This paper proposes a minimalist modeling approach for characterizing pesticide concentrations in runoff from agricultural catchments across spatial scales. The model proposed is of an intermediate level of complexity between traditional chromatographic separation models and the more complex dual-domain models. Parsimony in the model is achieved by assuming stationarity of catchment travel time distributions and by coupling a dual-domain source zone model that describes near-surface pesticide dynamics with the mass response function (MRF) approach, which describes catchment-scale solute transport. The model is evaluated by comparing predicted atrazine concentrations with measured values over a 5 yr period at two spatial scales (tile drain: 3-5 ha; river station : 69 km(2)) within an intensively managed agricultural catchment in Illinois, United States. Pesticide dynamics within the source zone provided the strongest control on leaching. Two parameters were calibrated at the tile scale, Gamma, which describes partitioning in the dual-domain surficial source zone, and k(e), which describes the mass transfer rate constant between the two domains. The initial peak of concentration was found to be sensitive to Gamma, while the later peaks were sensitive to k(e). The calibrated parameters at the tile stations were used to predict atrazine dynamics at the river station. Prediction errors are examined and related to the lack of detailed information about anthropogenic forcings across scales (e. g., land-use or soil/crop management practices).