Predicting pesticide transport in mulch-amended soils: a two-compartment model

With adoption of combine harvest technology, massive sugarcane (Saccharum Spp. Hyb.) residue is left on the soil surface following each harvest, which could affect the environmental fate and efficacy of applied pesticides in subsequent cropping seasons. The objective of this study was to quantify the retention characteristics and mobility of a commonly used herbicide, atrazine (2-chloro-4-ethylamine-6-isopropylamino-S-triazine), in soils when sugarcane mulch residue was present. Specifically, adsorption and desorption kinetic batch experiments were performed to quantify atrazine retention by the mulch residue over time. Atrazine sorption exhibited time-dependent behavior and was followed by slow release regardless of input concentration. A kinetic-equilibrium model based on a second-order two-site (SOTS) formulation was successful in describing atrazine adsorption versus time by the mulch residue. One set of model parameters was capable of describing atrazine release based on six successive desorption steps. As a test for the applicability of the model, data sets from two other experiments where sorption and release was measured for extended time periods (1348 and 2476 h) were successfully predicted by the SOTS model. Miscible displacement methods were used to measure the mobility of atrazine in packed columns where the mulch residue was mixed with a reference sand material or a Sharkey clay soil (very fine, smectitic, thermic Chromic Epiaquerts). Use of batch-measured model parameters did not adequately describe atrazine mobility in the sand-mulch column. In contrast, for a Sharkey-mulch soil column, a two-compartment SOTS model was successful in predicting atrazine breakthrough results where independently estimated parameters for the mulch residue and the soil matrix were used as the input parameters. Results also showed that mixing of mulch with soil created physical non-equilibrium condition in the columns, which was responsible for errors in model predictions.