HETEROGENEOUS SORPTION PROCESSES IN SUBSURFACE SYSTEMS .1. MODEL FORMULATIONS AND APPLICATIONS

The development of an appropriate model for sorption processes in natu ral systems has proven elusive. We hypothesize that standard determini stic models cannot capture the inherent variability in most natural sy stems. This work responds tb deficiencies in current models by formula ting two classes of sorption models: an extended set of deterministic models based upon a generalized multiple domain concept and a class of stochastic models. Stochastic sorption models are formulated to descr ibe microscale variability, which may be interparticle or intraparticl e in origin, in terms of both sorption equilibrium and rate properties . Stochastic model formulations treat sorption equilibrium and rate pr operties as continuously distributed random variables described by log normal or T probability density functions. Long-term sorption equilibr ium and rate data are shown for a herbicide on a subsurface material. Results from parameter estimation for 16 different rate models applied to the data set are summarized. Model-data comparisons show a range o f agreement betwe en individual model fits and experimental data, with several trends evident: diffusion models outperformed first-order mod els; the addition of an instantaneously sorbing fraction significantly improved model-data agreement; multiple-particle class models more cl osely fit the data than single-particle class models, without an incre ase in the number of fitted parameters; and stochastic first-order mod els agreed well with the data.