Currently used pesticides in water matrices in Central-Western Brazil

This study provides data on the presence of the pesticides atrazine, chlorpyrifos, α-endosulfan, β-endosulfan, flutriafol, malathion and metolachlor in water matrices in urban and rural areas of Campo Verde and Lucas do Rio Verde Cities, Mato Grosso State, Brazil. Surface, rain, and groundwater samples were collected in the rainy and dry seasons during 2007 and 2008 in these important grain-producing areas. The findings revealed a higher diversity of compounds and frequency of detection in the rain water than in surface and groundwater samples. Concentrations of atrazine, endosulfan and malathion above those permitted by Brazilian regulations were found in some surface and groundwater samples, and the degradation products DIA (deisopropylatrazine) and endosulfan sulfate, rather than their parental compounds, were found at higher levels in some samples. Our findings show the vulnerability of water systems in these areas and point to the risk of pesticide contamination in important headwater streams.


Introduction
Recent studies in Brazil have shown the presence of pesticides in water matrices, mainly in surface and ground water.However, considering the size of the country and the total amounts of pesticides applied, the number of studies describing pesticides in surface water [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20] and in ground water [16][17][18][19][20][21][22][23][24][25] is not large.Most of these studies have examined organochlorine pesticides, most of which have been banned in Brazil since the 1980s, except for endosulfan, which will be prohibited beginning in 2013. 26The main reason for these results is attributed to the large amounts of pesticides employed in agriculture both in the past and recently to maintain high productivity of different crops such as soybeans, sugar cane, corn and cotton.
The pesticides when used in agriculture can reach different environmental matrices as they are lost from fields through volatilization, runoff, leaching and atmospheric deposition. 27Water resources are among the final destinations for these compounds.
Rain water is also not free of pesticides since substances volatilized into the atmosphere can return to the soil, Vol. 23, No. 8, 2012   plants and surface water through wet and dry deposition. 28,29ome studies have associated the presence of pesticides in rain water with those in ambient air, occurring not only in rural areas where pesticides are widely applied, but also in locations where pesticides have not been used, such as remote environments and urban areas. 30,31he extensive use of these compounds is a matter of concern for governmental authorities and society at large.Recently, two monitoring programs to determine pesticides in food, mainly fruits and vegetable crops (rice and beans) are currently in place in Brazil, which aimed at evaluating compliance with national MRLs (maximum residue limits): the Program on Pesticide Residue Analysis in Food (PARA), coordinated by the Ministry of Health, through the National Sanitary Surveillance Agency (Agência Nacional de Vigilância Sanitária -ANVISA), and the National Residue and Contaminant Control Program (Plano Nacional de Controle de Resíduos e Contaminantes -PNCRC), coordinated by the Ministry of Agriculture, Livestock and Food Supplies (MAPA).Results of these two programs are presented elsewhere. 32However, humans become exposed not only through food, but also from drinking water, skin absorption and inhalation. 33esticide levels in drinking water in Brazil are regulated by the Ministry of Health (regulation 2914/2011), 34 while the Federal Environmental Agency controls surface water (regulation 357/2005) 35 and ground water (regulation 396/2008). 36However, only a limited number of compounds are regulated, and a systematic monitoring is not performed yet in the entire country.
The presence of pesticides in water and in sediment samples from Mato Grosso State has been reported in several studies, and attributed to the intensive use of these compounds in agriculture. 10,17,22,37This is a concern since this state is rich in water resources and the headwaters of major rivers of three important hydrographic basins (Paraguay, Amazon and Araguaia) are located within the state.Another important aspect is human exposure as intensive farming such as grain production occurs in most municipalities in the state, and the urban areas are surrounded by the fields.
The objective of our study is to identify whether there is a pattern of pesticides in water matrices (surface, rain and ground water) from urban and rural areas of two agricultural centers of the Mato Grosso State, Campo Verde and Lucas do Rio Verde Cities.It was tested for atrazine and its metabolites deethylatrazine (DEA) and deisopropylatrazine (DIA), chlorpyrifos, aand β-endosulfan and its metabolite endosulfan sulfate, flutriafol, malathion, metolachlor and methyl parathion.

