The emergency discharge of sewage to the Bay of Gdańsk as a source of bacterial enrichment in coastal air

The purpose of this research was to study the presence of potential pathogenic bacteria in the seawater and air in five coastal towns (Hel, Puck, Gdynia, Sopot, Gdańsk-Brzeźno) as well as the enrichment of bacteria from the seawater into the coastal air after an emergency discharge of sewage into the Bay of Gdańsk. A total of 594 samples of air and seawater were collected in the coastal zone between spring and summer (between 2014 and 2018). Air samples were collected using the impact method with a SAS Super ISO 100. The multivariate analysis, conducted using contingency tables, showed a statistically significant variation between the concentration of coliforms, psychrophilic and mesophilic bacteria in the seawater microlayer and air in 2018, after an emergency discharge of sewage into the Bay of Gdańsk, compared to 2014–2017. Moreover, we detected a marine aerosol enrichment in psychrophilic, mesophilic bacteria, coliforms and Escherichia coli. We also showed a statistically significant relationship between the total concentration of bacteria and humidity, air temperature, speed and wind direction. This increased concentration of bacteria in the seawater and coastal air, and the high factor of air enrichment with bacteria maybe associated with the emergency discharge of wastewater into the Bay of Gdańsk. Therefore, it is suggested that in the event of a malfunction of a sewage treatment plant, as well as after floods or sudden rainfall, the public should be informed about the sanitary and epidemiological status of the coastal waters and be recommended to limit their use of coastal leisure areas.


Results
Analysis of bacteria in the sea-surface microlayer in the 5 coastal towns of the Bay of Gdańsk in 2018 and during 2014-2017. The concentration of psychrophilic and mesophilic bacteria in the seasurface microlayer samples was higher in 2018 than in 2014-2017 at the seaside towns of Hel, Gdynia, Sopot and Gdańsk-Brzeźno. Using the non-parametric Friedman rank sum test, we demonstrated a statistically significant difference between the mean concentration of psychrophilic and mesophilic bacteria in the 5 seaside towns before and after the discharge (χ 2 = 5, df = 1, p = 0.025). The mean, median, and standard deviation values for the concentration of psychrophilic and mesophilic bacteria are presented in Table 1.
The results during 2014-2017 and in 2018 showed that in Hel, Sopot and Gdańsk-Brzeźno the mean concentration of P. aeruginosa and S. aureus was very close to the expected value, according to the use of the independence test. The result was statistically significant for the mean concentration of P. aeruginosa (χ 2 = 12.7, df = 4, p = 0.015). However, no statistical significance was demonstrated when comparing the average concentration of S. aureus bacteria in the water in 2014-2017 vs 2018 (χ 2 = 0.778, df = 4, p = 0.941) (Fig. 1E, F).  Table 4.

