Environ. Sci. Technol. 2002, 36, 1539-1545
Persistence of Antifouling Agents in the Marine Biosphere JOHANNES RANKE* UFT-Centre for Environmental Research and Environmental Technology, University of Bremen, D-28359 Bremen, Germany
Risk indicators provide an interesting way to compare chemical substances with respect to the risks of their largescaled release. The present study implies that, for antifouling agents used in commercial shipping, residence times in the marine biosphere are especially suitable to represent their inherent potential to cause exposure of organisms. A simple box model is described providing the possibility to estimate residence times in the marine biosphere from water-particle equilibrium partitioning constants and half-lives in water and sediment. Resulting residence times in the marine biosphere range from about 5 d (4,5-dichloro-2-n-octyl-4-isothiazolin-3-one) up to about 40 000 yr (copper). For an evaluation of the validity of the model, calculated values are compared with measured environmental concentrations. Remaining uncertainties are also discussed. The main purpose of the presented residence times is to serve as a basis for decisions in antifouling paint development or environmentally conscious purchasing of antifouling paints.
Introduction The conventional approach to evaluate the risk caused by the use of a certain substance implies the comparison of local exposure concentrations with an acceptable or a “noeffect” concentration. A complementary approach, aiming at a direct comparison of substances, is the use of risk indicators. Its main advantage is the increased transparency of the evaluation resulting from the possibility of separate comparisons of each of the indicators. In a previous study (1), release rate, spatiotemporal range, bioaccumulation, bioactivity, and uncertainty of the evaluation have been suggested as main indicators of biosphere risk when comparing different biocides. This paper presents the comparison of four antifouling biocides by means of the indicator “spatiotemporal range”. The method described can be applied to any biocide used for antifouling purposes if particle-water partitioning constants and degradation rates in water and sediments are known. The spatiotemporal range of a chemical substance is defined here as a general indicator of the relative tendency of a certain substance to cause exposure of organisms when released to the biosphere. This definition is inspired by the concepts of persistence (2, 3), spatial and temporal range (4, 5), characteristic travel distance (6), and long-range transport potential (7). The central tasks in assessing the spatiotemporal range of a substance are as follows: (i) building an operable multimedia fate model representing the spatial segments of the environment most important for the context, (ii) calculating fate parameters as cited above, and (iii) evaluating * Telephone: +49 421 218 9725; fax: +49 421 218 7643; e-mail:
[email protected]. 10.1021/es011155p CCC: $22.00 Published on Web 02/26/2002
2002 American Chemical Society
if the model estimates environmental concentrations with satisfying precision. Models that have been used for antifouling agents (8-10) generally aim at predicting environmental concentrations in confined areas so that the global fate of the substances remains unclear. Particularly the possibility of a deep sea contamination has largely been disregarded in fate modeling approaches. Nevertheless, the inclusion of this enormous biotope appears essential for fulfillment of task i for antifouling agents. The main step taken here to keep the complexity of the fate model tractable is to combine harbors used for commercial shipping, adjacent estuaries, shelf seas, the open epipelagic, and the deep sea as well as the pertaining sediment compartments to nine global bulk compartments that are assumed to be homogeneous and well-mixed. The low spatial resolution of the resulting model does not support the calculation of a characteristic travel distance (6) or a spatial range (4). Therefore, a global residence time in steady state (τss) was chosen as a measure for the spatiotemporal range of the antifouling agents (task ii) based on
τss )
Mtot Stot
(1)
where Mtot is the total steady-state quantity of the biocide in the model system and Stot is its constant total release rate. The spatial aspect of the spatiotemporal range is included in the persistence measure by the possibility of accumulation of the substances in the deep seawater and in all sediment compartments, which are remote from the surface water compartments. Fate model evaluation studies (task iii) (e.g., refs 11-13), despite their importance for science-based decision-support, are still rare. This study presents a comparison of simulated data for copper and tributyltin with environmental concentrations reported in the literature. In addition, Monte Carlo simulations with varying biocide parameters were performed in order to assess the variability of the results due to the variability of the input parameters. The primary aim of this study is the quantification of the inherent potential of antifouling biocides to produce exposure of living organisms in the marine environment. The comparison of the spatiotemporal ranges of the biocides can be approximated by a comparison of their residence times, even though there are considerable uncertainties about input parameters, model structure, and model parameters. Such a comparison is relevant for an environmentally conscious selection of antifouling agents aiming at a marine transport system with a lower environmental impact in the global future.
