Method development, validation, monitoring, seasonal effect and risk assessment of multiclass multi pesticide residues in surface and ground water of new alluvial zone in eastern India

A liquid–liquid extraction (LLE) method was validated as per SANTE/12682/2019 guidelines for gas chromatography–mass spectrometric (GC–MS) determination of thirty-six multiclass pesticides in environmental waters. Seasonal (summer, monsoon, and winter) effects on the magnitude of pesticide residues in environmental water (river, pond, and tube well) of six different urban areas of Nadia and North 24 Parganas districts (New alluvial zone, Eastern India) was monitored for subsequent risk assessment. Total 288 water samples (96 each of river, pond, and tube-well) irrespective of locations and seasons were monitored for multiclass multi pesticide residues during the experiment. Each sample (750 mL) was extracted with ethyl acetate/dichloromethane (8:2) liquid–liquid partitioning and filtration (0.22 μm nylon filter paper) and total residue was reconstituted in acetone (1 mL) for GC–MS analysis with developed and validated method resulting satisfactory recovery percentages (77.84–118.15%). The maximum no. of organochlorine (OC) and organophosphorus (OP) pesticide residues were dominated in river and pond water irrespective of seasons and monitoring sites. About 74% of river water samples were found to be contaminated with concerned pesticides in variable magnitudes. Monsoon (July to October) season was highly alarming with the highest presence of total pesticide residues in different types of environmental waters. Risk quotient (RQ) [acute and chronic] was also evaluated in pond and river water as sometimes used for drinking purposes. RQ value (5900) of total endosulfan indicates the highest risk of chronic toxicity to river fishes. Seven water samples from tube-wells were also monitored and found to be contaminated with butachlor and chlorpyriphos in non-significant amounts (< 0.1 ng mL−1), irrespective of seasons and sites, thus safe for consumption.


