Introduction

Persistent organic pollutants (POPs) are organic compounds characterized by their toxic properties, resistance to degradation, ability to be bioaccumulated and transported through air, water and organisms, across international boundaries and then deposited far from their place of release. Considering the necessity to take actions to prevent adverse effects caused by POPs at all stages of their life cycle, the Stockholm Convention (SC) of the United Nations Environment Program (UNEP) established a list of priority pollutants (UNEP 2017). Among them, the organochlorine pesticides are listed as part of the so-called “dirty dozen”.

Even though, the use of these pesticides has been banned since the 1970s (Japan, Germany, USA), 1980s (UE, Australia) and 90’s (Argentina), studies around the world continue to report elevated levels of OCs and their incidence in the environment can be considered on a global scale rather than a local case. For instance, there are several organochlorine pesticides still in use in emerging countries and tropical regions, which can be later mobilized by evaporation and subsequent atmospheric transport, to be ultimately deposited even in environmentally pristine areas such as polar zones (Rajendran et al. 2005). Although the new technological procedures have changed the effectiveness of the agrochemical application, it has been documented that in the past, even percentages lower than 0.1% reached the target goal, leaving over 99% elsewhere in the environment (Pimentel 1995). Once applied in the field, they can meet a variety of fates, such as volatilization, photodegradation, and carried by atmospheric mass transport or freshwater runoff. Thus, the soil is one of the environmental matrices that act as a sink for these chemical substances that reach its surface temporarily or permanent (Miglioranza et al. 2003a). In turn, OCPs in freshwaters have only short residence time due to their hydrophobic character and a strong sorption onto suspended particulate matter; therefore, they are carried downstream to coastal environment and finally trapped in marine sediments.

The gradual change in Argentina towards modern and intensive agricultural activities has led to an increase in the use of pesticides, but there are not reports related to the amount of OCs used before they were banned (Miglioranza et al. 2003b) and a scarcity of information for coastal strips-around less than 100 km from the ocean, where 61% of the gross domestic product (GDP) is generated (worldwide average) (MAyDS 2016a). In particular for the Argentine Atlantic coastline (along 4725 km) only six urban hubs exceed 100,000 inhabitants (MAyDS 2016b); the study area is included in such group and exhibits about 450,000 inhabitants according to the last National Population Census in 2010 (INDEC 2017). The main city in the area, Bahia Blanca city, is located on the north coast of the estuary that bears the same name and next to a significant agricultural-livestock zone, is one of the most important urban and industrial emplacements of Argentina. Precedent studies for the area were mainly focused on the marine and coastal zone, leaving a gap in the knowledge with regard to OCs in inland locations. In fact, the first baseline data for OCs in the region were reported by Sericano and Pucci (1984), Sericano et al. (1984) and Zubillaga et al. (1986) who studied these POPs in sediment and water of the estuary and its tributaries. After that, Farrington and Tripp (1995) continued the studies reporting OCs levels in mussel’s tissue and finally, Andrade et al. (2005), Arias et al. (2011) and Tombesi et al. (2014) reported levels of OCs in soils from the horticultural region next to Bahía Blanca city, sediments from the Bahía Blanca estuary, and air sampled in the city and surrounding region from the southwest of the Buenos Aires Province, respectively.

To fill the gap in regards to inland agriculture potentially impacted locations, twelve OCs, namely: p,p′-DDT, o,p′-DDT, p,p′-DDE, o,p′-DDE, p,p′-DDD, o,p′-DDD, the hexachlorocyclohexanes (α-HCH, β-HCH, δ-HCH and γ-HCH), and pentachlorobenzene (PeCB) and hexachlorobenzene (HCB) were studied along thirteen sampling points from the area. Then, the present study aims to assess the concentration levels, spatial distribution and putative input sources of these OCs, as well as to distinguish regions according to main current inputs of banned OC’s using principal component analysis.

Materials and methods

Study area

The area of study comprises the Bahía Blanca city and a surrounding area (about 100 km around) located on the southwest of Buenos Aires Province, Argentina (Fig. 1). This area bears an important industrial activity and is header from a leading agricultural export zone of the country by operating characteristics of their harbor system (MBB 2017). Moreover, the study area is located in the Pampas region of Argentina, an extensive plain with more than 52 million ha of land devoted mainly to livestock and farming activities (Viglizzo et al. 2004), including two sub-regions characterized by dry steppes of moderate continental climate (Dry Pampa) and humid temperate prairies (Humid Pampa) (Gutierrez and Martinez 2008). Especially, the study area is placed in the transition between dry and wet zones. The climate and geographic characteristics makes the region also favorable for wheat and barley farming with irrigation-assisted production to the south.