Model Description The evaluative antifouling-agent fate model (AFM) is a linear nine-box model (level III; 14) and can be understood as an extension of the two-box model of the oceans first described by Broecker (15). It is based on nine mass balance equations for the nine model compartments (boxes) given in the Supporting Information. The layout of the five global water compartments and the four global sediment compartments as well as input and transport fluxes taken into account are shown in Figure 1. Compartment abbreviations, volumes, surface areas, and particle concentrations are given in Table 1. The compartments representing the biotope referred to as “marine biosphere” were derived according to the folVOL. 36, NO. 7, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1539
FIGURE 1. Layout of the compartments and transfer processes in AFM.
TABLE 1. Abbreviations, Volumes (V), Surface Areas (A) (46, 47), and Particle-Water Ratios (rpw) (48, 16) of the Compartments; Summed Water Exchange Rates (Fwx,wy ) and Sediment Burial Rates (B); and Turnover Times of Water (τw) in the w Water Compartments and of Solid Matter in the Sediment Compartments compartment
abbrev
V (m3)
A (m2)
rpw (g L-1)
∑yFwx,wy (m3 s-1) w
harbor water estuary water shelf water epipelagic water pelagic water harbor sediment estuary sediment shelf sediment pelagic sediment
wh wt ws we wp sh st ss sp
1.2 × 1010 3.0 × 1010 2.2 × 1015 6.7 × 1016 1.3 × 1018 4.8 × 108 4.8 × 108 1.1 × 1013 1.3 × 1014
1.2 × 109 1.2 × 109 2.7 × 1013 3.3 × 1014 3.3 × 1014 1.2 × 109 1.2 × 109 2.7 × 1013 3.3 × 1014
0.01 0.01 1 × 10-3 7 × 10-5 2 × 10-6 600 600 800 800
5.5 × 103 1 × 105 7.5 × 108 7.8 × 108 3.4 × 107
lowing reasoning: Epipelagic water (0-200 m) and pelagic water (>200 m) of the sea are distinguished as commonly found in the literature because of their large differences in elimination, transport, and sedimentation of biocides. Water and sediment on the continental shelf are separated from the main ocean compartment because biocide input into shelf water and sedimentation rates are much higher on the shelf than in the deeper parts of the ocean. The sediment compartments are confined to a depth of 40 cm below the seafloor, considered as comprising the majority of the bioturbated layer (16). Estuaries in which harbors are located are also separated due to the difference of biocide input, volume, and water exchange as compared to shelf water. Commercial harbor basins have an even higher ratio of biocide input per volume and are therefore represented separately by a water and a sediment compartment. The mass loss due to photolytical, biological, and chemical degradation is described by kinetic rate constants kw and ks in water and sediment, respectively (see Biocide Data section). Neither temperature dependence nor other spatial or temporal differences in degradation processes are included in the present fate model because of a lack of suitable degradation data. This simplification however is regarded as justified considering scale and general purpose of the model. In the following, compartment abbreviations from Table 1 will be used as superscript indices for mathematical symbols. The first letter of each compartment abbreviation indicates if it is a water (w) or a sediment (s) compartment; the second letter gives the geographical characterization. Subscript indices give information about the physical phase that the symbol pertains to. For example, Fws,ss signifies the p flux of particles from the shelf water compartment to the shelf sediment compartment. The derivation of water exchange rates between the compartments and of particle sedimentation and sediment burial rates is documented in the Supporting Information. The water residence time as the average time a biocide molecule spends in the water compartments only is given by 1540
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 7, 2002
B (m yr-1)
τt
2 × 10-3 2 × 10-3 3 × 10-5 1.2 × 10-5
2.5 d 3.3 d 34 d 2.7 yr 1200 yr 200 yr 200 yr 13300 yr 33400 yr
its total quantity in steady state (M ˙ ) 0) in the water compartments divided by the biocide flux into the system. Total residence times, also derived from steady state, were calculated according to eq 1. Input parameters for the biocides tributyltin (TBT), copper, 2-tert-butylamino-4-cyclopropylamino-6-methylthio1,3,5-triazine (Irgarol 1051, shorten to Irgarol), and 4,5dichloro-2-n-octyl-4-isothiazolin-3-one (shorten to DCOI, active substance in Sea-Nine 211) were collected from scientific as well as administrative literature. Input Estimations. The magnitude of the estimated total release rate (Stot) from ship coatings to the marine environment does not affect the calculated residence times of the biocides since the steady-state quantity Mtot in eq 1 is directly proportional to Stot. Nevertheless, realistic approximations for release rates from ships were used in order to be able to compare calculated concentrations with concentrations measured in the environment. Antifouling biocides are released to the marine biosphere from biocide factories, during transport of biocides, during application of antifouling paints, and in the process of