Remediation Technologies - Environmental Investigation and Remediation - page 331

into dead-end pore spaces within the i ner-grained aquifer material. Diffusion into dead-end
pores will cause a reduction in monitored aqueous concentrations similar to the reduction caused
by chemical sorption of typical organic contaminants to organic materials in the aquifer. Fine-
grained clay and silt are present in most unconsolidated aquifers. Clays, though highly porous,
are relatively impermeable and hence restrict the free l ow of groundwater. Dead-end pores exist
that are connected to the more permeable materials of the aquifer, but are not subject to through-
l ow. Because water is not actually l owing into or through these dead-end pore spaces, the con-
taminant migration process is governed by molecular diffusion. The transfer of 1,4-dioxane into
dead-end pores is a slow process driven by the concentration gradient between the contaminated
free-l owing groundwater and the relatively clean water in the dead-end pore spaces. Unfortunately,
unlike chemical adsorption, where the contaminant may be permanently bound to the organic
material in the aquifer, this attenuation process involving diffusion into dead-end pores is a
reversible process. Back-diffusion from the dead-end pore spaces into the l owing groundwater
occurs after the concentration gradient reverses to cleaner water in the aquifer and more contami-
nated water in the dead-end pore spaces.
7.3.3 N ORTH C AROLINA E XAMPLE
Statistical trend analysis and a numerical l ow model were used to gain the approval of a natural
attenuation remedial approach for an industrial site in North Carolina (Chiang et al., 2006, 2007b,
2008). Active sources of 1,4-dioxane had been removed, and dissolved-phase plumes were pres-
ent in four distinct hydrogeologic layers beneath the site. Data collected from 60 wells at the site
since the early 1990s were evaluated yearly from 2003 through 2005 by using Mann-Kendall
trend analysis. The evaluation in 2005 noted that 12 wells (20%) exhibited a downward trend and
21 wells (35%) exhibited a stable trend. Twenty-six wells (43%) exhibited no discernable trend. An
upward trend was noted in only one well (2%). To demonstrate a continued attenuation trend into
the future, a Modular Three-Dimensional Groundwater Flow and Transport Model (MODFLOWT)
was constructed by using four hydrogeologic layers representative of the varying hydraulic con-
ductivities in subsurface sand versus clay or silt units. The 1,4-dioxane plume was assumed to be
more sensitive to advection than to hydrodynamic dispersion. (Advection refers to the process of
contaminant migration due to movement of groundwater. Hydrodynamic dispersion is the sum of
mechanical dispersion and diffusion.) The modeling effort simulated advective-dispersive trans-
port, but considered diffusion to be relatively insignii cant and therefore omitted this term from
the model.
The l ow model was calibrated by using measured heads from monitoring and production wells
at the site. Particle tracking and transport modeling were performed to estimate future concentra-
tions of 1,4-dioxane at a local surface-water body to the east. The modeling under an assumption of
no degradation underestimated the observed decline in 1,4-dioxane concentrations. To assess the
degradation rate, historical i eld measurements were evaluated at several monitoring wells within
different hydrogeologic layers. Time-series plots were prepared under the assumption of no degra-
dation versus degradation half-lives of three, i ve, and seven years. The optimal correlation between
observed and predicted values occurred with a seven-year degradation half-life. For each of the four
layers, Figure 7.3 presents the ten-year predictions, which show signii cant reductions in the distri-
bution of 1,4-dioxane as a result of using a 7-year half-life. The modeling results predict that 1,4-
dioxane migration at levels above North Carolina's surface-water standards would not reach the
surface-water body via any of the modeled hydrogeologic layers. Among the natural attenuation
me cha n ism s, Ch ia ng et a l. su r m ise d t hat 1,4 - d ioxa ne is more h igh ly i n l uenced by advection because
it is not readily sorbed to the aquifer matrix. Because no active sources of contamination existed at
the site and the combination of modeling and i eld observations demonstrated that the plume was
stable or shrinking, the proposed MNA approach was approved by the regulatory agency for imple-
mentation as the i nal remedy (Chiang et al., 2008).