Analytical method
The multi-residue extraction method was based on a procedure described elsewhere. 10Briefly, SPE cartridges were made by packing glass tubes (8 mL) with 1000 mg of Bakerbond C 18 (J.T. Baker, Grossgerau, Germany).The cartridges were conditioned with 10 mL of methanol followed by 10 mL of water.An aliquot of 500 mL of dissolved water phase (0.47 µm glass fiber filter), pH previously adjusted to 6.5-7.5, was transferred to the cartridge (ca. 5 mL min -1 ).Analytes were eluted with 10 mL portions of ethyl acetate and hexane:ethyl acetate (7:3) and a 5 mL portion of hexane.Different mixtures of solvent were required since the analytes presents not similar properties.Eluate was concentrated and the residue was redissolved in toluene (1 mL) with phenanthrene-d 10 employed as internal standard.Method accuracy ranged from 73 to 127%, and precision of the measurements was lower than 28%, except for malathion, for which recovery ranged from 106 to 146%.Limits of detection (LOD) and quantification (LOQ) ranged from 0.02 to 0.17 µg L -1 and from 0.02 to 0.84 µg L -1 , respectively (Table 1).LOD and LOQ criteria are presented in the Supplementary Information (SI) section.Samples were analyzed in duplicate and procedural blanks were analyzed with each set of 20 samples analyzed.
The column was set at 92 °C for 2.5 min, ramped at 15 °C min -1 to 175 °C (13 min hold time) and at 20 °C min -1 to 280 °C (9 min).The injector (250 °C) was operated in splitless mode (2.00 min) with injection volume of 1 µL.Helium (99,999%) was used as carrier gas at 1 mL min -1 .Transfer-line was set at 290 °C.The mass spectrometer was operated in electron impact ionization mode (EI) at 70 eV.Standards and samples were injected in the time-scheduled selected ion monitoring mode (SIM) using three ions for each compound (Table 1).Maximum tolerance for confirmation was specified as 20% of relative ion intensity response.

Study area
Campo Verde (CV) and Lucas do Rio Verde (LRV) Cities, distant 130 km and 350 km from the capital Cuiabá, respectively, are two important agricultural centers of Mato Grosso State, Central-Western Brazil, with populations of approximately 30,760 and 45,130, respectively.These areas have a rainy season from November to April and a dry season from May to October.Total precipitation was 1750 mm and 2000 mm in CV and LRV, respectively, during the study period (Figure 1).
Agriculture in these areas is dominated by intensive soybean, cotton and corn cultivation.Information about the pesticides studied here, including their chemical and physical properties, classification, use in crops and amounts of pesticides sold in LRV and CV during 2007 and 2008 is presented in Table 2 and Table 3.
The locations of the sampling points in urban (UA), rural (RA) and surburban (P) areas are presented in Table 4.
Total sampling sites for surface water (SW), ground water (GW) and rain water (RW) were three, seven and four, respectively in CV, and eight, ten and four, respectively in LRV.Depths of the wells ranged from 5 to 100 m in rural areas, and from 30 to 140 m in urban areas.In Campo Verde, surface-water samples were taken from the Casca River (SW52), São Lourenço River headwaters (SW57) and Lajes Stream.In Lucas do Rio Verde, samples were taken from the Lucas River (SW1 and SW2), Verde River (SW3, SW4, SW5 and SW6), Itambiquara River (SW12) and Cedro River (SW18).

Sample collection and treatment
In total, 96 water samples were collected from different water sources in the CV region, and 129 in the LRV region.The water samples were collected from October 2007 to May 2008, a period corresponding to an annual hydrological cycle, with sampling in both the dry and rainy seasons.
Water samples were collected in 1 L glass bottles, carefully filled to the brim to avoid trapping air in the sealed containers.Rain water collectors consisted of a 4 L glass bottle connected to a stainless steel funnel (30 cm diameter).Due to the sampling design, the bulk deposition of pesticides was measured, consisting of wet and dry deposition. 10Sampling was carried out every two months for surface and ground water, and rain water samples were collected after each rain event.When a rain water sample volume was not available for analysis, samples collected subsequently from the closest rain event in the same sampling site were combined and analyzed together.After filling, the bottles were sealed with Teflonlined screw caps, kept on ice in insulated containers and transported to the laboratory after 48 h maximum, where they were stored at 4 °C until extraction, which did not exceed 15 days.Because the distance between the study areas to the laboratory, rain water samples were stored at 4 °C after the rain events and transported to the laboratory once per month.

