Table 3.1
Physico-chemical properties of DCOI.

Table 3.2
Aerobic biodegradation of [¹⁴C]DCOI (0.05 mg/kg), polarity and distribution of metabolites in sediment and seawater. Data from Lawrence et al. 1991a.

Figure 3.1
Mineralization of [¹⁴C] DCOI (0.83 µg/g) in clayey sediment and seawater from the Sound (sediment LS). Aerobic conditions. Dotted curve represents ¹⁴CO₂ released by acidification.

Figure 3.2
Mineralization of [¹⁴C] DCOI (0.033 µg/g) in sandy sediment and seawater from the Sound (sediment SS). Aerobic conditions. Dotted curve represents ¹⁴CO₂ released by acidification.

Table 3.4
Anaerobic biodegradation of [¹⁴C]DCOI (0.05 mg/kg), polarity and distribution of metabolites in sediment and seawater. Data from Lawrence et al. 1991b.

Water-soluble metabolites from the transformation of DCOI constituted between 3.6% and 9.3% of the radioactivity added throughout the entire test period. Metabolites that were bound to the sediment and could not be extracted with methylene chloride/methanol constituted a constantly high part of between 40% and 67% of the ¹⁴C added (Table 3.4). Further extraction with HCl and NaOH showed that relatively water-soluble metabolites constituted <0.1% while fulvic and humic acids constituted 0.6% and 3.6%, respectively, of the ¹⁴C added after 365 days. Metabolites that were still bound to the sediment, probably to humin or clay, constituted 30% of the ¹⁴C added (Kesterson and Atkins 1992b). The formation of metabolites binding to humic acids, humin and clay minerals in the sediment is in agreement with the results in the aerobic biodegradation tests (Kesterson and Atkins 1992a).

Studies with Danish sediments
The anaerobic biodegradability of DCOI (0.83 µg/g) was examined by use of a clayey sediment and its seawater (Appendix 2), which was also used in the aerobic tests (cf. Section 3.2.2). Sediment and seawater was incubated under anaerobic sulfate-reducing conditions, which are normally prevalent in coastal marine sediments. The mineralization of [2,3-¹⁴C] DCOI into ¹⁴CO₂ constituted 14% of the ¹⁴C added after 56 days’ incubation at 15°C (Figure 3.3). The examination of the distribution of ¹⁴C in the clayey sediment at the termination of the test after 56 days showed that 45% of the ¹⁴C added was bound to humic acids, humin and clay minerals. Altogether, water-soluble substances in the water phase of the test system and hydrolizable compounds and water-soluble fulvic acids constituted 7% of the ¹⁴C added after 56 days. The mineralization of glucose, which was included as a readily biodegradable reference substance, constituted 59% of the ¹⁴C added after 56 days. Methods and results are described in detail in Appendix 2.

Figure 3.3
Mineralization of [¹⁴C] DCOI (0.83 µg/g) in clayey sediment and seawater from the Sound (sediment LS). Anaerobic conditions. Dotted curve represents ¹⁴CO₂ released by acidification.

Water and sediment samples from the tests were sampled at the start of the test and after 28 and 56 days. Chemical analyses of DCOI and metabolites in the water samples were made by Rohm and Haas (Spring House, Pennsylvania).

The analyses of water samples sampled at the termination of the test after 56 days showed that 4.0 ± 2.4% of the ¹⁴C added was present in the form of compounds with the same HPLC retention time as DCOI. In the same water samples, polar compounds constituted 13.7 ± 3.0% of the ¹⁴C added. The sediment samples were not analyzed as they contained 3-4 times less radioactivity than the sediment samples from the aerobic tests (Table 3.3).

3.2.4 Transformation and fate of DCOI in a harbour

An investigation of the spread and removal of DCOI was carried out in the vicinity of a freshly painted ship and of another ship that had been painted a couple of months earlier. Both ships were lying in Korsør Harbour where the investigations were made on 26 and 27 October 1998. Those days, the temperature of the water was approx. 10°C and varied very little according the depth of the water (Steen et al. 1999). The wind was southwesterly (between approx. 240 and 255° on 26 October and approx. 200° on 27 October). The wind velocity was approx. 8-10 m/sec with wind blasts of up to 15 m/sec on 26 October and a little more on 27 October (Danish Meteorological Institute 1999). The entrance of Korsør Harbour points in a north-easterly direction why the water must be expected to have been pressed out of the harbour.

