Interannual Survey on Polycyclic Aromatic Hydrocarbons (PAHs) in Seawater of North Nanao Bay, Ishikawa, Japan, from 2015 to 2018: Sources, Pathways and Ecological Risk Assessment

To improve the understanding of the emission sources and pathways of polycyclic aromatic hydrocarbons (PAHs) in the coastal environments of remote areas, their particulate and dissolved concentrations were analyzed on a monthly basis from 2015 to 2018 in surface waters of Nanao Bay, Japan. The concentration of the targeted 13 species of PAHs on the United States Environmental Protection Agency (USEPA) priority pollutant list in dissolved and particle phases were separately analyzed by high-performance liquid chromatography (HPLC) coupled to a fluorescence detector. Particulate and dissolved PAHs had average concentrations of 0.72 ng∙L−1 and 0.95 ng∙L−1, respectively. While most of the samples were lower than 1 ng∙L−1, abnormally high levels up to 10 ng∙L−1 were observed in the winter of 2017–2018 for particulate PAHs. Based on the isomer ratios of Flu to Flu plus Pyr, it was possible to determine that the pyrogenic loads were greater than the petrogenic loads in all but four out of 86 samples. The predominant environmental pathway for PAHs in winter was determined to be long-range atmospheric transportation fed by the East Asian winter monsoon, while for the summer, local sources were more relevant. By the risk quotients method, it was determined that PAHs in surface seawater presented a very low risk to marine life during the interannual survey.


Introduction
Polycyclic aromatic hydrocarbons (PAHs) are a family of pollutants formed by the imperfect combustion of fossil fuels and organic matter such as wood and grass; they can also naturally occur in petroleum derivates [1]. Because of their environmental hazards and health risks to humans and marine life [2][3][4][5][6], the US Environmental Protection Agency (USEPA) has listed them as priority pollutants [7,8]. Determining the environmental behavior and distribution of PAHs is important for the assessment, management and protection of wildlife health in various environments. PAHs enter marine ecosystems treatment.
Empore C18 disks (3M Company, St. Paul, MN, USA) were preconditioned by rinsing twice with 50 mL of dichloromethane, following which they were rinsed two times with methanol. The disks were stored in methanol at −20 °C until use. Prior to use in a sample filtration, the disks were rinsed with 100 mL of ultrapure water by being passed through it.

Seawater Sampling and Processing
From 2015 to 2018, surface seawater was sampled on a monthly basis at North Nanao Bay's most western coast (37°11'30.5"N 136°54'29.0"E) ( Figure 1) and analyzed to monitor PAH. To avoid photodegradation and any residual contamination, ~10 L sea water samples were collected and stored in a polished stainless well-closed container. To collect 10 L samples considering time, and thus avoid statistical error due to point emission events, either periodic sub-sampling or continuous filtration can be considered. However, in both cases, it would be necessary to use autosamplers that ensure the preservation of sample quality; that is, to keep samples from bio-and photo-degradation, and to prevent lower-ring PAH volatilization and contamination from plastic components inside the sampler. Therefore, to overcome the relatively large uncertainty associated with single grab-samples, we opted to perform the study in a larger time span.
Samples were passed through a GC-50 glass fiber filter (ADVANTEC, nominal pore size 0.5 μm). After adding a 100 μL aliquot of internal standard mix to the 10 L filtrate, dissolved PAHs were concentrated on an Empore C18 disk by passing the filtrate through it at a flow rate of approximately 200 mL•min −1 . The resulting glass fiber filter and Empore C18 disk were folded in half and dried in a desiccator over silica gel in the dark for more than three days. PAHs concentrated with the resulting glass fiber filter and Empore C18 disk were defined as particulate-phase PAHs and dissolved-phase PAHs, respectively. The amount of particle phase was estimated from the difference between the dryweight of the glass fiber filter before and after filtration. The dried filter membranes were stored at −20 °C until extraction.

