Summary: Carbonate lithology identification has long posed challenges in petroleum exploration due to overlapping well logging responses caused by compositional similarities and the thin-bedded nature of transitional lithologies. Although machine learning (ML) has provided new approaches to address this issue, the “black-box” nature of most ML models limits their geological interpretability and practical reliability. In this study, a high-precision lithology classification model was applied and coupled with the Shapley additive explanations (SHAP) framework to quantitatively evaluate the contribution and directionality of each logging feature to the model’s predictions. By combining K-means oversampling with the XGBoost algorithm, the training data were effectively balanced, and the robustness of the model was improved, ultimately achieving an overall classification accuracy of 92%. SHAP analysis further clarified the controlling influence of key logging parameters on lithology discrimination, revealing the model’s internal decision logic in a physically meaningful manner. This research establishes an interpretable and data-driven workflow for carbonate lithology identification and provides a theoretical and technical reference for the reliable application of ML in complex carbonate reservoir characterization.
Flute casts are diagnostic erosional sole structures of deep-water gravity-flow deposits. Yet, how reliably their dimensions and morphology record gravity-flow dynamics under the coupled influence of flow conditions and substrate properties remains poorly constrained. This study integrates field observations, morphometric measurements, sedimentological and grain-size analyses, and elemental geochemistry to investigate flute casts from the Upper Ordovician Lashizhong Formation, western Ordos Basin. The measured flute casts display substantial dimensional variability, ranging from 2.0 to 36.0 cm in length, 0.9 to 20.0 cm in width, and 0.1 to 5.5 cm in depth. Flute-cast width exhibits the strongest correlation with sandstone bed thickness (r = 0.672–0.862) and reflects the lateral extent of substrate erosion, whereas depth (r = 0.436–0.700) is more sensitive to substrate properties and local flow structures. Grain size records variations in flow energy and sediment transport yet does not directly determine flute-cast dimensions. From channel to lobe environments, flute casts generally decrease in size and transition from large and complex morphologies to narrow parabolic and spindle-shaped forms, reflecting flow attenuation and rheological evolution. Substrate erodibility further modulates their development. The temporal lag of morphological adjustment and the spatial offset between flow separation and maximum erosion indicate that preserved flute casts record successive stages of gravity-flow evolution rather than instantaneous flow conditions. Large flute casts in Member 3 may record episodic high-energy gravity flows generated within a tectonically influenced and potentially unstable slope system, although their specific triggering mechanism remains uncertain. These findings clarify the flow–substrate controls on flute-cast development in deep-water gravity-flow systems.