The concentration of DCOI in the water phase was measured along two transects: one perpendicular to the direction of the ships and the other in north-easterly direction, i.e. in the wind direction. Most of the samples were taken on 26 October. The samples were taken over a relatively short period of time (approx. 5 hours) and the measured concentrations can thus only be considered valid for the day in question. The highest concentrations measured of DCOI were <300 ng/L close to the ship’s side (£1 m) and decreased to <50 ng/L at a distance of approx. 30 metres from the ship. The concentration of DCOI in a distance of 2 metres from the ships (along the transect and perpendicular to the ships) varied very little according to the water depth why the vertical mixing was considered to be total.

Steen et al. (1999) have made model calculations in which Korsør Harbour was modelled as a one-dimensional box, in which the flow in and out of the harbour was neglected and in which the dispersion coefficient was varied between approx. 0.004-0.03 m²/s. This interval is stated to be the end points of the expected variation interval of the dispersion coefficient of the harbour. Apart from the spread, a first order disappearance kinetics is assumed for DCOI. The simulations were made with three different rate constants for this first order process: 0 day^-1, 1 day^-1, and 1 hour^-1. As a result of the winds on 26 and 27 October, the dispersion must presumably have been high in the basin. By way of comparison it may be mentioned that the horizontal dispersion coefficient in Danish coastal waters typically varies between 0.04 and 5 m²/s (Harremoës and Malmgren-Hansen 1989). As regards the two transects, the best correlation between the measured and the calculated DCOI concentrations was achieved by use of a rate constant of disappearance of between 1 hour^-1 and 1 day^-1.

With the rate constant, 1 hour^-1, a good correlation was achieved between measured and calculated values close to the ships for one transect while the concentrations of the other transect was underestimated at distances of more than approx. 8 m. For both transects, the calculated concentrations are lower than the measured concentrations at larger distances from the ships (approx. 30 m). When assuming a rate constant of disappearance of 1 day^-1, the calculated concentrations are higher than the measured concentrations close to the ships for both transects but lower than the measured concentrations farther away (approx. 60 m). There are thus indications that the rate constant of the disappearance of DCOI close to the ships is higher than the corresponding constant farther away from the ships. On this basis, the rate constant of disappearance of the whole basin is considered to be between 1 hour^-1 and 1 day^-1, which corresponds to a half-life of between approx. 0.69 and 16.6 hours. This half-life includes biological and abiotic transformation as well as processes like sorption to suspended matter, sedimentation, potential vertical mixing and potential imperfection in the calculation of the dilution in the harbour.

3.3 Bioaccumulation and aquatic toxicity

Studies on the bioaccumulation of DCOI in fish are available but not in other types of organisms (e.g. mussels). The ability of DCOI to bioaccumulate in fish has been examined in laboratory tests over 28 days by use of [¹⁴C]DCOI. Two studies including chemical analyses of water and tissue samples have been made (Forbis et al. 1985; Derbyshire et al. 1991). In all tests, [¹⁴C]DCOI was continuously added to a flow-through system. Chemical analyses showed that, at the final part of the tests, the concentration of DCOI was considerably lower than the nominal concentration (e.g. 4.5% and 0.55% of the ¹⁴C added after 21 and 28 days, respectively (Leak 1986)) while DCOI was hardly measurable in the second test (Derbyshire et al. 1991). Presumably, the principal part of the remaining ¹⁴C activity in the water represented one or more polar metabolites.

The BCF values found (measured as radioactivity) were more or less identical in the two studies. The BCF values were 130-200 for muscle tissue, 700-1,100 for internal organs and 600 for the whole fish (Forbis et al. 1985; Derbyshire et al. 1991). The chemical analyses demonstrated that only 1% of the radioactivity in the fish was intact DCOI (Leak 1986). In connection with the study by Derbyshire et al. (1991), HPLC as well as TLC was used for identifying ¹⁴C labelled substances accumulated in the tissue of the fish. These studies indicate that it was most likely a question of substances without an isothiazolone ring structure and that the substances were built into the protein of the fish. The results indicate that DCOI was transformed in the water, after which it was mainly polar and probably linear compounds that were taken up in the fish. This assumption is confirmed by the biodegradation of DCOI (cf. Section 3.2.2; Lawrence et al. 1991a). It is thus considered likely that the measured BCF values should rather be related to metabolites of DCOI but as only a few of these metabolites are identified, the importance of the recorded bioaccumulation of labelled ¹⁴C cannot be assessed.

