Heavy Metals in a High Arctic Fiord and Their Introduction with the Wastewater: A Case Study of Adventfjorden-Longyearbyen System, Svalbard

Longyearbyen is the largest settlement on Svalbard archipelago, with 2400 permanent residents and approximately 150,000 tourists visiting every year. The city annually releases approximately 285,000 m3 of untreated wastewater to the nearby Adventfjorden. To date, the environmental impact of this continuous input has been studied mainly regarding the sediments and benthic fauna in the fiord. Here, we present results from a study of raw wastewater entering Adventfjorden as well as heavy metals concentrations in the water column within the fjord itself. Two surveys were carried out in summer and autumn season 2018, to establish physical and chemical properties of water at various locations. Trace elements (V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Hg, As, Cd, Pb, U), total suspended solids (TSS) and total organic carbon (TOC) were measured. Our results show that Longyearbyen’s raw wastewater introduces low concentrations of heavy metals to the fiord, but due to the growing number of inhabitants and tourists, it should be monitored to avoid degradation of Adventfjorden ecosystem


Study Area
The study was conducted in the central part of West Spitsbergen island (Svalbard archipelago) in the vicinity of the largest settlement on the island, Longyearbyen (Figure 1), which discharges untreated wastewater (about 285,000 m 3 /year) to the nearby Adventfjorden. Adventfjorden is a part of the biggest fjord on the west coast of Spitsbergen, Isfjorden. Its wide and deep mouth facilitates the water exchange with the central part of Isfjorden due to tidal pumping and wind-driven surface currents [49].

Sample Analysis
Total organic carbon (TOC) concentration was determined by a TOC-VCSH/CSN Analyser (SHIMADZU, Japan) using the catalytic combustion method with non-dispersive infrared detection (NDIR). Total suspended solids (TSS) were determined in duplicates by the membrane filtration method according to the Standard Methods [52]. The concentrations of metals: arsenic (As), cadmium (Cd), chromium (Cr), cobalt (Co), copper (Cu), iron (Fe), lead (Pb), manganese (Mn), nickel (Ni), uranium (U), vanadium (V) and zinc (Zn) in water and wastewater samples were determined by inductively coupled plasma mass spectrometry, Thermo XSERIES 2 ICP-MS (Thermo Fischer Scientific, Germany). The certified reference material EnviroMAT Ground Water (Low ES-L-2 and High ES-H-2) was used to validate the method. Trace element concentration CVs of the obtained triplicate results ranged from 0.5% to 3.0%. Measurement range (MR), limit of detection (LOD) and limit of quantification (LOQ) are specified in Table 2.  The largest contributors of freshwater into Adventfjorden are two, predominantly glacier-fed rivers: Adventelva and Longyearelva. Thus, they discharge large amount of freshwater (between 0.4 and 0.8 m/year/km 2 , Nowak pers. comm) and suspended solids to the fiord (up to 1 g/L in the melting season, mostly silt) [50,51]. Rivers can be also enriched in trace elements due to the presence of mining sites around Longyearbyen. They reach the maximum flow rate in the summer months, which significantly influences the physicochemical conditions in the internal part of the fiord (near river mouth-points L and A), while in autumn, their inflow ceases.

Sampling
Sampling points in Adventfjorden ( Figure 1) were selected to reflect the inflow of wastewater to the fiord (sampling point P) and the impact of two glacial rivers: Adventelva and Longyearelva (sampling points: A and L, respectively). In addition, two reference points were chosen to include dominant water current in Adventfjorden ( Figure 1): station K (Karlskrona depth) located in the central part of Isfjorden, and station I, located inside Adventfjorden. Station K was assumed not to be impacted by the wastewater release, while station I could receive some of the pollution due to the water circulation in the fiord. Water samples in Adventfjorden were collected with a 10 L Niskin bottle at Water 2020, 12, 794 5 of 16 the surface (0-1 m) and bottom (~5-10 m above sea bed) of the water column. To minimize the risk of contamination, samples were collected using nitrile gloves and bottles were rinsed three times with sampled water and then filled completely without air bubbles. Simultaneously, untreated wastewater was collected prior discharge to the receiver-from the last pumping station on the sewer system (sampling point WW, Figure 1). All the samples were kept frozen until further analysis. Samples were collected twice: in summer (July) and in autumn (October) 2018. Preceding sampling, the average daily temperatures varied around 5-10 • C in July and from −12 • C to 5 • C in October. Precipitation did not exceed 0.4 mm/week prior to both of the sampling campaigns.