Fig. 1
figure 1

Site location for soils (green marks) and sediments (red marks) samples from the Bahía Blanca city and estuary (a) and the southwest of Buenos Aires Province (b)

Full size image

The study area also includes an estuarine ecosystem which has a total surface of 2300 km2 with 410 km2 of islands and 1150 km2 of intertidal area (Arias et al. 2011). The Bahía Blanca estuary is classified as a mesotidal system with very slight fluvial inputs sheltered by wide tidal flats and salt marshes (Piccolo and Perillo 1990). In general, the smallest channels flow from the inner plain to the largest channels, all of which drain into the Principal Channel route to the port complex in this region, where Bahía Blanca and Punta Alta cities are emplaced (Vecchi et al. 2013). Intensive anthropogenic activity at the north shoreline include oil, chemical and plastic factories, commercial harbors, and a fishing fleet of 12-m total length ships (Arias et al. 2009).

Sample collection

Representative soil and sediment samples were collected along a zigzag track (Tack and Verloo 2001) at the same sampling point. At each site, eight subsamples of a surface layer (1–5 cm) were taken and mixed in glass containers to homogenize them and generate composite samples (700–1000 g). They were air dried at room temperature for at least 72 h, isolated from the laboratory environment.

Oxic surface sediment samples-with light brown, flaky aspect and no sulfide smell-were taken by March 2011 at four locations (indicated as S1–S4) on the intertidal area, at the north coast of Bahía Blanca estuary (southwest of Buenos Aires Province, Argentina). The samples were sieved through a 250 mesh, retaining for analysis the particle size fraction less than 63 µm.

Soil samples were collected by October and November 2014 at nine sites, two located on Bahía Blanca city (labeled B1 and B2) and seven from an area 100 km around this urban cluster (identified as R1–R7). Table 1 shows the description and location of the sampling sites, and Fig. 1 illustrates the map of the region.

Table 1 Geographic coordinates and description of the sampling sites
Full size table

OCs analysis

Sediment and soil samples were analyzed for α-HCH, β-HCH, δ-HCH, γ-HCH, p,p′-DDT, o,p′-DDT, p,p′-DDE, o,p′-DDE, p,p′-DDD, o,p′-DDD, PeCB and HCB. High purity analytical reagents and chromatographic grade chemicals were used thorough the work. Organic solvents were supplied by J.T. Baker (Poland). Reference standards were provided by LGC Standards (© LGC Limited, UK), and by Wellington Laboratories Inc. (Canada).

Samples (5.00 g) were homogenized with anhydrous sodium sulfate to perform a chemical drying. Hereafter, they were spiked with surrogate recovery standards and introduced into previously acetone extracted cellulose cartridge (Whatman). Samples were Soxhlet-extracted for 8 h with hexane–acetone mixture (125 mL). Concentrated extracts were cleaned-up on a H2SO4 modified (44% w/w) silica column, and the analytes were eluted with 30 mL DCM/n-hexane mixture (1:1). The eluate was concentrated using stream of nitrogen in a TurboVap II concentrator unit and transferred into an insert in a vial. The syringe standard (native PCB 121) was added to all samples, the final volume was 100 μL.

GC–MS/MS was used for OCPs analysis. 7890A GC (Agilent, USA) equipped with a 60 m × 0.25 mm × 0.25 μm HT8 column (SGE, USA) coupled to 7000B MS (Agilent, USA) operated in EI + MRM was used. Injection was pulsed splitless 3 µL at 280 °C, He as carrier gas at 1.5 mL min−1. The GC temperature programme was 80 °C (1 min hold), then 40 °C min−1 to 200 °C, and finally 5 °C min−1 to 305 °C.

Quality assurance/quality control

To check for interference and cross contamination, a procedural blank consisting of all reagents and spiked samples (soils and sediments) were run for every set of five samples. The method detection limits (LOD) of OCs were described as 3:1 signal vs noise value. LOD of OCs ranged from 0.01 to 0.06 ng g−1, dry weight (dw).The spiked recoveries were higher than 75% for all samples for OCs. Recovery factors were not applied to any of the data. Recovery of native analytes measured for a reference material (soil) varied from 75 to 98% for OCs. The relative standard deviation ranged from 4 to 9%.

