Introduction

Polychlorinated biphenyls (PCBs) are persistent, bioaccumulative, and toxic (PBT) compounds that have the potential to move far from original sources, thus contaminating the environment on a global scale (UNEP 2015; Wania 2003). Therefore, routine monitoring of these compounds, in the vicinity of both point sources and at remote sites is important for identifying sources of contamination and in understanding the fate of these compounds in the environment. Using plants as bio-samplers of atmospheric contamination with PBTs is a relatively cheap and uncomplicated means of monitoring atmospheric contaminants that does not require equipment/samplers and hence, is particularly useful for sampling at remote sites that may be affected by long-range atmospheric transport (LRAT) (Bergknut et al. 2011; Grimalt and Van Drooge 2006, Klánová et al. 2009; Rappolder et al. 2007). Conifer species (e.g., spruce or pine) in particular have been used to monitor a range of PBT compounds, including PCBs because they are endemic and evergreen, and hence can be used for long-term monitoring over a broad spatial area (Al Dine et al. 2015; Bergknut et al. 2011; Chen et al. 2012; Di Guardo et al. 2003; Grimalt and Van Drooge 2006; Hanari et al. 2004; Klánová et al. 2009; Rappolder et al. 2007; Tian et al. 2008).

PCBs are expected to be present primarily in the gas phase of the atmosphere, depending on the molecular weight of the compound and the temperature (Die et al. 2015; Li et al. 2015; Lohmann et al. 2000; Simcik et al. 1996). PCBs are deposited on plant leaves primarily through dry gaseous deposition (Thomas et al. 1998). It is theorized that gas-phase lipophilic persistent organic pollutants (POPs) partition into the cuticle due to its lipophilic properties. In fact, PCBs have been found to have a greater affinity to one of the main components of the cuticle membrane, the cuticular waxes compared to other cuticle membrane components (Moeckel et al. 2008). PCBs are thought to sorb to the cuticular waxes and diffuse into internal leaf compartments (Barber et al. 2004). Therefore, plant leaves such as pine needles are akin to a natural passive sampler. However, while pine needles can be a useful indicator of atmospheric pollution, it is difficult to translate the information on levels of pollutants in needles to actual air concentration. Currently, there is not enough data on uptake rates to pine needles to calculate air concentrations of PCBs using needle contamination information. Comparison of active or passive air sampling to pine needle contamination serves to increase our understanding of the sampling efficiency of plants. By comparing profiles of PCBs in pine needles and active or passive samplers, we can also determine whether pine needles are a good indicator of atmospheric source contamination. In addition, active or passive sampler measurements can be used to assess whether pine needles can reflect temporal trends in atmospheric pollution, something which has been previously evaluated by Klánová et al. (2009).

Despite the research on PCB levels in conifer species over the past decades, there are still only a few studies reporting spatiotemporal patterns of PCB levels and congener profiles in pine needles, and many studies are conducted on a local (around a known point source), small regional or country scale (Al Dine et al. 2015; Chen et al. 2003; Falandysz et al. 2012; Tian et al. 2008; Tremolada et al. 1996). To date, there has also been limited comparison between levels of PCBs in pine species and conventional active air sampling or the increasingly used passive air samplers (Klánová et al. 2009; Romanič and Krauthacker 2007). Therefore, in the present study, we aimed to determine the spatiotemporal patterns in levels and congener profiles of PCBs in Scots pine (Pinus sylvestris) needles across three European countries (Sweden (SW), Czech Republic (CZ), and Slovakia (SK)) over two different sampling campaigns (autumn and spring), representing 0.5- and 1-year-old pine needles, respectively. In addition to this, we compared indicator PCB (ind-PCB) congener profiles from the European Monitoring and Evaluation Program (EMEP) active and passive air samples to those in pine needles.

