A Unified Approach to Characterizing Fracture-Flow Systems: Coupling Radar Tomography, Tracer Experiments, and Hydraulic Data

9705812戈雷里克水文地质学中最紧迫的问题之一是描述裂隙岩石系统的非均质性。为了建立水力水头变化和溶质运移的预测模型，需要对控制裂缝流动的水力特性进行准确的三维表征。这个问题一直很困难，因为通常使用单一类型的数据(例如水力试验)来推断裂缝渗流路径和渗流障碍的复杂模式。对数据集的单独分析可能会导致裂隙岩石几何形状和水力特性的“地图”彼此不一致。我们认为，同时反演多种类型的信息，从地球物理数据(本地和井间雷达)到水压试验再到示踪剂试验，提供了一种可以识别裂隙-岩石含水层非均质性的方法，包括裂隙带几何形状、连通性和传输特性。每种类型的数据都用适当的物理模拟模型进行了分析，包括关于波传播、水头变化和溶质迁移的模型。我们提出了一种基于一种非常灵活的优化过程的耦合反转过程，称为模拟退火法。该方法将产生与所有类型的数据(如水头、浓度、雷达)一致的图像，并显示裂隙带几何形状、水力传导性估计和空间可变的不确定性估计。它将被开发并应用于合成数据和现场数据。我们过去对沉积矿床的经验表明，合成数据分析练习可以促进“耦合”反演方法的快速发展和测试。一旦了解了该方法的优点和局限性，对复杂野外数据的应用提供了对该方法的真正检验。现场数据将来自新罕布夏州镜湖的岩石破裂实验场。到目前为止，美国地质勘探局已经收集了雷达层析成像、地震层析成像、钻孔地球物理记录S、水压试验数据和示踪剂试验数据。我们认为这是一个独特的机会，他们愿意向我们提供这些新收集的数据。该项目将深入了解可以从不同类型的数据中提取的信息、数据的相对价值，以及如何使用对不同类型的数据的联合分析来推断断裂带的结构。

9705812 Gorelick Among the most pressing problems in hydrogeology is describing heterogeneity in fractured-rock systems. Accurate 3D characterization of the hydraulic properties controlling fracture flow is required for predictive models of hydraulic head changes and solute migration. The problem has been difficult because typically a single type of data (e.g., hydraulic tests) has been used to deduce the complex pattern of fracture flow paths and flow barriers. Individual analysis of data set may lead to "maps" of fractured rock geometry and hydraulic properties that are inconsistent with each other. We suggest that simultaneous inversion of multiple types of information ranging from geophysical data (local and cross well radar) to hydraulic tests to tracer tests provides an approach that can identify fractured-rock aquifer heterogeneities including fracture zone geometry, connectivity, and transmissive properties. Each type of data is analyzed with the appropriate physical simulation models, including ones for wave propagation, hydraulic head changes, and solute migration. We propose a coupled-inversion process based on a very flexible optimization procedure known as simulated annealing. The method will yield images that are consistent with all the types of data (e.g., heads, concentrations, radar) and indicate fracture-zone geometry, hydraulic conductivity estimates, and spatially variable uncertainty estimates. It will be developed and applied to both synthetic and field data. Our past experience for sedimentary deposits has shown that synthetic data analysis exercises can promotes rapid development and testing of the "coupled" inversion approach. Once the benefits and limitations of the proposed method are understood, application to complex field data provides a true test of the method. The field data will be from the fractured-rock experimental site at Mirror Lake, New Hampshire. To date, the USGS has collected radar tomography, seismic tomography, borehole geophysical log s, hydraulic test data and tracer test data. In what we consider to be a unique opportunity, they are willing to make these newly collected data available to us. The project will provide insight into the information that can be extracted from different types of data, the relative worth of data, and how joint analysis of different types of data can be used to deduce fracture-zone architecture.