[1] The results of three-dimensional Navier-Stokes (NS) and Stokes simulations and twodimensional local cubic law (LCL) simulations of fluid flow through single rough-walled fractures are presented. Synthetic rough-walled fractures were created by combining random fields of aperture and the mean wall topography or midsurface, which quantifies undulation about the fracture plane. A finite volume formulation of the LCL that incorporates geometric corrections for fracture undulation is presented. Simulations of fluid flow through planar fractures with sinusoidal variation in aperture were compared to published results. The rough-walled fracture simulations demonstrated that the total flow rates predicted by the corrected LCL were within 10% of those predicted by the Stokes equations for all the fractures examined in this work. Differences between the NS and Stokes simulations clearly demonstrated that inertial forces can significantly influence the internal flow field within a fracture and the total flow rate across a fracture. By limiting the total flow rate differences between the NS and Stokes simulations, constraints for three kinematic parameters were determined. For all the fractures presented in this work, the corrected LCL was determined to be an acceptable approximation to the NS equations, provided that the kinematic and geometric constraints were met.

Batch and stop-flow column experiments were performed to estimate persulfate decomposition kinetic parameters in the presence of seven well-characterized aquifer materials. Push-pull tests were conducted in a sandy aquifer to represent persulfate decomposition under in situ conditions. The decomposition of persulfate followed a first-order rate law for all aquifer materials investigated. Reaction rate coefficients (k(obs)) increased by an order of magnitude when persulfate concentration was reduced from 20 g/L to 1 g/L, due to ionic strength effects. The column experiments yielded higher k(obs) than batch experiments due to the lower oxidant to solids mass ratio. The kinetic model developed from the batch test data was able to reproduce the observed persulfate temporal profiles from the push-pull tests. The estimated k(obs) indicate that unactivated persulfate is a persistent oxidant for the range of aquifer materials explored with half-lives ranging from 2 to 600 d.

Infiltration and dispersion (including molecular diffusion) can transport volatile organic compounds (VOCs) from urban air into shallow groundwater. The gasoline additive methyl-tert-butyl ether (MTBE) is of special interest because of its (1) current levels in some urban air, (2) strong partitioning from air into water, (3) resistance to degradation, (4) use as an octane-booster since the 1970s, (5) rapidly increasing use in the 1990s to reduce CO and O3 in urban air, and (6) its frequent detection at low microgram per liter levels in shallow urban groundwater in Denver, New England, and elsewhere. Numerical simulations were conducted using a 1-D model domain set in medium sand (depth to water table = 5 m) to provide a test of whether MTBE and other atmospheric VOCs could move to shallow groundwater within the 10−15 y time frame over which MTBE has now been used in large amounts. Degradation and sorption were assumed negligible. In case 1 (no infiltration, steady atmospheric source), 10 y was not long enough to permit significant VOC movement by diffusion into shallow groundwater. Case 2 considered a steady atmospheric source plus 36 cm/y of net infiltration; groundwater at 2 m below the water table became nearly saturated with atmospheric levels of VOC within 5 y. Case 3 was similar to case 2, but considered the source to be seasonal, being “on” for only 5 of 12 months each year, as with the use of MTBE during the winter fuel-oxygenate season; groundwater at 2 m below the water table became equilibrated with 5/12 of the “source-on” concentration within 5 y. Cases 4 and 5 added an evapotranspiration (ET) loss of 36 cm/y, resulting in no net recharge. Case 4 took the ET from the surface, and case 5 took the ET from the capillary fringe at a depth of 3.5 m. Net VOC mass transfer to shallow groundwater after 5 y was less for both cases 4 and 5 than for case 3. However, it was significantly greater for cases 4 and 5 than for case 1, even though cases 1, 4, and 5 were all no-net recharge cases. The mechanism responsible for this effect was the dispersion acting on each downward infiltration event, and also on the ET-induced flow. The ability of MTBE to reach groundwater in cases 2−5 is taken as evidence of the potential importance of urban air as a non-point source for VOCs in shallow urban groundwater. Two subcases were run for both case 4 and case 5:  subcase a (water and VOCs move with ET) and subcase b (water only moves with ET).

No abstract