Materials and methods

Site locations and sample collection

In brief, 13 sampling sites were selected within industrial (I), urban (U), or background (B) (i.e., distant from obvious PCB sources) land use areas. Sites were chosen across three different countries, including SW: 2 U (SW-UM and SW-LK) and 3 B (SW-AP, SW-GL, SW-KR) sites; CZ: 4 I (CZ-BB, CZ-BA, CZ-TR and CZ-HR), 2 U (CZ-PR and CZ-BR), and 1 B (CZ-KO) sites; and SK: 1 I (SK-KR) site. The details of each site and maps showing sampling locations are included in Table S1 of the supporting material (SM). Needles were taken from stand-alone trees or trees within a small group in order to standardize collection at all sites and to reduce the effect of tree canopy on air flow around the leaves, which may in turn affect PCB uptake (e.g., dense vegetation can decrease air movement around leaves and lead to decreased uptake of PCBs (Barber et al. 2002, 2004)). Pine needles were transported to the laboratory either wrapped in aluminum foil and within a plastic zip lock bag (CZ and SK) or within glass jars (SW) although this difference was not expected to have any effect on the levels of PCBs in the leaves. During transportation samples were kept at ∼5 °C. Twigs were defoliated manually using tweezers. For storage, samples were packed in air-tight glass jars and kept in the freezer (−20 °C).

Sample preparation and analysis

PCB analyses were carried out at the Research Centre for Toxic Compounds in the Environment (RECETOX) ISO/IEC 17025 accredited Trace Analytical Laboratories. Needle samples were oven dried (40 h at 50 °C), then ground and mixed with anhydrous sodium sulfate. An internal standard (IS) containing 12 dioxin-like PCBs (dl-PCBs—non-ortho PCBs 77, 81, 126, and 169 and mono-ortho PCBs 105, 114, 118, 123, 156, 157, 167, and 189) and 6 ind-PCBs (PCBs 28, 52, 101, 138, 153, and 180), all 13C12 labeled (Wellington Laboratories, USA) was added. Samples were then subject to accelerated Soxhlet extraction (B-811Buchi Extraction System) for 1 h using extraction solvent dichloromethane (DCM). Extracts were concentrated using a TurboVap nitrogen blowdown concentrator and an oven (103 °C, 1 h), and were then subject to clean-up. Clean-up included in the following order: a 25-g high-capacity sulfuric acid/silica (44 % w/w) column eluted with 450-ml hexane/DCM (9:1 v/v), a multilayer column packed with 4 g sulfuric acid/silica (44 % w/w), and 1.5 g potassium hydroxide/silica (36 % w/w) eluted with 120 ml of n-hexane, and a 3-g basic aluminum oxide column (ICN Alumina B Super I) eluted first with n-hexane (discarded) and then DCM/hexane (1:1, v/v), which was collected for analysis of PCBs. Fractionation occurred on a carbon column packed with AX-21 activated carbon dispersed on Celite 545. The first elution of cyclohexane/DCM/methanol (2:2:1, v/v) contained ind-PCBs, and the second elution of toluene contained dioxin-like PCBs. Sample extracts were concentrated, transferred to vials and recovery standard 13C12 PCB 162 was added. PCB congeners were separated using a 60-m SGE HT8 column (internal diameter of 0.25 μm) installed in a Thermo Fisher Scientific 1310 gas chromatograph (GC). Samples were splitless injected at 260 °C and helium was used as a carrier gas at a constant flow of 1.0 ml min−1. GC column temperature program for dl-PCBs was as follows: initial temperature 120 °C (1.5-min hold), then 120 to 200 °C at 30 °C min−1, 200 to 312 °C at 3 °C min−1, and finally, 312 to 325 °C at 30 °C min−1 (10-min hold). For the ind-PCBs, the GC column temperature program was as follows: initial temperature 120 °C (1.5-min hold), then 120 to 200 °C at 30 °C min−1, 200 to 299 °C at 3 °C min−1, and finally, 299 to 325 °C at 30 °C min−1 (8-min hold). PCBs were detected and quantified using the isotope dilution method and a DFS high-resolution mass spectrometer (HRMS) (Thermo Fisher Scientific, USA) with positive electron ionization at electron energy of 47 eV. The selected ion monitoring mode at a resolution of >10,000 (10 % valley) was used for obtaining the responses of two most abundant molecular ions of PCB analytes. The GC/HRMS records were evaluated using the TargetQuan 3.2 software (Thermo Fisher Scientific, USA).