Data treatment
The data are presented by descriptive statistics: minimum, maximum, mean, median and frequency of detection.Wet deposition fluxes of individual pesticides (mg m -2 ) were calculated for each pesticide in each month, multiplying the concentration of pesticides in the samples for a given whole month (sum of concentrations of different samples collected in the month) (mg L -1 ) by the amount of rainfall in each month (Figure 1).

Results and Discussion
Occurrence of pesticides in surface and ground water Nine of the eleven analytes were found in surfacewater and groundwater samples from CV and LRV, as summarized in Table 5. Figure 2 presents chromatograms of a standard solution (a) and of an example of a real sample (e), with the mass spectrum of the identified analytes atrazine, β-endosulfan and flutriafol in the solution (b, c and d) and in the real samples (f, g and h).Mean concentration distributions of pesticides among samples were strongly biased toward low values, considering the wide range of concentration and the low frequency of detection.Samples collected from rural and urban areas in CV and LRV did not differ significantly (p < 0.05).
Pesticides were detected with higher frequency in surface and groundwater samples from LRV than from CV. Higher values were observed for metolachlor and flutriafol (31%) followed by endosulfan sulfate (29%) in surface water and β-endosulfan (22%) and a-endosulfan (18%) in ground water.Although the amounts of pesticides applied  differed among compounds and years, it was not known if the rain events had occurred immediately following pesticide applications in these areas.Atrazine, DEA, a-endosulfan and β-endosulfan were found in surface-water samples collected in both CV and LRV.Malathion, metolachlor, chlorpyrifos, flutriafol and endosulfan sulfate were found only in surface-water samples from LRV.
Higher levels of atrazine and endosulfan, exceeding the guidelines for surface water established by the Brazilian Federal Environmental Agency (Instituto Brasileiro do Meio Ambiente e dos Recursos Naturais Renováveis -IBAMA; 2 mg L -1 for atrazine and 0.056 mg L -1 for the total of a-endosulfan, β-endosulfan and endosulfan sulfate) 35 were observed in LRV and CV.Excessive levels were found for atrazine during sampling in March and April at sites in CV (SW57) and LRV (SW18), respectively.Endosulfan exceeded the guideline levels in all the samples from LRV, where the isomers and degradation product were found.At least one sample from each sampling site exceeded the guideline levels for these compounds.Malathion also exceeded the guidelines for surface water, established as 0.1 mg L -1 , 35 at sampling site SW2 in LRV in October 2008.In LRV, two degradation products were found in surface water, DEA and endosulfan sulfate.Notably, DEA was detected in the absence of the parental compound at SW1, and its concentration was higher (1.26 mg L -1 ) than that of atrazine (0.20 mg L -1 ) at SW2.Both compounds were detected in similar frequencies, 11% (Table 5), suggesting that atrazine applications in the area were both recent and non-recent.
Endosulfan sulfate was detected at all the sampling sites except SW18.Throughout the study period, endosulfan sulfate was found in the absence of its parent compound in almost all the samples; this may reflect old applications in the area.When aand β-endosulfan and endosulfan sulfate were found simultaneously in the same sample, the endosulfan sulfate concentration was lower than those of the isomers, which may indicate a recent application.Endosulfan sulfate was more frequently detected (29%) in surface water than a-endosulfan (23%) and β-endosulfan (23%), possibly because it is less absorbed in soil compared to the isomers.
Similarly to the surface water, a-endosulfan and β-endosulfan were found in groundwater samples collected in LRV and CV.These isomers differed from the other compounds in showing the highest K OC values and low water solubility (Table 2), which probably result in low rates of leaching.However, for oxisols, the common soil type in the study area, preferential flow is likely to be an important mechanism enabling some pesticides to move down through the soil profile. 39In general, oxisols are well drained and have average permeability, favoring leaching, as observed in Primavera do Leste City, Mato Grosso State. 40The presence of endosulfan compounds in ground water can be enhanced by their high persistence in the environment.In this study, endosulfan levels were not as high as other compounds analyzed in groundwater samples from CV, but high endosulfan concentrations were found in LRV.
As in surface water, atrazine also exceeded (18.9 mg L -1 ) the allowable level established for ground water in Brazil 36 (2 mg L -1 ) in samples collected in CV (RGW4, 5 m well).The sampling point RGW4 is located in the same area where an excessive level of atrazine was also found in surface water (SW57).Both sites (SW57 and RGW4) are in the degraded headwater area of a tributary of the São Lourenço River, with no buffer strips of natural vegetation between the fields and the stream.This area has a high percentage of cropland and crops planted near the stream.In some places, the stream is located at the lower edges of fields, which can facilitate runoff of pesticides, soil and nutrients.A review study concluded that the presence of vegetation such as grass buffer strips along a stream is effective in reducing pesticide runoff and erosion losses. 41These results highlight the risk of pesticide contamination in headwaters with no protective plant cover and intensive use of pesticides.
The presence of a vegetated buffer strip along streams and avoiding application in seasons with a high probability of occurrence of runoff events (due to highintensity rainstorms or saturated soils) are some possible measures to mitigate the entry of pesticides into an aquatic ecosystem. 42However, these mitigation measures are not usually practiced in the areas studied, as the crops are planted at the beginning of the rainy season, together with intensive application of pesticides, and most of the headwaters have no surrounding vegetation.
Metolachor and flutriafol were also found in samples from CV and LRV.The higher solubility and greater persistence of flutriafol may explain the high levels found in ground water from CV and LRV since this compound is applied in smaller amounts compared to the other pesticides such as metolachlor.Flutriafol was found in wells up to 120 m deep.Although flutriafol was found at higher levels in groundwater samples from CV than the other compounds (57.1 mg L -1 in RA-GW3), this pesticide is not subject to Brazilian regulation.
Atrazine, DEA and metolachlor have been previously found in groundwater samples in an area close to CV, Primavera do Leste. 17,37The levels reported were lower than those found in the present study, but may indicate non-recent use of these compounds in the area and also a history of pesticide detection in water samples in Mato Grosso State.Pesticides found in surface water and in ground water in this study have also been identified in other locations in Brazil, as compared in Table 6.