Sample Analysis
Total organic carbon (TOC) concentration was determined by a TOC-VCSH/CSN Analyser (SHIMADZU, Japan) using the catalytic combustion method with non-dispersive infrared detection (NDIR). Total suspended solids (TSS) were determined in duplicates by the membrane filtration method according to the Standard Methods [52]. The concentrations of metals: arsenic (As), cadmium (Cd), chromium (Cr), cobalt (Co), copper (Cu), iron (Fe), lead (Pb), manganese (Mn), nickel (Ni), uranium (U), vanadium (V) and zinc (Zn) in water and wastewater samples were determined by inductively coupled plasma mass spectrometry, Thermo XSERIES 2 ICP-MS (Thermo Fischer Scientific, Germany). The certified reference material EnviroMAT Ground Water (Low ES-L-2 and High ES-H-2) was used to validate the method. Trace element concentration CVs of the obtained triplicate results ranged from 0.5% to 3.0%. Measurement range (MR), limit of detection (LOD) and limit of quantification (LOQ) are specified in Table 2. MR-measurement range; LOD-limit of detection; LOQ-limit of quantification.

Data Analysis
To detect pair-wise relationships among the metals, salinity, TOC and TSS concentration in the investigated water samples, Pearson's correlation coefficients (r) were calculated using Excel 2010. Statistical significance of correlation coefficients was assessed at a significance level of p < 0.05. The statistical significance of correlations was verified using the Student's t-test.

Heavy Metals Concentrations in the Wastewater and Adventfjorden
Wastewater generated in Longyearbyen contained higher concentrations of all heavy metals than the recipient, with the exception of U and Cd ( Table 3). The highest concentrations were noted in the case of iron (226 ± 79 µg/L) and manganese (176 ± 62 µg/L) and the lowest were in case of Pb (1.12 ± 0.53 µg/L), U (0.11 ± 0.02 µg/L) and Cd (0.03 ± 0.01 µg/L) ( Table 4). The order of heavy metals abundance in the untreated wastewater collected prior discharge to the Adventfjorden (sampling point WW) was as follows:  The concentrations of heavy metals in Adventfjorden investigated in this study followed the general order: Fe ≥ Zn ≥ Mn ≥ As ≥ V ≥ U ≥ Cu ≥ Ni ≥ Cr ≥ Pb > Co > Cd, which only to some extent follows the order represented by wastewater presented above. Moreover, in this study, the differences in heavy metals concentrations in the marine water were not as pronounced as in the wastewater. None of the metals in the recipient exceeded concentration of 5 µg/L (Figure 2). In the summer season, Pb was below detection limit (<0.10 µg/L). The concentrations of cobalt, nickel, cadmium and iron in summer season were the highest at the Adventelva outflow (sampling point A). The concentrations of vanadium in October were higher than in July and bottom samples presented higher values than surface water samples. On the contrary, manganese concentrations were on average higher in surface than bottom water samples and were the highest for surface samples in July. A similar relation with the depth was noted for Ni, Co, Fe and Zn. Fe levels also increase near the river outlets (points L and A, see Figure 2), in contrast to As. Arsenic presents higher concentrations in the bottom water samples, which is more pronounced in July.
abundance in the untreated wastewater collected prior discharge to the Adventfjorden (sampling point WW) was as follows: Fe > Mn >> Ni ≥ Zn > Cu ≥ As ≥ V ≥ Co ≥ Cr ≥ Pb > U > Cd.
The concentrations of heavy metals in Adventfjorden investigated in this study followed the general order: Fe ≥ Zn ≥ Mn ≥ As ≥ V ≥ U ≥ Cu ≥ Ni ≥ Cr ≥ Pb > Co > Cd, which only to some extent follows the order represented by wastewater presented above. Moreover, in this study, the differences in heavy metals concentrations in the marine water were not as pronounced as in the wastewater. None of the metals in the recipient exceeded concentration of 5 µg/L ( Figure. 2). In the summer season, Pb was below detection limit (<0.10 µg/L). The concentrations of cobalt, nickel, cadmium and iron in summer season were the highest at the Adventelva outflow (sampling point A). The concentrations of vanadium in October were higher than in July and bottom samples presented higher values than surface water samples. On the contrary, manganese concentrations were on average higher in surface than bottom water samples and were the highest for surface samples in July. A similar relation with the depth was noted for Ni, Co, Fe and Zn. Fe levels also increase near the river outlets (points L and A, see Figure. 2), in contrast to As. Arsenic presents higher concentrations in the bottom water samples, which is more pronounced in July.
The difference between heavy metals concentration in wastewater and recipient was the highest for Mn and Fe (over 100 times higher concentration in wastewater than in marine waters), Ni and Co (33 times and 18 times higher, respectively). The smallest difference between sewage and recipient was noted for As and Cd (mean value in wastewater was around 150% of the mean concentration in marine waters). U concentrations were on average higher in Adventfjorden than in the wastewater.  The difference between heavy metals concentration in wastewater and recipient was the highest for Mn and Fe (over 100 times higher concentration in wastewater than in marine waters), Ni and Co (33 times and 18 times higher, respectively). The smallest difference between sewage and recipient was noted for As and Cd (mean value in wastewater was around 150% of the mean concentration in marine waters). U concentrations were on average higher in Adventfjorden than in the wastewater.

