Partition of pentachlorophenol (PCP) into particulate and dissolved phases in the waters of the Vistula River mouth at the Gulf of Gdańsk

The research was carried out at the Kiezmark (Tczew) station, the Vistula River mouth cross-section, the Swibno station - the Vistula River mouth and in the coastal zone of the Gulf of Gdansk, within the seaside resorts - Gdynia, Sopot, Gdansk (Fig. 1). The Vistula River is the longest river in the Baltic catchment area (1022 km long) with the mean water discharge equal to 1080 m³ s^-1 (Niemirycz 2008). Eighty seven perecent of its drainage area is located in Poland (Niemirycz 2008, 2011). The river sections close to its mouth reflect the pollution caused by discharges from the river drainage area of 194 400 km² (Niemirycz 2008).

Figure 1

Samples of SPM from the Kiezmark station (the Vistula River mouth cross-section), used for the determination of long-term temporal fluctuactions, were collected in 2002-2006 in seasonal cycles (twice per month).

Surface river water (0-5 m) was collected (1 dm³) and filtered through the Whatman GF/F 0.45 μm membranes (47 mm filter diameter). Prior to use in the analysis, membranes were combusted at 550°C in a furnace to eliminate any traces of organic contamination. SPM concentration was calulated gravimetrically.

River (Kiezmark) and sea water (Gdansk, Sopot, Gdynia, Swibno) was collected from March to December 2014 with the Niskin bottle from the surface layer (0-5 m). The sample collected in March form the Kiezmark station had to be excluded due to the damage during transportation.

The 2 dm³ volume of a sample was divided into two replicates to determine the concentration of SPM in a sample and PCP in either dissolved (PCP_diss) or particulate phase (PCP_spm), after filtration, GF/F 0.45 μm. The dissolved fraction was analyzed within 24 h, while filters with SPM were deep-frozen at 28°C, then freez-dried and analyzed after 48 h. Concentration of SPM was measured gravimetrically.

Basic parameters such as salinity, temperature, oxygen saturation and pH were determined in the whole sample using a multifunctional measurement device (EC410 ExStik II Conductivity/TDS/Salinity Kit). Chemical parameters: nitrates (NO₃^-) and ortho-phosphates (PO₄^3-) were determined using spectrophotometric methods in order to examine a possible impact of primary production on PCP dissociation or adsorption on SPM.

Data (daily mean) on the water flow rate (Q) from the Kiezmark station from 2002-2006 and 2014, used as an additional parameter, were obtained from the Institute of Meteorology and Water Management in Warsaw, Poland.

Number of samples and characteristics of the collected environmental material are given in the Table 1.

In order to determine long-term changes (annual, seasonal) in the concentration of SPM in the Vistula River, the collected data were related to a mean value of the water flow rate (Q) [m³ s^-1]. Based on the collected data, seasonal differences in the water flow rate and the SPM concentration were confirmed by the statistical ANOVA Kruskall-Walis (K-W) test (p<0.05). No statistical differences were observed during the study period (2002-2006, 2014) regardless of the season.

For further interpretation of the collected data, the load of SPM discharged into the Gulf of Gdansk by the Vistula River was calculated according to the equation presented below (Niemirycz 2008).

where:

L_y – Q_i load (kg year^-1) m – a unit coefficient n - the number of measurements C_i – instantaneous concentration (g m^-3)

Qi – instantaneous water flow (per second corresponding to the concentration C_{i (m}³ _s^-1₎

The estimated SPM load significantly correlated with the mean water flow rate (Q) in the Vistula River, resulting in the following regression equation with the satisfactory and statistically significant correlation coefficient (r=0.7982, p<0.05), (Fig. 2). This fact confirms the thesis on the relationship between the SPM load and the water discharge rate observed by others (Jansson 1982; Walling & Amos 1999; Walling 2000). However, this phenomenon may differ in rivers depending on the catchment soil type, climate conditions, the existence of upstream lakes, which would act as sediment traps, land-use, vegetation, etc. (Håkanson et al. 2004).