Occurrence of pesticides in rain water
Among the matrices analyzed, more compounds, with higher frequencies of detection, were observed in rain water (Figures 3 and 4).Both regions have high mean air temperatures (monthly means from 22 to 27 °C), and soil temperatures can be extremely high in the surface layer (up to 65 °C), favoring degradation, but also volatilization.All the compounds found in rain water were also found in surface and ground water, although DIA and methyl parathion were found only in rain water.Similar compounds were also found in rain water samples from CV and LRV, except DIA and chlorpyrifos, which were found only in samples from LRV, and malathion, which was found only in samples from CV.
Higher total pesticide concentrations were observed in February and March, at the end of the wet season (Figures 3 and 4).Intensive rain events that occurred in Vol.23 the beginning of the rainy season may have caused the dilution of pesticides in samples collected from November to January, the period of most intense pesticide application.Atrazine showed the highest concentrations in the study period.This compound also showed the highest frequency of detection in CV (54%), followed by aand β-endosulfan (46%) (Table 7).In LRV, the frequency of detection of the compounds that were most often detected followed the order: β-endosulfan (82%) > endosulfan sulfate (79%) > metolachlor (78%) > a-endosulfan (67%) > atrazine (65%).
A wide range of pesticide concentrations was observed in rain water, from 0.11 to 75.4 mg L -1 in CV (Figure 3) and from 0.02 to 47.2 mg L -1 in LRV (Figure 4).
Table 7 shows the maximum estimated deposition fluxes (bulk deposition) and frequency of detection for individual pesticides in the study period, for all the sampling sites.Atrazine showed the highest maximum values in both CV (10.8 mg m -2 ) and in LRV (35.9 mg m -²).The bulk deposition of atrazine was about 5 to 8 times higher than that observed for DEA, the compound with the secondhighest deposition rate in LRV, and about 2 to 5 times higher than the values observed for β-endosulfan and metolachlor in CV.These findings can be attributed to the larger amounts of atrazine applied (Table 3) and to its ability to undergo long-range transport; atrazine is the most commonly identified pesticide in studies determining pesticides in rain water and bulk deposition. 43nterestingly, the maximum precipitation rate for β-endosulfan was higher than that for a-endosulfan in all the sampling sites from CV.In LRV, the difference among the isomer concentrations was not great; however, the frequency of detection for β-endosulfan was higher (82%) than for a-endosulfan (67%).Technical grade endosulfan contains approximately 70% a-endosulfan and 30% β-endosulfan, and published information regarding the stability of these compounds in the atmosphere is inconsistent. 44,45β-endosulfan was also more frequently detected and at higher concentration than a-endosulfan in rain water samples collected in Mato Grosso by Laabs et al. 10 Measured levels of endosulfan in rain water in the study areas as well as further studies could be useful to monitor the phase-out of endosulfan use in Brazil by 2013.
There was a notable similarity among the sampling sites from rural and urban areas, with respect to the frequency of detection and concentrations of pesticides.Because of the small size of the towns and the proximity of urban areas to farms, the presence of pesticides in rain water calls attention to the possibility of exposure of the population to these compounds.The presence of pesticides in rain water from a few locations in Mato Grosso (São Vicente, Cuiabá, Jaciara, Rondonópolis, Barão de Melgaço Cities) was reported by Laabs et al., 10 including atrazine (0.012-0.148 mg L -1 ), a-endosulfan (0.004-0.322 mg L -1 ) and β-endosulfan (0.004-0.147 mg L -1 ).][48][49][50] These results highlight the possible impacts of pesticide deposition on aquatic and terrestrial ecosystems.Considering that pesticides can be dispersed in the atmosphere, the intensive use of pesticides in the study area could affect other remote and important natural ecosystems, such as the Pantanal.