Analysis of bacteria in
The concentration of coliforms in the sea spray aerosol (SSA) samples after an emergency discharge of raw sewagewas higher in 2018 than in 2014-2017 at the seaside towns of Hel, Gdynia, Sopot and Gdańsk-Brzeźno. The non-parametric Friedman rank sum test showed that the results were statistically significant (χ 2 = 5, df = 1, p = 0.025).
Puck 3.9 × 10 1 ± 8; 3.5 × 10 1 ± 1.1 × 10 1 1 3 ± 5; 0 (0-8) 0.00 3 Gdynia 2 × 10 1 ± 1 × 10 1 1 ± 1 20 0.00 0.00 - Gd.-Brzeźno a 6.7 × 10 1 ± 6.8 × 10 1 ; 5.   An attempt was made at the next stage of our research to assess the impact of temperature, humidity, wind speed and wind direction on the presence of bacteria in the air of the coastal towns of Hel, Puck, Gdynia, Sopot and Gdańsk-Brzeźno, after an emergency discharge of sewage into the Bay of Gdańsk. There was a positive relationship between airborne bacteria (psychrophilic andmesophilic bacteria, coliforms, E. coli, S. aureus) and the air temperature. It is possible that UV radiation can quickly destroy airborne bacteria. However, the outdoor atmospheric bacteria can tolerate extreme sunlight and moderately high temperatures due to their spore form and pigments [40][41][42] . According to Aller et al. (2005) some bacteria and viruses are likely embedded in transparent gel-like organic particles that can provide some degree of physical protection against UV radiation and drying 43 .
Moreover, there wasa statistically significant relationship between relative humidity and the concentration of psychrophilic and mesophilic bacteria in coastal towns. Recent studies of bioaerosols by other authors have also shown that the relative humidity of the air can affect the species composition of microorganisms and their survival in the air 44,45 . Brągoszewska and Pastuszka showed thathigh relative humidity may result in cell clumping, which possibly increases the odds of microorganism survival 45 . Our study showed a statistically significant, positive relation between a higher concentration of psychrophilic and mesophilic bacteria, coliforms and S. aureus and the wind blowing from the north and north-east (i.e. from the sea towards the land). In microbiological air tests conducted by Montero et al. (2016) along the quay in Flushing Bay, Queens, New York (USA), correlations were found between the concentration of microorganisms in the coastal air and the wind direction. The total concentration of bacteria detected in air samples taken along the coast was higher when the wind was blowing from the sea. When the wind was blowing from the land, a higher concentration of fungal mould spores was detected 46,47 . Our study indicates a statistically significant relationship between the wind speed and the concentration of psychrophilic and mesophilic bacteria after an emergency discharge of sewage into the Bay of Gdańsk. The results of our team's research have shown for the first time ever that air velocity plays a key role in bacterial emissions from marine coastal water of the Bay of Gdańsk to coastal air [48][49][50] . What's more, it has also been shown that the process of breaking wind waves is also influenced by wind speed, which causes the formation of bubbles in seawater 48,51 . The bubbles selectively accumulate hydrophobic matter and microorganisms, which are transported towards the water surface and are partially emitted into the air [ [51][52][53][54][55]. For example Uetake et al. showed that the wind speed was positively correlated with wave height in Tokyo Bay. High wind speeds above the bay's surface correlated with greater aerosol production by the bubble bursting process during breaking waves, especially on the shore 55 . Other researchers found thatthe viable and total bacterial concentrations was significantly higher when the wind speed was greater than 5.4 m s −1 (exceeding a Beaufort force of 3, at which wave breaking can occur) 53  And finally, our study showed a clear tendency of SSAenrichment with psychrophilic, mesophilic bacteria, and E. coli and P. aeruginosa in Sopot and Gdańsk-Brzeźno. The high, although not statistically significant, factor of air enrichment with coliforms was detected in 2018 in Hel, Gdynia, Gdańsk-Brzeźno and Sopot. The high factor of air enrichment with mesophilic bacteria, coliforms, E. coli, P. aeruginosa or S. aureus may be associated with the emergency discharge of wastewater into the Motława River flowing into the Bay of Gdańsk. Therefore, it is suggested that in the event of a malfunction of a sewage treatment plant, as well as after floods or sudden rainfall, the public should be informed about the sanitary and epidemiological status of the coastal waters and be recommended to limit their use of coastal leisure areas.
In our previous studies, at the mouth of the Vistula River, we showed a 12-fold greater enrichment rate in SSA for mesophilic, potentially pathogenic bacteria compared to psychrophilic bacteria 48 . We also provided evidence that oxygen supersaturation in the surface water may contribute to enhanced bubble-mediated seato-air bacteria transport , in particular during the presence of a summer phytoplankton bloom in the Gulf of Gdańsk 48 . Blanchard and Syzdek (1970) were the first to study the phenomenon of enrichment of marine aerosols drops with bacteria. The authors demonstrated that air bubbles breaking at the air-water interface can remove www.nature.com/scientificreports/ bacteria that concentrate in the surface microlayer and eject them into the atmosphere. The bacterial concentrations in the drops ejected from the bubbles may, depending on the drop size, be from 10 to 1000 times that of the water in which the bubbles burst 54 . In turn, other researchers showed 15-25 fold enrichment in bacteria and viruses during transport from SML into the atmosphere. The data support the idea that the SML is a major source of microorganisms entering the atmosphere from water bodies 43 . It should be kept in mind that studying aerosols in nature is extremely difficult as, confounding factors such as ocean and atmospheric circulation patterns prevent convolution of terrestrial and marine sources of airborne microbes 38,56,57 . The mechanism is based on the mutual attraction of negatively charged bacterial cells by cationic vortexes formed below the bubbles. The scavenging of microorganisms in seawater depends on the bubble' size, the value of the negative charge of anions accumulated on the outer membranes and the charge of the cations accumulated in the vortex under the bubble 52,58 . Further research is needed, both in the laboratory and marine waters, to understand the "transport" of microorganisms in detail.

Study limitations.
The present study has some limitations that we would like to address. First, the studies presented were based on the assessment of the concentration of bacteria grown on microbiological media (described in the Materials and methods section). Not all live bacteria may grow on the microbiological plates, so the concentration of live bacteria in the surface microlayer and sea spray may have been greatly underestimated. Second, most studies on the microorganisms in the sea-surface microlayer have been conducted in the open ocean and coastal water, but studies from the Bay of Gdańsk after an emergency discharge of raw sewage are not available.

