Using Multiple Environmental Tracers to Estimate Field-Scale Longitudinal Dispersivity

In this study, we seek to reduce parameter uncertainty in groundwater modeling systems, particularly in reactive transport models, by quantifying effective field-scale longitudinal dispersivity using anthropogenic environmental tracers. We generate synthetic aquifer fields and model transport of atmospheric tracers and test whether tracers can be used to determine an effective aquifer-scale dispersion coefficient. We generate synthetic datasets by simulating transport of chlorofluorocarbons (CFC11, CFC12, CFC113), sulfur-hexafluoride (SF6) and tritium (3H) with input functions derived from known atmospheric concentrations, through a three-dimensional, stochastic, heterogeneous synthetic aquifer developed using sequential Gaussian simulation using the PFLOTRAN reactive transport model. Flux-averaged concentrations calculated from model output are used as synthetic observation datasets to calibrate effective dispersivity for simplified homogeneous models with the PEST parameter estimation software. Tracer-derived effective dispersivity values are compared with theoretical and empirical values reasonable for our stochastic structure. We assess the ability of our homogenous model with tracer-derived effective dispersion coefficients to reproduce transport of a synthetic contaminant through the heterogeneous 3D field with two new boundary conditions. The ratio of CFC11/SF6 displays less than a 10% difference between the full (4.12m) and single-time (4.43m) series derived effective dispersivity. The ratio of CFC12/SF6 displays less than a 10% difference between the full (4.09m) and single-time (4.43m) series value. While all tracer-derived values from both the full-time and single-time series (1.96 m to 10.75 m) were within reason compared to the theoretically and empirically derived values (1.01 m to 5.32 m), dispersivity values derived from CFC11/SF6 and CFC12/SF6 for the full and single-time series display quantitatively smaller residuals compared to our heterogeneous truth model for our new boundary conditions. Our results indicate that environmental tracers can be useful in estimating effective dispersion coefficients for reactive transport models over longer length and time scales than traditional applied tracer studies. This new method of utilizing multiple environmental tracers over a limited time series could be an easy, inexpensive, and effective solution in quantifying field-scale longitudinal dispersivity and reduce parameter uncertainty in groundwater/contamination transport models