Quality assurance/quality control

An in-house reference material (pork lard spiked with six native ind-PCBs (28, 52, 101, 138, 153, and 180) and 12 native dl-PCBs (77, 81, 126, 169, 105, 114, 118, 123, 156, 157, 167, and 189) and a procedural blank were used in the pine needle analysis to assess the efficacy of the extraction method and potential contamination during sample processing in the lab (extraction, clean-up, and fractionation), respectively. As with the samples, IS was added to an aliquot (∼5 g) of reference material before its clean-up. Analytes were identified and confirmed based on retention time relative to internal standards, as well as quantification and confirmation ion ratio (±15 % of theoretical values). The limit of quantification (LOQ) was based on a signal-to-noise ratio of 8 for the quantification ion signal. Internal standard recoveries for pine needle samples ranged from 1 to 139 % (average 50 %). LOQs and internal standard recoveries for individual samples are listed in the supplementary material (Table S2, SM). For samples with very low internal standard recoveries (<10 %; as marked in Table S2, SM), namely, SK-KR_0.5 and SW-KR_1, the results were excluded from data summaries and data analysis in order to decrease uncertainty in results interpretation. Amount of dl-PCBs and ind-PCBs in the blank were on average <4 % of the amount detected in the samples (Table S3). However, the blank levels of PCB 52 were >20 % in some samples, namely SK-KR_1, CZ-PR_1, CZ-KO_1, SW-AP, and SW-GL_1. Despite this, it was decided that SK-KR_1, CZ-PR_1, CZ-KO_1, and SW-AP would not be removed from data analysis in order to not further decrease the sample size and, because QA/QC for all other PCBs was acceptable. It was also decided that the values in the present study would not be blank corrected given that the blank levels were acceptable for all PCBs apart from PCB 52.

Data analysis

Within the context of the present paper, PCB 118 was included in the ∑dl-PCBs, although this PCB can be classed as both a dl- as well as an ind-PCB. Levels of dl- and ind-PCB presented in the results sections below and any data analysis include lower bound concentrations (LOQ = 0). Toxic equivalents (TEQs) were calculated using middle bound concentrations and WHO 2005 toxic equivalency factors (TEF) for mammals (van den Berg et al. 2006). Statistical analysis (Wilcoxon signed-rank sum (WRS) test for comparing paired samples; α = 0.05) was carried out using Statistica 12 to detect any significant temporal differences in PCB concentrations in pine needles. Principal component analysis (PCA) and hierarchical cluster analysis (HCA) were conducted using SIMCA 14 to aid in source identification and to show potential spatiotemporal patterns between sites. PCA model PCA-X was used with default model options. HCA was applied to a score matrix (T) of observations as recommended in the SIMCA program. For HCA, the default similarity criterion (Ward) and sorting principal (size) were used. The clustering was conducted using the first and second components of the active PCA model.

Results and discussion

Spatiotemporal patterns of PCBs in pine needles

A summary of the PCB levels in pine needles from all sites and separate for the 0.5- and 1-year-old needles is shown in Fig. 1 and Table S4, SM. Ʃdl-PCB concentrations ranged from 27 to 410 ng kg−1 dry weight (dw) and were highest in pine needles from industrial sites in SK and the CZ, especially at CZ-TR, as well as at urban sites CZ-BR and CZ-PR, the latter being a site in the largest city in CZ, Prague. Levels of ind-PCBs ranged from 200 to 5100 ng kg−1 dw, and were (like the dl-PCBs) highest at SK and CZ industrial (with highest levels in CZ-TR) and urban sites CZ-BR and CZ-PR. The dl-PCB levels measured in pine needle samples at background sites and at SW urban sites were lower than those from the SK and CZ industrial and urban sites. Middle-bound TEQ levels ranged from 0.045 to 1.7 ng TEQ kg−1 dw (Fig. 1b and Table S4, SM) and were also highest at industrial sites, particularly at SK-KR and CZ-BA, as well as urban site CZ-PR. Non-ortho substituted PCBs contribute by far the greatest amount (>95 %) to the TEQ in pine needles at all sites (Fig. 1b). Comparison of levels in pine needles was not always possible due to differences in reporting (e.g., some studies report levels on a wet weight basis) or it was not clear which TEF values used to calculate TEQ. For those studies where comparisons could be made, levels of ind-PCBs in pine needles from this study were in the range of those from other studies (Table S5). Yet, the dl-PCBs and TEQ are lower than industrial sites in Germany (Rappolder et al. 2007) (Table S5).