Extraction
Concentrated PAHs were extracted from the dried filter membrane by sonicating with two portions of 50 mL dichloromethane. To evaluate the recoveries of particulate PAHs, a 100 μL aliquot of ISTD mix was added into the glass fiber (G50) filter before the extraction process. The fragments Samples were passed through a GC-50 glass fiber filter (ADVANTEC, nominal pore size 0.5 µm). After adding a 100 µL aliquot of internal standard mix to the 10 L filtrate, dissolved PAHs were concentrated on an Empore C18 disk by passing the filtrate through it at a flow rate of approximately 200 mL·min −1 . The resulting glass fiber filter and Empore C18 disk were folded in half and dried in a desiccator over silica gel in the dark for more than three days. PAHs concentrated with the resulting glass fiber filter and Empore C18 disk were defined as particulate-phase PAHs and dissolved-phase PAHs, respectively. The amount of particle phase was estimated from the difference between the dry-weight of the glass fiber filter before and after filtration. The dried filter membranes were stored at −20 • C until extraction.

Extraction
Concentrated PAHs were extracted from the dried filter membrane by sonicating with two portions of 50 mL dichloromethane. To evaluate the recoveries of particulate PAHs, a 100 µL aliquot of ISTD mix was added into the glass fiber (G50) filter before the extraction process. The fragments of membranes were removed by passing the extract through a filter (ADVANTEC No. 6, Japan), and 100 µL of DMSO was added to the eluent in order to avoid the volatilization of PAHs during the later concentration process. Dichloromethane was removed using a rotary evaporator until only the DMSO remained; then, the DMSO fraction was then reconstituted to 1 mL with acetonitrile. After this, the resulting solution was filtered with a membrane disk (HLC-DISK3, pore size 0.45 µm, Kanto Chemical Co., Tokyo, Japan), and an aliquot of the solution was then injected into the HPLC for the analysis of PAHs. The extraction process was similar for dissolved PAHs; however, instead of sonication, simple vacuum filtration was utilized.

HPLC Analysis
The 13 PAHs out of the 16 priority PAHs listed by the USEPA determined in the study were acenaphthene (Ace), fluorene (Fle), anthracene (Ant), fluoranthene (Flu), pyrene ( Chromatographic separation was performed using an Inertsil ODS-P column (diameter 4.6 mm, length 250 mm, 5 µm) kept at 30 • C. The mobile phase was a mixture of acetonitrile and water operated under a gradient elution, starting at 55% and increasing to 99% acetonitrile over 65 min at a flow rate of 1 mL·min −1 . Accordingly, 13 PAHs were characterized by referred to the 16 PAHs in the EPA 610 Mix diluted 104 times with acetonitrile. As there were chromatographic interferences with Nap and Phe, they could not be clearly distinguished on the HPLC spectra and were thus excluded from the final analysis. The concentrations of dissolved PAH and particulate PAH were added to provide the total PAH content (total PAHs) in the marine samples. More detailed conditions were described in our previous paper [15].

Quality Assurance
The limit of detection (LOD) was evaluated based on a signal-to-noise ratio cutoff of 10. It ranged between 9.93 pg·L −1 (Pyr) and 105.96 pg·L −1 (Ace), with the correlations of linear calibration curves over 0.998. Sample blanks, as well as Millipore water blanks, were run with each batch to ensure that there was no contamination between samples. Overall recovery rates for dissolved PAH internal standards were 97.5 ± 9.62% for Phe-d10 and 86.4 ± 5.39% for Pyr-d10. Recovery rates for particulate PAH internal standards were 89.1 ± 3.49% for Phe-d10 and 94.6 ± 4.73% for Pyr-d10. To calculate the loss by extraction procedures, recoveries of Phe-d10 was utilized for Ace, Fle, Ant, and recoveries of Pyr-d10 for Flu, Pyr, BaA, Chr, BbF, BkF, BaP, DBA, BPe, and IDP.

