Optimal two-stage pump-and-treat of groundwater impacted by nonequilibrium desorption, heterogeneity and persistent contaminant sources

The goal of this work is to improve our understanding of the effects of nonequilibrium desorption, low permeability units and continuous contaminant sources on aquifer remediation. The first two processes result in mass transfer limitations that reduce the effectiveness of pump-and-treat (PAT), while the latter results in PAT process inefficiencies by providing persistent sources of contamination. The work focuses on implementing the understanding of these processes to improve dynamic PAT schemes. Mass transfer limitations are considered at two scales: that associated with the diameter of the soil particles, and that associated with relatively impermeable layers. The problem is studied in terms of contaminant sources of finite and infinite durations resulting from dissolution of nonaqueous phase liquids (NAPL).; Mass transfer limitations resulting from nonequilibrium desorption and heterogeneities suggest that PAT can be viewed as being composed of two distinct periods. Initially, the remediation is characterized by efficient extraction resulting from advective removal of the contaminant's mobile phase. At later times, however, rate-limited mass transfer at the intra-particle and macroscopic scales substantially reduce the contaminant extraction efficiency. This change in the behavior of the cleanup effort suggests the need to break the operation into two stages to arrive at an optimum dynamic pumping strategy. In stage 1, the goal is to remove the maximum amount of mass to reduce aqueous concentrations and generate a strong desorption gradient. In stage 2, the goal is to maintain an adequate desorption gradient between the immobile and mobile phases that captures the plume without excessive pumping. The scheme proves beneficial over classical steady pumping in terms of total volume extracted under scenarios dominated by mass transfer limitations, heterogeneities or reduced-strength NAPL sources. Conversely, the strategy is less beneficial under conditions of local equilibrium sorption in homogeneous aquifers or when high-strength contaminant sources are present. The research was conducted using a groundwater flow and transport simulation model (Kong and Harmon, 1996), which was coupled with the optimization model MINOS (Murtagh and Saunders, 1995).