Introduction
West Bengal is the fourth largest populated state in India. The population density was 1100 people km −2 in 2020. Barrackpore and Kalyani are two important subdivision towns of West Bengal having population densities of 11,500 people km −2 and 7500 people km −2 , respectively. The present study was carried out in these two locations because they are situated in the new alluvial zone of river Hooghly (Ganga) and have been developed agriculturally as well as industrially. Three important sources of environmental water (river, pond, and tube well) are used for domestic, agricultural, and industrial consumption. The water of river Hooghly is used by the municipalities especially for drinking purposes after treatment in various water treatment plants. Besides, the river water is also used in agriculture and industries directly (KMDA 2020).Water is very crucial for sustaining life. About 50% of urban and 80% of rural people in India are affected by water pollution, and therefore a scarcity of pure drinking water is being created (Hegde 2012). Environmental waters are usually polluted from point and nonpoint source of contaminants. However, a major source of water contamination comes from non-point sources (Calhoun 2005). Many research studies (Ongley 1996;Rickert et al. 2002;Kaushik et al. 2008Kaushik et al. , 2012Maurya and Kumar 2013;Toccalino et al. 2014;Székács et al. 2015;Hassaan and Nemr 2020) have reported the contamination of environmental waters. Due to dense population of Barrackpore and Kalyani, chances of water contamination from domestic sources are relatively high. In India, during 2014-2015 to 2019-2020, consumption of chemical pesticides in agriculture has been increased by 7% in comparison to biopesticides by 34%. Awareness is therefore growing towards quality assurance and bio-safety in agriculture. Along with all other classes of pesticides, long persistent pesticides may have important roles in contamination and/or pollution of environmental waters. Considering the fact, thirty-six pesticides (12 organochlorines, 7 organophosphates, 7 synthetic pyrethroids, and 10 others) were chosen based on the screening of literature and farmers' use in these localities.
Some possible problems occur at application sites of pesticides to non-target regions due to the movement of pesticides, which include environmental contamination, economic loss, and inefficient pest control (Duttweiler and Malakhov 1977;Waite et al. 2002). Transportation of pesticides and their degradation products to surface water depends on several factors, including environmental weather condition, soil characteristics, topography, agricultural practices, and chemical properties of individual pesticides (Leonard and Knisel 1988;Arias-Estévez, et al. 2008;Skoulikidis et al. 2011). As there are huge variations in temperature, rainfall, and humidity with seasons like summer, monsoon, and winter ( Fig. 1), seasonal effects are also considered for investigation under the present study. Though some literatures of environmental effects on contamination in surface water were reported (Mondal et al. 2018;Kaushik et al. 2008), the monitoring of environmental water contamination variation with seasons is not available specifically for target areas (Barrackpore and Kalyani). So, there is a requirement of residue data for possible correlation between seasonal variation and non-point source of contamination of surface water. Various methods like solid phase extraction (SPE) (Tsochatzis et al. 2012;Sangchan et al. 2012;Kouzayhaet al. 2012;Maurya and Kumar 2013;Glinski et al. 2018) and liquid-liquid extraction (LLE) (Singh and Mishra2009;Kaushik et al. 2012;Erkmen et al. 2013;Saadaoui et al. 2020) are available to analyze pesticide residues in water. To analyze pesticide residues in water samples, solid-phase extraction (SPE) is usually applied as it is an easy and fast process (Hatrík and Tekel 1996). Liquid-liquid extraction (LLE) method is more reliable than SPE method (Tan 1992;Awofolu and Fatoki 2003;Hosseini et al.2013;Freitas et al. 2020) also recommended LLE methodology as a selective and precise method for water sample analysis. Campanale et al. (2021) also indicated the merits of LLE-based analysis in water matrix through a review study. Commercial SPE cartridges or disks have produced extraneous peaks during instrumental analysis which might be due to phthalate esters contained in the housing materials of the used cartridges (Awofolu and Fatoki 2003). In present experiment, LLE method with three different solvent mixtures using hexane, ethyl acetate (EA), and dichloromethane (DCM), namely, S 1 (EA:DCM 8:2), S 2 (Hexane/DCM 8:2), and S 3 (100% DCM), was considered to check extraction efficiency of different pesticide residues from collected water samples. Total 36 pesticides of different classes were monitored in the samples collected from Hooghly River, ponds, and tube-wells maintaining proper sampling protocol. Ecological risk of pesticides from non-point sources in close proximities has huge connectivity with surrounding areas with limited dilution potentials (Karaouzas et al. 2018;Schulz 2004).Therefore, ecological risk assessment is very important for the evaluation of toxic effects of pesticides on non-target organisms present in aquatic system. Based on the results found in the monitoring study, risk assessment was investigated. Barrackpore (22°76´N and 88°37′E) and Kalyani (22°58′N and 88°26′E), two sub-divisional metropolitan towns of respective districts of North 24Parganas and Nadia, have occupied major industrial and agricultural areas. Nadia covers 79.48% of agricultural areas (Matirkatha 2021), whereas Barrackpore is one of the dense industrial areas of West Bengal. Both are situated at the bank of Hooghly River (Ganga River) and use river water for drinking, agriculture, industrial use, and aquaculture (KMDA 2020). In these two areas, diversified cropping pattern is found throughout the year.

Sampling
River water (RW), pond water (PW), and tube-well water (TW) samples were collected for consecutive 48 weeks to monitor the pesticide residues (Fig. 2). 96 samples each of river, pond, and tube-well water from two locations were collected throughout the year covering winter (November to February), summer (March to June), and monsoon (July to October) seasons. A total of 288 (144 from Barrackpore + 144 from Kalyani) samples were tested throughout the year under the present study. Every week, we have altered the sampling points under target locations. Ponds and tube-wells chosen for sample collection are located within a 2-3-km radius under domestic, agricultural, and industrial activities. Amber glass bottles (2 L) with stopper cap were used for sampling after washing with detergent Rankleen (Avantor) in hot water followed by rinsing with deionized water-acetone and dried in an oven at 100 °C. Samples were collected from 1 ft depths of river and pond surfaces using Kemmerer water sampler and stored in an icebox at 4 0 C. Sodium azide (NaN 3 ) solution (2.0%) was used for the protection of water samples from microbial growth. Samples were extracted immediately and processed subsequently.

Chemicals and reagents
A total of 36 pesticides including insecticides, herbicides, fungicides, and acaricides of different classes were selected for the present study (Table 1). All certified reference materials (CRMs) were purchased from Dr. Ehrenstorfer (GmbH, Augsburg, Germany) with purity > 97%. LC-MS grade dichloromethane (DCM), hexane, acetone and ethylacetate (EA) were obtained from J.T. Baker, Avantor, USA. Analytical grade anhydrous magnesium sulfate (MgSO 4 ) and sodium chloride (NaCl) were purchased from Merck, Darmstadt, Germany. MgSO 4 was heated at 450 0 C for 5 h to remove phthalates and stored in a desiccator.

