Representing Complex Geologic Environments
The 14-compartment model of Sale and Newell (2011) assigns “low- permeability” compartments to both the source and plume domains, highlighting the potential role of back-diffusion in both domains. Such an approach is conceptually similar to the classification scheme proposed by NRC (2005), which included a hierarchy of five geologic environments ranging from nearly uniformly homogeneous, unconsolidated porous media (Type I) to fractured rock and carbonate aquifers (Types IV and V). While both schemes distinguish between contaminants in “mobile” and “immobile” groundwater, the five-region classification recognizes two subtle but potentially significant differences not captured by the 14-compartment model. First, the diffusion rate and storage capacity of contaminants in low-permeability geologic materials can differ substantially among clays, fractures, and/or intrinsic porosity of indurated rock. Second, in addition to providing potential sinks for diffusive exchange of contaminants, some complex domains (highly heterogeneous unconsolidated porous media, fractured rock, karst) are often characterized by large variations in the groundwater velocity. Hence efforts to characterize “complexity” understood in terms of spatial variability must consider both groundwater flow and contaminant transport within and between discrete compartments, regardless of how such compartments are delineated.
Differences in the diffusion process are relatively straightforward to account for, but require appropriate specification of the geometry and diffusion characteristics of the low-permeability material. In some cases, the necessary information is provided by field characterization, but for many problems of interest, such as diffusion out of thin clay lenses, the relevant diffusion path length is difficult to determine. Similarly, accounting for variation in advective transport pathways typically requires a very detailed conceptualization of the groundwater flow field, particularly the low-permeability features. For example, spatial variations in the hydraulic conductivity of unconsolidated media can lead to preferential pathways in aquifers over significant distances, similar to characteristics associated with fractured rock and karst formations. Such paths of preferential groundwater flow often control the distribution of contaminant mass in both source areas and downgradient plumes, and must be properly considered in the design and implementation of containment and remediation strategies. Chapman et al. (2010) present an example of how information from detailed site characterization can be incorporated into a remedial design that yields good performance despite the presence of preferential flow paths. However, while available modeling tools are increasingly capable of incorporating detailed descriptions of hydraulic conductivity heterogene-