Characteristic and Relative Environmental Risk of Disinfection by Products Associated with Simple Glucose or Naturally Occurring Algal Organic Matter as Tested in Ballast Water Treatment System

Abstract

To prevent the invasion of alien species, the International Maritime Organization and the United States Costal Guard require that a ballast water management system (BWMS) be installed on ships to treat the ballast water before discharging it. BWMS technologies use active substances, which create disinfection by-products (DBPs) during ballast water treatment. This study compared the characteristics of DBPs generated in the treatment of commercial glucose with those of algal organic matter (AOM) derived from field-collected phytoplankton using NaOCl as the active substance. During the treatment of AOM, a greater variety and higher concentrations of DBPs were generated than for glucose. For AOM in freshwater, bromoform and dibromoacetic acid were dominant because of the bromine ions present in the phytoplankton. During the treatment of glucose, the ratio of the predicted environmental concentration to the predicted no-effect concentration of dibromoacetonitrile and chloropicrin exceeded 1, indicating a potential environmental risk. Whole effluent toxicity (WET) testing showed that the chronic toxicity of phytoplankton and the total DBP concentration were highest in marine water with AOM. In addition, the results of WET testing suggested that the concentrations of haloacetic acids (HAAs) and haloacetonitriles (HANs) were important indicators for the evaluation of environmental risk. Therefore, to evaluate the risk of DBPs in international ports where phytoplankton outbreaks frequently occur, it is important to monitor not only the total DBP concentration but also the total HAA and HAN concentrations.

1. Introduction

The International Maritime Organization (IMO) and the United States Coast Guard (USCG) regulate the discharge of treated ballast water by mandating the use of a ballast water management system (BWMS) on international ships to prevent the invasion of foreign species through the ballast water [1,2]. To install a BWMS on a ship, government type approval must be obtained and, for this purpose, land-based and shipboard tests must be performed. Land-based tests are conducted in fresh, brackish, and marine water in accordance with the BWMS Code and Environmental Technology Verification (ETV) protocol, and the discharge requirements (Regulation D-2) must be satisfied five times consecutively for each level of salinity [3,4].

BWMS technologies used in BWMSs include electrolysis, ultraviolet (UV) irradiation, ozonation, filtration, and thermal treatment. As of November 2019, a total of 45 BWMSs had received final approval from the IMO. Approximately 85% of these approved BWMSs employ active substances for treatment such as NaOCl, NADCC, and other oxidizing agents [5]. Because treatment technologies that employ active substances generate disinfection by-products (DBPs), risk assessment must be performed [6]. The IMO outlines the test method for assessing this risk in accordance with the Procedure G9, while the USCG requires the use of the whole effluent toxicity (WET) method outlined by the US Environmental Protection Agency (EPA) [4,7].

Past research has shown that physicochemical factors such as the water temperature, pH, and contact time strongly influence the formation of DBPs in BWMSs that employ active substances and seawater desalination [8,9,10,11,12]. The concentration and properties (e.g., molecular size, molecular bonding form, etc.) of dissolved organic matter (DOM) in the ballast water also affect the type and concentrations of DBPs that are generated [11,13]. To meet the DOM concentration standards for the test water used in land-based tests, test organizations typically use commercially available glucose, lignin, cellulose, and methylcellulose [10,13].

Algal organic matter (AOM) originating from phytoplankton has been reported to play an important role in the dissolved organic carbon (DOC) pool of aquatic environments [14,15]. AOM consists of a heterogeneous mixture of carbohydrates (including mono-, oligo-, and polysaccharides), nitrogenous compounds (amino acids, proteins, and polypeptides), lipids (fatty acids), and organic acids (glycollate, tricarboxylic acid, hydroxamate, and vitamins) [14]. In general, coastal environments, including harbors, contain many nutrient sources and favorable physical conditions for growth, thus they are susceptible to frequent phytoplankton blooms [16]. Phytoplankton release different intracellular and extracellular components depending on their growth stage, and the composition of these components differs between seawater and freshwater species [17,18,19]. While protein is the major cellular component in exponentially growing cells, carbohydrates are the most abundant extracellular products, followed by proteins and amino acids [17]. In addition, some AOM moieties (e.g., dihydroxybenzene compounds) have been identified as key DBP precursors [18,20].

Because past research on DBPs in BWMSs that employ active substances has primarily been conducted in the context of land-based tests to obtain BWMS type approval, the formation of DBPs from commercially used organic materials has received significant attention [5,10,11,13,21]. Some studies have investigated the generation of DBPs from phytoplankton-derived AOM in relation to drinking water [10,22,23,24], and a few studies have looked at the DBPs generated from the treatment of AOM derived from cultured phytoplankton with active substances in fresh and marine water [18,19,22]. However, there have been few studies on DBPs formed during the treatment of ballast water using phytoplankton species collected from natural environments and with no additional additives to increase DOM [25,26].

In this study, both glucose, which is widely used as the DOM in land-based tests, and AOM derived from phytoplankton collected in the field were treated with active substances to identify the characteristics associated with their DBP formation. This information can be used to understand the environmental risks posed by DBPs that arise from BWMSs that employ active substances in real port environments.

2. Materials and Methods

Land-based testing was carried out following the procedures for BWMS approval specified in the BWMS Code and ETV protocol. To understand the relationships between the environmental parameters and DBP formation, principal component analysis (PCA) was conducted using SPSS Statistics 12.0 (IBM, Armonk, NY, USA). The variables in the PCA were explained using Pearson correlation with varimax rotation.

2.1. Preparation of the Test Water and the Sampling Period

The tests were carried out at test facilities designed for the type approval of BWMSs at the South Sea Research Institute of the Korea Institute of Ocean Science & Technology in Korea. To conduct the land-based testing to evaluate BWMS performance, test water and natural plankton assemblages were also collected near the testing facility (Table S1).