removal of antifouling paints from ships in drydocks. Not all of these release pathways are negligible, but they can be technically controlled to a high degree. On the contrary, release from coatings of ships in service is desired for the purpose of fouling control. Because of its essential role in fouling protection of ships, only this release is being considered in the present residence time study. This must be kept in mind for the model evaluation described below. For TBT and the organic biocides, a mean release rate of 3.5 µg cm-2 d-1 was assumed for traveling vessels, as was estimated for the release of TBT from state-of-the-art selfpolishing copolymer coatings (17). For copper, experts from an European Community project (8) used release rates of 50 µg cm-2 d-1. For stationary vessels in harbors, a reducing factor of 10 was applied as reported in ref 17. For the calculation of the total release rates, a total surface area of the world commercial fleet of 3 × 107 m2 was estimated
TABLE 2. Geometric Mean of Particle-Water Sorption Constants in Water (Kpw) and Degradation Rates in Water (kw) As Well as In Sediment (ks) with the Number of Data Points (n) Used for Distribution Estimations, the 5th Percentile (q0.05), and the 95th Percentile (q0.95) of the Estimated Log-normal Distributions substance TBT+/TBTOH Cu2+ Irgarol DCOI
parameter
Kpw kw ks Kpw Kpw kw ks Kpw kw ks
n
mean
q0.05
13 13 4 4 1d 1d 1d 2d 5 1d
4.7 × 0.061 0.24 45.7 × 103 3.1 × 103 0.0054a 0.086 1.1 × 103 b 0.44 6.1 × 103 c
0.28 × 0.042 0.062 10.3 × 103 0.36 × 103 0.0024 0.038 0.13 × 103 0.07 2.7 × 103
103
q0.95 103
79 × 0.088 0.90 204 × 103 26 × 103 0.012 0.19 9.3 × 103 2.7 14 × 103 103
unit kg-1
L d-1 yr-1 L kg-1 L kg-1 d-1 yr-1 L kg-1 d-1 d-1
source(s)
49-53 54-57 58-61 62-65 66 calcd from ref 40, as in ref 26 67, as in ref 27 68 and 69, as in refs 28 and 29 70, as in ref 9 71, as in refs 28 and 72
a Only used for surface water compartments. b Geometric mean of the two values was used as estimator. c In ref 71, 4.8-0.7% of the applied radioactivity as DCOI in aerobic water from days 15-30 was not identified and around 60% was not extractable from the sediment. Despite the high degradation rate claimed by the manufacturers in ref 39, 4.8 and 2.6% of radioactivity was found extractable as DCOI from anaerobic sediments on days 14 and 61, respectively. d No maximum likelihood estimates were carried out.
from the ships with gross tonnage >300 in 1998 (18) and average wetted hull areas of eight types of ships (19) as shown in the Supporting Information. On the basis of a very basic understanding of global traffic patterns, it was assumed that at any time 10% of this wetted hull surface area is stationary in harbors, 1% is traveling in estuaries with harbors, 64% is traveling above the continental shelf, and 25% is traveling in the open sea. For the case of all ships using TBT, the global release estimation from antifouling paints results to 1900 t yr-1. If all ships use copper, a release from their hulls of 27 × 103 t yr-1 is estimated. Because of the large nonantifouling inputs of copper into the marine environment, these were estimated from literature values. On this basis, the potential contribution of copper antifouling paints on ships to the total copper input was additionally evaluated. These background inputs where only used for the model evaluation. The residence times (Table 4 and Figure 3) used for the comparison of the biocides were calculated with the release pattern only from ship hulls. For the total copper flux from rivers into the oceans, a geometric mean of a value from a recent study (20) and three older values (21-23) gave 5 × 105 t yr-1. One percent of this amount was apportioned as input into the wt compartment, since the river water Fwr,wt of the model estuaries amounts w to about 1% of the total riverine water as shown in the Supporting Information. Ninety-nine percent of the copper flux from rivers was apportioned as input into the shelf water compartment. The total atmospheric input into the oceans is estimated by a geometric mean of values from refs 23 and 24 to be 51 × 103 t yr-1. This amount was split as input into the surface water compartments according to their surface areas (Table 1). All modeled copper fluxes into the system are given in Table 3. Degradation and Partitioning Data. For the biocidespecific parameters kw (degradation rate constant in water), ks (degradation rate constant in sediment), and Kpw (partitioning constant between suspended particulate matter and water), strongly varying data were encountered in the literature. These parameters were described probabilistically using log-normal distributions (Table 2). From Kpw, kw, and ks values for tributyltin and Kpw values for copper from literature data, it was possible to do maximum likelihood estimations of log-normal distributions (25). The variation was attributed mainly to variability of the observing systems, which corresponds to but is not equivalent to the variability in the environment. For the degradation of tributyltin in sediments, rate constants estimated from depth profiles were used as well as rate constants from sediment microcosms.