PCA analysis

To obtain an outline of correlations of the obtained data, a principal component analysis was performed. As an exploratory technique, PCA allows to study of main sources of variablity present in the data sets, to detect clustering, and to found relationships between samples (objects) and compounds (variables) (Jolliffe 1986). Statistical analysis was implemented with XLSTAT 19.03 (free version, 2017).

Results and discussion

The OCs concentrations (ng g−1 dw) obtained for sediment and soils samples are shown in Table 2. Average concentration levels for ∑OCs in soil/sediments ranged between 0.206 and 1040.4 ng g−1 dw (mean = 82.4 ng g−1 dw) and from 0.0805 to 1039.9 ng g−1 dw (mean = 81.3 ng g−1 dw), and 0.0858 to 0.876 ng g−1 (mean = 0.43 ng g−1 dw) for DDXs and HCHs, respectively.

Table 2 OCs concentrations (ng g−1 dw) found in sediment and soil samples from the study area
Full size table

HCHs

HCHs concentrations ranged from 0.410 to 0.876 ng g−1 dw in sediments and from 0.0858 to 0.589 ng g−1 in soils. The highest HCHs concentrations were found for S1 sediment (Puerto Cuatreros, a meat factory old harbor) and the lowest for R2 soil (Hilario Ascasubi, a rural and agricultural-livestock area). The relative composition of HCH isomers at each sample is shown in Fig. 2. In regards to HCHs, two commercial products have been commercially available at the market: technical HCH (usually 55–80% α-HCH, 5–14% β-HCH, 12–15% γ-HCH and 2–10% δ-HCH) (Kim et al. 2002) and lindane (principally γ-HCH). Technical HCH was firstly forbidden in Argentina by 1980 for agricultural and livestock use, and completely banned in 1998. On the other hand, lindane-as other antiparasitics for external use containing chlorinated hydrocarbons as active material- had its first restriction for livestock use in 1968 by the National Service of Agri-Food Health and Quality of Argentina (SENASA); however, it was still permitted until 2011 for human medical treatment of pediculosis and scabicides (ANMAT 2011). Consistently, the more abundant compounds in the commercial products (α-HCH and γ-HCH or only γ-HCH) were found in all the samples with values above the limit of detection (Table 2).

Fig. 2
figure 2

HCHs relative composition at each sampling site

Full size image

A significant proportion of γ-HCH (around 60% in average; Fig. 2) was recorded for most of the samples (S2, S3, S4, B1, B2, R1, R2 and R7) and could be related to the lindane usage. To differentiate fresh usage of a technical mixture from an aged signature of technical product and/or lindane usage (mainly γ-HCH), the α-HCH/γ-HCH ratio was calculated for each sample. While a fresh input usually yields 4 to 7, and aged/lindane usage one generally obtains lower values (Shen et al. 2004). In general, the α-HCH/γ-HCH ratios ranged from 0.11 to 0.22 and from 0.08 to 0.31 in sediments and soils, respectively, pointing to aged technical mixture signatures and/or lindane usage.

β-HCH-secondary product of technical mixture (5–14%) or unintentionally generated during the production of lindane-showed high percentages in sediments (16–53%) and in B1, B2, R1, R2 and R7 soil samples (25–44%). As this isomer is considered to be the most refractory due to its chemical structure (Decision POPRC-2/10, 2006), these results are consistent with the hypothesis of an extended and continuous use of HCH pesticides at those sites.

Finally, soil samples R3 to R5 showed a high proportion of δ-HCH. The prevalence of δ-HCH has been observed in other agriculture zones (Rajendran et al. 2005) and indicated as a degradation product of γ-HCH (Manz et al. 2001). Then, the δ-HCH occurrence could be interpreted as old and degraded inputs of that compound (Arias et al. 2011).

The highest lindane levels (γ-HCH) were recorded in sediments (0.235–0.447 ng g−1) while for soils, the values ranged from 0.0482–0.212 ng g−1 with the maximum corresponding to B1 (Bahía Blanca city). A rationale for this could be explored in a water stream irrigating the area: the Sauce Chico river, which cross an extensive area of crops after ending in the Bahía Blanca Estuary.

