Future use of the sole-source aquifer near Fresno in the eastern San Joaquin Valley, California, will depend, in part, on how long 1,2-dibromo-3-chloropropane (DBCP), an agricultural fumigant banned from use since the late 1970's, persists at concentrations greater than the maximum contaminant level of 0.2 micrograms per liter (mg/L). Field data indicate that DBCP concentrations in ground water have decreased since the late 1970's. Laboratory experiments by earlier investigators show that DBCP transformed to 2-bromoallyl alcohol (BAA) under conditions similar to in situ conditions, with an estimated half-life ranging from 6.1 (pH 7.8, 21.1 degrees Celsius) to 141 years (pH 7.0, 15 degrees Celsius). For this current study, a detailed hydrogeologic investigation was done to assess the relative importance of chemical transformation, dispersion, and ground-water pumping and reapplication of irrigation water in affecting DBCP concentrations. Ground-water samples were collected from 20 monitoring wells installed along a 4.6-kilometer transect. DBCP concentrations in these samples ranged from less than the detection limit of 0.03 mg/L to a maximum of 6.4 mg/L. Results of chlorofluorocarbon (CFC) age dating indicate that DBCP occurs in water that ranges in age from about 2 to 41 years. The primary transformation product BAA, which was identified during previous laboratory studies, was not detected at or greater than 0.03 mg/L in any of the 20 ground-water samples. The lack of detection of BAA indicates that transformation to BAA is insignificant relative to other processes controlling DBCP concentrations. Results from this current study indicate that the in situ hydrolysis half-life for DBCP to BAA is much greater than the laboratory-determined values. Estimated initial concentrations of DBCP, calculated using CFC-estimated travel times and a half-life of 6.1 years, indicate that maximum initial concentrations are consistent with maximum measured concentrations in ground water. In contrast to initial DBCP concentrations, the estimated initial nitrate concentrations indicate that nitrate concentrations in recharge water have increased with time. A conceptual two-dimensional numerical flow and transport modeling approach was used to test hypotheses addressing dispersion, transformation rate, and in a relative sense, the effects of ground- water pumping and reapplication of irrigation water on DBCP concentrations in the aquifer. The flow and transport simulations, which represent hypothetical steady-state flow conditions in the aquifer, were used to refine the conceptual understanding of the aquifer system rather than to predict future concentrations of DBCP. Results indicate that dispersion reduces peak concentrations, but this process alone does not account for the apparent decrease in DBCP concentrations in ground water in the eastern San Joaquin Valley. Ground-water pumping and reapplication of irrigation water may affect DBCP concentrations to the extent that this process can be simulated indirectly using first-order decay. Transport simulation results indicate that the in situ 'effective' half-life of DBCP caused by processes other than dispersion and transformation to BAA could be on the order of 6 years.