For Irgarol and DCOI, mainly values from commercially performed studies as cited in administrative reports (9, 2630) were used as estimators of geometric means of the corresponding log-normal distributions. Except for the degradation of DCOI in water, where a maximum likelihood estimate could be performed, the geometric standard deviations expressing the parameter variability for these two organic biocides were estimated as intermediate between the geometric standard deviations for tributyltin (Kpw, kw, and ks) and Cu (only Kpw) as shown in Table 2. The predominant species of tributyltin (TBT) in seawater will be the neutral TBTOH since its acidity constant pKa(TBT+) is 6.25 at standard conditions (31). The vapor pressure of TBTOH was estimated with the MPBP software v1.30 of Syracuse Research Center to be 2 × 10-3 Pa, but it is not included in the group of volatile organotin compounds given in a recent study (32). The vapor pressures of the two other organic biocides are even lower. Thus, except for copper, none of the substances meet the criterion for being involatile given by Mackay et al. (14) (vapor pressure 0.45 µm to a significant degree in the particulate-rich harbor, estuary, and shelf compartments (Figure S.1 in Supporting Information). As a consequence, scavenging is significant and the uncertainty in sorption to particles leads to a much higher variation of modeled concentrations than for TBT. Nonetheless, the estimated concentration ranges in the generic harbor, harbor estuary, and shelf compartments are in reasonable agreement with data from the literature. The comparison of measured and calculated copper concentrations for harbor, estuarine, and shelf water shows that the uncertainty in the sorption parameter Kpw results in similar variation in concentration values as the temporal and spatial variability of the real world (Figure 2c). The model deviation in epipelagic and pelagic water could be caused by several reasons: (a) the particle export to bottom sediments could be underestimated since the particle flows are quite uncertain; (b) the sorption constant for deep sea particle sorption of copper might be too low since it has been derived for surface water and the chemistry of the particles changes through their degradation and through redissolution of calcium carbonate during passage down the water column (34); and most importantly, (c) the oceans did not have enough time to reach steady-state concentrations with the increased copper inputs (35) since industrialization started. A dynamic analysis of the model (data not shown) suggested that it would take the system more than 20 000 yr to approximate the steady-state concentrations in the large sediment compartments only half the way in the case of copper. The overestimation of copper concentrations in pelagic sediments amounts to almost 2 orders of magnitude while it is only about 1 order of magnitude for pelagic water. This indicates, contrary to hypothesis b given above, that copper
sorption to deep sea particles is lower than to surface water particles. Except for the TBT concentrations on the continental shelf, the deviations are small considering the low spatial and temporal resolution of the model. Remaining Uncertainties. (a) Model Parameters and Model Structure. For the size of the global harbor basin compartment, it was impossible to come up with an exact estimation. The current estimation based on the number of major ports and an estimated mean harbor basin size might deviate within the order of magnitude. Similarly, the apportionment of the underwater hull area of the world commercial fleet to the different surface water compartments is only a rough approximation as mentioned above. The estimation of the total surface area itself is based on detailed data as shown in the Supporting Information (Table S.2) and is therefore not significantly adding to the uncertainty. Uncertainties due to the structure of the model primarily concern the steady-state character of the model. The periodical tidal flow in estuaries is generally represented in the model by the water exchange rates of the estuary compartment with the shelf sea. Again, this will be quite different for each individual estuary, so no conclusion about the concentration in a single estuary should be drawn. Fluctuations in biocide input will take place on the time scale of years. These input fluctuations are relevant for the biocides with residence times of the same order of magnitude or longer. Especially for those biocides (e.g., Cu in the chosen subset), a dynamic model would be preferable. Another deficiency inherent in the model structure is that only substance-specific parameters were modeled with probability distributions. A probability distribution for water residence times in harbor basins and for the particle settlement rates in all water compartments would lead to a much more realistic depiction of variability in the model output. (b) Biocide Parameters. The dependency of degradation rates on temperature, redox potential, and concentration and activity of biomass could not be included in the model, because the available experimental data was not sufficiently detailed. Concerning the sorption constants to particles Kpw, the difference between terrigeneous particles (predominant in estuaries and shelf) and biogenic particles (predominant in epipelagic and pelagic water) was neglected because no appropriate data were available. If the model structure is reasonable, as suggested by the satisfying simulation of the copper concentrations, the TBT degradation rates must largely overestimate the real degradation taking place in the environment. Possible reasons could be the lack of correcting the degradation rates for temperature, the much lower metabolic activity of plankton in the shelf compartment, or even a systematic error of previous degradation observations due to uptake of TBT into algae during the experiments (36). Recent measurements suggest that the total degradation of TBT in North Sea water is slower by a factor 7 than the value from Table 2. This would partly explain the model deviation. Another source of errors is the assumption that the sediment compartment is well-mixed. Samples are likely to be taken from the uppermost part of the sediment profile, whereas the modeled concenVOL. 36, NO. 7, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1543