Preparation of stock and working standards
Stock solutions of each CRM (100 µg mL −1 ) were prepared in a 100-mL volumetric flask with hexane/toluene (1:1) solvent mixture. Working standard mixtures of different concentrations (10-10,000ngmL −1 ) were prepared from the stock solution using dilution technique. All the stock and working solutions of 36 standards were stored under refrigerated condition (− 4 °C) and protected from sun light.

Sample preparation
The conventional liquid-liquid extraction (LLE) method was optimized in three different solvent mixtures, namely, S 1 (EA/DCM 8:2), S 2 (Hexane/DCM 8:2), and S 3 (100% DCM), to test extraction efficiency and compare recovery% (Fig. 3). Collected water samples were filtered through Whitman glass fibro filter (GF/F, 0.45 mm) to remove suspended particles before extraction. Sample (750 mL) was taken in a separatory funnel (1 L), and NaCl (150 g) was dissolved properly in water and extracted thrice with 70, 40, and 40 mL of each solvent mixture (S 1 , S 2 and S 3 ) in a separating funnel shaker at 210 rpm. Altogether, 150 mL (70 + 40 + 40) of organic solvent layer was collected in a conical flask after passing over anhydrous Na 2 SO 4 and evaporated to dryness in a rotary vacuum evaporator at 40 0 C. The dried residues were reconstituted in 5-mL hexane and transferred in a Tur-boVap tube for evaporation to dryness using TurboVap LV and residues again reconstituted in 1-mL acetone. After vortexing (30 s) and sonicationing (5 min), the sample was syringe filtered with 0.2-µm nylon 6,6 membrane (Ultipor) and transferred to 2-mL vial for gas chromatography-mass spectrometric (GC-MS) analysis.

Instrumental analysis
The final extracted samples were analyzed using GC-MS, QP 2010 Plus (Shimadzu Corp., Kyoto, Japan), with a mass selective detector (MSD). Initial temperature was held at 40 °C for 1 min, raised progressively 25 °C min −1 to 130 °C, again 12 °C min −1 to 180 °C, and finally 3 °C min −1 to 280 °C, with a hold time of 7 min. The injector temperature was set at 250 °C. Helium was used as a carrier gas (purity 99.999%). The ion source temperature was optimized at 250 °C, and the interface temperature was set at 280 °C. Instrument was operated in spit mode with split ratio of 1:10 and injection volume of sample was 2 µL. The MS conditions included solvent delay of 6 min, scan rate of 0.50 s −1 , and scanned mass range of 50-500 m/z. All samples were analyzed in selected ion monitoring (SIM) mode (Fig. 4), and data were acquired and processed by GC-MS Lab Solution Software (version 4.45). The pesticide-specific retention times, m/z ions, and molecular mass for the identification, confirmation, and quantification are represented in Table 1.

Method validation
The method was validated following SANTE/12682/2019 guidelines including linearity, limit of detection (LOD), limit of quantification (LOQ), specificity, accuracy (% recovery), and precision (% RSD) (SANTE 2019). LOD was determined based on signal-to-noise ratio (S/N) of 3:1, whereas S/N ratio of 10:1 was used to set LOQ. A 5-point (10, 20, 50, 100, and 250 ng mL −1 ) calibration curve of mixture-standard solution was extrapolated to examine linearity with regression coefficient (R 2 ). A comparative recovery experiment in triplicates was carried out with blank water sample at fortification levels of LOQ, 2 × LOQ, and 5 × LOQ. The results of repeatability, reproducibility, and trueness were expressed as relative percentage deviation (%RPD).

Risk assessment
Risk quotient (RQ) was assessed and curtained following EPA's level of concern (LOC) for aquatic animals (e.g., fish and invertebrates), for presuming the potential risk associated with the presence of pesticide residues in the aquatic environment (USEPA 2020). The RQ ratio reflects low-and high-risk state of the aquatic ecosystem and was evaluated by dividing a point estimate of exposure by a point estimate of effects by considering two representative taxons like fish and algae as an aquatic animal and semiaquatic plant, respectively.

