Use of quantitative models to design microbial transport experiments in a sandy aquifer.

A suite of numerical models was applied to the problem of designing field tracer and bacterial injection experiments in a sandy surficial aquifer near Oyster, Virginia. The models were constructed based on the integration of diverse characterization data including hydrologic, geophysical, geological, geochemical, and biological information. A one-dimensional particle-tracking model was used to analyze laboratory transport experiments conducted using intact core samples to prescribe transport parameters describing solute dispersion and bacterial fate. A geostatistical model of three-dimensional hydraulic conductivity variations was developed, conditioned on in situ measurements of hydraulic conductivity and interpretations of geophysical data, and used to generate alternative aquifer descriptions. A regional-scale, two-dimensional flow model was used to design pumping rates of a forced-gradient hydraulic control system. Information from these various models was then combined into a high-resolution, three-dimensional flow and transport model for the prediction of field-scale solute and bacterial transport. Model predictions were used in an iterative experimental design process to specify: (1) the locations of multilevel samplers for monitoring transport; (2) frequency and timing of sample collection during bromide tracer injection experiments; and (3) frequency and timing of sample collection during a bacterial injection experiment. At each stage of the design, information gained during the previous stage was used to refine the model and target subsequent experimentation.