3.3.2 Toxicity towards aquatic organisms

Aquatic organisms
The toxicity of DCOI has been examined in standard laboratory tests with a number of aquatic organisms living in fresh water and in seawater:

Fresh water:

	Selenastrum capricornutum, green algae (Forbis 1990)
	Daphnia magna, crustacean (Burgess 1990; Ward and Boeri 1990)
	Oncorhynchus mykiss, rainbow trout (Shade et al. 1993)
	Lepomis macrochirus, bluegill sunfish (Shade et al. 1993)

Seawater:

	Skeletonema costatum, green algae (Debourg et al. 1993)
	Mysidopsis bahia, mysid, crustacean (Boeri and Ward 1990)
	Penaeus aztecus, brown shrimp, crustacean (Heitmuller 1977)
	Cyprinodon variegatus, sheepshead minnow, fish (Shade et al. 1993)
	Paralichthys olivaceus, Japanese flatfish, fish (Kawashima 1997a)
	Pagrus major, red sea bream, fish (Kawashima 1997b)
	Crassostrea virginica, oysters (Roberts et al. 1990)

Furthermore, tests with one mussel and protozoans are quoted by Shade et al. (1993) and Debourg et al. (1993), respectively. In some of the tests, problems with maintaining a constant exposure concentration have been reported and not all results have been calculated on the basis of measured concentrations (see below). The result of these irregularities is an overestimation of the effect concentrations - resulting in an underestimation of the toxicity of the substance.

The results, which are compiled in Appendix 4, show that there was no big difference in the sensitivity of freshwater and marine organisms. Table 3.5 summarizes the effects on the different groups of organisms.

The results from the algal tests performed (Forbis 1990) are calculated on the basis of the nominal concentration. The report on one of the tests shows that the concentration of DCOI decreased during the whole test period. Only 48% of the nominal concentration was left after 48 hours and, at the end of the test after 72 hours, it was only possible to measure the substance in the test vessels containing the highest concentration (Forbis 1990). The EC50 values stated are thus too high.

A 21-day reproduction test with daphnids (Ward and Boeri 1990) was conducted in such a way that is difficult to draw certain conclusions. This is due to the use of various concentrations of a solvent in relation to the addition of DCOI and to large variation in the data. The NOEC value stated represents the lowest concentration tested but the way in which the test has been conducted does not exclude that effects of DCOI may have occurred at this concentration as the effect may be dimmed as an unintentional result of the solvent. As the result of this test is the lowest NOEC value found in the tests, this value forms the basis of the calculation of PNEC for DCOI.

N-(n-octyl) malonamic
acid
The acute aquatic toxicity of N-(n-octyl) malomanic acid, which is an important metabolite from the transformation of DCOI (cf. Section 3.2.2), has been investigated in tests with fish and daphnia. The toxicity of N-(n-octyl) malomanic acid was tested in static tests and the calculations are based on measured mean concentrations of the substance. The effect concentrations for N-(n-octyl) malomanic acid are given for daphnia (48 h): EC50 = 260 mg/L, NOEC = 16 mg/L (Sword and Muckerman 1994b) and for rainbow trout (96 h): LC50 = 250 mg/L, NOEC = 160 mg/L (Sword and Muckerman 1994a). In the daphnia test, there is large variation in data and the basis of the calculation of the result is not clearly defined. Daphnids lying on the bottom of the test vessels do not seem to have been included as "immobile", which they should according to the method description used. The actual EC50 is estimated to be in the interval of 90-160 mg/L rather than 260 mg/L as stated in the report (Sword and Muckerman 1994b).

Even with the above reservations, it must, however, be concluded that N-(n-octyl) malomanic acid is several orders of magnitude less toxic than DCOI.

In connection with the investigations of N-(n-octyl) malomanic acid, QSAR calculations have been made of the toxicity of this metabolite and some substances with similar structure, which are important metabolites from the microbial transformation of DCOI. The results are given in Table 3.6.

The results in Table 3.6 indicate that the probable metabolites from the transformation of DCOI are neither particularly toxic nor bioaccumulative in aquatic organisms.

Sediment-living
organisms
Results of a 10-day test with the marine sediment-living crustacean, the amphipod Ampelisca abdita are available: LC50 = 320 mg/kg and NOEC = 6.9 mg/kg dry weight (Putt 1994). The test was made with ¹⁴C-labelled DCOI and the concentrations were measured as radioactivity. At the end of the test, approx. 90% of the ¹⁴C activity was attached to the sediment while the remaining part was distributed in the ratio of approx. 8:2 of pore water to the water above the sediment. No chemical analyses were made and the authors draw attention to the fact that the measured radioactivity is probably owing to metabolites and not to DCOI.