PAH Emission Source Characterization
The isomer ratios of Flu  [20][21][22]. It must be noted that BaA presented considerably higher photo-degradation rates than Chr, which affected the source appointment by the isomer ratios method. Also, since the method was based on statistical analysis, the PAH sources which directly interfered with the studied area may not be properly reflected within the set ranges.

Atmospheric Mass Long Range Transportation
To evaluate the emission origins of atmospheric PAHs, the Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) model developed by the National Oceanic and Atmospheric Administration Air Resources Laboratory, USA, was used to calculate the backward trajectory of air masses reaching the monitoring station (http://www.arl.noaa.gov/HYSPLIT_info.php) [23]. The trajectory frequency over a grid cell was calculated and then normalized by the total number of trajectories during the 32-day backward trajectories at 500 m above sea level.

Ecological Risk Assessment
Risk quotients (RQ) are commonly used to assess the potential ecological risk of given chemicals on aquatic organisms [24][25][26][27]. The risk levels of a given PAH are characterized by its RQ: where [PAH] i is the concentration of the given PAH in each water sample (dissolved + particulate phase) and QV i is the corresponding quality value. Two sets of quality values, the negligible concentrations (NCs) and the maximum permissible concentrations (MPCs), were calculated by Kalf et al. [25]. Therefore, for any given PAH, the values of RQ NCs(i) and RQ MPCs(i) can be defined as follows: where RQ NCs(i) and RQ MPCs(i) represent the quality value of the NCs and MPCs for the given PAH, respectively. Based on RQ NCs and RQ MPCs values, risk assignment parameters were determined [26]. Quality values for 10 of the 16 USEPA priority PAHs were originally reported [25] (Supplementary Materials). For the remaining six, we utilized toxic equivalent factors (TEF) to define the missing quality factors [27]. PAHs with similar TEFs should have similar negligible concentrations (NCs) and maximum permissible concentrations (MPCs).
To determinate the overall risk of the 13 studied PAHs, the RQ values for (13) total PAHs were calculated as defined by Cao et al [28]: It is known that many PAH metabolites such as 6-nitrochrysene, 1,6-dinitropyrene, dibenzo[a,i]pyrene, etc., have TEFs of 10 or higher (relative to BaP) [28]. On the other hand, the presence of humic substances would reduce the biotoxicity of PAHs by adsorbing them [29]. Although these and many other factors were not considered, the method proposed by Cao et al. provided a simple method for the risk assessment based solely on the concentrations and toxic equivalent factors of each PAH, granting valuable information on the corresponding monitoring sites.

General Results
The

Comparison of PAH Values with Other Remote Areas Worldwide
The results of this study were compared with previous data from different parts of the world. The data were selected from recent (2005~) research articles reporting PAH concentrations in surface seawater for both dissolved and particulate phases [14,[30][31][32][33][34]. From the collected data, concentrations of Ace, Fle, Ant, Flu, Pyr, BaA, Chr, BbF, BkF, BaP, BPe, IDP and DBA were added and are presented in Table 1 for comparison with the results presented in this study. The average total PAH concentration at North Nanao Bay was the lowest of the collected data (1.67 ng·L −1 ). High concentrations in and around the Mediterranean Sea (245-605 ng·L −1 ) can be attributed to its enclosed nature and its highly populated coasts. In the case of the East China Sea (79.6 ng·L −1 ), the high energy demand due to its rapid economic growth and highly dense population are responsible for PAH emissions, but dilution and mixing from the strong Kuroshio Current must be expected. The northwestern Japan Sea is a place known for Russian ports linked to oil extraction activities, and therefore higher pollution levels are expected (7.9 ng·L −1 ).

