Formatted Title
“My Site is a Sandbox” and Other Myths: Heterogeneity in Fluvial Aquifers and the Scale at Which it Matters
Background/Objectives
“This site is a sandbox” is an often-heard phrase in groundwater remediation. Poor core recovery, low resolution lithologic logging methodologies, and only basic understanding of clastic depositional environments and the scales of heterogeneity observed within them has contributed to the use of this dismissive and misleading site description. Simplification of site lithologic characterization is unfortunately also harmonious with the pervasive simplifying assumptions of homogeneous and isotropic permeability distribution in aquifers often used in modeling and initial remedial design. Industry history shows that many remedies have fallen short of predicted outcomes due to the inherent heterogeneity in the permeability architecture of aquifers. Are there really any sites where “The Sandbox Assumption” applies? The objective of this presentation is to demonstrate the scale of heterogeneity present in some of the sandiest of depositional environments encountered in Midwestern alluvial valley aquifer systems, braided fluvial deposits, and then use site examples to investigate “does this heterogeneity matter” for remediation?
Approach/Activities
We use a database of lithologic logs, high resolution site characterization (HRSC) data (electrical conductivity-hydraulic profiling tool, geophysical logs), geotechnical samples, and traditional hydraulic characterization (slug and pump test data) from multiple sites and literature review to address four questions:
- What is the scale and lithologic heterogeneity present in braided fluvial systems (both modern analogues and midwestern site examples)?
- Does the magnitude of grainsize variation present result in observable and significant changes in hydraulic properties that would alter preferential pathways in the subsurface?
- What are the appropriate strategies and scales for the correlation of hydrostratigraphic units (HSUs) in braided fluvial settings?
- What is the most effective strategy for communicating the observed and hydraulically significant heterogeneity to inform remedial design and/or the groundwater models?
Results/Lessons Learned
HRSC and lithologic logs from braided fluvial deposits in the Missouri River Valley, Michigan, and Nebraska show that within the overall sand-rich aquifer 10 to 25 ft thick fining-upward packages ranging in grainsize from sandy gravel to sandy silt are readily identifiable deposits associated with channel-axes and complex bar formation. Initial results suggest that hydraulic conductivity across such deposits ranged from less than 10 ft/day to > 300 ft/day. Permeable bar deposits are typically no greater than hundreds to 1000 ft in length and tens to hundreds of ft in width. However, complex cross-cutting relationships between successions of bars and active channels result in a depositional fabric that promotes flow in the direction of initial deposition. Preferential pathways consist of multiple bar and axial channel deposits imposed on one another. Correlation of HSUs using a combination of interpolated HRSC data and facies-based stratigraphic interpretation such as environmental sequence stratigraphy (ESS) can effectively define, correlate, and communicate the variations in HSUs that can impact remediation design (e.g., permeable reactive barrier and/or extraction well placement) and timeframes. These examples show that in aquifers previously referred to as “sandbox” there exists more than two orders of magnitude of hydraulic conductivity depending on the facies. Incorporating this knowledge into the remedy design greatly improves predictability of outcomes and remedy timelines.