Figure 2

Pentachlorophenol was identified in all SPM samples collected in 2014. The Predicted No-Effect Concentration (PNEC) value established for PCP adsorbed on SPM was exceeded in all samples by more than an order of magnitude (Muir & Eduljee 1999).

Values of PCP concentrations in both fraction in river (Kiezmark), estuarine (Swibno) and sea (Gdynia, Sopot, Gdansk) water are presented below (Table 2).

Additionally, statistics on river (Kiezmark), estuarine (Swibno) and sea (Gdynia, Sopot, Gdansk) water parameters were presented, as well as SPM concentration, which was measured in order to interpret the results.

The analysis of PCPspm from the Kiezmark station revealed that the concentration of PCP adsorbed on SPM is not correlated with the sorbent (SPM) measured at this station, but the statistically significant and negative relationship (r=-0.616667) was observed for the particulate phase of PCP (PCP_spm) and the mean water flow rate (Q) at the same station in 2014 (Fig. 3, Table 3). This fact may indicate that the hydrological characteristics, dynamics of the Vistula River appear to significantly affect the partitioning of PCP into dissolved and particulate phases (Fig. 3). Since the analysis of the SPM concentration (mg dm^-3) and the mean water flow rate (Q) revealed a significant positive, strong correlation, the concentration of PCP does not appear to be related to the SPM concentration. However, the Spearman rank correlation test revealed a significant (p<0.05) negative correlation between PCPspm and the mean water flow rate (Q), thus the lowest values of this form at the Kiezmark station were observed during the period of the highest water flow rate (April). This phenomenon may indicate that there are different processes or factors affecting the sorption of PCP on SPM in the water of the Vistula River.

Figure 3

The ANOVA Kruskal-Wallis (K-W) test revealed significant (p<0.05) differences between the concentrations of PCPdiss at selected stations during 2014. No significant differences were reported in PCP in the particulate fraction (PCPspm) applying the same statistical test.

Given that the sampling stations and the material collected (riverine, estuarine and marine) may have different properties, the U-Mann-Whitney test was carried to verify the statistical differences between the mean values of PCPdiss and PCP_spm in all water types. The test proved significant differences between the analyzed water types (p<0.05).

Statistically significant seasonal differences in the occurrence of pentachlorophenol, also in the dissolved phase were observed. These fluctuations could possibly be related to the unstable form of PCP in the dissolved phase due to its varying solubility in water depending on temperature and redox conditions (Andersen et al. 2005). The distribution of PCP between both phases in water is apparently controlled by the complex physicochemical properties of the compound, such as solubility, vapor pressure and hydrophobicity. Due to the fact that hydrophobicity expressed by the K_ow (octanol-water) coefficient increases with a decrease in pH, the redox conditions may be responsible for the reactions of PCP sorption on SPM and fluctuations in partitioning of this chemical.

Concentrations of PCP_spm tend to increase during summer months until September to decrease in the autumn (Fig. 4), which could possibly be explained by the increase in the total suspended particulate matter concentration, observed during summer months in the Gulf of Gdansk (Krężel & Cyberski 1983). However, no significant correlation between these two parameters was observed to confirm this thesis.

Figure 4

The episode with the highest PCP_diss concentration was recorded in the Vistula River (Kiezmark) and in Swibno in September (120.45 ng dm^-3, 94.21 ng dm^-3, respectively), the highest concentration of PCPspm was observed in the sea water station-Gdynia in April. In order to verify the potential cause of this fact, a hypothesis proposed by Schwarzenbach et al. (1993) was tested, which assumes that fluctuations in salinity affect the partitioning of hydrophobic substances into particulate and dissolved forms.

Partitioning is expected to increase with salinity, primarily due to the fact that solubility of organic substances decreases with increased salinity. The significant negative correlation between PCP_diss and salinity determined in this study confirms the presented hypothesis (Table 4). The correlation between PCP_spm and salinity revealed a contrasting (positive) trend which is likely to be related to the increased association of pollutants with suspended solids in the transition from the river to marine environment (Zhou et al. 1999). Additionally, a significant correlation between PCP in both phases and pH of the collected water was observed, confirming that the solubility of PCP increases with pH of water and sorption of PCP on SPM is enhanced while the hydrophobicity of the compound is intensified by a reduction in this parameter (Table 4).