Fig. 1
figure 1

a ΣPCB levels in pine needles from the two sampling campaigns, conducted during autumn (0.5-year-old needles) and spring (1-year-old needles) and b TEQ levels for non-ortho and mono-ortho PCBs. ΣPCBs includes 12 dl-PCBs and 6 ind-PCBs; dl dioxin-like; ind indicator; _0.5 and _1 signifies pine needles collected in autumn (0.5 years old) and spring (1 years old), respectively

Full size image

Levels of dl- and ind-PCB from 0.5- and 1-year-old needles were available for two industrial (CZ-BA and CZ-TR), one urban (SW-LK), and one background site (CZ-KO). Differences in levels of dl- and ind-PCB from 0.5- and 1-year-old needles were significant (WRS test (α = 0.05)) for pine needles from CZ-BA and KO. Levels of all PCB congeners were higher in the 1-year-old needles collected in spring at these sites (Fig. S1, SM), which indicates an accumulation in PCB levels at CZ-BA and CZ-KO over time. The accumulation of PCBs in 1-year-old pine needles may have been enhanced by lower temperatures prior to the spring sampling campaign, i.e., lipophilic organic compounds are known to partition to plants when temperatures were lower (Simonich and Hites 1995). Given that atmospheric levels of PCBs are generally higher in the warmer months of summer and spring (Holoubek et al. 2007), it is unlikely that the higher levels in 1-year-old pine needles is due to an increase in atmospheric levels of PCB. There was no significant difference found between TEQ of 0.5- and 1 year-old pine needles, using the WRS test (α = 0.05).

PCB congener profiles (contribution of each congener to the sum of dl- and ind-PCBs or (percent contribution to Σ18PCB, where Σ18PCB included 77, 81, 126, 169, 105, 114, 118, 123, 156, 157, 167, 189, 28, 52, 101, 138, 153, 180) are presented in Fig. 2. Overall, ind-PCBs contribute the greatest amount to ∑PCBs in pine needles at all sites with PCBs 28 and 153 dominating among the ind-PCBs. PCB 118, which is also classed as an ind-PCB, dominated amongst the dl-PCBs in the pine needle samples. Aside from PCB 118, PCBs 105 and 156 generally made a greater contribution to the ∑PCBs in the pine needles than other dl-PCBs. However, there were some differences in PCB congener profiles in pine needles samples based on site characterization (i.e., industrial, urban, or background), as well as between countries. For example, highest contributions of most toxic dl-PCBs (PCBs 126 and 169) were in pine needles from industrial site SK-KR. dl-PCBs 126 and 169 contribute 0.65 and 0.19 % at SK-KR compared to <0.23 and <0.067 % at other sites. The contribution of these PCBs to the TEQ at SK-KR, despite being relatively small compared to the other PCBs, was significant due to the relatively high toxicity of these congeners. Pine needles at Swedish urban and background sites were distinguished from those in the CZ by a higher contribution of dl-PCBs 118 and 105 to the ∑PCBs, as well as a lower contribution of 156 to the congener profile and a more ubiquitous contribution of each of the ind-PCBs to the PCB fingerprint. The PCBs congener profiles were not significantly different at any of the sites between 0.5- and 1 year-old pine needles (WRS; α = 0.05).