To prepare the land-based test water (LBTW), approximately 4 tons of field-collected plankton, commercial glucose, and starch were added to the test-water tank to meet the test water condition (

Table 1. Minimum criteria for land-based (LB) IMO (BWMS code) and USCG (ETV protocol) requirements (freshwater: F; brackish water: B; marine water: M).

Parameter		Test Water Condition						Discharge Water
		IMO			USCG			Control Water	Treated Water (D-2)
	unit	F	B	M	F	B	M
Salinity		<1	10–20	28–34	<1	10–20	28–36
POC	mg/L	1	5		4			-	-
DOC		1	5		6			-	-
TSS		1	50		24			-	-
Escherichia coli	cfu/100 mL	Measured			Measured			-	<250
Intestinal Enterococci		Measured			Measured			-	<100
Vibrio cholera (O1&O129)		Measured			Measured			-	<1
Heterotrophic bacteria	cells/ml	104			103			-	Measured
Organisms of ≥10 & <50 μm	organisms/mL	103			103			>100	<10
Organisms of ≥50 μm	organisms/m3	105			105			>100	<10

), and after the LBTW was mixed for 30 min. Subsequently, 250 tons of the prepared test water was treated using the BWMS (maximum dose of total resident oxidant (TRO): 10 mg/L as Cl₂), which killed the living organisms via the production of active substances using electrolysis (the side-stream method), and then stored in a treated tank. The remaining 250 tons were stored in a control tank without treatment. The test methods of LBTW are according to the methodology of IMO G9, BWMS code, and ETV protocol [3,4,7]. A schematic diagram of the test process is shown in Figure 1A.

To prepare the AOM test water (AOMTW), a portion of the field-collected phytoplankton samples prepared for the land-based testing was transferred to the laboratory and sonicated for 30 min to increase the AOM concentration. After filtering the water with GF/F (<0.7 μm), the DOC concentration was measured, and the prepared AOM was injected into 100 L of water taken from the test-water (550 tons) to adjust the DOC concentration to 6 mg/L. The field-collected plankton and starch were then added to the prepared AOMTW according to the test water conditions for land-based testing. NaOCl (Junsei, CAS No. 7681-52-9) was injected into the AOMTW so that the TRO concentration was about 10 mg/L as Cl₂. In order to ensure that the environmental conditions of the two test waters were consistent, the AOMTW prepared in the laboratory was divided into 12 sterilization packs (1 L each) and stored in a 250-ton control-water tank. The LBTW and AOMTW were simultaneously sampled for environmental factor analysis, DBP analysis, and whole effluent toxicity (WET) testing on Days 0, 2, and 5. A schematic diagram of the test process is shown in Figure 1B.

2.2. Water-Quality Parameters

Water temperature, salinity, and dissolved oxygen were measured in situ with an EXO II water-quality monitoring platform (YSI, Yellow Springs, Inc., Yellow Springs, OH, USA). A pocket Colorimeter^TM II analysis system (Hach Co., Loveland, CO, USA) was used to measure the TRO concentration in conjunction with the N,N-diethyl-p-phenylenediamine (DPD) method, which produced a red solution. Samples for DOC analysis were gravity-filtered directly from a glass syringe before acidification to 1 ≤ pH ≤ 2 using H₃PO₄ or H₂SO₄. Immediately after acidification, the sample was stored at a temperature of less than −20 °C and analyzed within 14 d of sampling. Analysis of the DOC was carried out using a TOC analyzer (TOC-V_CPH, Shimadzu Co., Kyoto, Japan), while the specific UV absorbance (SUVA₂₅₄) was determined by normalizing the mean UV spectrophotometer data (Optizen 3220UV, Mecasys Co., Daejeon, Republic of Korea) to the mean DOC concentration (UVA/DOC). Active chlorophyll-a was measured using a chlorophyll fluorometer (PHYTO-PAM, Heinz WALZ Gmb Co., Effeltrich, Germany).

2.3. Relevant Chemicals

A total of 41 potential DBPs related to the BWMS, as defined in the MEPC 67/INF.17 guidelines from the Marine Environment Protection Committee (MEPC) of the IMO, were investigated (

Table 2. Acronyms and mark of relevant chemicals indicating above detection limit.

Compounds	Acronyms	Mark *	Compound	Acronyms	Mark
Trihalomethanes	THMs		Trichloroacetic acid	TCAA	A4
Haloacetic acids	HAAs		Dichloroacetic acid	DCAA	A5
Haloacetonitriles	HANs		Monochloroacetic acid	MCAA	A6
Halogenated propionic acid	HPA		Dibromochloroacetic acid	DBCA	A7
Halogenated nitroalkane	HN		Dichlorobromoacetic acid	DCBA	A8
Bromoform	TBM	M1	Bromochloroacetic acid	BCAA	A9
Chloroform	TCM	M2	Dibromoacetonitrile	DBAN	N1
Dichloromethane	DCM	M3	Monobromoacetonitrile	MBAN	N2
Dibromochloromethane	DBCM	M4	Dichloroacetonitrile	DCAN	N3
Dichlorobromomethane	DCBM	M5	Monochloroacetonitrile	MCAN	N4
Tribromoacetic acid	TBAA	A1	Bromochloroacetonitrile	BCAN	N5
Dibromoacetic acid	DBAA	A2	Chloropicrin	CP	CP
Monobromoacetic acid	MBAA	A3	Dalapon	DP	DP

* Mark for PCA diagram.

