Pesticidas e recursos hídricos : validação de método analítico e avaliação da distribuição em dois pólos agrícolas de Mato Grosso

Abstract

Campo Verde and Lucas do Rio Verde, Mato Grosso state, Brazil, are large farmers (cotton, corn and soybeans) and consumers of pesticides. Once applied, pesticides can achieve different environmental compartments, such as surface and groundwater that may be used for human consumption. This study aimed to evaluate the distribution of some pesticides commonly used in these two regions in surface water, groundwater and rainwater samples collected from October 2007 to May 2008. Two multi-residue methods were employed, one was developed and validated to determine eight pesticides and three metabolites based on

solid phase extraction (SPE) and gas chromatography coupled to mass spectrometry (GC/MS- SIM), and the other was a method already described in the literature, which uses SPE and

HPLC/DAD. Limits of detection of the validated method ranged from 0.02 to 0.17 g L-1 . Both methods were employed to analyze 41 surface water samples, 62 groundwater samples and 120 rainwater samples collected in the urban and rural areas of Campo Verde and Lucas do Rio Verde. Potential of contamination of the analyzed matrices was observed in the region, taking into account the presence of atrazine, chlorpyrifos, α- and β-endosulfan, flutriafol,

malathion, metolachlor, parathion methyl and three metabolites DEA-deethylatrazine, DIA- deiopropylatrazine and endosulfan sulfate in the analyzed samples (0.02 e 75.43 μg L-1

), including sites far from pesticides application. Levels of atrazine α- and β-endosulfan and endosulfan sulfate in some surface and groundwater samples exceeded the values established by Brazilian legislation for these matrices. Principal component analysis and hierarchical cluster analysis provided information about the pesticides distributions in the studied sites, indicating the compounds and the sampling sites which contribute to explain data variance. When applied to a higher number of variables, including besides the pesticides deposition rates, chemical and physical properties and the amount applied, PCA resulted in 3 factor

explaining 95.3% of data variance. Considering the 15 variables analyzed, just pesticides half- life values did not contributed to these factors.