Materials and methods
Collection of aerosol and seawater samples. In the years 2014-2018, a total of 408 air samples and 186 sea-surface microlayersamples were collected in 5 coastal towns (Hel, Puck, Gdynia, Sopot and Gdańsk-Brzeźno) in the Bay of Gdansk. In the years 2014-2017, the air samples were collected between the 22nd of May and 22nd of July, every 20 days between 9:00 a.m. and 2:00 p.m. In 2018, after an emergency disposal of raw sewage, the air samples were collected between the 22nd of May and 22nd of July, every 11 days between 9:00 a.m. and 2:00 p.m. Air and water samples were not collected under rainy conditions. The air samples were collected for 10 min by impaction with a SAS Super ISO 100 (Milan, Italy) sampler. The nozzle of the sampler was positioned perpendicularly to the wind direction. The sampler automatically collected 100-L samples of air. The microorganisms passed through small holes in the sampler, directly on to Petri dishes containing an agar medium appropriate for each type of organism. The maximum efficiency of collection is for particulate matter with a d50 = 2-4 μm. The flow rate is 90 lpm. All removable parts of the air sampler were sterilized by autoclaving before sampling and the sterilized sampler head was cleaned between samples with a 70% ethanol solution 17 . The samples of seawater were collected from the microlayer (SML) ≤ 100 μm with a sterile 3 mm thick glass plate sized 50 cm × 50 cm. The plate was immersed in the seawater at an angle of 45 degrees and once the surface of the water was stable the plate was pulled out with a vigorous movement. Water from both sides of the plate was removed with a rubber wiper into sterile glass bottles 59,60 . The samples of water from the sea microlayer were stored in a cooling container at 4 °C and delivered to the laboratory within 4 h from collection for further analysis 36 .
Microbiological analysis of bioaerosols. where Pr is the revised colony in stage, N is the concentration of sieve pores, and r is the concentration of viable particles counted on the agar plate. The concentration of bacterial colonies (CFU/m 3 ) was calculated using the following Eq. (2) where C -airborne bacteria concentration; CFU -colony-forming unit; T -total colonies after application of the Pr statistical correction; t -sampling time and F -airflow rate. Enumeration of bacteria in the air was conducted according to the Polish Standard (PN-89 Z-04111/02-Air purity protection, Microbiological testing, Determination concentration of bacteria in the atmospheric air (emission) with sampling by aspiration and sedimentation method 1989) 62 . We also used theNIOSH Manual  Model evaluation of bacterial emission from SML to atmospheric air. In order to assess bacterial emission from seawater to atmospheric air we calculated the enrichment factor (EF) according to the following formula 72,73 : where L A -oncentration of bacteria in sea aerosols (CFU/1 ml), L W -concentration of bacteria in the sea surface microlayer in the Bay of Gdańsk (CFU/1 ml). The concentration of bacteria in marine aerosol (L A ) was calculated from the formula: where L-concentration of bacteria in the air (CFU/m 3 ), A-volume of sea derived droplets in air for a given wind speed u(10), A = exp((-2.62 + 0.59 u(10)) 0.142857 (ml/m 3 ). Where: 0.142857 = (1000/7)10-3, 7‰ is the mean salinity of the Baltic Sea within (7-8‰), the value of 7 ‰ salinity fits to the range of calculations considered in our model, while 10-3 results from the conversion of µg/ ml, assuming the density of water at 1 g/ml. u(10) is the wind speed in m/s measured 10 m above the surface of water, under the conditions of a wind fetch over the sea of at least 5-10 km.

Statistical analysis.
A Pearson's chi-squared test for independence was performed to assess significant differences between the categorical variable: for one "sample collection years" (two subgroups; 2014-2017 and 2018), and for the second categorical variable:"place" (five subgroups: Hel, Puck, Gdynia, Sopot and Gdańsk-Brzeźno). The contingency tables obtained this way are presented in the form of association figures with the values of Pearson's residuals and the p value of the independence test 74 . Association plots visualize the table of Pearson residuals: each cell is represented by a rectangle that has a height proportional to the corresponding Pearson residual and width proportional to the square root of the expected value. Thus, the area is proportional to the raw residuals. The rectangles representing each cell in the table are positioned relative to a line representing independence. Cells with an observed frequency greater than expected are shown above the line and cells with an observed frequency lower than expected are shown below the line 75 .
In each of the association figures, the blue colour means that there are more observations than would be expected if the data was random. Negative Pearson residuals (the red colour) means that the cell values were smaller than expected. The grey colour represents the data where the concentrations are close to the expected, i.e., the null hypothesis of the independence test is true. To assess the differences between the mean concentration of bacteria detected in seawater and air in the two time periods, 2014-2017 and in 2018, the non-parametric www.nature.com/scientificreports/ Friedman rank sum test was used. To assess the differences between bacterial aerosol concentrations and meteorological parameters, the Kruskal-Wallis ANOVA non-parametric test was used, and the Spearman correlation was determined. In order to compare the bacterial enrichment of the seawater to air between 2014-2017 and 2018, we used the following tests: a two-sample t-test (where the Shapiro-Wilk test showed that the compared groups are from a normal distribution population) and the Wilcoxon rank sum test with continuity correction (for all other cases). The statistical significance of the differences between the groups was set at p < 0.05. Statistical analysis of the results was carried out using software R (2018) 76 .