Fig. 2
figure 2

PCB congener profiles (percent contribution to Σ18PCB) for pine needles from the two sampling campaigns, conducted during autumn (0.5-year-old needles) and spring (1-year-old needles) at a industrial, b urban, and c background sites. dl- and ind-PCBs have been separated onto different scales due to the high contribution of ind-PCBs to the PCB profile. Samples SK-KR_0.5 and SW-KR_1 have been removed due to poor internal standard recoveries

Full size image

Potential sources of PCBs in pine needles

PCBs were intentionally produced in the past (e.g., heat exchange fluids), but are also formed unintentionally from a number of processes (e.g., chemical manufacturing or incomplete combustion of waste) (Breivik et al. 2004; Jansson et al. 2011b; UNEP 2015). PCB mixtures have been phased out since the 1970s and 1980s and are no longer produced. Therefore, most of the PCBs found today in the environment originate from legacy sources (e.g., release from transformers or capacitors still in use, building materials, stored waste, or contaminated soils) or as unintentional by-products of combustion processes (e.g., waste incineration) (Breivik et al. 2002, 2007).

Multivariate analysis was used to aid in tracing potential sources of PCB contamination in pine needles. PCA was undertaken using PCB congener profiles, as well as, site latitude and total level of ∑PCBs. Both (autumn and spring) sampling campaigns were included in the data analysis and lower bound PCB values were used. Needle samples collected at SK-KR_0.5 and SW-KR_1, were excluded from the PCA due to low internal standard recoveries (<10 %; as marked in Table S2, SM). Principal components (PCs) 1, 2, and 3 collectively explain 80 % of the total variance (Table S6). PC 1 explained 36 % of the variation underlying the separation between industrial, urban, and background sites. PC 2 distinguished the needles collected at industrial sites from each other and explains 27 % of the variation. PC 3 explained only an additional 17 % of the variation and did not further distinguish pine needles from different sites. Results of the PCA (the score and loading plots for PCs 1 and 2) are included in Fig. 3a, b, respectively). The HCA dendrogram, which ranks clusters based on size (Fig. 3c), was used to further illustrate the separation of sites based on profiles, ∑PCBs, and latitude. The dendrogram distinguishes needles collected at urban and industrial sites (groups 2 and 3) from those collected at background sites (group 1). The separation of pine needles from the different industrial sites and between SK/CZ and SW was also highlighted.

Fig. 3
figure 3

a Score and b loading plots generated from PCA, as well as c dendrogram from HCA, of PCBs in pine needle samples at each site, over two sampling campaigns

Full size image

The PCA (score plot—Fig. 2a) and HCA (dendrogram—Fig. 2c) showed clear groupings of pine needle sampling according to site category—industry/urban or background. Pine needles from the industrial/urban sites were further split into three main groupings in the PCA and the HCA. In the PCA metallurgical industry, sites in SK and CZ (SK-KR and CZ-BA) were separated from other industrial/urban sites where thermal processes occurred. Other industrial/urban sites included in CZ: CZ-PR, CZ-TR, CZ-HR, CZ-BB, and CZ-BR and in SW: SW-UM and SW-LK. At CZ-PR, CZ-TR, CZ-HR, CZ-BB, and CZ-BR, activities such as municipal waste incineration (MWI), hazardous waste incineration (HWI), power generation (using coal or oil), steel making, and using an electric arc furnace (EAF) were carried out. CZ and SW background sites CZ-KO, SW-AP, and SW-GL are grouped together with SW urban site SW-LK_1 in both the PCA and HCA (Fig. 2a, c).

The contribution of non-ortho, co-planar dl-PCBs 81, 126, and 169 to the congener profile in pine needles were positively correlated (Fig. 3b); along with the congener fractions of mono-ortho dl-PCBs 114 and 157. There also appeared to be a gradient of the non-ortho PCBs based on chlorination state along the horizontal axis of the loading plot (from right to left: 77, 81, 126, and 169). The congener profile contributions of PCBs 28, 52, 105, and 118 in pine needles were also positively correlated to each other; and there was an inverse correlation between the congener contributions of these PCBs and the high chlorinated ind-PCBs (138, 153, and 180) and ∑PCB levels (Fig. 2b).

