1、American Petroleum Ins ti tute Toxicity Bioassays on Dispersed Oil in the North Sea: June 1996 Field Trials Regulatory Analysis and Scientific Affairs PUBLICATION NUMBER DR 342 JUNE 2002 Toxicity Bioassays on Dispersed Oil in the North Sea: June 1996 Field Trials Regulatory Analysis and Scientific A
2、ffairs API PUBLICATION NUMBER DR 342 JUNE 2002 GINA M. COELHO DON V. AURAND ECOSYSTEM MANAGEMENT to determine the degree of toxicity in the regions of the slick that had the highest oil or dispersed oil concentrations, and to determine how this toxicity varied over depth; to collect at-sea toxicity
3、data for future comparisons to laboratov test results; and to develop recommendations concerning protocols appropriate for defining ecological impacts in the water column through a combination of field and laboratory data. o o o Four different toxicity tests were chosen for use during the sea trials
4、. Two of the tests, a bioluminescent bacterial assay (referred to as the “microbial test”) and a rotifer acute toxicity test (referred to as the “rotifer test”) are commercially available, rapid toxicity test kits. They were selected because of their publicized “ease of use” in the field (Le., at se
5、a). The other tests, oyster embryo development and copepod lethality tests, utilize a traditional 48-hour acute exposure. The microbial test showed statistically significant effects for dispersed-oil water samples at both 1 meter (m) and 5 m depths. The corresponding total petroleum hydrocarbon valu
6、es determined by gas chromatography (TPH-GC) beneath the chemically dispersed oil (CDO) slick were ES- 1 approximately 22 ppm and 2 ppm, respectively. No significant effects were observed for samples taken beneath the untreated oil slick, in which TPH-GC values were reported as 1.4 ppm and 0.5 ppm f
7、or 1 m and 5 m depths, respectively. The test results were consistent with the results of the copepod tests and showed a rapid decrease in toxicity over increasing depth beneath the slick. The test is rapid, easy to learn, inexpensive (except for the initial purchase of the equipment), easi!y transp
8、orted, and requires little work space. There was no difficulty conducting the tests at sea under the conditions encountered during this shdy. Preliminary results were available within one hour during this study, and it is the only test evaluated during this study which offers the potential of near r
9、eal-time information. Before it can be used in any real-time capacity for oil spill response, however, the test must be effectively calibrated against the existing laboratory data to ensure that the results can be properly interpreted. Survival was low in the reference, control, and test copepod bio
10、assays. This is attributed to the poor shipboard conditions that caused undue stress on the animals. When ANOVA techniques were utilized, the only significant difference noted from the copepod 48-hour lethality test was for the sample water taken at 1 m depth beneath the CDO slick. The TPH-GC value
11、for this sample was 22 mg/L. Although there was a trend that indicated lower survival in the 5 m depth CDO sample (2 ma) versus the seawater control, the difference was not statistically significant with the low number of replicates and organism numbers that were used during the bioassay. Neither th
12、e rotifer acute toxicity test nor oyster embryo tests were successfid, but the microbial test and copepod toxicity test results allowed many of the project objectives to be met. Even if successful, neither the rotifer test nor the oyster bioassay would have had application as a real- time monitoring
13、 assay for spill countermeasure assessment because of their time requirements. In summary, the study was successful in collecting data to use in comparison with laboratory results, and supports the conventional belief that the zone of biological effects within a dispersed oil plume is limited both s
14、patially and temporally. ES-2 Section 1 INTRODUCTION BACKGROUND It is generally acknowledged that the use of dispersants in oil spill response can provide protection for shoreline resources, but at the expense of increasing exposure to oil for organisms in the water column (or on the bottom, in shal
15、low water). If this is the case, then it becomes very important to define the level of effect anticipated in the potentially affected habitats, in order to identif+ the best course of action in terms of planning for response. For the purposes of this discussion, areas of concern for both dispersed a
16、nd undispersed oil are divided into two categories: 1) shoreline (intertidal) and 2) nearshore subtidal and water column. - - - Shoreline (Intertidal) and Nearshore Subtidal Effects Shoreline effects of oil spills have been well studied and are generally predictable if the climate, hydrographic cond
17、itions, and habitat type are known. The literature on this habitat is extensive for untreated oil, and is based on experimental work and on observations of effects at actual oil spills (see National Research Council (NRC), 1985 or GESAMP, 1993 for summaries). For dispersed oil, there is little infor