Comparison of PAH Values with Other Remote Areas Worldwide
The results of this study were compared with previous data from different parts of the world. The data were selected from recent (2005~) research articles reporting PAH concentrations in surface seawater for both dissolved and particulate phases [14,[30][31][32][33][34]. From the collected data, concentrations of Ace, Fle, Ant, Flu, Pyr, BaA, Chr, BbF, BkF, BaP, BPe, IDP and DBA were added and are presented in Table 1 for comparison with the results presented in this study. The average total PAH concentration at North Nanao Bay was the lowest of the collected data (1.67 ng•L −1 ). High concentrations in and around the Mediterranean Sea (245-605 ng•L −1 ) can be attributed to its enclosed nature and its highly populated coasts. In the case of the East China Sea (79.6 ng•L −1 ), the high energy demand due to its rapid economic growth and highly dense population are responsible for PAH emissions, but dilution and mixing from the strong Kuroshio Current must be expected. The northwestern Japan Sea is a place known for Russian ports linked to oil extraction activities, and therefore higher pollution levels are expected (7.9 ng•L −1 ).

PAHs Source Determination
As shown in Figure 4,

PAHs Source Determination
As shown in Figure

Environmental Pathways
To identify the distribution pathways of PAHs in the North Nanao Bay, a detailed analysis of the relative distributions of the 13 components in the dissolved and particulate phases was performed. In addition, the temporal variability of each component and their relative concentrations in each phase were analyzed.

Long-Range Transportation of Pyrogenic PAHs
In a previous study [19], seasonal behavior was reported for all nine PAHs, presenting maximums in winter (ca. 300 pg•m −3 ) and minimums in summer (ca.10 pg•m −3 ) for aerosol PAHs monitored at the Noto Peninsula between 2007 to 2014. It was reported that these aerosol PAHs had pyrogenic origins [19], and the seasonality was attributed to long-range transport from Northeast China to Japan in the cold seasons and domestic sources in the warm season [19].
In this study, it was observed that, as a major component, the behavior of four-ring PAHs (Figures 2 and 3) greatly influenced the levels of PAHs in Nanao Bay. Figure 5 shows the variation of the relative abundance of Flu, which is a four-ring PAH, in the dissolved phase. Seasonal variations were apparent in the relative compositions of Flu in dissolved-phase-surface seawater samples, with high values in winters (ca. 40%) and low values in summers (ca. 20%) ( Figure 5), consistent with the previous report on aerosols [19]. Similar seasonal trends, although less clear, can be observed also in

Environmental Pathways
To identify the distribution pathways of PAHs in the North Nanao Bay, a detailed analysis of the relative distributions of the 13 components in the dissolved and particulate phases was performed. In addition, the temporal variability of each component and their relative concentrations in each phase were analyzed.