The high levels of PCBs in needles collected from urban site CZ-PR, which was located near to a MWI and industrial site CZ-TR, which was in the vicinity of a HWI and a power station powered by brown coal affected the PCA grouping for these sites (Fig. 2b—grouped under thermal I)). Despite having similarly high levels of PCB, the congener profiles of PCBs in pine needles at the CZ-PR and CZ-TR were different. PCB 28 contributes more to ∑PCBs in pine needles at CZ-PR, whereas pine needles at CZ-TR had a greater dominance of PCBs 138, 152, and 180. The differences in congener profiles at PR and TR are likely due to different source influences—at PR MWI and at TR HWI and coal-fired power station. Incineration of different wastes (e.g., levels of precursors) at the MWI and HWI may also have influenced the PCB congener profiles in pine needles as has been observed in the case of waste emissions from incineration processes (Abad et al. 2006; Jansson et al. 2011a).

Pine needles from CZ industrial sites CZ-HR, CZ-BB, CZ urban site CZ-BR, as well as SW urban sites SW-UM and SW-LK, were fairly closely grouped together in the score plot with CZ-PR (Fig. 2b), indicating that the PCB congener profiles and hence sources for the needles were analogous. PCB contamination of pine needles at CZ-BR, CZ-BB SW-UM, and SW-LK were therefore also assumed to be mainly influenced by primary thermal sources. The congener profile of dl-PCBs in pine needles from CZ-HR, where steel making with an electric arc furnace (EAF) occurs showed some similarities to those for emissions from steel-making plants with the same processes—PCB 118 dominates, followed by the contribution of 105 (Antunes et al. 2012). There was a boiler house powered by mazut, a heavy fuel oil, located in the vicinity of site CZ-BB but the levels of PCBs at CZ-BB were lowest of all the industrial sites. There was no information on the congener profile of emissions from the burning of mazut although burning of low-grade mineral oils (e.g., shale oil) is not considered a major source of PCB emissions and is known to emit mainly PCB 28 amongst indicator PCBs (Roots 2004). Use of building materials could be another source of PCBs in pine needles from urban sites, because CZ-BR and CZ and CZ-PR are large sites with a high concentration of building and paints manufactured with PCBs using Delor® 106, were was dominated by higher chlorinated congeners, particularly PCBs 138 and 153 were used in CZ (Taniyasu et al. 2003).

Needles from site SK-KR were separated from those collected at other industrial sites by a greater dominance of PCBs 77, 81, 126, and 169 (Fig. 2b—grouped under thermal II). Site SK-KR is in the vicinity of a secondary copper smelter, where raw material used in the smelting process included reclaimed materials (Kovohuty 2007). Copper smelting can emit significant levels of PCBs, particularly secondary copper smelting, and the different raw materials used in secondary copper smelting can affect PCB emissions (Ba et al. 2009; Hu et al. 2013). The dl-PCB congener profile in emissions from a Portuguese copper plant was also dominated by PCBs 126 and 169 in SK-KR (Antunes et al. 2012). Pine needles from site CZ-BA, near to steelworks had the most comparable profile to needles from SK-KR, with slightly higher fractions of non-ortho, co-planar PCBs (particularly PCBs 77 and 126) than other sites. There were some differences between congener profiles in needles at CZ-BA and needles near to another steel-making source, CZ-HR. Apart from the slightly higher contributions of non-ortho co-planar in pine needles at CZ-BA described earlier, there were much higher fractions of ind-PCBs 28 and 52 and lower fractions of PCBs 138, 153, and 180 in pine needles from CZ-BA. Although little is known of the steel-making process at CZ-BA, CZ-HR uses an EAF. There are two main steel-making processes (Jackson et al. 2012), but little information on the PCB patterns/fingerprints in the emissions of these processes. As with CZ-HR, the dl-PCB congener profile in needles at CZ-BA, showed some similarities to those for emissions from steel making plants with EAF—dominance of PCBs 118 and 156 (Antunes et al. 2012). The study by Antunes et al. (2012) also shows that there were some slight differences in the contributions of dl-PCBs in emissions from different plants, as was observed in the present study. To our knowledge, no information was available on ind-PCB fingerprint from steel-making emissions.

