The purpose of this project was to determine the extent to which munitions constituents (MC) could migrate in the subsurface due to colloid migration and/or via macropore flow. Both of these processes have the capability to bypass large proportions of subsurface media, resulting in faster or deeper migration that might be expected. The finding of a large plume of RDX at the Massachusetts Military Reservation (MMR) was surprising because MC are sparsely distributed at low concentrations on surface soils at operational ranges, and because the vadose zone at MMR is quite deep. Consequently, there was a need to explain the findings at MMR and to determine if groundwater contamination with MC was important at other operational ranges.

The research goals were to investigate the transport of particulate and dissolved MC in intact samples of multiple soil types. MC particles distributed heterogeneously at operational range sites from low-order denotations were suspected to be the source of contamination to underlying aquifers. The physical and chemical environments of these particles were investigated to determine how particles were dissolved and mobilized through natural soil samples. This information was applied to predict the transport of MC in various operational ranges to determine the risk of underlying aquifers to contaminated flux from the vadose zone.

Experiments were designed to enhance knowledge of the multiple processes influencing the transport of munitions particles, including the range of particle sizes over which colloids are mobile, rates of dissolution and reactive transport, and extent of sorption. Undisturbed soil samples containing natural root traces, fractures, and macropores were used to determine the importance of preferential flow in each soil type under realistic vadose zone conditions, e.g., saturated and unsaturated flow. Munitions particles were applied to the soil columns to determine the mobility of particulate MC and the evolution of dissolved-phase munitions constituents. Dissolved MC were applied to duplicate soil columns to isolate the control of particle dissolution. Following application of particulate MC, the soil columns were flash frozen and sliced as a function of depth. To the extent possible, the slices were examined to determine the physical relationship of MC particles with preferential flow features and the chemical relationship of particles with mineralogical and biological components of the soil. The end result has detailed knowledge of the range of particle sizes that were mobile in different soil types. Proportions of relatively mobile fractures and macropores susceptible to preferential flow and relatively immobile micropores that were long-term reservoirs for dissolved contaminants, and mass transfer kinetics between these two types of flow domains were evaluated using a dual-domain reactive transport model coupled with nonlinear inverse solution code. All the reactive transport parameters obtained from these experiments were applied to site-specific subsurface models of the DoD facilities from which the soil samples were collected to determine the risk of contamination in underlying aquifers.

Since the dissolved and particulate MC transport was fully coupled, researchers needed to quantify the sorption and dissolved MC transport in order to quantitatively separate the dissolved MC transport from particle transport. The team conducted a variety of batch isotherm experiments to determine the role of reactive soil components, including OC and soil minerals (clays, iron oxides) in soils on the sorption of dissolved MC compounds. Initial results suggested that the range soil samples exhibited two major behaviors with respect to sorption of dissolved TNT and RDX that appeared to be dependent on the concentration of soil OC. For TNT, the Freundlich coefficient ranged from 40-70 for soils containing 3-6% OC (two sites at JBLM, and MMR-E horizon), while the Freundlich coefficient ranged from 25-30 for soils containing <1% OC (MMR-B horizon, Aberdeen B horizon). Similar results were observed for RDX, but with linear distribution coefficients having much lower magnitudes than observed for TNT, Kd ~7 and 3.4 for MMR-E and MMR-B, respectively. Subsequent and ongoing experiments are investigating the sensitivity of sorption to the type of organic matter (natural organic matter, humic acid (high aromaticity), and fulvic acid (high aliphaticity) and the type of clay (illite, kaolinite). Smectite may be considered in future experiments based on IPR panel response.

Ultimately, the generalized project findings and parameter estimates could be used to build models to assess potential MC subsurface contamination at operational ranges based on existing site characterization. The assessment of the significance of particle transport at various site conditions (recharge, permeability, depth to groundwater, etc.) provide insight into what might occur at the field scale.