TSS, TOC and Correlations with Heavy Metals
In July, TSS concentration in the recipient reached values much higher than in October (up to 835.11 mg/L versus 11.40 mg/L, respectively). The highest values were recorded in July at the Adventelva outlet (point A), followed by points L and P in the inner Adventfiorden basin (53.60-62.2 mg/L in the surface water samples and 16.1-29.5 mg/L in the bottom layer). The lowest TSS was noted for outer stations K and I (2.47-10.8 mg/L). In October, when the river discharge into the fiord decreases, TSS concentration did not exceed 12 mg/L at all stations ( Figure 3).
Water 2020, 12, x; doi: FOR PEER REVIEW www.mdpi.com/journal/water (sampling point WW), than in the recipient (mean of 39 mg/L in wastewater versus mean 0.9 mg/L in marine waters). TOC values in Adventfjorden oscillated from 0.85 ± 0.06 mg/L in October to 0.95 ± 0.28 mg/L in July ( Figure. 3) and the highest values were found and wastewater discharge pipe and at the mouth of Adventelva river in July (sampling points P and A, respectively). In October, no significant correlation was found between TSS and TOC, even though the correlations in July and for the whole set of data were strong (0.69 mg/L and 0.71 mg/L, respectively, see Table 5).  Total organic carbon (TOC) concentrations were around 40 times higher in wastewater samples (sampling point WW), than in the recipient (mean of 39 mg/L in wastewater versus mean 0.9 mg/L in marine waters). TOC values in Adventfjorden oscillated from 0.85 ± 0.06 mg/L in October to 0.95 ± 0.28 mg/L in July ( Figure 3) and the highest values were found and wastewater discharge pipe and at the mouth of Adventelva river in July (sampling points P and A, respectively). In October, no significant correlation was found between TSS and TOC, even though the correlations in July and for the whole set of data were strong (0.69 mg/L and 0.71 mg/L, respectively, see Table 5). Table 5. Values of Pearson correlation coefficient (r) for total concentrations of investigated elements (selected metals nonmetals and total organic carbon concentration). Strong and very strong correlations (0.6 < |r| ≤ 1) are given in bold. Grey background emphasize very strong correlations (0.8 < |r| ≤ 1). Values of Pearson's correlation coefficients in the range (0.6 < |r| ≤ 1) are statistically significant at p < 0.05.