18、mation fiom actual spills, although Lunel (1997) attempted to estimate the shoreline benefits accrued by oil dispersed at sea during the Sea Empress spill in Milford Haven (United Kingdom). Summaries of real response operations in which oil was successfully dispersed have not made note of adverse sh
19、oreline impacts (see Table in Appendix in Lewis and Aurand, 1997). Although this lack of evidence on shoreline effects does not prove that there have been no effects, it does provide some indication that such impacts have not been severe. The most meaningful field experiments defining the consequenc
20、es of dispersant use in intertidal and nearshore areas were a series of three experiments implemented in the early 1980s. The first was conducted at a site on Baffin Island in the Canadian Arctic (BIOS Study; Sergy, 1985), the second, at a small embayment near Searsport, Maine (Gilfillan et al., 198
21、3, 1985), and the third, in a tropical embayment containing mangroves, sea grass beds, and corais in Panama (TROPICS 1-1 Study; Ballou et al., 1989; Dodge et al., 1995). In all three studies, both pre-dispersed and untreated crude oil were released slightly offshore or subtidally, and ailowed to str
22、and on the intertidal zone. In all studies, significant shoreline contamination was seen in the high intertidal zone with the untreated oil, but not with the pre-dispersed oil. The ecological effects of this untreated oil contamination varied widely, depending upon the intertidal species, but were s
23、evere in the mangrove community in Panama. Initial impacts in the water column and in benthic organisms were higher in the presence of the pre-dispersed oil, but the effect was temporary and conditions returned to normal within one year (in the arctic) to several years (in Panama). The Searsport stu
24、dy was only short-term. Monitoring at the BIOS site continued for several years and documented long-term contamination of the sediments from erosion of beach sediments at the untreated oil site. In Panama, the TROPICS sites were monitored several times during the two years immediately following the
25、experimental spill, and then revisited at the ten year point (Dodge et al., 1995). At the end of two years, the authors concluded that the undispersed oil had severely damaged or killed the mangrove trees in the test area, while there was little or no damage to trees fi-om the pre-dispersed oil. Sea
26、 grasses were not affected by either treatment, but invertebrates present in the sea grass beds were severely affected initially by the pre-dispersed oil. Recovery appeared complete after two years. Corals were also significantly affected by the pre-dispersed oil in the first two years, but not by t
27、he untreated oil. The evaluation at ten years confirmed the significant effect of the untreated oil on mangroves, since the area now contained even fewer living trees (approximately half of the original population). In contrast, no explicit mortality of trees exposed to the pre-dispersed crude oil w
28、as observed. The corals, which had shown an effect from the pre-dispersed oil at the two-year point, had now recovered and no significant difference existed between sites. Benthic habitats immediately adjacent to oiled shoreline may be affected by erosion of contaminated sediment from the intertidal
29、 zone. Contamination of intertidal sediment is reduced or eliminated by the use of dispersants. Effects on birds and marine mammals are well understood, but are more difficult to quantiQ except when connected to specific geographic locations (e.g., haul-out areas for marine mammals or rookery sites
30、for nesting birds). It is generally accepted that reducing oil on the water surface will protect marine mammals and birds 1-2 at sea. Preventing oil from stranding in nesting or haul-out areas is an obvious benefit of dispersant use. Water Column Effects Based on our knowledge of the fate and behavi
31、or of oil, water column effects are much less significant than intertidal effects, except under unique conditions; therefore, they have been much less studied (NRC, 1989). There have been a number of field studies to examine the _- physical fate of dispersed versus untreated oil at sea, and this pro
32、cess is well-understood. In the late 1970s, the U.S. Environmental Protection Agency (EPA) and the American Petroleum Institute (API) conducted at-sea studies on both the East and West Coasts of the United States to evaluate dispersant performance, the effects of various application methods, and the
33、 fate of both treated and untreated oil (McAuliffe et al., 1980; McAuliffe et al., 198 1). Throughout the 1980s and 1990s, sea trials have continued in the North Sea by researchers from both Norway and the United Kingdom (UK) (for examples of the most recent work see Lunel, 1994, 1995; Lewis and Dav