TABLE 4. Geometric Mean Values for Water Residence Times (τw) and Overall Residence Times (τ) in the System substance
τw
τ
unit
DCOI TBT+/TBTOH Irgarol Cu2+
4.55 16.5 10.2 8350
4.76 16.9 10.2 42700
d d yr yr
tration represents all the sediment up to a depth of 40 cm where it has aged considerably. The fast disappearance of DCOI and other isothiazolones from aqueous biotic systems seems to be caused by its reaction with thiol groups of intracellular nucleophiles such as glutathione or of proteins (37). The high amount of inextractable residues in DCOI degradation studies, where organic matter is present, supports the hypothesis of covalent binding. However, reactions of isothiazolones with thioles and other nucleophiles have been shown to be reversible (38), and the presence of DCOI found in an anaerobic sediment-seawater system after 61 d (39) suggests that DCOI might be remobilized from sediments. Photodegradation of Irgarol is slow, the suggested decay rates in surface water (40) and sediment (27) have been extrapolated from observation times of 15 and 30 d, respectively. In the sediment, 99.3% of the radioactivity was still found as Irgarol after the observation time, which illustrates the uncertainty of this degradation rate. Biological degradation studies of TBT and DCOI (41, 42) have shown that in filtered seawater (1.6 µm) the biological activity decreases much slower than would be expected from the cited chemical studies. A further improvement of the quantification of spatiotemporal ranges should also take transformation products into account. This can be done by explicit modeling of their fate (43, 44) but also by means of biological techniques of degradation monitoring (41, 42), which integrate the degradation of released substances and their transformation products simply by means of the measurement method. All the compromises made in order to obtain an operational model produce uncertainties of the model results. Therefore, environmental concentrations predicted by this model should not be used in the context of conventional risk analysis, i.e., a comparison of calculated concentrations with some kind of unacceptable environmental concentrations derived from toxicity tests. However, the model provides a couple of relevant insights into the fate of these substances which are summarized below. Residence Times and Their Implications. For each of the four biocides, the mean residence times in the marine water (τw) and the mean overall residence times τ are given in Table 4. For copper, Whitfield and Turner (45) give a mean ocean residence time (MORT) of 4.4 × 103 yr based on the relatively low copper input estimates from Collier and Edmond (22) (see input estimations above). This is in approximate agreement with the geometric mean water residence time of 8.4 × 103 yr calculated here. The difference between total residence time and water residence time is much smaller than 1% for Irgarol and DCOI because they are mainly present in the water compartments. While all sediment compartments surprisingly only add about 2.5% to the residence time of TBT, this difference is much more significant for copper. In fact, for copper it is only a little bit smaller than the turnover time of the pelagic sediment. The geometric mean values of calculated residence times in Table 4 allow a ranking of the biocides. The high variability of data on the degradation of DCOI (Figure 3), however, leads 1544
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 7, 2002
FIGURE 3. Histograms of total residence times in the system according to Monte Carlo simulations. to some uncertainty wether this favorable evaluation of its spatiotemporal range will be kept in the future. The relevance of the very long residence times of copper in the system for environmentally conscious selection of antifouling biocides is limited because only about 5% of the estimated total copper input into the sea seem to stem from antifouling paints on commercial ships. Independent from the comparison of biocides, the model results indicate that copper concentrations in pelagic sediments could increase by more than 2 orders of magnitude if current input rates remain constant. This model result is quite robust since it is dominated by only two quantities: the global input rate and the global removal by pelagic sediment burial. There are two main results of the present study: an indication of a risk due to current inputs of copper into the pelagic sediments, which requires further investigation and discussion, and an improved quantification of the spatiotemporal range of antifouling biocides, which can be applied to other antifouling biocides as well. This quantification should not be used as a single basis for biocide choices. It is especially suitable for use in risk characterizations by multi-criterial methods such as the classification of “PBT” substances which are persistent, bioaccumulating, and toxic or like the method previously presented (1).
