An Oil Spill Affecting Coral Reefs And Mangroves On The Caribbean Coast Of Panama

John D. Cubit, Charles D. Getter¹, Jeremy B. C. Jackson, Stephen D. Garrity,² Hugh M. Caffey, Ricardo C. Thompson, Ernesto Weil, Michael J. Marshall

Smithsonian Tropical Research Institute
Apartado 2072, Balboa, Republic of Panama

¹Also, 915 Academy Street N.E., Salem, Oregon 97303 2.
²Also, Department of Zoology, University of Rhode Island, Kingston, Rhode Island 02881

ABSTRACT: In April 1986, more than 50,000 barrels of medium-weight crude oil were spilled into the largest complex of coral reefs and mangroves on the central Caribbean coast of Panama. Considerable amounts of oil came ashore at Punta Galeta, where a long-term environmental sciences program of the Smithsonian Tropical Research institute provided extensive baseline information for investigating the effects of the oil spill. Immediate mortality was most apparent for organisms living at the seaward edge of the reef flats and on drying substrata above mean water level. By June 1986, a number of reef species were reduced in abundance, and a bloom of microalgae grew over much of the reef flat that had been directly exposed to the oil at low tide. The abundances of some fast-growing algae appeared to have recovered by September 1986, but the slower-growing corals, zoanthids, and calcareous algae were at the lowest abundances recorded.

Defoliation and mortality of mangroves, particularly Rhizophora mangle, was severe on windward coasts, and other areas where the oil penetrated into sediments around the mangrove roots. Oysters and other organisms living on mangrove roots also suffered severe mortality. The types of habitats and organisms affected were obviously dependent on the particular weather conditions during the oil spill. Studies are continuing to follow succession and other changes in seagrass meadows, coral reefs, mangrove forests, and associated habitats that were affected by the oil.

On April 27, 1986, an estimated 50,000 barrels of medium-weight crude oil spilled into the sea on the Caribbean coast of Panama. By September 19$6, the oil had contaminated coral reefs, algal flats, seagrass beds, mangrove forests, small estuaries, and sand beaches. The biological reserve and research areas at the Smithsonian Tropical Research Institute’s laboratory at Punta Galeta received substantial amounts of oil. Extensive environmental monitoring data and other detailed information about the area before the spill provided a rare opportunity for examining the ecological effects of the oil. In this paper we briefly describe the spill and give a preliminary account of its effects.

The spill occurred at Refineria Panamá (a subsidiary of Texaco, Inc.), located on Payardi Island (Lat. 9″ 24′ N, Long. 79° 49′ W), about 12 km northeast of the City of Colón (Figure 1). About 240,000 barrels (1 barrel = 159.6 liters = 42 gal) of oil drained from a ruptured storage tank into a surrounding impoundment. About 140,000 barrels broke through a containment dike and overwhelmed separators and a retaining lagoon. Refinery personnel originally estimated that about 35,000 to 50,000 barrels escaped into Bahía Cativá, the bay on the west side of Isla Payardi (quantities from R. Morales and A. Lasday, Texaco, Inc.). The recovery of more than 48,000 barrels of cleaned oil from the sea (R. Morales, pers. comm.), plus the probable losses from evaporation, suggest the initial spill was at least 55,000 to 60,000 barrels. An undetermined amount of oil continued to escape into the sea from seepage through the porous, coral landfill beneath the refinery. The oil type was 70 percent Venezuelan crude and 30 percent Mexican Isthmus crude, with a specific gravity of 27° API at 15.6° C (≈0.89 gm/cm3 (R. Morales, pers. comm.)).

For six days, onshore (northerly) winds pushed the oil into the head of an adjacent bay, Bahia Cativik. On May 4, 1986, refinery personnel reported that rainfall and shifting winds moved a large quantity of oil out to sea, where it could not be controlled by means that were available locally (R. Morales, pers. comm.). For the next two weeks, onshore winds concentrated the oil along the coast, at the heads of the bays, and in estuaries where land-based crews could contain and collect it. Lack of rain minimized flushing from the estuaries. Refinery officials estimate that more than 50 percent of the oil lost to sea was recovered during these favorable conditions. Eventually, shifts of winds moved part of the unrecovered oil out to sea and along the coast. Oil sheen extending from the spill was observed on the open sea over the region from Bahia Lim6n to the town of Nombre de Dios, a distance of about 60 km.

