Spectral decomposition reveals Hidden Fluvial channels: Implications for geothermal potential in the Karlsruhe Area, Upper Rhine Graben, Germany

This study employed advanced seismic analysis techniques to enhance our understanding of subsurface geology for geothermal utilization. Focusing on a rectangular 10 km × 4 km area encompassing the Campus North of the Karlsruhe Institute of Technology (KIT) and including the former Leopoldshafen oil field, we investigated the ca. 250 ± 25 m thick late Oligocene (Chattian) Niederrödern Formation to analyze the geometries of up to 15 m thick sand layers. The initial interpretation of a pre-stack depth-migrated (PSDM) seismic cube using conventional attributes failed to resolve these small-scale features. To overcome these limitations, we implemented spectral decomposition techniques that convert seismic signals into their constituent frequencies using a short-time Fourier transform (STFT) algorithm. The applied spectral decomposition covered a frequency range of 10–60 Hz, with optimal imaging achieved at 10, 26, and 40 Hz. A total of 25 fluvial channels were identified within the Late Oligocene interval at depths between 178 and 1155 m. Individual channels exhibit widths of approximately 25–100 m and spectrally derived thicknesses of 15–24 m, which were previously unrecognized using conventional seismic interpretation techniques, spectral decomposition improved lateral channel detectability by more than one tuning thickness, enabling the resolution of features below the nominal seismic vertical resolution. Particular emphasis is placed on analyzing the geometry and dimensions of these channels to enable more specific thermo-hydraulic (TH) modeling and assess their suitability and boundary conditions for thermal storage. The fluvial channels appear to be stacked within the generally >200 m thick stratigraphic succession, distributed across the stratigraphic record and the three tectonic blocks. In our small survey, the apparent decrease in sinuosity with time may reflect an increase in stream power over time, possibly due to the initiation of subsidence in the Heidelberg Basin, which is the most distinct late Oligocene-Neogene depocenter in the northern URG. • First use of seismic spectral decomposition to image Oligocene channels in Karlsruhe. • Detection of 25 meandering channels as thin as 25 m in the Niederrödern Formation. • RGB-blended STFT improves thin-bed resolution beyond conventional seismic attributes. • Channel networks refine geothermal reservoir models and thermal storage assessment. • Workflow applicable to rift basins and other complex geothermal settings worldwide.