) [27]. THMs concentrations were determined using a headspace with a gas chromatography/mass selective detector according to US EPA method 8260C (Agilent Co, Santa Clara, CA, USA). HAAs and HANs concentrations were determined using a gas chromatography electron-capture-detector system according to US EPA 552.2:2003 and US EPA 551.1:1995. A DB-624 column (60 m long, 0.25 mm id, 14 μm stationary phase) was used for the THM, a DB-1701 column (30 m long, 0.32 mm id, 0.25 μm stationary phase) for the HAAs, and a DB-5 column (30 m long, 0.25 mm id, 0.25 μm stationary phase) for the HANs. The THM, HAA, and HAN standards were acquired from AccuStandard (New Haven, CT, USA). Samples were quantitated using calibration standards, following the procedures for external and internal calibration standards or as otherwise specified in the method. All reagents used for extraction were high-performance liquid chromatography (HPLC) grade or higher. The method detection limit (MDL = t × sd, where sd is the standard deviation of the data and t is the compensation factor from a Student’s t-test with n − 1 degrees of freedom at a confidence interval of 95%) and the method quantification limit (MQL = 10 × sd) were calculated using the lower spike concentration. The relative standard deviations (RSDs) of the laboratory control samples (LCSs) for the compounds used to construct the calibration curves met the condition |RSD| < 15%.

2.4. Environment Risk Assessment

The Marine Anti-Foulant Model to Predict Environmental Concentrations (MAMPEC) was used to calculate the predicted environmental concentrations (PECs) of relevant chemicals in GESAMP-BWWG Model Harbors (MAMPEC-BW Model 3.1.0.3). This model can be modified to calculate the concentrations of substances released into water by specific processes, such as ballast water treatment, making it a useful tool for environmental impact studies. In calculating the PECs using the MAMPEC-BW model, the worst-case scenario was assumed, in which 100,000 m³ of ballast water is discharged and the chemicals in this discharge do not subsequently decay in the water of the harbor. The PECs under general conditions for all of the relevant chemicals were taken to be the highest concentration (i.e., the maximum discharge) observed on Days 0, 2, and 5 of water treatment after neutralization. The PECs were calculated for fresh, brackish, and marine water. The predicted no-effect concentration (PNEC) values are normally derived from acute and/or chronic aquatic toxicity results for relevant aquatic species by dividing the lowest available effect concentration with an appropriate assessment factor [7]. The values of the assessment factors (AF) depend on the composition of the ecotoxicity data for the PNEC derivation (Table S2). If the ecotoxicity dataset is properly constructed considering the tropic level, the lower value of AF is considered; otherwise, the higher value of AF is considered. The factors are inversely proportional to the availability and quality of toxicity data. Due to their simplicity and minimal data requirements, methods that use AFs are globally used in preliminary ecological risk assessments [28]. For the predicted no-effect concentration (PNEC), we used the values provided by IMO’s Global Integrated Shipping Information System (GISIS) to establish the ecotoxicity for each chemical (http//gisis.imo.org, assessed on 21 November 2022). If the PEC/PNEC ratio under general conditions was 1 or higher for those DBPs present above the detection limit, then they were considered potentially harmful to the environment.

2.5. Whole Effluent Toxicity Testing

In order to determine the acute and chronic toxicity for microalgae, invertebrates, and fish, samples of discharged untreated (control) and treated water after neutralization of the active substances were collected in 4 L sterilized water bags and 20 L polyethylene containers with no headspace to protect the samples from sunlight and contact with the atmosphere. The sample volumes for the untreated and treated ballast water were 100 L and 40 L, respectively. Samples were placed in the refrigerated cargo space of a transport vehicle to minimize temperature changes and were transported to the testing laboratory. Upon arrival at the laboratory, the transported samples were immediately filtered through a 1-μm CP filter (Chisso Filter, Osaka, Japan) to remove debris before processing. Test water for WET testing was prepared at concentrations of 100%, 50%, 25%, 12.5%, 6.25%, and 0% (control) of the discharged water using control water, which was used after acclimating to the test temperature for one day. At least 4 replicate chambers per concentration were placed in the acute toxicity and microalgal growth inhibition test, and 8 to 10 replicate chambers per concentration were placed in chronic test. The acute and/or chronic test data are used in the point estimation for median lethal (effective) concentration (L(E)C50). All ecotoxicity tests were performed to meet the test acceptability criteria required by the standard. In addition, tests using reference toxicants for the precision of the test were conducted, and the results were within the range suggested by the standard. Ecotoxicity bioassays and evaluations of reference toxicants were performed using USEPA or International Standard Methods (OECD, ISO). The WET tests were conducted with reference to the following standard protocols: OECD 201 and USEPA-821-R-02-013 for freshwater microalgae, OECD 201 and ISO 10253 for marine microalgae, and USEPA-821-R-02-012, USEPA-821-R-02-013, and USEPA-821-R-02-014 for invertebrate and/or vertebrates. The test species and test methods are summarized in

Table 3. Test species and test references used for whole effluent toxicity (WET) testing.