Meteorological parameters
Meteorological data for the entire study period (temperature, relative humidity, and rainfall) were also collected from the Department of Agricultural Meteorology and Physics, Bidhan Chandra Krishi Viswaavidyalaya (BCKV), Kalyani, West Bengal (Fig. 1). These data were used to correlate and discuss the seasonal variation of the occurrence of pesticide residues in environmental water samples.

Method standardization
The extraction efficiency of pesticides with different polarities depends on the polarity of the extraction solvent (Allen et al. 2015). Single and binary solvent systems were used to standardize the best solvent combination for the extraction of pesticides from water samples. Three solvent mixtures namely, S 1 (EA/DCM 8:2), S 2 (Hexane/DCM 8:2), and S 3 (100% DCM), were considered for extraction. Among these three solvent mixtures, the highest % recovery (77.85 to 110.5) was achieved in the case of S 1 solvent mixture at a fortification level of 0.10 ng mL −1 , whereas other two solvent mixtures (S 2 and S 3 ) were not able to extract all pesticides satisfactorily. Even though all solvent mixtures showed an average percent recovery from 11.52 to 258.03 (Fig. 3). Sibali et al. (2008) used hexane (100%), DCM/methanol (1:1), DCM/hexane (1:1), and DCM (100%) as extracting solvent mixtures and found 100% DCM as the best one, producing recoveries above 70%. Using the standardized solvent mixture S 1 (EA:DCM 8:2), the method was validated following SANCO/11813/2017guidelines. The LOQ of the individual pesticide was calculated and found to be within the range of 0.011-0.087 ng mL −1 . A common LOQ of 0.1 ng mL −1 was set for all 36 pesticides to execute recovery experiment as the standard value by EC (2016) and in accordance with Govt. of India (Sankararamakrishnan et al. 2005;Raghuvanshi et al. 2014). Average percent
Residues of malathion and profenofos were not detected in any pond water in any season, although those two pesticides were quite frequently noticed in river water, whereas in pond water, pp´DDD was monitored in summer and monsoon seasons and β-endosulfan only in monsoon. Chlorpyriphos and butachlor residues were recorded in tube-well water samples irrespective of seasons. No residues of other pesticides under study were detected in any of the environmental water samples irrespective of seasons. A total of 96 tube-well water samples (48 from each location) were analyzed in the present study. Seven (7.29%) tubewell water samples were contaminated only with 2 (5.55%) pesticides (chlorpyriphos and butachlor) and the remaining 89 (92.71) samples were not contaminated in any season or below the limit of quantification (LOQ) of 0.1 ng mL −1 . Alarmingly, the river water was contaminated with 18 pesticides including 11 organochlorines of the restricted use category (UNEP 2003) and low water affinity (Singh et al. 2012).
The mean concentration of total organochlorine (T OC = Total HCH + total DDT + total endosulfan) residues in river water was quantified as 2.96 ng mL −1 (Table 2) despite of recommended for restricted uses. These values are higher compared to those reported in other Indian rivers, viz., Yamuna (Kaushik et al. 2008) and Kuano (Singh and Mishra 2009), but obviously lower than the reported T OC level in the Sembrong lake basin in Malaysia (Sharip et al. 2017). In this study, DDT isomers were detected more frequently in river water compared to pond water. Although DDT is banned in agriculture, the occurrence of its residues might be due to its uses in National Malaria Eradication Programme (NMEP) in India (NVBDCP 2016). Traces of DDTs and HCHs were getting high in the river water during the monsoon which might be due to their wide uses in wetlands to control mosquito larvae, followed by transportation to adjacent aquatic systems via runoff and leaching after rains. Moreover, the slow rate of degradations of DDT isomers in nature by biological and chemical methods leads to its persistence in the environment (Mondal et al. 2018). Values of total HCH (sum of HCH isomers) in the river system indicate the non-point source of contamination. It is assumed as an old source of contamination due to the isomer's least reactivity and the most persistent nature among HCH isomers (Wang et al. 2007).
