Remediation of groundwater in fractured rock aquifers is notoriously difficult. Both consultants and clients face many challenges meeting regulatory requirements at their sites, and often, regulators are not much assistance. In fact, regulators tend to play it safe, following decades-old guidelines even when such approach is scientifically not justified. Furthermore, regulators expect results from inadequate investigation technologies or quantitative analyses based on the assumption that porous media is continuous and Darcy’s Law, which was developed for unconsolidated sediments, is applicable.
During phases of remediation, from investigation to feasibility to design, misuse of groundwater models is often problematic causing skepticism from clients and regulators. The source of skepticism surrounding numeric groundwater flow models is vast, largely because models are not commonly applied. Similarly, computer programs for simulating correct nature of groundwater flow in fractured rock aquifers are scarce. As such, industry stakeholders and regulators are not familiar with proven models considered to be the best approach versus what is not proven. Therefore, they cannot fully judge the validity of what has been produced and submitted.
It is possible to use familiar numeric groundwater flow and fate and transport (F&T) models while concurrently representing the conceptual site model in a physically defensible way. However, this requires very fine model discretization and millions of cells, which then extends the process. Nevertheless, such models can be developed to demonstrate concepts and analyze “what if” remedial scenarios especially where it matters, such as source areas and some distance downgradient. For example, an area of several hundred feet by several hundred feet as deep as 200 feet, can be feasibly simulated with the most widely used numeric models for flow and F&T such as MODFLOW and MT3DMS, which includes several vertical and horizontal fractures half an inch wide or less, spaced adequately apart.
Beyond the familiarity of these models, the advantage is that all model input parameters and assumptions can be tested with a variety of available model packages such as: recharge, extraction rates at wells, correct placement of well screens (lengths and elevations), presence of surface water features, incorporation of matrix diffusion, contaminant degradation rates and other F&T parameters, transfer from NAPL phase to dissolved phase, and the list goes on. With that said, it is mostly not feasible to upscale these numeric models developed at conceptual levels to some more realistic site scales due to restrictions of computing power and time, but they can be very powerful in conveying to regulators what may work and what may not. This includes demonstrating technical impracticability of remediating a site to maximum contaminant levels (MCL) when contaminants where released into the area decades ago and have diffused into the rock matrix making it difficult to both identify and remove.