understand variations in ground and air surface temperature in the geological past. The inference of ground surface temperature histories from borehole temperature data is an ill-posed inverse problem that can be approached via a range of numerical methods, broadly divisible into (1) regularisation and (2) Bayesian techniques. Inversion results are dependent on the method chosen and the details of the implementation, as well as the geological constraints built into the inversion. A number of authors have compared different approaches exploring, for example, the impact of variations in geological and geophysical parameters such as borehole depth and uncertainties in measured thermal data on inversion results. We add to this body of work looking at the sensitivity of different inversions of realistic synthetic temperature data. On the basis of this, we invert real temperature data collected from two purpose-drilled and well-characterised boreholes in southeastern Australia: Tynong-1, southeast of Melbourne (total depth of 400m) and Stavely-07, in western Victoria (total depth of 500m). Results suggest that inversion may result in over-interpretation of the data and that it may be difficult to distinguish equally plausible temperature histories.