Remarkably, OP pesticides were predominant in pond and tube-well water, located in neighboring places within a 1-km radius, whereas the abundance of OC residues in river water systems was monitored. A mean concentration of chlorpyriphos (0.13 ng mL −1 ) was found, and the insecticide was frequently detected in pond water in all seasons. In tube-well water, it was the second most noticed pesticide next to butachlor irrespective of seasons. Intensive agricultural practices adjacent to the water system were also the probable reason for the recommended OP pesticide contamination in pond water (Agrawal et al. 2010), but the traces of banned OC (PPQS 2021) residues leads to unknown sources of contamination. The residues of OPs in pond water are probably due to contamination from agricultural pesticides used in nearby cultivated land (Reddy and Reddy 2010;Lari et al. 2014). The mean concentration of T OC was much lower in pond water (0.32 ng mL −1 ) compared to river water throughout all seasons. Interestingly, no residues of synthetic pyrethroid (SP) insecticide were detected in any water throughout the study period. This is probably due to the photo instability of synthetic pyrethroids (Kocoureket al. 1987). Moreover the persistence and leaching capacity of synthetic pyrethroids in moist alluvial soil are relatively low (Gupta and Gajbhiye 2002) which supports our results of water samples of the new alluvial zone.
Residues of pesticides found in different seasons were influenced hugely by environmental parameters like rainfall, temperature, and relative humidity (Fig. 1). In the monsoon season, mean rainfall (50.9 mm) was three times higher than summer and the average residues found in different aquatic systems were also the highest in monsoon season. In the tube-well water sample, insecticide chlorpyrifos (0.157ngmL −1 ) and the herbicide butachlor (0.065 ng mL −1 ) were detected in monsoon period predominantly. In river and pond water samples, pesticides were found in winter comparatively lower than summer, because the average temperature of summer was expectedly higher with the least rainfall, lower the leaching rate. Although winter season is the major agricultural period of Bengal, this scenario also indicates some non-point sources of contaminations. Prevalence of agricultural pesticides in neighboring pond water systems may endorse to the run-off monsoon water from the subsequent agricultural land (Reddy and Reddy 2010;Lari et al. 2014). Seasonal variation of detected residues of OP and OC pesticides in aquatic system (Fig. 5) showed different behavior (Fig. 6). Abundances of OC pesticides in river water and OPs in pond water were higher in monsoon season. Nguyen et al. (2019) also reported the maximum OC pesticide concentration in the Dong Nai River during the rainy season. The rising of OCs' contamination in river water was related to run-off rain water to river from nonspecific sources. Moreover, OCs have an affinity to gather in the river sediments due to their hydrophobic nature and high octanol-water partition coefficient (Yang et al. 2005;Masiá et al. 2013).
Most of the study areas consist of well-drained sandyloam soil, and the ground water levels are 35-50 ft deep from the surface. This low depth ground water (tube-well sample) showed seasonal contamination due to rapid transport of soil holding pesticides in monsoon (butachlor and chlorpyrifos) with low water solubility and high Log K ow (octanal-water partition coefficient) (Yang et al. 2005;Masiá et al. 2013). Table 3 lists the physicochemical properties of the pesticides inspected.
In summer there is reduction of water volume in river and pond with the increase of temperature (Sharip et al. 2017;Kaushik et al. 2012).Total pesticide concentration was much higher in monsoon than summer and winter period (Fig. 5). The huge rainfall and temperature variation were considered as the major factors influencing the movement of pesticides entering the water systems (Zhang et al. 2017). The increase in rainfall during the monsoon season facilitates runoff from catchment areas (which are predominantly agricultural lands) and leaching of surface water which altogether contributed to high concentrations. The results of higher residues may also reflect a rather higher application of pesticides for main crops in monsoon. Tanabe et al. (2001) reported the maximum residues in Shinano River in the month of May when the most applications were conducted.