Taxon	Condition	Species	End-Point	Test Duration	References
Algae	Freshwater	Raphidocelis subcapitata	Population growth Inhibition	72 h	OECD 201 USEPA 2002b EPA-821-R-02-013 (Method 1003.0)
	Seawater/ Brackish water	Isochrysis galbana *			OECD 201 or ISO 10253; 2006
Invertebrate	Freshwater	Daphnid (Daphnia magna)	Mortality	48 h	USEPA 2002a EPA-821-R-02-012 (Method 2021.0)
		Daphnid (Ceriodaphnia dubia)	Survival and reproduction	8 days	USEPA 2002b EPA-821-R-02-013 (Method 1002.0)
	Seawater/ Brackish water	Mysid (Neomysis awatschensis)	Mortality	48 h	USEPA 2002a EPA-821-R-02-012 (Method 2007.0)
		Mysid (Neomysis awatschensis)	Survival and growth	7 days	USEPA 2002c EPA-821-R-02-014 (Method 1007.0)
Vertebrate (Fish)	Freshwater	Oryzias latipes	Mortality	96 h	USEPA 2002a EPA-821-R-02-012 (Method 2000.0)
			Survival and growth	7 days	USEPA 2000b EPA-821-R-02-013 (Method 1000.0)
	Seawater/ Brackish water	Cyprindon variegates *	Mortality	96 h	USEPA 2002a. EPA-821-R-02-012 (Method 2000.0 and 2004.0)
			Survival and growth	7 days	USEPA 2002c EPA-821-R-02-014 (Method 1004.0)

* Test species is not mentioned in the standard protocol presented references, but it is well-known in ecotoxicity testing and used in many studies, including the examples as below: 1. Fisher, D., Yonkos, L., Ziegler, G., Friedel, E., and Burton, D. 2014. Acute and chronic toxicity of selected disinfection byproducts to Daphnia magna, Cyprinodon variegatus, and Isochrysis galbana. Water Research; 55, 233~244. 2. Emerging Ballast Water Management Systems, Proceedings of the IMO-WMU Research and Development Forum 305 pp, 26–29 January 2010, Malmö, Sweden, Globallast Partnerships Programme and The Global Industry Alliance. 3. Matheickal, J.T. and Raaymakers, S. (Eds.) 2004. 2nd International Ballast Water Treatment R&D Symposium, IMO London, 21~23 July 2003: Proceedings. GloBallast Monograph Series No. 15. IMO London.

.

3. Results and Discussion

3.1. Comparison of Environmental Parameters

The TRO concentration in the treated LBTW and AOMTW on Day 0 was 8.00 ± 0.26 mg/L and 10.10 ± 0.24 mg/L, respectively, with the residual TRO concentration in the treated water decreasing over time (

Table 4. Environmental parameters measured for the land-based test water (LBTW) and algal organic matter test water (AOMTW) in control (C) and treated (T) water (The values presented in the table are mean values, and the values in parentheses are standard deviations.)

Parameters	Time (Day)	LBTW						AOMTW
		Fresh Water		Brackish Water		Marine Water		Fresh Water		Brackish Water		Marine Water
		C	T	C	T	C	T	C	T	C	T	C	T
Temperature (°C)	0	14.0	14.0	10.3	10.9	22.2	22.4
	2	12.5	12.5	9.6	9.6	23.1	22.8
	5	12.4	12.7	8.8	8.3	23.8	23.5
Salinity (PSU)	0	0.17	0.17	17.7	17.7	32.0	32.0
TRO (mg L−1) n = 2	0	-	7.90(0.14)	-	7.80(0.1)	-	8.37(0.1)	-	9.90(0.00)	-	10.4(0.14)	-	10.0(0.00)
	2	-	4.10(0.14)	-	2.47(0.01)	-	1.57(0.00)	-	0.67(0.01)	-	4.95(0.00)	-	0.01(0.01)
	5	-	2.30(0.00)	-	1.03(0.01)	-	0.67(0.14)	-	0.09(0.00)	-	3.14(0.00)	-	0.01(0.00)
DOC (mg L−1) n = 3	0	6.03 (0.12)	6.63 (0.05)	6.14 (0.14)	6.57 (0.02)	6.63 (0.11)	6.98 (0.12)	7.43 (0.18)	11 (0.24)	5.37 (0.08)	6.18 (0.11)	6.14 (0.06)	9.39 (0.17)
	2	5.47 (0.07)	7.00 (0.14)	8.56 (0.13)	7.53 (0.04)	6.32 (0.09)	9.01 (0.02)	6.76 (0.02)	14.78 (0.37)	8.35 (0.24)	7.55 (0.14)	3.61 (0.07)	9.99 (0.09)
	5	4.03 (0.09)	7.07 (0.11)	3.73 (0.07)	6.61 (0.10)	7.11 (0.21)	7.05 (0.05)	4.67 (0.17)	9.33 (0.19)	5.19 (0.04)	7.93 (0.01)	3.16 (0.05)	6.23 (0.01)
SUVA254 (m−1 of absorbance per mg/L of DOC)	0	0.437	1.176	0.342	2.261	0.196	1.526	1.306	1.109	0.87	1.401	0.741	1.128
	2	0.548	2.372			0.206	0.416	1.316	0.880	0.425	1.106	0.499	0.095
	5	0.678	1.278	0.870	2.459	1.350	1.248	1.886	1.826	0.694	1.192	0.680	0.385
Active chlorophyll-a (μg L−1) n = 3	0	33.6 (0.22)	4.9 (0.11)	21.9(0.85)	1.1 (0.04)	19.5 (0.24)	2.3 (0.36)	77.1 (1.70)	7.0 (0.23)	26.1 (0.85)	0.4 (0.20)	61.4 (1.94)	1.9 (0.19)
	2	22.4 (1.10)	0.7 (0.07)	25.7 (0.40)	0.0	0.3 (0.03)	0.0	22.0 (0.12)	2.0 (0.12)	27.1 (1.27)	0.0	3.1 (0.11)	0.0
	5	18.7 (0.30)	0.5 (0.10)	22.7 (0.47)	0.0	0.3 (0.02)	0.0 (0.00)	24.7 (0.25)	0.1 (0.02)	26.2 (1.37)	0.0	0.7 (0.07)	0.0
Heterotrophic bacteria (CFU/mL) n = 3	0	207,333 (3055)		74,000 (2291)		141,667 (7638)		-	-	-	-	-	-
	5	351,667 (10,408)	3983 (126)	281,667 (2887)	1983 (126)	403,333 (60,277)	4500 (1000)	-	-	-	-	-	-

