Abstract Geothermal heat flow (GHF) is a critical parameter for understanding the thermal structure and dynamics of the lithosphere, providing critical insights into lithospheric thermal evolution and geothermal energy potential. This study investigates the spatial variability of GHF in Germany by applying a Bayesian Markov Chain Monte Carlo method to estimate key thermal parameters, including crustal and mantle thermal conductivities, crustal heat production, and mantle heat flow. The analysis integrates data on surface heat flow, surface temperatures, and the lithosphere‐asthenosphere boundary depth. To address the limitations posed by the sparse and uneven distribution of direct borehole measurements, comprising only 595 GHF records, we incorporated a wide range of geophysical and geological constraints, such as gravity, magnetics, seismic velocity, topography, and proximity to faults and volcanic regions. These data sets were analyzed using a Quantile Regression Forest approach that enabled robust GHF estimations, while accounting for uncertainties and providing reliable prediction intervals. This methodology significantly improves upon traditional Curie Point Depth‐based methods, providing a more accurate and comprehensive GHF model for Germany. The probabilistic multi‐observable approach enhances GHF estimates in Germany, improving constraints on geothermal resources and the lithospheric thermal state.
Sobh et al. (Wed,) studied this question.