The subsurface temperature field and the geothermal conditions in sedimentary basins are frequently examined by using numerical thermal models. For those models, detailed knowledge of rock thermal properties are paramount for a reliable parameterization of layer properties and boundary conditions. Despite the state-of-the-art in other research fields (e.g. hydrogeological ground-water models) where the spatial permeability variations within geological layers is often considered, parameterization of the major rock thermal properties (in particular thermal conductivity, to minor extent radiogenic heat production and specific heat capacity) is almost always conducted by applying constant parameters for each modelled layer. Moreover, initial parameter values are usually obtained from only few core measurements and/or literature values, which raise questions for their representativeness. Only some rare studies have considered detailed lithological composition or well log information, still with constant layer properties.This study presents a thermal-modelling scenario analysis in which we demonstrate how the use of both different parameter input type (from literature, lithology and well logs) and parameter input style (constant or spatially varying layer values) affects model temperature predictions in sedimentary basins. It is a case study located in the Danish-German border region at the northern margin of the North German Basin. To conduct the scenario analysis, rock thermal properties are determined from lithological descriptions and standard petrophysical well logs for several boreholes in the area of study. Statistical values of rock thermal properties are derived for each geological formation at each well location and, furthermore, for the entire dataset. The thermal model is validated against known observed temperatures of good quality.Results clearly show that the use of location-specific well-log derived rock thermal properties and the integration of laterally varying input data (reflecting changes of lithofacies) significantly improves the temperature prediction. The parameterization from boreholes always prevails over the parameterization based on literature values, and it allows for reducing uncertainty of model temperatures by up to 80%.
Read full abstract