). In the marine LBTW and AOMTW, the TRO concentration decreased fairly rapidly, which was assumed to be due to the elevated temperature and high DOC concentration of the test waters during the summer. In particular, a detection limit of 0.01 mg/L for the TRO and high value of DOM were observed from Day 2 in the marine AOMTW. Moreover, in freshwater, the TRO concentration in the AOMTW decreased more sharply from Day 2 than did that of the LBTW, which is thought to be due to the relatively high DOC concentration on Day 2 for the AOMTW. In this study, the main factors affecting the TRO concentration are considered to be water temperature and DOM concentration, which is consistent with the results of previous studies [5,13,29]. This means that the TRO concentration in the treatment can be consumed quickly if the BWMS is operated during the summer period when phytoplankton bloom occurs in the harbor. Thus, since the BWMS using active substances kills organisms even with TRO remaining in the ballast tank, a low TRO concentration may weaken the performance of BWMS to kill unwanted organisms.

The average DOC concentration in the control LBTW and AOMTW was 6.27 ± 0.32 mg/L and 6.31 ± 1.04 mg/L on Day 0, 6.79 ± 1.59 mg/L and 6.24 ± 2.41 mg/L on Day 2, and 4.96 ± 1.87 mg/L and 4.34 ± 1.05 mg/L on Day 5, respectively. The average DOC concentration in the treated LBTW and AOMTW was 6.73 ± 0.22 mg/L and 8.86 ± 2.46 mg/L on Day 0, 7.85 ± 1.04 mg/L and 10.8 ± 3.68 mg/L on Day 2, and 6.91 ± 0.26 mg/L and 7.83 ± 1.55 mg/L on Day 5, respectively (Table 4). The chlorophyll-a concentration on Day 0 in the control AOMTW indicated that phytoplankton were present at relatively high levels. When the phytoplankton cells were killed by the TRO, DOM leaked into the treated water, leading to the high DOC concentration on Day 2.

The DOC concentration in the control water decreased over time in the fresh and brackish waters and the levels of heterotrophic bacteria increased two-fold, indicating that the DOC was consumed by these bacteria. It has previously been reported that these heterotrophic bacteria play an important role in the DOC pool [30,31]. However, the DOC concentration in the control marine water increased, possibly due to the rapid death of the organisms (indicated by the low chlorophyll-a concentration) in response to the high water temperature. In the treated water, the DOC concentration increased on Day 2 for all levels of salinity before decreasing on Day 5, with the exception of the brackish water. As explained above, the increase in the DOC concentration on Day 2 was likely due to the death of the organisms from the treatment, while the decrease on Day 5 was the result of the consumption of DOC by heterotrophic bacteria. In the brackish water, the number of heterotrophic bacteria was lower than for the other levels of salinity, suggesting that the increase in DOC due to cell death had a greater effect than DOC consumption by heterotrophic bacteria.

The dominant species in the fresh and brackish test water was Cyclotella spp., which accounted for 97% and 92% of the identified organisms, respectively, whereas Melosira spp. dominated the marine test water (85%) (Table S3). Phytoplankton differ in their internal and external constituents depending on their growth stage, and the constituents of the DOM also differ between species. For example, Biddanda and Benner (1997) reported a positive correlation between photosynthesis and the release of DOM [14], while Myklestad (2000) found that protein accounted for 50% of the DOM during the growth phase of phytoplankton, with carbohydrates dominating during the stationary phase [17]. It has also been reported that the DOM concentration increases during phytoplankton blooms in ports, and the highest DOM concentration is observed during the last stages of the bloom [19]. Thus, the concentration and characteristics of the DOM in a port vary depending on the phytoplankton species and growth stage. In general, the concentration and type of DOC in test water have a significant correlation with the increase in the concentration of DBPs [8,13]. Therefore, the concentration and species of DBPs generated from DOM as a precursor differ when ballast water is treated using active substances in the port during a phytoplankton bloom.

SUVA provides rapid information about DOM aromaticity and represents an important property related to DOM reactivity [32]. SUVA at 254 nm has been widely used as a simple indicator for DBP precursors because of their strong correlation [22,32,33]. However, the correlation between SUVA and DBP formation behavior is strongly dependent on the SUVA of the water. In general, a high SUVA can predict possible DBP species, whereas it is quite difficult to predict DBPs with a low SUVA [34,35]. In both the control LBTW and AOMTW, the SUVA followed the order freshwater > brackish water > marine water, the result of organic matter of terrestrial origin, which has a high SUVA, undergoing photodegradation over time to become organic matter with a low SUVA [36]. The SUVA of the control water on Day 0 was higher in the AOMTW (mean 0.97) than in the LBTW (mean 0.32) and was the highest in the fresh AOMTW (mean 1.31), but a SUVA lower than 2 is considered low. The low SUVA of the AOMTW was due to the fact that the DOM originates from phytoplankton-derived AOM, which is known to have a SUVA of <2, indicating high biodegradability and a low aromatic content [22,23]. The reason that the SUVA of the LBTW was lower than that of the AOMTW is thought to be because the DOM is derived from glucose, a simple carbohydrate that is highly biodegradable and has a low SUVA [37,38].

The chlorophyll-a concentration was high in the control water and fell after treatment with the active substance, indicating that the organisms were killed during the treatment process (Table 4). The concentration of chlorophyll-a also decreased over time in the control water, indicating that the phytoplankton population decreased due to natural death in a tank with no access to light. Based on the chlorophyll-a levels, two and three times more phytoplankton were present in the control AOMTW than in the LBTW in the fresh and marine water tests, respectively, whereas the brackish water test exhibited no significant difference. As discussed above, the difference in the concentration of the injected phytoplankton was related to the increase in the concentration of DOC and the decrease in the concentration of the TRO in the treated water.

