Atrazine movement in soil: Comparison of field observations and PRZM simulations

The predictive capability of the pesticide root zone model (PRZM) was investigated for herbicide atrazine 2‐chloro‐4‐(ethylamino)‐6‐(isopropylamino)‐s‐triazine] in corn production under no‐till (NT) and conventional‐till (CT) management practices. Simulation values of atrazine residues obtained using our site‐specific soil and environmental data were compared with the actual values measured in soil samples taken from the root zones of the NT and CT plots during three growing seasons: 1986, 1987, and 1988. The mean concentration of atrazine in soil at each sampling time and depth after application, for each tillage treatment plot (NT or CT), was estimated based on the type of distribution (i.e., normal or lognormal). Overall, the PRZMs simulated concentrations for the top 10 cm of soil compared well with the atrazine residues measured in the CT plots, but overestimated measurements in NT plots. For example, in 1986 the mean atrazine concentration measured in soil samples taken 6 d after application from the top 10 cm of CT plots was 548 μg/kg (S.E. 198 μg/kg), and the PRZM predicted value was 690 μg/kg. In contrast, the mean atrazine concentration for the same soil depth increment in NT plots was 385 μg/kg (S.E. 154 μg/kg), with a PRZM predicted value of 674 μg/kg. Although the PRZM prediction was closer to the measured mean for atrazine concentrations in the top 10 cm of the CT system, the model did not transport atrazine to the lower soil depths, as the actual values have indicated in all 3 years. The results of this model comparison, especially for the lower soil depths (20 to 30 cm) in the NT practice, indicated that the PRZM model does not account for the preferential transport of, and, consequently, underestimates the atrazine residue levels in the lower soil profile under NT management systems.