The studies also demonstrated that approximately 80% of PCP is associated with the particulate fraction in water and only about 10% of PCP dissolved in water. There were no changes in this distribution during the year.

Analysis of two PCP forms independently (PCPdiss, PCPspm) does not give a complete picture of seasonal variations of this substance, contrary to the analysis of both forms together. This suggests the probability of an unstable PCP molecule bound to SPM, and thereby the continuous reaction of sorption and desorption (Czaplicka 2004), related to various environmental conditions.

Based on the obtained results, the statistical Spearman’s rank correlation test (p<0.05) was carried out for the variables. The test showed significant correlations for PCP_diss (NO₃^-, PO₄^3-, pH, T). This may suggest a relationship between PCP and the parameters of the water, while in the case of the particulate form, its sorption is more likely related to the properties of the sorbent (SPM) such as the content of organic matter, nitrogen, phosphorus or the presence of various functional groups (Pu & Cutright 2006; Luo et al. 2011).

In order to identify the substances that could represent a relevant risk to the environment, an environmental risk evaluation was carried out using two different approaches: first of all (PCP_diss + PCP_spm), the concentrations were compared with the Environmental Quality Standards (EQS). The purpose of this comparison was to present the threat to the aquatic environment caused by the release of contaminants. This approach was applied to substances having an EQS established by the European Commission in the Directive 2008/105/ EC (EC 2001), referring to the concentration in the whole water sample (at 0.45 μm).

In accordance with the objective of the study, the analyses were performed on both the dissolved and particulate phases to assess the partition of contaminants. As there are no available EQS for both fractions, PCPdiss + PCPspm concentrations were compared to EQS. This approach gives an indicative rate of dilution necessary in the environment to prevent any pollution impact. The dilution factor (DF) was therefore calculated as follows:

However, this method needs to be considered with some caution when one of the PCP fraction values is below LOQ. The dilution factor established for PCP within the Water Frame Directive in the rangeof DF < 0 ≤ 1 and the EQS value included in the groudwater and surface water assessment criteria is 100 ng dm^-3.

The table below presents the data, including mean values of the total PCP in both fractions, the EQS value established by the EU Commission for PCP in surface waters and the calculated DF. The calculated values of DF exceed the required level of PCP dilution in water which would prevent any environmental risk and it remains similar in both water types (Table 5).

The results on PCP in sea (the Gulf of Gdansk), estuarine (Swibno-the Vistula River mouth) and river water (Kiezmark - the Vistula River cross section) obtained during this study are similar to those obtained, respectively, in the surface waters of the Gulf of Gdansk (Kot-Wasik et al. 2004), the Elbe estuary, Germany (Penta Task Force 1997), the north-east coast of England (Muir & Eduljee 1999). The maximum concentration of PCPdiss determined in the estuarine water from the mouth of the Vistula (Swibno) is almost four times lower than the maximum concentration of this substance in the mouth of the Elbe River (410 ng dm^-3), but almost two times higher than the level indicated in the water of the Rhine River (Germany), where the maximum concentration does not exceed 50 ng dm^-3 (ICRP 1991).

The average annual load of PCP delivered to the Gulf of Gdansk by the Vistula River in 2014 was estimated using the results of PCP determined in the water on the basis of the equation (Equation 1) approved by HELCOM and proposed by Niemirycz (2008). The calculated load of PCP_diss and PCP_spm is equal to 64 kg and 131 kg, respectively, which is similar to loads estimated by Niemirycz (2008) for other organic pollutants (ΣDDT-230 kg year^-1, ΣPCBs - 260 kg year^-1, ɣ-HCH-380 kg year^-1), discharged into the Gulf of Gdansk by the Vistula River.