Pine needles at CZ and SW urban sites (CZ-PR, CZ-BR, SW-UM, and SW-LK) and background (CZ-KO, SW-GL, and SW-AP) sites are clearly separated in the score plot of PC1 and PC2 as a result of differences in congener profiles. In general, for the background site CZ-KO (B) we observed greater fractions of PCB 118 in pine needles than at CZ-PR (U) and CZ-BR (U). In SW background sites, SW-GL and SW-AP, there was a pattern of decreasing contribution of ind-PCBs from 28 to 180, which was unlike what was observed at urban sites SW-UM and SW-LK. The difference in pine needles in the PCA score plotting of SW background sites SW-AP and SW-KR was due to latitude rather than PCB profile or ∑PCB levels. Pine needles at background sites were most likely influenced by secondary sources of PCBs. PCBs were formally manufactured in Slovakia and used throughout the Czech Republic so for CZ-KO it was theorized that re-volatilization from soils, building materials, and paints was the source of PCBs in pine needles (Dvorska et al. 2008; Holoubek et al. 2007; Taniyasu et al. 2003). Given the distance of SW-GL and SW-AP from known sources of PCBs, LRAT and subsequent atmospheric deposition were assumed to be the source of PCBs at this site (Newton et al. 2014).

A comparison of PCB profiles in pine needles to those in active and passive air sampling measurements

We compared congener profiles of ind-PCBs (i.e., the contribution of each congener to the ∑ind-PCBs or percent contribution to Σind-PCBs, including 28, 52, 101, 138, 153, 180), in pine needles and those from high-volume active air sampler (HVAAS) and passive sampling (polyurethane foam disk passive air sampler (PUF-PAS)). This comparison was done for sites for which there was available data, namely the background sites, (a) Košetice in the CZ (CZ-KO) and (b) Aspvreten in SW (SW-AP), using the samples taken in spring (1-year-old needles). Both CZ-KO and SW-AP were considered to be distant from major PCB point sources. CZ-KO is located at a Czech Hydrometeorological Institute (CHMI) monitoring site within a rural area in the centre of the Czech Republic. Aspvreten is a rural, coastal site approximately 70 km from Stockholm. Active and passive sample measurements methods for these sites are taken as part of EMEP and are described in more detail in Aas and Bohlin Nizzetto (2015). Further detail on the active and passive air sampling methods used at Košetice and Aspvreten, as well as analytical methodology for PCB analysis of these samples are included in Dvorska et al. (2008), Klánová et al. (2008), Liu et al. (2014), and Halse et al. (2011). In brief, PUF-PAS were deployed for 28 days and 3 months at Košetice and Aspvreten, respectively (Dvorska et al. 2008; Halse et al. 2011; Klánová et al. 2008; Liu et al. 2014). The AAS at Košetice and Aspvreten were both high volume but were deployed for different time periods; HVAAS were deployed for 24 h and 1 week at Košetice and Aspvreten, respectively (Bohlin et al. 2014; Dvorska et al. 2008). For the comparison with 1-year-old needles, HVAAS and PUF-PAS measurements were averaged for the year prior to when the pine needle samples were taken.

Ind-PCB congener profiles in HVAAS and PUF-PAS measurements and pine needles are included in Fig. 4. At Košetice, there was a greater contribution of lower chlorinated PCBs 25 and 52 to the average annual congener profile in HVAAS and PUF-PAS measurements than in the 1-year-old pine needles. Thus, higher chlorinated PCBs (e.g., 138 and 180) were less dominant in the average annual congener profiles of HVAAS and PUF-PAS compared to the congener profile of the 1-year-old pine needles from this site. The same was observed in the comparison of profiles for HVAAS, PUF-PAS, and pine needles from Aspvreten, SW. For both Košetice and Aspvreten, the contribution of PCBs 28 and 52 to the congener profile varied throughout the year for HVAAS and PUF-PAS with higher contributions to the fingerprint in winter than in summer. Yet, the difference between the ind-PCB profiles in pine needles and HVAAS and PUF-PAS at Košetice, CZ-KO, and Aspvreten were not significant (p < 0.05) according to the WRS test for comparing paired samples.

