The development of methods to extract information from landscape analysis to refine risk assessment is becoming increasingly important. This paper presents results from a pesticide surface water exposure assessment at the watershed scale, based on a combination of edge of field studies, large-scale monitoring studies, and modeling activities with GIS-based landscape analysis methodologies covering an area of approximately 3200 ha surrounding the Simeto River in Sicily (Italy). The dynamic behavior of the pesticide chlorpyrifos-methyl was modeled in two different steps: calculation of the fraction of the application rate that is deposited beyond the field edge and simulation of the fate and persistence of the pesticide in the aquatic environment. Drift loads showed high spatial variability. Considering spray drift deposition as a fraction of the pesticide application rate, 60% of the results were < or = 0.02 (equal to 0.04 mg/m2). Only 8.5% of the results were above 0.5. The highly variability of the landscape factors was reflected in the results. More than 60% of the predicted pesticide concentrations were less than the limit of quantification (0.05 microg/L), affecting about 75% of the total length of the river tract analyzed. Predicted pesticide concentrations were higher than 0.1 microg/L in 23% of cases, but this corresponded to an insignificant portion of the river (1.2% of the total length). These results suggest that management options, such as increased no-spray zones, could provide further protection for surface water. These could be modeled to illustrate their overall impact. As an alternative, the introduction of a 20-m no-spray zone clearly reduced potential exposure, and 92% of the water body was protected. Estimated data are in agreement with data collected during a field monitoring study.