Heavy Metals in the Wastewater
Wastewater generated in Longyearbyen contained higher heavy metals concentrations than the Advenfjorden, thus, it can be considered as a point source of pollution. Apart from nickel and manganese, heavy metals concentrations in wastewater from Longyearbyen were lower than in typical municipal wastewater [25,26,59,61] and wastewater generated at the Antarctic stations [8,9,58], see Table 4. The higher concentrations in the municipal wastewater should be attributed to the industrial contribution, while at Antarctic stations the main explanation may be the lack of dilution, as those settlements usually operate at a minimum water consumption, with a less efficient potable water supply system than in Longyearbyen.
The order of heavy metals abundance in the untreated wastewater (Fe > Mn >> Ni ≥ Zn > Cu ≥ As ≥ V ≥ Co ≥ Cr ≥ Pb > U > Cd) is mostly in agreement with the relative abundance reported by Chipasa [61] and Karvelas [60] in municipal wastewater, apart from Ni and Mn abundances, which are higher in case of Longyearbyen wastewater.
In domestic wastewater, trace elements can originate from toilet paper, detergents, medicines cosmetics and other personal care products [25][26][27]29], see Table 1. Moreover, food leftovers can contain trace elements-note that in Longyearbyen, they are milled in sinks and directed to the sewer. Sörme and Laderkvist [26] indicated household emissions, such as food and artistic paints, as the second largest contributor of Cd in the wastewater. It has been estimated that approximately 100 tonnes of food is discharged to the fjord every year. Food waste constitutes about 0.4 m 3 daily, which can be of a significant importance with regard to local organic load and the supply of certain heavy metals [63].
Plumbing material, especially copper piping, which is also used in the Longyearbyen water supply system, was pointed out by Comber and Gunn [27] and Sörme and Lagerkvist [26] as an additional source of trace elements (mainly Cu, Pb and Zn) in wastewater. Taps, heat exchangers in housing, iron valves in the pumping stations and slightly acidic pH may also contribute to heavy metals load in Longyerbyen water and then wastewater. Indirectly, it is confirmed by a large corrosion on the metal pipes that has been reported in town (Kjersti Olsen Ingerø, Longyearbyen Lokalstyre, pers. comm.).
In Longyearbyen, the drinking water system is supplied by two meltwater and rain fed reservoirs in Gruvedalen and Isdammen watershed. Nowak and Hodson [64] indicated the influence of acid mine drainage on the quality of drinking water sources, especially at the beginning and end of the summer. In spring, heavy metals are delivered by melting snowpack and can be leached from the underlying mine waste rock. At the end of the summer, the surface runoff is lower and the rock-water contact is prolonged, also resulting in higher trace elements concentrations. Heavy rainfalls during the year may deliver contamination from the hillslopes, but also dilute them. In Longyearbyen's drinking water reservoirs (Isdammen and Gruvedalen) concentrations of Mn, Fe, Cr, Co, Cd, Cu, Zn and Ni were noted to be higher than those in the wastewater (Table 3). For example, in Isdammen reservoir over 70% of the analyses in years 2005-2017 revealed manganese concentrations higher than allowed for drinking water [53], indicating the need of pre-treatment. In this case, drinking water may contribute to a significant share of heavy metals in the wastewater, especially Mn, Fe, Cr, Co, Cd, Cu, Zn, Ni, Pb (Table 3).

