Theme: 05. Fractured Rock and Complex Geology

Topic Area: 5d. Process-Based Conceptual Site Models (CSMs) for Informing Remediation

Formatted Title:

Process-Based Conceptual Site Model Development for Complex Fractured Sedimentary Bedrock Sites

Background/Objectives:

To develop conceptual site models (CSMs) at sites impacted by chlorinated solvents in fractured bedrock environments, a processed-based approach should be used. In general, geologic conditions inform hydrogeologic processes which inform the physical contaminant transport, fate, and natural attenuation processes (e.g., advection, dispersion, and matrix diffusion). It is critical to consider geologic processes associated with depositional environment or tectonic history that can form preferential groundwater flow pathways and groundwater flow inhibitors. Establishing a link between key geologic processes, hydrogeologic features, and transport and fate processes allow for the development of CSMs with predictive capabilities. These processes occur at different scales and therefore data should be collected at commensurate scales. High-resolution site characterization (HRSC) data sets are preferred for development of predictive CSMs. However, at many sites, significant four-dimensional traditional datasets already exist. Therefore, we typically design focused HRSC programs to augment these historical datasets to develop robust CSMs with predictive capabilities. This talk will discuss how we successfully developed and implemented processed-based CSMs using HRSC and traditional data sets at fractured sedimentary bedrock sites.

Approach/Activities:

Our process-based approach for developing predictive CSMs for fractured bedrock sites combines the use of an extensive literature review of the local and regional geology, and integration of regionally specific hydrogeologic models (e.g., Leaky-Multiple-Aquifer-system [LMAS] and slope-aquifer system [SAS]), when available, and implementation of the Discrete Fracture Network - Matrix (DFN-M) characterization toolbox approach, as needed. We first develop the three-dimensional geologic and hydrogeologic models by applying environmental sequence stratigraphy (ESS) to interpret various types of geologic data, including data derived from local and regional geologic literature, outcrops, surface geophysics, borehole geophysics, and rock cores, as available. The geologic model is then used to develop an initial hydrogeologic model using our understanding of relevant scientific principles. High-resolution vertical contaminant profiling is completed in select bedrock boreholes using Discrete Fracture Network (CORE^DFN) and/or FACT (FLUTe Activated Carbon Technique) technologies to evaluate the relationship between contaminant mass and key hydrogeologic features. Using the geology and hydrogeology models and high-resolution vertical contaminant distribution data, we design and construct a monitoring well network. Temporal data collected from the monitoring well network can be used to develop a dynamic groundwater flow and contaminant distribution model. We then use chemical fingerprinting, forensics analysis, and sometimes compound-specific isotope analysis (CSIA) to further refine the model.

Results/Lessons Learned:

We will discuss how site-specific process-based CSMs were used to predict aquifer and aquitard geometry, contaminant plume architecture and extent, and differentiate plumes originating from onsite and offsite sources. We will discuss how our process-based CSMs were implemented to avoid additional characterization requirements and achieve expedited regulatory closure. We will discuss how processes observed at different scales informed a dynamic remediation strategy for an expedited site closure timeline. We will discuss how this approach can be used to identify natural attenuation processes that provide insights on the assimilation capacity of fractured sedimentary bedrock aquifers. We will discuss site-specific limitations for effective implementation of DFN-M tools. We will also discuss results that did not agree with the hypothesis made using the predictive CSM.

Menu

Maddie Hayes