Acknowledgments I would like to thank B. Escher, T. Frische, O. MosbachSchulz, U. Mu ¨ ller-Herold, M. Scheringer, and P. Stepnowski for valuable comments on earlier versions of the manuscript. Special thanks go to my advisor, Prof. Bernd Jastorff, and for financial support to the University of Bremen.
Supporting Information Available Additional derivations of model parameters, mass balance equations, tables of environmental water, and sediment concentrations of TBT and copper and histograms of dissolved fractions. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) (2) (3) (4) (5) (6)
Ranke, J.; Jastorff, B. Environ. Sci. Pollut. Res. 2000, 7, 105-114. Stephenson, M. E. Ecotoxicol. Environ. Saf. 1977, 1, 39-48. Klo¨pffer, W. Environ. Sci. Pollut. Res. 1994, 1, 107-116. Scheringer, M. Environ. Sci. Technol. 1996, 30, 1652-1659. Mu ¨ ller-Herold, U.; Nickel, G. Ecol. Modell. 2000, 126, 191-200. Bennett, D. H.; McKone, T.; Matthies, M.; Kastenberg, W. E. Environ. Sci. Technol. 1998, 32, 4023-4030. (7) Beyer, A.; Mackay, D.; Matthies, M.; Wania, F.; Webster, E. Environ. Sci. Technol. 2000, 34, 699-703. (8) Utilisation of more ‘environmentally friendly’ antifouling products; EC Project 96/559/3040/DEB/E2; European Council of Paint, Printing Ink and Artists' Colours Industry: 1999.
(9) Madsen, T.; Samsøe-Petersen, L.; Gustavson, K.; Rasmussen, D. Ecotoxicological Assessment of Antifouling Biocides and Nonbiocidal Antifouling Paints; Environmental Project 531; Miljøstyrelsen, Danish EPA, DHI Water and Environment: Copenhagen, 2000. (10) Boxall, A.; Comber, S.; Conrad, A.; Howcroft, J.; Zaman, N. Mar. Pollut. Bull. 2000, 40, 898-905. (11) Wania, F.; Mackay, D.; Li, Y.-F.; Bidleman, T. F.; Strand, A. Environ. Toxicol. Chem. 1999, 18, 1390-1399. (12) Berding, V.; Schwartz, S.; Matthies, M. Environ. Sci. Pollut. Res. 2000, 7, 147-158. (13) Schwartz, S.; Berding, V.; Matthies, M. Chemosphere 2000, 41, 671-679. (14) Mackay, D.; di Guardo, A.; Paterson, S.; Kicsi, G.; Cowan, C. E. Environ. Toxicol. Chem. 1996, 15, 1618-1626. (15) Broecker, W. S. Quat. Res. 1971, 1, 188-207. (16) Lalli, C. M.; Parsons, T. R. Biological Oceanography: An Introduction, 1st ed.; Pergamon Press: Oxford, 1993. (17) Schatzberg, P. Measurement and significance of the release rate for tributyltin. In Organotin: Environmental Fate and Effects; Champ, M. A., Seligman, P. F., Eds.; Chapman and Hall: London, 1996. (18) Shipping Statistics Yearbook; Institute of Shipping Economics and Logistics: 1998. (19) Isensee, J.; Watermann, B.; Berger, H.-D. Dtsch. Hydrogr. Z. 1994, 46, 355-365. (20) Poulton, S. W.; Raiswell, R. Mar. Chem. 2000, 72, 17-31. (21) Broecker, W. S.; Peng, T.-H. Tracers in the Sea; Lamont-Doherty Geological Observatory: Palisades, 1982. (22) Collier, R.; Edmond, J. Prog. Oceanogr. 