Fig. 4
figure 4

PCB congener profiles in pine needles (1-year-old needles), active and passive air samplers, from a Košetice and b Aspvreten. PN pine needle. HVAAS high-volume active air samples. PUF-PAS polyurethane disk passive sampler. In a(i–ii) and b(i–ii) bars on the left represent all available data points for HVAAS and PUF-PAS; bars on the right are the average congener profile over 1 year for HVAAS and PUF-PAS compared to the 1-year-old pine needles (PN) from the spring sampling campaign = CZ-KO_1 and SW-AP_1

Full size image

There has been very limited comparison of pine needles with both active and passive samplers in the past, and this research shows that there is a strong similarity between patterns of PCB levels in passive or active samplers and pine needles (Klánová et al. 2009; Romanič and Krauthacker 2007). Klánová et al. (2009) previously found that patterns in ind-PCB levels in pine needles (P. sylvestris) were comparable to levels measured in HVAAS and PUF-PAS at Košetice in a study to determine whether pine needles could be used to assess long-term trends of semi-volatile organic compounds. Patterns in levels of PCBs have also been found to be comparable in pine needles collected in urban sites in Croatia and active air samples from Zagreb (Romanič and Krauthacker 2007). However, unfortunately, the methods for the study by Klánová et al. (2009) were different to the present study and there is limited information on the methods used by Romanič and Krauthacker (2007). For example, Klánová et al. (2009) compared levels in 3.5-year-old pine needle and mean HVAAS and PAS over 3.5 years and, needles were not ground prior to extraction, as they were for the present study.

It is possible that the reason that there was a greater contribution of lower molecular weight PCBs in HVAAS and PUF-PAS compared to pine needles, observed in the present study, was due to the greater influence of environmental factors on uptake and accumulation of PCBs in pine needles. The PUF-PAS and HVAAS are designed so that passive or active sampling media (PUF disk and PUF plug for PUF-PAS and HVAAS, respectively) are largely protected from wind, rain, and ultraviolet (UV) light. Yet, pine needles are directly exposed to the elements, and hence, PCBs can be photodegraded by UV, with the lower molecular weight PCBs more effectively photodegraded than higher molecular weight PCBs (Barber et al. 2004; Chang et al. 2003). Another factor that may affect the levels of lower chlorinated PCBs in plants is the metabolism of these compounds by the plant (Barber et al. 2003, 2004). However, the fact that the profiles of ind-PCBs are not significantly different between pine needles and conventional methods of air sampling shows that pine needles may be an acceptable indicator of spatiotemporal patterns of air pollution with POPs.

Conclusions

Higher levels of PCBs were detected in pine needles close to both industrial, as well as urban sites in the vicinity of a waste incineration plant and a combination of waste incineration plant and a coal-fired power station. Levels of ind-PCBs are relatively high at these sites compared to most recent studies of PCB levels in pine needles but dl-PCB levels were less than industrial sites in the past, suggesting that these are still important sources of PCBs but in general, levels of PCBs are reducing, possibly due to the implementation of policy measures to reduce their emissions. Differences in profiles of PCBs in pine needles at SW and CZ urban sites most likely reflect the influence of local combustion sources. Results from the present study also show that levels of PCBs in pine needles are affected by distance from known sources with lower levels at sites classed as background than at urban or industrial sites. It has been theorized that LRAT and subsequent deposition is the most likely source of PCBs in pine needles at background sites in Sweden. Profiles of active and passive samplers show similarities. Yet, our comparison of these sampler profiles to those of pine needles indicates that lower chlorinated congeners may be degraded on the pine needle or metabolized by the plant leading to lower contributions of these compounds in pine needles. Given the potential usefulness of pine needles at bi-monitors of atmospheric pollution congruity with PAS and AAS needs to be further investigated with a greater amount of samples than the present study. Despite increasing of PAS in the last decade, which has improved the spatial coverage of atmospheric monitoring sites, there are still countries or regions with no information. Using needles in a standardized way (e.g., consideration of the age of needles to collect and collection across the same season) and interpreting results with caution (taking into account limitations/enhancements to uptake into leaves) can help in filling these gaps.