Risk assessment
Aquatic as well as human ecosystems are very much scarring due to the biotransformation (Le Du-Lacoste et al. 2013) and bioaccumulation (Luna-Acosta et al. 2015) of pesticide residues in aqueous systems. In India, due to the persistent nature of DDT (since 1993) and HCH (since1997), the restriction was imposed in its agricultural (UNEP 2003;PPQS 2021) as well as domestic use (UNEP 2003;NVBDCP 2016). The remarkable presence of nonagricultural pesticides in water bodies is very alarming in spite of their restricted uses. OC pesticides are also the key suppliers of global persistent organic pesticides (POP) circulation (Yadav et al. 2015) and are also known as global pollutants. Mankind directly depends on water and aquatic life, so the risk is connected with contaminated water systems and needs to be evaluated in terms of humans (EC 2016) and aquatic health (USEPA 2020). EC (European Commission) limit for single pesticide and T-pesticide in drinking water are 0.10 ng mL −1 and. 0.50 ng mL −1 , respectively (EU 1998). It was observed that the permissible limits of all pesticides in drinking water were not set by agencies, i.e., World Health Organization (WHO), Indian Standard Institution (ISI), Central Pollution Control Board (CPCB), and Indian Council of Medical Research (ICMR). Interestingly, more than 50% of monsoon samples have exceeded the limit for both pond and river water. In 14% of contaminated river water samples, the levels of pp´-DDT were exceeded above EC limit of drinking water, whereas, in the case of pond water, about 10% of contaminated samples have exceeded the EC limit for chlorpyrifos. Pollution of pesticides is related to cancer, obesity, and other diseases in humans (Upadhayayet al.2020). Therefore, proper safety measures should be adopted for these water resources for community drinking. Conversely, tubewell water was safe for drinking as water contained residues below the total pesticides limits irrespective of seasons.
To evaluate the possible ecological risk (acute and chronic) associated with the pesticides, contaminated river and pond water risk quotient (RQ) was assessed ( Table  2). The RQ level of concern (LOC) of semi-aquatic (e.g., marimo and algae) plants was 1, but our calculated RQ values (< 1.0) showed no possible risk on the aquatic ecosystems as it was below LOC. In river water, the estimated RQ value (0.434) of T-HCH for aquatic animals (e.g., fish; Oncorhynchusmykiss) was very closer to the limiting marks of EPA's (United States Environmental Protection Agency) LOC for acute toxicity in terms of acute high risk (0.500). Chen et al. 2020 also indicated OC pesticides as a Fig. 6 Frequency of occurrence of pesticides in River and Pond water with three seasons moderate risk group in their study at Shanghai river, China. In river water fish, RQ of total endosulfan (0.295) crossed acute toxicity level in terms of acute endangered species (LOC = 0.050) and also indicated high chronic toxicity (RQ = 5900 compared to LOC = 1). So, the potential risk of river water on animals (e.g., fish) was expected by persistent OC pesticides, predominantly endosulfans. Kapsi et al. (2018) and his team also reported α-endosulfan as a medium risk at Louros river, Greece. Profenofos also indicated chronic toxicity threat in river water (RQ = 7.5) for aquatic animals. However, in the pond water system, there was no ecological risk (RQ < 1) for total-HCHs, total-DDTs, and total endosulfan. Although this study exhibited that the borderline RQ value of chlorpyrifos (0.928) comparing the LOC for chronic risk (1.0) in pond water fish, indicating the future topic of concern. Very high RQ values of chlorpyriphos and total-endosulfan were (Jabali et al. 2020) far alarming for their toxicity to fish. Ccanccapa et al. (2016)  that chlorpyrifos is the most frequent occurring pesticide in Ebro river water. Subsequently, the outcomes of this study are originated very serious for aquatic habitats around the study area.

Conclusion
In this study, an effective and sensitive method involving liquid-liquid extraction (LLE) and GC-MS detection was standardized for the determination of 36 selected pesticides in the surface and groundwater of Barrackpore and Kalyani, West Bengal. EA and DCM (8:2, v/v) mixture solvent system was found to be the best compared to other solvent systems for LLE. The developed method was validated in accordance with EC (European Commission) guidelines and successfully applied for monitoring of 36 pesticide residues (OCs, OPs, and others) in different aquatic systems with seasonal variations. The distribution of OPs and OCs was evaluated in river and pond water systems with variable environmental conditions. Comparatively, higher concentrations of pesticide residues were occurred during monsoon period than summer and winter. Despite the restricted use of persistent OCs, the river water samples were found to be contaminated with high amounts of OC residues which reflect non-point sources of contamination.
In most of the samples, the pesticide residue levels spotted were higher than EC recommended drinking water quality standards. Risk assessment in the aquatic system was also assessed where the potential risk on aquatic bio-network in terms of acute and chronic toxicity was observed for OC pesticides in river aquatic ecosystems. Therefore, this study initiates reference line data for the possible execution of pollution control and governing decision-making policies.