3.2. Comparison of the Disinfection By-Products

The DBP concentration detected in the AOMTW and LBTW was divided by the DOC concentration and the results plotted over time (Figure 2). The total DBP concentration in the freshwater AOMTW and LBTW exhibited almost no difference, whereas that in the brackish and marine AOMTW was more than twice as high as in the LBTW (Figure 2). In the AOMTW, the total THM and HAA concentration followed the order marine water > brackish water > freshwater, and tended to increase over time, and this trend was also observed in the LBTW (Figure 2). As a result, higher concentrations of DBPs were produced in DOM derived from phytoplankton, and among them, TBM and DCAA concentrations were high.

In the brackish water, the THM concentration in the AOMTW was significantly higher than in the LBTW, and this trend was even stronger in the marine water (Figure 2). In fresh AOMTW, the THM concentration was high on Day 0, while that of fresh LBTW was high on Days 2 and 5 (Figure 2). The concentration of THMs in the marine and brackish water was about two times higher in the AOMTW than in the LBTW (Figure 2). Of the individual THMs, the TCM concentration was high in the freshwater, while the TBM concentration was high in the marine and brackish water, especially in seawater due to the difference in the concentration of bromine ions in the test water. Chlorine can cleave aromatic rings, producing both chlorinated and oxygenated by-products, while bromine can be substituted into the ring structures without cleavage. Thus, the reaction rate of bromine (HBr, OBr-) is about 10 times faster than that of chlorine (HOCl, OCl-), so TBM appears in high concentrations in brackish and particularly marine water [39].

The initial TRO concentration in the AOMTW was about 25% higher than that in the LBTW, which could affect their THMs concentration (Table 4). Moreover, the concentration of HAAs in brackish AOMTW was about 3 times, and that in marine AOMTW was about 1.6 times higher than that in the marine LBTW except for Day 5, respectively. However, in the fresh LBTW, the concentration of THMs and HAAs was about 2 times and 1.8 times higher than in the AOMTW, respectively, and HANs were not readily detected in the fresh AOMTW (Figure 2). Thus, the increase in the TRO may affect the THM and HAA concentrations in the marine and brackish water, which is consistent with reports that an increase in active substances increases THM and HAA concentrations [9,10,40]. Additionally, as the concentration of THMs increases as the pH increases, brackish and marine water are likely to result in higher THM concentrations than freshwater [18].

On Day 5, the total concentration of HAAs fell in the marine AOMTW; in particular, DBAA and DCAA decreased significantly, which was considered to be related to the residual TRO concentration. The marine water tests were conducted in early summer when the water temperature was high, thus the DOM and phytoplankton concentration of AOMTW were also high, leading to the rapid consumption of TRO. The residual TRO concentration was found to be 0.01 mg/L on Day 2, which was the detection limit for the TRO device (Table 4). On the other hand, although the LBTW had a low TRO concentration, it remained at 0.67 mg/L on Day 5 (Table 4). In addition, the concentration of HAAs did not decrease on Day 5 in the fresh and brackish water, in which the residual TRO concentration remained high. Jang et al. (2020) reported that DBAA can maintain a constant concentration over long periods of time (20 d) if the residual TRO concentration is maintained in test water treated with an active substance [5]. However, further analysis of the change in DBP characteristics with respect to AOM and residual TRO concentrations is required.

Of the analyzed DBPs, 12, 13, and 16 species were above the detection limit in the fresh, brackish, and marine LBTW, respectively, compared to 17, 19, and 19 species in the AOMTW. There was little difference in the abundance of individual THM and HAA species between the marine and brackish LBTW and AOMTW. However, a difference was observed in the freshwater samples, with the AOMTW dominated by TBM (40.5%), DBCM (32.5%), and DCBM (19.0%), whereas the LBTW was characterized by high levels of TCM (35.8%), DCBM (40.7%), DBCM (12.3%), and DCM (9.1%) (Figure S1). In the fresh LBTW, only DCAA (85%) and MBAA (15%) were observed, whereas eight species were found in the AOMTW, including DBAA (27%) and BCAA (18.6%) (Figure S1). In particular, the concentration of Br-DBPs was high in the freshwater AOMTW where only TBM and DBAA were found (Figure 2). This is thought to be influenced by the characteristics of DOM in the freshwater AOMTW. The field-collected phytoplankton sonicated to increase the DOM in the AOMTW were primarily Cyclotella spp (98%), which can survive in both fresh and brackish water. Therefore, it was presumed that the bromine ions in these species were released together with the DOM during the sonication process. Because Cyclotella was added to meet the conditions for the LBTW, a small amount of bromine ions may have been generated, leading to the formation of DCBM and DBCM in freshwater LBTW.

Of the HANs, DCAN and CP exhibited their highest concentrations on Day 2 in the fresh LBTW, DBAN was highest on Day 5 in the marine LBTW, and MBAN was highest on Day 5 in the brackish AOMTW (Figure 2). DBPs are potentially toxic, with HANs in particular known to be highly toxic, especially brominated HANs (br-HANs) [41]. In this study, br-HANs were primarily detected in the brackish and marine LBTW and AOMTW (Figure 2). HANs were frequently observed in the LBTW, which had a relatively lower SUVA than the AOMTW. However, high DBAN concentrations were detected in the brackish AOMTW despite the low SUVA of LBTW, which indicates that it is important to have a low SUVA with N-rich organic matter in the test water. Hua et al. (2020) reported that a high SUVA only produced high yields of carbonaceous DBPs (e.g., trichloromethane, HAAs, and haloketones), whereas low-SUVA N-rich precursors yielded high levels of both C and nitrogenous DBPs (e.g., HANs and chloropicrin) [34]

