Pesticides in surface and groundwater in three micro-basins of the São Lourenço River, state of Mato Grosso, Brazil: grab and passive sampling

This study aimed to assess the occurrence of pesticides in São Lourenço River basin, next to soybean, corn and cotton plantation, using complementarily grab and passive sampling techniques. São Lourenço River is an important affluent of Paraguay River which run to the Pantanal, the greatest wetland in the world with high environmental importance to the region. Methods for determining pesticides retained in the solid phase of passive samples and grab samples were developed and validated, which showed good precision and accuracy for the determination of 30 and 34 pesticides, respectively. A total of 72 surface and 55 groundwater samples using grab sampling and 137 surface water passive samples using POCIS were collected over a 1-year period. Pesticides atrazine, azoxystrobin, carbendazim, carbofuran, cyproconazole, clomazone, chlorantraniliprole, chlorpyrifos, difenoconazole, malathion, and metolachlor were detected in the three types of samples. Eleven pesticides were detected at concentrations ranging from 0.0132 to 1.31 µg L⁻¹ for grab samples (surface waters), 0.0132 to 0.986 µg L⁻¹ in shallow groundwater and 0.0011 a 0.204 µg L⁻¹ in surface waters using passive samplers. The passive sampling technique applied to surface water confirmed the presence of the 15 analytes detected by grab sampling, allowed the identification of another 7 pesticides which were not detected in grab samples, and increased the incidence of detection of the analytes. While at least one pesticide was detected in 34% of grab samples, in the POCIS samplers the detection frequency was 95%. POCIS also provided a better temporal understanding of contamination levels showing that the higher detections and concentrations occurred in the rainy period and the period of more intensive application. Therefore, passive sampling showed to be highly effective for detecting pesticides at low concentrations as a complement to grab sampling, indicating that the joint use of both sampling strategies provides consistent monitoring data.