NONLINEAR-REGRESSION OF BREAKTHROUGH CURVES TO OBTAIN RETARDATION FACTORS IN A NATURAL GRADIENT FIELD-STUDY

An investigation of the fate and transport of trace quantities of comm only detected herbicides and herbicide degradates in groundwater under aquifer conditions was conducted in a shallow sand and gravel aquifer . Herbicides and a conservative tracer (bromide) were injected into th e aquifer and the plumes monitored for 2 to 3 mo using an array of dow ngradient multilevel samplers (MLSs). Breakthrough curves were fitted to the advection-dispersion and exponentially modified Guassian (EMG) equations. Coefficients of determination were high for both models (ge nerally >0.90). Comparison of the breakthrough curves for the conserva tive tracer and each herbicide provides a measure of retardation in th e aquifer. A retardation factor (R) may be calculated from the first c entral moment of the data or from peak parameters obtained from the fi tted curves, such as: (i) the time required to reach half-maximum conc entration, (ii) the centroid of the curve, or (iii) model parameters s uch as the apparent transport velocity. Comparison of the retardation factors shows small differences due to the method of calculation both within and between models. The similarities in the retardation factors and the fitted curves suggest that both the EMG and the advection-dis persion model describe the data well. The detection of differences in herbicide retardation factors varied with the method of calculating R.