Algal communities
Acute and chronic effects of DCOI have been examined on communities of natural phytoplankton (planktonic algae) and epipsammon (micro algae living on grains of sand). Acute and chronic effects of DCOI on communities of phytoplankton have been found at a concentration of DCOI of 0.0003 mg/L (the lowest concentration in which effects were observed, Lowest Observed Effect Concentration, LOEC) (Arrhenius 1997). The acute effect of DCOI was a stimulation of the activity of the algae while the chronic effect was an adaptation to DCOI in a few days. Inhibition of the photosynthesis occurred at higher concentrations (EC50: 0.05-0.1 mg/L (95% confidence interval)). The study concludes that the effect of DCOI was still significant at the end of the test after 7 days. Communities of epipsammon were extremely tolerant to DCOI and the effect concentrations were several orders of magnitude higher than those for phytoplankton.

Effects of degradation of
DCOI on aquatic toxicity
Laboratory tests have been made in which the effects of degradation on the toxicity towards aquatic organisms were tested for a number of antifoulants including DCOI, Irgarol 1051 and Diuron (Callow and Finlay 1995; Callow and Willingham 1996). In these tests, the substances were incubated in seawater, seawater enriched with of marine bacteria and in sterilized seawater. Changes in the toxicity as the result of degradation of the active substances were tested towards marine bacteria (counting of colonial bacteria), diatoms (Amphora coffeaeformis) and crustaceans (Artemia salina). The degradation tests were started at concentrations of the substances causing 80% effect on the algae (EC80 = 0.5 mg/L for DCOI) so that a potential decrease of the toxicity of the solutions could be traced. The results showed that the toxicity practically did not decrease in sterilized seawater, and that the transformation of the active substance into metabolites with low toxicity progressed most rapidly in the bacteria-enriched seawater. The diatom test showed that e.g. the toxicity of DCOI had been considerably reduced (from approx. 80% to approx. 20% inhibition) after two weeks in natural and in bacteria-enriched seawater and that tests incubated for 4, 6 and 8 weeks in these two types of seawater caused 10% or no significant inhibition (Callow and Finlay 1995). The half-life of the toxicity of DCOI was calculated at 8.5 days in natural seawater and 3 days in bacteria-enriched seawater (Callow and Finlay 1995).

The relation between degradation and sorption of DCOI and the acute toxicity towards the marine crustacean Acartia tonsa has been examined in the present study. The tests were performed in systems with the sandy sediment and its seawater from the Sound (Appendix 2), which was also used in the biodegradation tests. DCOI was added in a concentration of 100 µg/kg to the sediment-seawater systems. Water phase and sediment were separated 20 min. after dosing and use of the water phase in tests with A. tonsa caused a mortality corresponding to 35% of the test organisms. Stationary incubation in the dark at 20-25°C resulted in the fact that there were no mortal effects on A. tonsa after day 1 (Figure 3.4). Similar results were achieved when the sediment-water systems were incubated in the light at an intensity corresponding to 340 µmol/m² × s. Measurements made by VKI in the Sound show that, in the period from May to October 1998, the average light intensity was 420 µmol/m² × s in a depth of approx. 1 metre. The light intensity used was thus approx. 80% of the mean value calculated on the basis of the measurements in 1998. The test results with A. tonsa show that DCOI sorbs to sediment or is transformed into metabolites with a considerably lower toxicity than the parent compound. The methods used are described in detail in Appendix 3. A parallel test was made with zinc pyrithione (cf. Section 4.4).

Figure 3.4
Effects of degradation of DCOI (100 µg/kg) dosed to sediment and seawater on the acute toxicity to Acartia tonsa (test performed in the dark).

3.4 Risk assessment of DCOI

Calculation of exposure
concentrations (PEC)
In order to calculate the exposure concentrations (PEC, Predicted Environmental Concentration), a model was established, based on principles normally used for exposure assessments (EC 1996). The exposure assessments were made for two scenarios:

	A pleasure craft harbour (on the basis of the conditions in the pleasure craft harbour of Jyllinge)
	A busy navigation route (on the basis of the conditions at the Kronprins Frederiks Bro, Frederikssund)

The model and the two scenarios are described in detail in Appendix 1. For the parent compound and the most essential metabolites, the following exposure concentrations were calculated for each of the two scenarios:

	PEC (water column)
	PEC (sediment)
	PEC (sediment-pore water)

The three exposure concentrations were defined as the steady-state concentration of the sub-environment in question. I.e., the concentration which the calculated concentrations eventually approach when a continuous leaching of the parent compound to the water environment is simulated. The calculations of PEC have been made by use of realistic worst-case scenarios, which means that the parameters used in the model are based on realistically conservative assumptions, which results in the fact that, in practice, the calculated PEC values are seldom exceeded. The model used is not validated towards measured concentrations in harbour environments or navigation routes. More of the assumptions that form part of the simulation are of vital importance to the result of the calculations:

The background concentrations for both the parent compound and the metabolites were assumed to be zero.

