Exhumation affects sedimentary basin evolution by influencing structural, pressure and temperature dynamics, thereby impacting energy resource formation. Compaction‐based methods are widely used to quantify exhumation, utilising sonic and porosity data to track sediment uplift from its maximum burial depths. However, uncertainties arise from applying empirical compaction models developed for specific geological regions, highlighting the need for region‐specific models. Even such region‐specific models contain uncertainties, which can compromise exhumation estimates. We, therefore, develop a probabilistic compaction model for the Northwest Shelf Basins using sonic data from normally compacted and unexhumed shales from the Northern Carnarvon Basin (NCB). The model's robustness is estimated using MCMC, and uncertainty propagation analysis is employed to assess the impact of model uncertainty on the model's predictive applications. The model shows exponential porosity reduction with depth, demonstrating rapid compaction from the surface to ca. 2 km and slower compaction thereafter. The model is then applied to interpret new datasets from the Canning, Gippsland and NCB regions. The results reveal that while some parts of the NCB exhibit normal compaction without exhumation, others were significantly exhumed. Conversely, Canning and Gippsland Basin data indicate signs of significant exhumation, as suggested by previous studies, thereby confirming the model's effectiveness outside the Northwest Shelf. Since the model could not explain data from exhumed regions, we inferred new models incorporating “exhumation” parameters to interpret the complex compaction histories of these areas, and the best‐fitting models were selected using the Bayes Factor method. Uncertainty analysis revealed that the impacts of model uncertainty on exhumation estimates are consistent across wide depth ranges. Our findings highlight the need to refine compaction models for better predictive reliability and informed resource exploration in sedimentary basins.

中文翻译：

---

使用贝叶斯推理为北卡那封盆地开发概率压缩模型


挖掘通过影响结构、压力和温度动态来影响沉积盆地的演变，从而影响能源资源的形成。基于压实的方法被广泛用于量化挖掘，利用声波和孔隙度数据来跟踪沉积物从其最大埋藏深度的隆起。然而，应用为特定地质区域开发的经验压实模型会产生不确定性，这凸显了对特定区域模型的需求。即使是这种特定于地区的模型也包含不确定性，这可能会损害挖掘估计。因此，我们使用来自北卡那封盆地 （NCB） 的正常压实和未挖掘页岩的声音数据，为西北大陆架盆地开发了一个概率压实模型。使用 MCMC 估计模型的稳健性，并采用不确定性传播分析来评估模型不确定性对模型预测应用的影响。该模型显示孔隙度随深度呈指数级减少，表明从地表到约 2 公里的快速压实，此后压实速度变慢。然后应用该模型来解释来自 Canning、Gippsland 和 NCB 地区的新数据集。结果表明，虽然 NCB 的某些部分在没有挖掘的情况下表现出正常的压实，但其他部分则被大量挖掘。相反，正如以前的研究所表明的那样，坎宁和吉普斯兰盆地的数据表明有大量挖掘的迹象，从而证实了该模型在西北大陆架之外的有效性。由于该模型无法解释来自挖掘区域的数据，因此我们推断出包含“挖掘”参数的新模型来解释这些区域复杂的压实历史，并使用贝叶斯因子方法选择最佳拟合模型。 不确定性分析表明，模型不确定性对挖掘估计的影响在较宽的深度范围内是一致的。我们的研究结果强调了改进压实模型的必要性，以提高沉积盆地的预测可靠性和明智的资源勘探。