In general, sediment concentrations of HCHs were comparable or lower to those reported by Arias et al. (2011) (< LD to 3.20 ng g−1 dw) and Sericano and Pucci (1984) (0.3–14.2 ng g−1 dw) corresponding to sediments from the same area of study. In regards to HCH’s levels in soils, there are not previous data for the area of study, with the exception of lindane (up to nearly 800 ng g−1) from a punctual area close to Bahía Blanca city which has been intensively farmed for the last six decades in horticultural production (Andrade et al. 2005). For this reason, a comparison with levels of HCHs in soils found in other worldwide locations is included in Table 3. The HCHs concentrations observed in this study were lesser than the reported in literature, included that observed by Miglioranza et al. (2003a) for other region of Argentina, and only similar to the values found by Zang et al. (2015) in Antarctic Peninsula.

Table 3 Comparison of OCPs levels in surface soils (ng g−1 dw) at worldwide locations
Full size table

DDXs

Widely applied for agriculture purposes as an insecticide in the past, DDT has been totally banned in Argentina since 1998. Once released into the environment, the main component of DDT, p,p′-DDT is transformed onto p,p′-DDE. The relative abundance of parent and metabolite is frequently used to discriminate between recent use (DDT/DDE > 1) and an old mark (DDT/DDE < 1) (Miglioranza et al. 2003a; Tombesi et al. 2014; Zhang et al. 2015, 2018). In this study, DDXs concentrations ranged from 0.32 to 7.29 ng g−1 dw in sediments and from 0.0805 to 1039.9 ng g−1 dw in soils. Both the highest and the lowest DDXs concentrations were found in soil samples: R5 (Dorrego, a rural, agricultural-livestock area close to agrochemical supplier for aerial fumigation) and R3 (La Chiquita, a new and small seaside resort) respectively. For sediments and soils, DDT/DDE ratio was < 1 at all cases, with the exception of two samples B1 and B2 (1.73 and 1.06 for urban and industrial soil, respectively, Fig. 3). The DDT/DDE ratio recorded at R5 (a DDT hotspot ~ 1040 ng g−1 dw) was 0.24 and pointed to old DDT inputs and a probable reductive dechlorination process of DDT to DDE under aerobic conditions. Meanwhile, in B1 and B2 sites the DDT/DDE ratios were just above 1 and could be attributed to a more recent input of DDT compared with the rest of the sites. As stated by Miglioranza et al. (2003b) a putative present source of DDT could be attributed to commercial Dicofol, a widespread acaricide which is still in use in Argentina and has been reported to have a substantial amount of DDT as an impurity of the production process.

Fig. 3
figure 3

DDXs relative composition at each sampling site

Full size image

To compare with precedent p,p′-DDXs data for sediments of the Bahía Blanca estuary (Arias et al. 2011), the calculation of the contribution of these isomers has been made. In this study, p,p′-DDXs levels ranged between 0.25 and 6.09 ng g−1 (S2 and S4, respectively) which were comparable to those found by Arias et al. (2011) (0.39–3.65 ng g−1).

Moreover, total DDXs levels found in soils in this study were overall below that the only data available and previously reported by Andrade et al. (2005) up to 12,000 ng g−1 dw for soils corresponding to horticultural belt of Bahía Blanca city. DDXs levels found in other sites of the world are showed in Table 3. Except the site R5, DDXs concentrations observed in this study were comparable to the observed in Antarctic Peninsula (Klánová et al. 2008; Zhang et al. 2015) and similar or lower than the detected for other agricultural, urban and industrial areas from other region of Argentina (Miglioranza et al. 2003a), Albania (Mukaj et al. 2016), China (Meng et al. 2013; Sun et al. 2016b), India (Kumar et al. 2011; Mishra et al. 2012), Italy (Qu et al. 2016), Kenia (Sun et al. 2016a), Mexico (Orta-García et al. 2015) Pakistan (Alamdar et al. 2014) and USA (Bidleman and Leone 2004). Although the highest DDXs level found in this study is located close to agrochemical supplier for aerial fumigation, the concentration (~ 1040 ng g−1 dw) is similar or lower than the reported in other agrochemical handling sites as in Brazil (Rodrigues et al. 2017) and Pakistan (Alamdar et al. 2014), and that observed in special sprayed areas for disease control as Oman (Booij et al. 2016).