1984, 13, 113-199. (23) Mackenzie, F. T.; Lantzy, R. J.; Paterson, V. J. Int. Assoc. Math. Geol. 1979, 11, 99-142. (24) Nriagu, J. O.; Pacyna, J. M. Nature 1988, 333, 134-139. (25) Morgan, M. G.; Henrion, M. Uncertainty; Cambridge University Press: Cambridge, 1990. (26) Bard, J.; Pedersen, A. Ecotoxicological evaluation of the antifouling compound 2-(tert-butylamino)-4-(cyclopropylamino)6-(methylthio)-1,3,5-triazine Irgarol; Technical Report; KEMI (National Chemicals Inspectorate Sweden): Solna, Sweden, 1992. (27) Bard, J. Supplement 1 to the ecotoxicological evaluation of the antifouling compound 2-(tert-butylamino)-4-(cyclopropylamino)6-(methylthio)-1,3,5-triazine Irgarol; Technical Report; KEMI (National Chemicals Inspectorate Sweden): Solna, Sweden, 1992. (28) Bjo¨rk, M.; Karlsson, S. Ecotoxicological evaluation of 4,5-dichloro2-n-octyl-4-isothiazolone-3-one (RH-287, RH-5287); Technical Report; KEMI (National Chemicals Inspectorate Sweden): Solna, Sweden, 1992. (29) Dryselius, E. Ecotoxicological evaluation of the biocide 4,5dichloro-2-n-octyl-4-isothiazolone-3-one (RH-287, RH-5287): Supplementary documentation; Technical Report; KEMI (National Chemicals Inspectorate Sweden): Solna, Sweden, 1998. (30) Shade, W. D.; Hurt, S. S.; Jacobson, A. H.; Reinert, K. H. Ecological risk assessment of a novel marine antifoulant; Standard Technical Publication; American Society for Testing and Materials: Philadelphia, 1993. (31) Arnold, C. G.; Weidenhaupt, A.; David, M. M.; Mu ¨ ller, S. R.; Haderlein, S. B.; Schwarzenbach, R. P. Environ. Sci. Technol. 1997, 31, 2596-2602. (32) Amouroux, D.; Tessier, E.; Donard, O. F. X. Environ. Sci. Technol. 2000, 34, 988-995. (33) Ranke, J.; Stock, F.; Grimmer, C. Database UFT_SAR. http:// eckehaat.uft.uni-bremen.de/UFT_SAR. (34) Ricken, W. Sedimentation as a Three-Component System: Organic Carbon, Carbonate, Noncarbonate; Springer: Berlin, 1993. (35) Nriagu, J. O. Science 1996, 272, 223-224. (36) Brandsch, R. University of Bremen, personal communication, 2001. (37) Collier, P. J.; Ramsey, A.; Waigh, R. D.; Douglas, K. T.; Austin, P.; Gilbert, P. J. Appl. Bacteriol. 1990, 69, 578-584. (38) Crow, W. D.; Leonard, N. J. J. Org. Chem. 1965, 30, 2660-2665. (39) Kesterson, A.; Atkins, R. Supplemental study on the anaerobic aquatic metabolism of [13/14C]RH-5287; PTRL Report 1313; Rohm & Haas Technical Report 34-92-46; Pharmacology and Toxicology Research Laboratory: Richmond, KY, 1992. (40) Doyle, R. C. Photodegradation of 14C-Irgarol 1051 in Sterile Seawater; Technical Report; IIT Research Institute: 1991. (41) Callow, M. E.; Finlay, J. A. Biofouling 1995, 9, 153-165. (42) Callow, M. E.; Willingham, G. L. Biofouling 1996, 10, 239-249.