Except for the areas of deeper water in Las Minas Bay and Cativá Bay, collection of the oil was mostly limited to areas that are accessible by road. Much of this region is roadless, and pockets of oil remained in lagoons and mangrove inlets that were too shallow to be reached by conventional boats. A Hercules C130 transport plane and a Cessna crop duster also sprayed the dispersant Corexit 9527 (Exxon Chemical Americas, Houston, Texas) on oil slicks. The total amount of Corexit sprayed has not yet been determined, but refinery officials estimated the quantity at less than 21,000 liters (A. Lasday, pers. comm.).

In surveys on foot, in small boats, and from low-flying aircraft, we have seen oiled shoreline from the Chagres River to the north shore of Isla Grande, a distance of about 85 km (Figure 1). The most heavily oiled shoreline was between Isla Margarita and the Islas Naranjos; as of August 1986, only two areas within this region appeared not to be contaminated with oil. Both were mangrove-lined lagoons, one between Isla Margarita and Isla Galeta, and the other in the uppermost reaches of the unnamed bay south of Isla Pena Guapa. Natural barriers, augmented by floating booms at Isla Margarita, protected these areas.

Habitats and species affected

The oil refinery is located within a complex maze of coral reefs and coral islands with shallow lagoons, sand beaches, coconut groves, and mangrove forests. Small estuaries exist along the shore. Much of the shoreline is drying (intertidal) reef flat. This is the largest area of mangrove-reef complex on the central Caribbean coast of Panama. The total area of land and water encompassed by the oil spill was about 40 km², which includes about 16 km² of mangroves and 8 km² of coral reefs. The actual area or length of coastline covered by oil is difficult to determine precisely because of the intricate channels and shorelines, and the undetermined penetration of oil into the mangrove forests.

Mangroves

Of the four common mangrove species that occur in the area of the oil spill, the red mangrove (Rhizophora mangle) is the most seaward in occurrence, and makes up nearly all of the fringing forest. The prop roots of these trees form thickets that line much of the shoreline within the range of the spill. Over flights in June and July 1986 showed that more than 15 km of this shoreline between Isla Margarita and Maria Chiquita was oiled. On the open coast, a tarry residue coated the roots and lower leaves of these trees between lowest low water and the upper reaches of the splash zone (about 60 cm vertical extent). White mangroves were also exposed to oil, especially where the oil penetrated deeper into the forests. Our studies have concentrated on two aspects of the mangroves: the communities of plants and animals that inhabit the roots of the red mangroves and the mangrove trees themselves.

Sessile organisms on mangrove roots

The labyrinthine prop roots of the red mangroves serve as firm substrata or refugia that support a number of plants and animals. For example, at Punta Galeta, the prop roots of the red mangrove are the habitat for 127 species of animals and 43 species of algae, some of which are specialized for this habitat.’ Before the oil spill the abundances of oysters, mussels, algae, and other organisms living on mangrove roots had been surveyed in three types of habitats (open coast, channel/lagoon, and riverine), each with a distinctive biotic community. In each habitat, approximately 100 roots (submerged but not firmly anchored to the bottom) were randomly chosen and the relative cover of organisms estimated. After the spill, the surveys were repeated in both oiled and unoiled sites. Before the spill (1981-1982), the mussel Mytilopsis domingensis covered about 50 percent of the root surface in the riverine habitats (Table 1). The barnacle, Balanus improvisus, was present in each survey, but varied in mean cover from a low of 2 percent in June 1982 to a high of 18 percent in January 1981. The overall mean of algal coverage was 6-9 percent, which was formed of 3 to 10 species. Bare space averaged 20-40 percent. In unoiled sites after the spill (July 1986), the average space occupancy in most categories was roughly similar to that of 1981-1982. Only Mytilopsis was less abundant (mean cover = 32 percent) than in 1981 or 1982. In oiled sites, the amount of bare space (microbial slime and oil are included in this category) averaged 85 percent. The rest of the space was occupied by dead and decomposing organisms (Table 1).

