Ludovic Schorpp ^* Julien Straubhaar Philippe Renard

• Centre d'hydrogéologie et de géothermie, Faculté des sciences, Université de Neuchâtel, Neuchâtel, Neuchâtel, Switzerland

The Stratigraphic Pile (SP) is one of the foundation of most geological studies. It represents, in a compact and practical way, a vertical succession of depositional events over geological time.Accurate definition of the SP is crucial for geological modeling, yet challenges arise when relying on borehole data in the absence of clear biostratigraphic indicators or chronostratigraphical data. This manuscript introduces an algorithm designed to automatically determine the SP using borehole unit sequences. The algorithm also addresses the complexities associated with incomplete sedimentological records and subjective geological interpretations. The algorithm was tested on various datasets, taking into account differences in the number of boreholes and available information. The efficiency of the algorithm was demonstrated through real-world applications, providing a basis for a comprehensive discussion of its advantages, limitations, and potential applications. The proposed methodology assumes that each borehole contains a single occurrence of a stratigraphic unit, taking into account possible interpretation errors and inconsistencies. The algorithm is capable of: automatically determining one or an ensemble of plausible stratigraphic sequences, identifying potential misinterpreted wells, quantifying the vertical relationships of the stratigraphic units, and assisting in the data preprocessing step and in building the geologic concept of the modeling area. In particular, this ensemble of SPs and identified inconsistencies provide valuable insights into the geological history and concepts for a particular area. This research contributes to the refinement of geological modeling workflows and provides a valuable tool for automatic refinement of SP selection.