Long-Range Transportation of Pyrogenic PAHs
In a previous study [19], seasonal behavior was reported for all nine PAHs, presenting maximums in winter (ca. 300 pg·m −3 ) and minimums in summer (ca.10 pg·m −3 ) for aerosol PAHs monitored at the Noto Peninsula between 2007 to 2014. It was reported that these aerosol PAHs had pyrogenic origins [19], and the seasonality was attributed to long-range transport from Northeast China to Japan in the cold seasons and domestic sources in the warm season [19].
In this study, it was observed that, as a major component, the behavior of four-ring PAHs (Figures 2  and 3) greatly influenced the levels of PAHs in Nanao Bay. Figure 5 shows the variation of the relative abundance of Flu, which is a four-ring PAH, in the dissolved phase. Seasonal variations were apparent in the relative compositions of Flu in dissolved-phase-surface seawater samples, with high values in winters (ca. 40%) and low values in summers (ca. 20%) ( Figure 5), consistent with the previous report on aerosols [19]. Similar seasonal trends, although less clear, can be observed also in BbF and BkF. Although no reference is available for the direct derivation of BbF and BkF from Flu, their similar atomic structure and the known chemical reaction pathways for the inflamed growth of PAHs suggest that the seasonality of the relative abundance of the three compounds is linked to pyrogenic events [35,36]. BbF and BkF. Although no reference is available for the direct derivation of BbF and BkF from Flu, their similar atomic structure and the known chemical reaction pathways for the inflamed growth of PAHs suggest that the seasonality of the relative abundance of the three compounds is linked to pyrogenic events [35,36]. In the other hand, relatively high levels of particle PAHs were observed in January (9.06 ng•L −1 ) and February (8. [19,[38][39][40]. At the same time, the lesser influence of dry northern winds could be responsible for the larger snow falls discussed above helping to deposit the long-range transported PAHs. The higher number of high PM10 events registered in Seoul, Korea for the 2017/2018 winters over the same 2015/2016 and 2016/2017 seasons matches the characteristics expected for pyrogenic activities [41]. The pyrolytic sources, together with the high PM10 events, explain the high PAH levels observed solely in the particulate phase for the surface seawater samples of north Nanao Bay. Road dust washout after snow melting should be certainly considered among the local inputs; however, road dust has a petrogenic character, disagreeing with the observed data.
In conclusion, a portion of the pyrogenic emissions from East Eurasia would be transported long-range in the atmosphere to be finally deposited into the coastal aquatic environments of Japan, appearing in the seasonal trends in Flu, BbF and BkF as a percentage of dissolved PAHs, as well as contributing to the abnormally high concentrations of dissolved PAHs after the snowfalls of the winter of 2017-2018. Finally, the fact that yearly total PAH averages remained relatively constant, with 1.  In the other hand, relatively high levels of particle PAHs were observed in January (9.06 ng·L −1 ) and February (8. [19,[38][39][40]. At the same time, the lesser influence of dry northern winds could be responsible for the larger snow falls discussed above helping to deposit the long-range transported PAHs. The higher number of high PM 10 events registered in Seoul, Korea for the 2017/2018 winters over the same 2015/2016 and 2016/2017 seasons matches the characteristics expected for pyrogenic activities [41]. The pyrolytic sources, together with the high PM10 events, explain the high PAH levels observed solely in the particulate phase for the surface seawater samples of north Nanao Bay. Road dust washout after snow melting should be certainly considered among the local inputs; however, road dust has a petrogenic character, disagreeing with the observed data. In conclusion, a portion of the pyrogenic emissions from East Eurasia would be transported long-range in the atmosphere to be finally deposited into the coastal aquatic environments of Japan, appearing in the seasonal trends in Flu, BbF and BkF as a percentage of dissolved PAHs, as well as contributing to the abnormally high concentrations of dissolved PAHs after the snowfalls of the winter of 2017-2018. Finally, the fact that yearly total PAH averages remained relatively constant, with 1.  , it can be seen that gasoline cars' lubricating oils had the greatest influence. Clean oil for gasoline cars was the petrogenic source which presented the closest Pyr/Chr ratio (0.34) to the one observed in this study (0.41). Used oil from gasoline cars was the second (0.52) most relevant petrogenic source. Therefore, the petrogenic source found in our study can be characterized as a mixture of clean and used lubricating oil for gasoline engine vehicles. Since the discussed petrogenic sources are liquid, the lack of correlation with the particulate PAHs is to be expected. , it can be seen that gasoline cars' lubricating oils had the greatest influence. Clean oil for gasoline cars was the petrogenic source which presented the closest Pyr/Chr ratio (0.34) to the one observed in this study (0.41). Used oil from gasoline cars was the second (0.52) most relevant petrogenic source. Therefore, the petrogenic source found in our study can be characterized as a mixture of clean and used lubricating oil for gasoline engine vehicles. Since the discussed petrogenic sources are liquid, the lack of correlation with the particulate PAHs is to be expected.
Therefore, it was possible to determinate that the high PAHs in the dissolved phase in July, October and November 2016, and December 2017 not only had petrogenic origins but also that the source was a mixture of clean and used lubricating oil for gasoline engine vehicles. For the rest of the samples identified as mainly having pyrogenic origins (Section 3.2), it was not possible to establish a more precise identity even of a minor-if any-petrogenic source. Therefore, it was possible to determinate that the high PAHs in the dissolved phase in July, October and November 2016, and December 2017 not only had petrogenic origins but also that the source was a mixture of clean and used lubricating oil for gasoline engine vehicles. For the rest of the samples identified as mainly having pyrogenic origins (Section 3.2), it was not possible to establish a more precise identity even of a minor-if any-petrogenic source.