34、ies, 1996; Lewis et al. 1995; and Brandvik et al. 1996a,b). None of these studies focused on long-term biological effects in the water column, although they provided data on water column concentrations of hydrocarbons, which can be used to make estimates of potential effects. Soule and Oguri (1 985)
35、 did conduct an evaluation of short-term water column biological effects in conjunction with the 1979 EPNAPI experiment off the U.S. East Coast. The authors collected water samples from beneath dispersed and undispersed oil slicks (Murban and La Rosa crude oils) at depths of one and three meters. Us
36、ing net-collected zooplankton samples from uncontaminated control areas, they conducted shipboard toxicity tests in ambient temperature water baths for either three or twelve hours, and then used a vital stain to identi living zooplankton in test and control samples. The distribution of zooplankters
37、 was highly variable, and the survival patterns were inconsistent, leading to inconclusive results. However, the authors reported that there was a tendency for higher mortality beneath the dispersed oil slicks, based on survival of adults of the three most common species, all calanoid copepods (Cent
38、ropages typicus, Paracalanus parvus, and Acartia tonsa). Of the three, A. tonsa was the most resistant, not showing any statistically significant mortality. Some significant results were obtained for the other two species. Chemical data were available from water samples aicen at epproximatey the 1-3
39、 same time (approximately one-half hour after release and dispersion), with no “oil” concentrations reported in excess of 0.5 mg/L (pprn). This value is lower than any of the LC50 (96 or 72-hour) values reported for copepods in NRC (1985, Table 5-5). Given that this exposure concentration was mainta
40、ined for only a brief time in companson to the laboratory tests, the basis for the observed effects is unclear. Baker (1 993) used the ppm-hour concept (Anderson et al., 1984; 1987), in which the units are the product of the measured mean concentration and time. She estimated that the lowest reporte
41、d LC50 value was about 8 ppm-hour, a value much higher than the apparent exposure in the Soule and Oguri (1985) study. It is important to note that these early studies that examined water column effects did not utilize highly sophisticated means of chemical analysis (such as gas chromatography). Man
42、y of the early studies relied on infrared or fluorescence methods to quanti Celewycz and Wertheimer, 1996). There are no reported water column studies to date at spills where dispersants were used. In the case of the MV Bruer spill off the Shetland coast, the severe weather and chemical and physical
43、 properties of the spilled oil (Gullfaks crude) essentially led to a total, natural dispersion of the approximately 85,000 tonnes of cargo, along with a small volume of fuel oil. In the three months after the incident, over 800 water samples were taken for hydrocarbon analysis. Concentrations of tot
44、al hydrocarbons in the water column within 100 to 200 yards of the tanker were as high as 10,000 ppm (a 1% oil-in-water dispersion). However, by the time the plume was 2 km away from the wreck, concentrations were approximately 100 ppm, a level that was observed as far as 24 km north of the wreck (T
45、homas and Lunel, 1993). These levels were apparently maintained over much of the six days it took for the wreck to break up, and then declined rapidly. Within another three days, concentrations in Garth Ness (the small embayment where the wreck grounded) were only 25 ppm (Thomas and Lunel, 1993). Ri
46、tchie and OSullivan (1994) 1-4 characterized concentrations in the water column “near” the tanker as “in the range of several hundred ppm soon after the grounding.” They also summarized concentrations over the next few days as around 50 ppm in areas affected by the spill. Obviously, concentrations w
47、ere highly variable in both time and space but it is clear that dilution was quite rapid after the release ended and as the plume moved away from the site. Laboratory toxicity tests are the most frequently encountered data when planners attempt to evaluate potential risks and benefits of dispersant
48、use. One must keep in mind that crude oil is a complex mixture of thousands of compounds of varying volatility and a wide range of solubilities in water, and that the addition of dispersants can create a multi-phase system (e.g., water, oil, dispersant, dispersed oil droplets), all of which have a d
49、ifferent affinity for the compounds present in the oil. Because of these characteristics of oil and dispersant interaction, it should not be surprising that test results have been greatly influenced by factors such as the nature and degree of agitation, the type of test container, whether the test containers are open or closed, the oi1:water ratio, the source of the crude oil, or its weathering state, to name only a few (Aurand, 1995). Laboratory toxicity tests do not relate directly to field conditions unless the exposure regimes are comparable. Standard acute toxicity tests with a stat