PCA was conducted to determine the associations between physicochemical environmental factors and DBP formation (Figure 3). In the AOMTW and LBTW, the two main components one and two (C1 and C2, respectively) accounted for 66.7% and 72.8% of the total variance, respectively (n = 9). In the AOMTW, C1 explained 42.0% of the variance and C2 explained 30.8%. C1 exhibited a strong positive correlation with consumed TRO, BCAA, temperature, MBAA, dalapon, TCAA, DBCA, DBCM, and bromoform, and a strong negative correlation with SUVA. C2 exhibited a strong positive correlation with MBAN, BCAN, TBAA, and DCBA, and a strong negative correlation with TCM, DCBM, and MCAA. DBPs found at relatively high concentrations in freshwater were located in the third quadrant, brackish water in the second quadrant, and marine water in the first quadrant, suggesting that bromine ions are an important determinant of DBP concentration. Many types of DBP were located in the first quadrant, which was associated with an increase in the concentration of DBPs from freshwater to marine water. This trend was similar for the LBTW, though it was less pronounced.

The DOC concentration in the LBTW had a positive correlation with the DBPs, while the AOMTW exhibited a negative correlation (Figure 3). This is because the concentration of DOC may increase with the deaths of organisms during treatment with the active substance. In the fresh AOMTW, the highest concentration of DOM was detected in the treated water due to the high concentration of injected phytoplankton, but the concentration of DBPs was lower than in the other test waters. For this reason, the concentration of DOC in the AOMTW had a negative correlation with DBPs and was located in the third quadrant of the PCA plot. The SUVA exhibited a positive correlation with BCAN in the AOMTW with DCAN in the LBTE, but a negative correlation with MCAN and DCAN in AOMTW and MCAN in LBTW. It remains difficult to determine the mechanisms for HAN formation using only the SUVA. As mentioned above, nitrogen-rich AOM, rather than bromine ions, is an important factor in the formation of HANs. Therefore, parameters that can measure nitrogen-rich precursors (e.g., fluorescence measurements or dissolved organic nitrogen analysis) should be investigated in the future.

3.3. Comparison of MAMPEC Results

In the AOMTW, the PEC/PNEC ratio for the DBPs did not exceed 1, and only DBAN and chloropicrin in the LBTW had a PEC/PNEC value higher than 1, despite their low concentrations (

Table 5. Predictive environmental concentration (PEC) for the surrounding conditions as calculated using MAMPEC 3.1, the predictive no-effect concentration (PNEC) as calculated using the lowest toxicity data and the assessment factor (AF) from the IMO, GISIS, and the PEC/PNEC ratio of THMs, HAAs, and HANs detected in both the land-based test water (LBTW) and algal organic matter test water (AOMTW).

				LBTW		AOMTW
Compound	Chemical Group	Asseseement Factor	PEC	PNEC	PEC/PNEC	PEC	PNEC	PEC/PNEC
TBM	THMs	5.0 × 10	0.38 × 10	9.6 × 10	4.0 × 10−2	0.98 × 10	9.6 × 10	1.0 × 10−1
TCM		5.0 × 10	3.0 × 10−1	1.5 × 102	2.0 × 10−3	3.9 × 10−2	1.5 × 102	2.6 × 10−4
DCM		5.0 × 10	2.0 × 10−2	1.2 × 102	1.7 × 10−4	4.4 × 10−3	1.2 × 102	3.7 × 10−5
DBCM		1.0 × 10	1.6 × 10−1	0.63 × 10	2.5 x10−2	3.3 × 10−1	0.63 × 10	5.3 × 10−2
DCBM		1.0 × 10	1.8 × 10−1	7.8 × 101	2.3 × 10−3	1.0 × 10−1	7.8 × 10	1.3 × 10−3
TBAA	HAAs	1.0 × 10	2.7 × 10−1	1.4 × 104	2.0 × 10−5	1.3 × 10−1	1.4 × 104	8.9 × 10−6
DBAA		1.0 × 10	8.7 × 10−1	6.9 × 103	1.3 × 10−4	7.0 × 10−1	6.9 × 103	1.0 × 10−4
MBAA		1.0 × 102	4.0 × 10−2	1.6 × 10	2.5 × 10−3	4.3 × 10−2	1.6 × 10	2.7 × 10−3
TCAA		1.0 × 10	2.7 × 10−3	3.0 × 102	9.0 × 10−6	9.2 × 10−2	3.0 × 102	3.1 × 10−4
DCAA		1.0 × 103	9.3 × 10−1	2.3 × 10	4.1 × 10−2	0.20 x10	2.3 × 10	8.8 × 10−2
MCAA		1.0 × 10	2.0 × 10−3	5.8 × 10−1	3.5 × 10−3	3.6 × 10−3	5.8 × 10−1	6.1 × 10−3
DBCA		1.0 × 10	3.2 × 10−1	3.0 × 102	1.1 × 10−3	1.6 × 10−2	3.0 × 102	5.2 × 10−5
DCBA		5.0 × 10	3.3 × 10−2	6.0 × 10	5.4 × 10−4	1.3 × 10−2	6.0 × 10	2.1 × 10−4
BCAA		1.0 × 102	4.9 × 10−2	1.6 × 10	3.0 × 10−3	5.1 × 10−2	1.6 × 10	3.2 × 10−3
DBAN	HANs	1.0 × 104	0.11 × 10	5.5 × 10−2	1.9 × 10
MBAN		1.0 × 103	1.8 × 10−1	2.3 × 10	8.0 × 10−3	5.5 × 10−1	2.3 × 10	2.4 × 10−2
DCAN		1.0 × 103	4.5 × 10−3	2.4 × 10	1.9 × 10−4	1.1 × 10−3	2.4 × 10	4.4 × 10−5
MCAN		1.0 × 104	5.4 × 10−2	1.6 × 10−1	3.4 × 10−2	1.1 × 10−2	1.6 × 10−1	6.8 × 10−2
BCAN		1.0 × 103	4.9 × 10−2	6.9 × 10−1	7.2 × 10−2	1.8 × 10−3	6.9 × 10−1	2.6 × 10−3
DP	HPA	1.0 × 103				2.5 × 10−3	1.1 × 10	2.2 × 10−4
CP	HN	1.0 × 102	3.4 × 10−2	2.5 × 10−2	1.4 × 10	1.9 × 10−4	2.5 × 10−2	7.7 × 10−3