70% of the pleasure craft was assumed to have been painted with paint containing DCOI.

The leaching rate of DCOI from bottom paints was calculated at 13 mg/m2/day in harbours and 25 mg/m2/day when sailing (Appendix 1).

The primary biological transformation of DCOI into the expected metabolite N-(n-octyl) malonamic acid was assumed to proceed with a half-life of 14 hours in surface water at a temperature of 12°C.

The biological half-life of DCOI of 14 hours, which was assumed in the simulation, is established on the basis of an experimentally determined half-life in seawater at 12°C (Jakobson and Kramer 1999). The half-life of DCOI in seawater and not in seawater and sediment was chosen as the result of the simulation is exposure concentrations at a continuous leaching of DCOI to seawater after steady state was achieved. When the pleasure craft are taken out of the water at the end of the sailing season, DCOI will probably be rapidly eliminated as DCOI is either transformed in the water phase or sorbs to the sediment, in which it is transformed with a very short half-life (cf. Sections 3.2.2 and 3.2.3).

The exposure concentrations calculated for DCOI and its metabolites are approx. 50 times higher in the pleasure craft harbour than in the busy navigation route outside the harbour (Table 3.7).

Calculation of Predicted No-Effect Concentration (PNEC)
The Predicted No-Effect Concentrations (PNEC) are estimated for DCOI and N-(n-octyl) malonamic acid. The other stable metabolites from the transformation of DCOI are considered to have aquatic toxicity corresponding to that of N-(n-octyl) malonamic acid.

The available studies on the aquatic toxicity of DCOI are considered representative and the data include long-term studies with fish and crustaceans. The algal test may be interpreted as a short-term as well as a long-term test (EC 1996).

For DCOI, three NOEC values from long-term tests (fish, crustaceans and algae) are available, including the groups of organisms most sensitive in short-term tests (fish). On this basis, PNEC is calculated by dividing the lowest NOEC value, which is 0.00063 mg/L for crustaceans (Ward and Boeri 1990), by an assessment factor of 10 (EC 1996). This results in a PNEC of 0.00006 mg/L = 0.06 mg/L for DCOI. As already mentioned in Section 3.3.2, no unambiguous NOEC value can be derived from this long-term test with crustaceans (Ward and Boeri 1990). If this study is ignored, the results from one single long-term study with fish are available, in which NOEC was 0.006 mg/L. In this case, an assessment factor of 100 is applied, which results in a PNEC value calculated at 0.00006 mg/l = 0.06 µg/L, which is identical with the above calculated value.

Calculation of PNEC for N-(n-octyl) malonamic acid is based on the lowest effect concentration. As data primarily originates from short-term tests, an assessment factor of 1,000 is used for the lowest effect concentration. For N-(n-octyl) malonamic acid, the lowest reported LC50 = 250 mg/L for rainbow trout while the value for daphnia (EC50 = 260 mg/L) as already mentioned above is a moot point. For the calculation of a PNEC for N-(n-octyl) malonamic acid, an LC50 value of 90 mg/L (towards daphnids) is used as this value is considered the actual effect concentration in the tests performed (cf. Section 3.3.2). This results in a PNEC value of 0.09 mg/L = 90 µg/L. The calculated PNEC for N-(n-octyl) malonamic acid is assumed to be representative of the other metabolites from the transformation of DCOI. The two calculations of PNEC are shown in Table 3.8.

Risk quotient
On the basis of the above calculated PEC (water) values for DCOI and its metabolites given in Table 3.7 and PNEC values for DCOI and N-(n-octyl) malonamic acid, the risk quotients Rq = (PEC/PNEC) can be calculated as shown in Table 3.9.

*  N-(n-octyl) malonamic acid

The stated risk quotients are calculated on the basis of realistic worst-case scenarios (Appendix 1), which are i.a. based on the assumption that 70% of the pleasure craft are painted with a bottom paint containing DCOI. On the basis of the assumptions made in the simulation and of the calculated PEC values, it is considered likely that a risk of chronic ecotoxic effects within the pleasure craft harbour may exist as, presumably, DCOI will constantly be applied by leaching from bottom paints. The risk quotient for DCOI out of the harbour is less than 1 and here, the risk of ecotoxic effects is considered to be low.

Within as well as out of the pleasure craft harbour, a very small risk of ecotoxic effects of metabolites from the transformation of DCOI is considered to exist.