Ice Borehole Thermometry: Paleo-Climate reconstruction using Bayesian modeling

In the quest of understanding the past and present climate system, we aim to reconstruct paleo-climate using Bayesian inversion from Antarctic borehole temperature profiles. We aim to develop a refined process for reconstructing the temperature evolution of Antarctica over the last century and millennia. Initially, a forward heat transfer model is implemented that simulates borehole temperature profiles for time-dependent surface temperatures. The forward model provides an approximate notion of the borehole depth at which a signal from the past may be obtained. Using forward simulations, a greedy approach is employed for the optimal placement of temperature sensors in the borehole to record temperature effectively. We invert the forward model to reconstruct past surface temperature evolution from borehole temperature measurements. For this, we apply Bayesian inference to optimally account for the uncertainty in the various influencing quantities. We model known uncertainties as priors and obtain the reconstructed surface temperatures with connected uncertainty information.