). In particular, DBAN had a PEC/PNEC ratio of 19, indicating a significant risk; however, due to insufficient toxicity data for this substance, an assessment factor of 10,000 was applied to the PNEC, leading to a very low value for the PNEC and a PEC/PNEC ratio higher than 1. However, Liu et al. (2018) reported that brominated HANs such as DBAN, which is relatively stable, were mostly responsible for the toxicity, and greater formation of DBAN leads to higher cytotoxicity [18]. In order to accurately determine the toxicity of this compound, it is important to obtain more toxicity data [5].

3.4. Comparison of the WET Test Results

In the acute toxicity tests performed after neutralizing the treated water, the L(E)C50 and no observed effect concentration (NOEC) were 100% or more, indicating no potential toxicity. However, in the chronic toxicity tests using phytoplankton, all of the levels of salinity showed potential toxicity. In particular, the L(E)C50 of the marine AOMTW was 33.4% on Day 2, representing the strongest toxicity, and the total DBP concentration was the highest during this time (

Table 6. Water quality parameters for the samples used for WET tests and subsequent results on Days 0, 2, and 5 using microalgae exposed for 72 h to treat the land-based test water (LBTW) and algal organic matter test water (AOMTW).

Environmental Condition		Freshwater			Brackish Water			Marine Water
	Day	0	2	5	0	2	5	0	2	5
LBTW	Temp. (°C)	5.4	4.8	5.3	6.7	5.6	5.2	7.9	8.3	6.5
	Sal. (PSU)	0.2	0.2	0.3	17.9	17.6	17.7	32.3	32.4	31.7
	pH	7.02	7.20	7.22	8.07	8.10	7.76	8.01	7.98	7.90
	DO (mg/L)	7.49	7.29	7.29	8.14	8.20	8.57	6.58	6.43	6.74
AOMTW	Temp. (°C)	4.8	2.2	2.8	4.5	3.3	2.4	5.9	6.4	4.3
	Sal. (PSU)	2.0	1.9	1.9	23.1	23.1	23.2	33.2	33.1	33.2
	pH	7.59	7.60	7.59	8.32	8.21	8.33	7.98	7.96	7.94
	DO (mg/L)	5.92	5.54	5.68	8.98	8.68	8.60	7.69	7.65	7.53
					LBTW			AOMTW
Water Type	Test Species	End point		Day	NOEC (%)		LC50 or EC50 (%)	NOEC (%)		LC50 or EC50 (%)
Fresh Water	Raphidocelis subcapitata	Algal growth inhibition (72 h)		0	50		>100	100		>100
				2	25		>100	<6.25		>100
				5	0		>100	<6.25		>100
Brackish Water	Isochrysis galbana	Algal growth inhibition (72 h)		0	100		>100	62.5		>100
				2	25		>100	25		57.6
				5	25		76.7	25		60.9
Marine Water	Isochrysis galbana	Algal growth inhibition (72 h)		0	100		>100	25		96.0
				2	62.5		>100	12.5		33.4
				5	50		>100	25		86.0

). Although the toxicity of DBPs varies from substance to substance, toxicity tends to follow the order HANs > HAAs > THMs and Br-HANs > Br-DBPs > Cl-DBPs [40,41]. In this study, the concentration of HAAs and HANs decreased on Day 5 in the AOMTW and the potential toxicity was lower, indicating that the HAA and HAN concentration affects the potential toxicity. Considering the PEC/PNEC values, the LBTW was considered to be more toxic than the AOMTW, but chronic toxicity of the phytoplankton was stronger in the AOMTW. Because WET tests evaluate the overall toxicity and do not address individual compounds, the total DBP concentration is assumed to play an important role in the evaluation of toxicity, indicating that it is also related to the cocktail effect of toxic substances [5,10]. Therefore, the results of this study show that the monitoring of individual DBP concentrations such as DBAN and chloropicrin in a port environment is important, but the monitoring of total HAA and HAN concentration is also important [42].

4. Conclusions

In this study, the characteristics of the DBPs generated when using an active substance to treat glucose, which is widely used in land-based testing for the type approval of BWMS and AOM derived from field-collected phytoplankton, were investigated. The following conclusions were drawn.

• A greater number of DBP species in freshwater and higher concentrations of DBPs in brackish and marine water could be generated in real aquatic environments.
• Br-DBPs can be formed in relatively high concentration when ballast water is treated using active substances in a real freshwater environment.
• DBAN and CP exceeded a PEC/PNEC ratio of 1 only in the LBTW test, but the WET results showed that the chronic toxicity of phytoplankton was highest in the AOMTW, indicating that the total concentration of DBPs may be more important than individual concentrations.
• The WET testing showed that the concentrations of HAAs and HANs play an important role in environmental risk. Therefore, it is necessary to monitor the concentration of DBPs, especially HAAs and HANs, in major international ports where AOM concentrations may increase due to frequent phytoplankton outbreaks.