Heavy Metals in the Adventfjorden
In the summer, the highest concentrations of Co, Ni, Cd and Fe were noted at the Adventelva outflow (sampling point A), which suggests the impact of the river. This is supported by Bazzano et al. [65], who found that natural input from the glaciers is the major source for Co, Fe and V in another fjord system in Svalbard, Kongsfjorden. In case of vanadium, however, this trend in our data cannot be seen. This suggests that other factors rather than river discharge can play a more significant role in the distribution of vanadium concentrations. Manganese and cobalt seem to depend on the sampling depth in the fiord, but also reveal higher concentrations at the river mouth in the summer (similarly to Fe), which suggests the contribution of freshwater to discharge these metals into the fiord. Additionally, the physical mixing of the fresh water with seawater masses influences the concentration of heavy metals, which was confirmed for dissolved arsenic in estuarine waters [66][67][68][69][70]. The increase of As concentrations with the increasing salinity (freshwater versus seawater) observed in our dataset was also recorded in Krka Estuary [71]. The lowest As values in our data was observed at the stations with the greatest freshwater influence in the period when river discharge is the highest. Higher concentrations in the bottom water masses are especially visible in data from the summer (Figure 2) when the stratification in the fiord is the strongest. Separation of water masses slows down mixing and therefore, restricts vertical migration of pollutants that are not bound to solid particles and therefore, are not prone to settling. According to Bazzano et al. [65], Cr, Cu, Mn, Ni, Pb and Zn in the Arctic fiord were mainly introduced with the inflow of Atlantic waters. This may be reflected by the higher concentrations of Cr, Cu and Zn observed at Longyearelva mouth (sampling point L), which is located on the way of general water current in Adventfiorden (see Figure 1).
The sources of heavy metals in the fiord, apart from wastewater discharge, can also be associated with other human activities, which makes tracking of pollution origin in the fiord a complex task. For instance, allochthonous sources of heavy metals Longyearbyen may include ships and airplanes, which may contribute to heavy metals emission into the environment. In addition, old waste dumping sites have been frequently mentioned as a heavy metals pollution source in the vicinity of human settlements in polar regions [8]. In the case of Longyearbyen, the dumpsite at the seashore was closed in 1990 [63] and moved further from the fiord; however, the risk of leaching heavy metals from the old dumping site or coal mining constructions remnants still exists. In the case of Cd, however, there is little evidence that its concentration in the marine waters has been impacted by human activities [3,72], which is in contrary to nickel. Its relatively high concentration in biota around Longyearbyen has been attributed to the mining activity in the area [35]. As mentioned above, some contaminants could be potentially delivered into Adventfjorden via river discharge.

Correlations with TSS and TOC
Total suspended solids values correlated with the Adventelva and Longyearelva inflow, as they introduce a large amount of coarse silt. This relation has also been noted in other studies in the area [50,73]. In contrary to the wastewater, TSS originated from rivers is mainly in mineral form [Kalinowska, unpublished data]. This may be of importance, as heavy metals usually show correlations with the organic matter content [74]. Wastewater is the main contributor of TOC and heavy metals to the recipient. Suspended solids, which are mainly of the glacial river origin, do not contribute strongly to the heavy metals input to the fiord. The influence of the surface runoff or remobilization of heavy metals from the sediments has not, however, been studied.

Conclusions
We found that untreated wastewater discharged by Longyearbyen to the nearby Adventfjorden delivers heavy metals in concentrations similar or lower to typical domestic wastewater, also those from polar scientific stations. Despite the low pollutants concentration, wastewater can be considered a local pollution point source, as it contains slightly higher concentrations of heavy metals that the waters of its marine recipient.
No spatial or seasonal pattern in distribution in the recipient was found for all the elements analyzed. Nevertheless, some patterns, specific for individual metals, were noted: concentrations of manganese, cobalt, nickel, cadmium, iron and arsenic seem to be impacted by river discharge into the fiord, which is the strongest during the summer. Rivers, however, can be influenced by the vicinity of former or still existing mining sites. Chromium, copper and zinc concentrations can partly depend on Atlantic waters inflow into the Advenfjorden. Taking into account the relatively small volume of wastewater introduced to the Adventfjorden-Isfjorden system and dynamic exchange with water outside the fiord, we conclude that the wastewater is diluted and the concentrations of heavy metals in Adventfjorden are influenced by other factors, including water masses dynamics in the fiord, oceanic currents, river inflow and mixing.
Nonetheless, further wastewater monitoring and its impact on the fiord is recommended due to the growing number of Longyearbyen's visitors and inhabitants, resulting in the increase of wastewater volume in the following years. Special attention should also be given to substances of emerging concern, such as pharmaceuticals and personal care products (PPCP), together with antibiotic resistant bacteria (ARBs) and genes (ARGs). Data collected throughout the years may serve as a basis for the local authorities to undertake the appropriate research, monitoring as well as proper prevention and mitigation action.