The characterization of the mechanical properties and stress state of a hydrocarbon or geothermal reservoir and its overburden is crucial for its optimal exploration and exploitation. We present a geomechanical modeling study from the Lower Magdalena Valley Basin in northern Colombia that integrates various borehole and seismic data-sets to build a 3D model for a robust prediction of the mechanical property distribution and the pre-production stress state. The workflow includes the analysis of log and core data as well as hydraulic tests leading to detailed 1D mechanical earth models (MEM) for each of the five wells currently available in the study area. Subsequently, this information is upscaled to and integrated with a structural model derived from 3D seismic interpretation. For this 3D MEM we test two different property population methods and compare the resulting stress prediction. The first method uses a geostatistical approach based exclusively on well data for property interpolation, whereas the second method additionally utilizes seismic inversion techniques to account for the vertical and horizontal differences in mechanical rock properties. The results include a mechanical characterization of the subsurface as well as the complete stress tensor for the reservoir and encompassing formations in the model domain. The simulation results show that the dominant stress regime in the study area is a normal faulting regime with a governing orientation of $S_{Hmax}$ in the WNW–ESE direction. At reservoir depth, the vertical stress gradient (${\mathit{S}}_v$) has a mean value of 23.29 MPa/km, $S_{Hmax}$ is on average 1.08$\times$Shmin and 0.8$\times$ ${\mathit{S}}_v$. Comparison of the two property population methods shows substantial enhancements by using the seismic-driven approach to distribute log-derived elastic properties. This worked geomechanical modeling example from Colombia illustrates the potential of 1D and 3D MEM’s for mechanical and stress characterization and can be used as template for similar studies elsewhere.

碳氢化合物或地热储层及其覆盖层的力学性能和应力状态的表征对于其最佳勘探和开发至关重要。我们提出了一项来自哥伦比亚北部下马格达莱纳谷盆地的地质力学建模研究，该研究整合了各种钻孔和地震数据集，从而建立了3D模型，以可靠地预测机械性能分布和生产前的应力状态。该工作流程包括对测井和岩心数据的分析，以及对测试区域中当前可用的五口井中的每口井进行详细的一维机械地球模型（MEM）的水力测试。随后，此信息将按比例放大并与从3D地震解释中得出的结构模型集成在一起。对于此3D MEM，我们测试了两种不同的特性填充方法，并比较了结果应力预测。第一种方法使用专门基于井数据的地统计方法进行属性插值，而第二种方法另外利用地震反演技术来解决机械岩石属性的垂直和水平差异。结果包括地下力学特征以及储层和模型域内地层的完整应力张量。仿真结果表明，研究区的主导应力状态是正常的断层状态，其控制方向为。而第二种方法还利用地震反演技术来解决机械岩石特性的垂直和水平差异。结果包括地下力学特征以及储层和模型域内地层的完整应力张量。仿真结果表明，研究区的主导应力状态是正常的断层状态，其控制方向为。而第二种方法还利用地震反演技术来解决机械岩石特性的垂直和水平差异。结果包括地下力学特征以及储层和模型域内地层的完整应力张量。仿真结果表明，研究区的主导应力状态是正常的断层状态，其控制方向为。${\mathrm{小号}}_{H米\mathrm{一种}X}$沿WNW–ESE方向。在油藏深度处，垂直应力梯度（${\mathit{小号}}_v$）的平均值为23.29 MPa / km， ${\mathrm{小号}}_{H米\mathrm{一种}X}$ 平均为1.08$\times$Shmin和0.8$\times$ ${\mathit{小号}}_v$。两种属性填充方法的比较显示，通过使用地震驱动的方法来分布对数派生的弹性属性，可以显着提高性能。这个来自哥伦比亚的地质力学建模实例说明了1D和3D MEM在机械和应力表征方面的潜力，可以用作其他地方类似研究的模板。