PeCB and HCB

Polychlorobenzenes were developed in the second half of the nineteenth century for seed treatment. Precisely, hexachlorobenzene (HCB) was used as a highly specific weapon to control wheat bunt (Tilletia foetida and T. caries) (Matanguihan et al. 2011). For Argentina, the HCB was first restricted by 1993 and banned in 2000 (SAGPA 2000) while there is not specific regulation on PeCB usage. On the one side, PeCB ranged from 0.035 to 0.658 ng g−1 dw in sediments and from < 0.0017 to 0.0154 ng g−1 dw in soils, while, in the other side, HCB ranged from 0.145 to 6.36 ng g−1 dw and from 0.0139 to 0.0534 ng g−1 dw in sediments and soils, respectively. PeCB showed maximum values in sediments and the highest was registered for S3, a petrochemical park effluent’s discharge area. A rationale for this could be found taking into account that this compound may be generated as by-product or impurity in the manufacture of several chlorinated compounds as vinyl chloride, which in fact, is produced in the area (EEA 2005). This is the first record for PeCBs as no previous data were reported in sediment/soils from the area of study.

Principal component analysis (PCA)

To characterize the OCs occurrence patterns and sampling locations a PCA was applied. Then, the biplot of principal component (PC) analysis (loading plots and score plots) is presented in Fig. 4.

Fig. 4
figure 4

Score plot of the first three components and explained variance: a PC1 vs PC2 (73.66%); b PC1 vs PC3 (66.34%) and c PC2 vs PC3 (36.06%)

Full size image

Results showed that the three first principal components explained 88.03% of the total variance present in the data set. PC1 accounted for the major proportion of the total data variance (51.97%) while PC2 and PC3 explained 21.69 and 14.37%, respectively and the following PCs progressively explained smaller amounts of data variation. Firstly, PC1 was mainly associated with the content of DDXs and unequivocally linked to R5; in this way, this component was interpreted as a DDT usage factor.

PC2 was able to distinguish between the group of sediment samples (S1, S2, S3 and S4)—which scored positively—from the rest of soil samples (B1, B2, R1, R2, R3, R4, R6 and R7)—which were negatively scored. In addition, HCHs (except δ-HCH), PeCB and HCB were positively clustered to this component; then, this PC could be interpreted as resuming the input of technical HCH mixture signatures.

Finally, for PC3, the absolute loadings of HCB and PeCB were significant. Taking into account the most probable source for these compounds in the area, this PC was interpreted as chlorinated petrochemical by-products inputs. The strongest spatial association for this component was S3, a petrochemical park effluent’s discharge area.

Figure 4 shows the score plot of the first three components. These scores allowed the characterization of the sampling stations according to the different explained variance components. As stated before, several locations showed high scoring at different PCs. Thus, PC1 (DDT usage) highest scores corresponded to R5, located at rural, agricultural-livestock area close to agrochemical supplier for aerial fumigation. Although there is not enough information, based in both levels and age ratios, it can be hypothesized that these depositions belong to an extensive chronic spills which lasted for years.

PC2 highest scores (technical HCH mixture signatures) were for a group of sediment samples (S1, S2, and S4) receiving continuous freshwater inputs of Sauce Chico stream and Maldonado channel (Napostá stream branch) after running through hundreds of kilometers of intensive agriculture lands. These results are consistent with previous findings (Arias et al. 2009) and reveal the possible breakage of the technical HCH banning law. At last, PC3 (pesticides from petrochemical by products) significant scores were mostly contributed from S3, where the surrounding area counts an industrial park zone which includes the production of chlorinated compounds such as PVC.

Conclusions

The present study fills a gap in the knowledge in regards to the occurrence of banned pesticides at both coastal and inland locations at the Southwest Buenos Aires region, Argentina. For the first time several DDT, HCHs and PeCB/HCB hotspots were uncovered, revealing spatial distribution occurrence and putative input sources for these OCs. All detected OCs are presently banned by the Argentinean legislation. On the one side, a technical HCH and lindane signature was found to be widespread in the area, in particular for urban (Bahia Blanca and Médanos cities) and rural locations (Ascasubi, Napostá). In addition, a strong coastal pesticides sign revealed their possible input through runoff and tributaries. On the other side, DDXs were preeminent in inland locations related to rural/agriculture activities and the signatures were tagged as old at all cases. On the opposite, Bahía Blanca urban locations showed quite recent DDT inputs. Additionally, a hotspot showing levels of warning was identified at one of the rural/agricultural-livestock sites (Dorrego). Further, PeCB and HCB occurrence in the area was linked to by product or impurity in the manufacture of several chlorinated compounds as vinyl chloride. Finally, the Principal Component Analysis provided an overview of pesticides distribution, allowing the distribution of sampling site along three main components involving the three major groups of pesticides found: DDXs, HCHs and PeCB/HCB.