(43) Fenner, K.; Scheringer, M.; Hungerbu ¨ hler, K. Environ. Sci. Technol. 2000, 34, 3809-3817. (44) Quartier, R.; Mu ¨ ller-Herold, U. Ecol. Modell. 2000, 135, 187198. (45) Whitfield, M.; Turner, D. R. The role of particles in regulating the composition of seawater. In Aquatic Surface Chemistry; Stumm, W., Ed.; John Wiley and Sons: New York, 1987; Chapter 17, pp 457-493. (46) Sverdrup, H. U.; Johnson, M. W.; Fleming, R. H. The Oceans: Their Physics, Chemistry, and General Biology; Prentice Hall: Englewood Cliffs, NJ, 1942. (47) Neshyba, S. Oceanography: Perspectives on Fluid Earth; John Wiley and Sons: New York, 1987. (48) Kranck, K. Sedimentation processes in the sea. In Handbook of Environmental Chemistry, Vol. 2A.; Hutzinger, O., Ed.; Springer: Berlin, 1980. (49) Maguire, R. J.; Tkacz, R. J. J. Agric. Food Chem. 1985, 33, 947953. (50) Harris, J. R. W.; Cleary, J. J. Particle-water partitioning and organotin dispersal in an estuary. In Proceedings of the Oceans, Vol. 4; Marine Technology Society: New York, 1987. (51) Valkirs, A. O.; Seligman, P. F.; Lee, R. F. Butyltin partitioning in marine waters and sediments. In Proceedings of the Oceans, Vol. 4; Marine Technology Society: Washington, DC, 1986. (52) Valkirs, A. O.; Stallard, M. O.; Seligman, P. F. Butyltin partitioning in marine waters. In Proceedings of the Oceans, Vol. 4; Marine Technology Society: New York, 1987. (53) Adelman, D.; Hinga, K. R.; Pilson, M. E. Q. Environ. Sci. Technol. 1990, 24, 1027-1032. (54) Harino, H.; Fukushima, M.; Kurokawa, Y.; Kawai, S. Environ. Pollut. 1997, 98, 157-162. (55) Harino, H.; Fukushima, M.; Kurokawa, Y.; Kawai, S. Environ. Pollut. 1997, 98, 163-167. (56) Seligman, P. F.; Valkirs, A. O.; Lee, R. F. Environ. Sci. Technol. 1986, 20, 1229-1235. (57) Seligman, P. F.; Adema, C. M.; Grovhoug, J.; Fransham, R. L.; Valkirs, A. O.; Stang, P. M. Persistence and fate of tributyltin in aquatic ecosystems. In Organotin: Environmental Fate and Effects; Champ, M. A., Seligman, P. F., Eds.; Chapman and Hall: London, 1996; Chapter 21, pp 429-458. (58) de Mora, S. J.; King, N. G.; Miller, M. C. Environ. Technol. Lett. 1989, 10, 901-908. (59) de Mora, S. J.; Stewart, C.; Phillips, D. Mar. Pollut. Bull. 1995, 30, 50-57. (60) Dowson, P. H.; Bubb, J. M.; Lester, J. N. Estuarine Coastal Shelf Sci. 1996, 42, 551-562. (61) Stewart, C.; de Mora, S. J. Environ. Technol. 1990, 11, 565-570. (62) Grassi, M. T.; Shi, B.; Allen, H. Colloids Surf. A 1997, 120, 199203. (63) Pohl, C.; Hennings, U. Mar. Chem. 1999, 65, 41-53. (64) Helmers, E. Mar. Chem. 1996, 53, 51-67. (65) Balls, P. W. Neth. J. Sea Res. 1989, 23, 7-14. (66) Tolosa, I.; Readman, J. W.; Blaevoet, A.; Ghilini, S.; Bartocci, J.; Horvat, M. Mar. Pollut. Bull. 1996, 32, 335-341. (67) Schmidt, J. M.; Head, L. L. Aerobic aquatic metabolism of 14CIrgarol 1051; ABC Final Report 38889; Analytical Bio-Chemistry Laboratories, Inc.: Columbia, MO, 1991. (68) Warren. Determination of adsorption/desorption constants of C-9211/14C-RH-5287; ABC Report 32116; Rohm & Haas Technical Report 310-86-31; Analytical Bio-Chemistry Laboratories, Inc.: Columbia, MO, 1986. (69) Gene, L.; Olson, M. S.; Lawrence, L. J. Soil adsorption/desorption of [14C]RH-5287 by the batch equilibrium method; PTRL Report 1246; Rohm & Haas Technical Report 34-90-28; Pharmacology and Toxicology Research Laboratory: Richmond, KY, 1991. (70) Jacobson, A.; Kramer, V. Additional infromation on watersediment partitioning and half-life of DCOI in a harbor; Note prepared for VKI, 1999. (71) Lawrence, L. J.; et al. Aerobic aquatic metabolism of [13/14C]RH-5287; PTRL Report 1291; Rohm & Haas Technical Report 34-91-01; Pharmacology and Toxicology Research Laboratory: Richmond, KY, 1991. (72) Willingham, G. L.; Jacobson, A. H. Am. Chem. Soc. Symp. Ser. 1996, No. 640, 224-233.
Received for review July 23, 2001. Revised manuscript received December 10, 2001. Accepted December 21, 2001. ES011155P
VOL. 36, NO. 7, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1545