Ecological Risk
As expected from such low concentrations, most of the 13 PAHs had risk quotients lower than 1, representing a very low risk in North Nanao Bay for the 43    Therefore, it was possible to determinate that the high PAHs in the dissolved phase in July, October and November 2016, and December 2017 not only had petrogenic origins but also that the source was a mixture of clean and used lubricating oil for gasoline engine vehicles. For the rest of the samples identified as mainly having pyrogenic origins (Section 3.2), it was not possible to establish a more precise identity even of a minor-if any-petrogenic source.

Ecological Risk
As expected from such low concentrations, most of the 13 PAHs had risk quotients lower than 1, representing a very low risk in North Nanao Bay for the 43

Ecological Risk
As expected from such low concentrations, most of the 13 PAHs had risk quotients lower than 1, representing a very low risk in North Nanao Bay for the 43 [14,[30][31][32][33][34], it was observed that Phe can be considered to be a major component, especially in the dissolved phase. However, even after subtracting Nap, Phe and Acy from the cited articles in Table 1, the higher-lower relationships of total PAHs are clearly observable. For the environmental risk analysis, Phe RQs would be highly valuable information, but without trustable data, the author does not consider it appropriate to present the calculations.

Conclusions
From 2015 to 2018, particulate and dissolved PAH average concentrations and their relative compositions were analyzed with an interannual monitoring survey performed in surface waters of Nanao Bay, Japan with the aim of understanding the emission sources and environmental pathways of these pollutants in the coastal environments of remote areas. To ensure the sustainability of oyster production in Nanao Bay, the monitoring of PAH levels is of utmost importance. The concentrations of 13 USEPA priority PAHs were measured with a HPLC-fluorescence system with monthly surveys. At the end of the study, the following results were obtained:

1.
Particulate and dissolved PAHs had average concentrations of 0.72 ng·L −1 and 0.95 ng·L −1 , respectively. While most of the samples were Lower than 1 ng·L −1 , abnormally high levels up tõ 10 ng·L −1 were observed in the winter of 2017-2018 for particulate PAHs. In general, dissolved PAHs were 1.3 times greater than particulate PAHs, with the three-ring and four-ring PAHs being the dominant species in dissolved PAHs, in the range of 40%−50% and 45%−55%, while four-ring and five-ring PAH were the major components for particulate PAHs, with ranges of 45%−50% and 20%−30%, respectively. Total PAH concentrations were below 2.0 ng·L −1 for most samples, showing isolated spikes in April, July and November 2016; and December 2017 and January, February 2018. The relative abundance in dissolved PAHs of Flu showed clear seasonal trends, presenting clear highs in winter (ca. 40%) and lows in summer (ca. 20%). Chr and Pyr, in dissolved PAHs, presented an overall high correlation (r 2 = 0.95) for the 43 months of the survey.

2.
Based on the isomer ratios of Flu to Flu plus Pyr, it was possible to determine that pyrogenic sources were greater than petrogenic sources in all but three samples (July, October and November 2016; dissolved PAHs). In addition, from the same isomer ratios of Flu to Flu plus Pyr, it was also visible that petrogenic sources were slightly more relevant in the particulate PAHs than in dissolved PAHs.

3.
The seasonal trends in Flu as a percentage of dissolved PAHs, as well as the abnormally high concentrations of dissolved PAHs in the winter of 2017-2018, indicate that a portion of the pyrogenic emissions from East Eurasia are transported long-range in the atmosphere to be finally deposited into the coastal aquatic environments of Japan. In addition, it was determined that the high PAHs in the dissolved phase in July, October and November 2016 and December 2017 have petrogenic origins. Such petrogenic sources were further identified as a mixture of clean and used lubricating oil for gasoline engine vehicles.

4.
Total PAH concentrations in surface seawater varied from 0.5 ng·L −1 to 10 ng·L −1 in the monthly sampling from 2015 to 2018, representing a very low risk to marine life during the 3 years of the interannual survey.