Conclusions
The distribution of pesticides in surface water, ground water and rain water from urban and agricultural areas of LRV and CV showed similarities among pesticides in both matrices and regions, presumably from non-point agricultural sources.Although a few pesticides, atrazine, metolachlor and endosulfan were the main pollutants in the samples analyzed, other pesticides in current use that are not included in Brazilian legislation were also found in the samples.The findings of sites exceeding Brazilian regulatory limits and the high levels of pesticides in rain water highlight the necessity for further monitoring studies in those regions of Brazil where pesticides are intensively used and no native vegetation protects the rivers.This study revealed the vulnerability of water systems in Mato Grosso State, and calls attention to the risk of pesticide contamination of important headwater streams in this area.

Table 1 .
Analytical data and limits of detection (LOD) and quantification (LOQ) for pesticide analysis by GC-MS a Ions used for substance quantification (first ion) and identification (second and third ions) during the selected ion monitoring (SIM); b internal standard; r: coefficient of correlation.Figure 1.Mean precipitation in Campo Verde and Lucas do Rio Verde.Vol. 23, No. 8, 2012

Table 2 .
Chemical and physical properties of pesticides in the study area 38P vap : vapor pressure; b K H : Henry's Law constant; c K OC : soil organic carbon sorption coefficient; d t 1/2 : half-live; e S w : water solubility; I: insecticide; H: herbicide; F: fungicide; -: metabolite; Pesticide Properties Database (PPDB).38

Table 3 .
Application of pesticides in the study area in 2007 and 2008 a Mato Grosso Institute for Agropecuary Defense (INDEA-MT, Brazil).

Table 5 .
Pesticides in surface-and groundwater samples from Campo Verde and Lucas do Rio Verde dt: detected (< LOQ); nd: not detected (< LOD); fd: frequency of detection.

Table 6 .
, No. 8, 2012 Levels of pesticides in surface and ground water in Brazil