On channel roots before the spill. Myalopsrs formed a mean cover of 8-12 percent and the edible oyster Crassostrea rhizophorae covered most space (mean cover = 46-62 percent). Two other bivalves oyster, Isognomon alatus and the mussel, Brachidontes exustus) were always present in relatively low abundance (Table 1). Balanus improvisus averaged from 9-15 percent cover, and the smaller barnacle, Chthamalus sp., covered an average of 2-5 percent space on roots. Other sessile organisms (sponges, bryozoans, hydroids, algae) averaged only 1-3 percent cover. Prior to the oil spill, the bare space on roots in this habitat averaged 10-13 percent of the total area. Post-spill changes were less obvious on channel roots. Crassostrea’s cover decreased from 1981-1982 levels in both oiled and unoiled areas, but more so where roots were oiled. Most other invertebrates also declined from 1981-1982 levels of abundance (Table 1). Algae increased in relative cover in both oiled and unoiled areas, compared with 1981-1982 data.

Table 1.
Abundances of major organisms on Rhizophora prop roots before and after the oil spill
(N is the number of roots samples.)

Before the oil spill, open coast mangrove roots had about 30 percent bare space on their roots (Table 1). These differed from roots in other habitats by the absence of a dominant bivalve; instead, a diverse array of foliose algae covered a mean of 29-47 percent cover. Several species of hydroids and sponges averaged from 3-5 percent and 5-6 percent cover respectively. Open coast mangroves are presently being monitored. No unoiled roots have yet been examined; however, data from 100 oiled roots suggest a diatom bloom and a concomitant decrease in sponges and in some species of algae and hydroids (Table 1). The shells of the mangrove snail Liaorina angulifera were coated with oil, but the snails moved out of the oily areas and into the higher parts of the mangrove trees. Additional measurements are being made of size-frequency distributions and ash-free dry weights of the species living on the roots in all mangrove habitats.

Mangrove trees

Aerial surveys in June and July of 1986 showed defoliated mangrove trees and trees with yellowing leaves along the borders of Bahía Cativá and on the northern portions of the islands and headlands between Punta Muerto and Punta Galeta. The stress and defoliation of the mangroves tended to be concentrated where the trees were rooted in a berm of intertidal sediments. In most places, the berm apparently intercepted and partially absorbed the oil, blocking further movement of oil into the forests. Monthly surveys are being made of mangroves in oiled and unoiled habitats. The variables used to quantitatively describe the structural characteristics of the mangrove communities are individual tree location, species composition, diameter at breast height (DBH), tree height, leaf area index (LAI), phenology, canopy density, growth of respiratory organs, and rate of litter fall. In addition, the leaves from randomly chosen trees are evaluated for phenology, longevity, structure, herbivory, and deformities. Groups of red mangrove propagules also have been planted and are being monitored monthly for sprouting success, height, leaf number, phenology, and leaf structure. Transect surveys were made of the mangrove forests in September 1986. In heavily oiled sites, there was complete defoliation of trees in inner fringe and outer fringe forests that were rooted in the intertidal sediments. Trees of the inner fringe, but not the outer fringe, that were rooted in subtidal sediments suffered less defoliation. As of September 1986, the trees rooted in the supralittoral sediments of the oiled region had suffered less defoliation than the trees rooted at lower levels (Table 2).

Reef flats

In the heavily oiled region, platforms of fringing reefs form extensive shallow habitats covered with algae, sea grasses, and invertebrates. In the early days of the oil spill, between 10 and 19 May, extreme low tides exposed the reef flats above water level during warm, sunny weather. Driven by onshore winds, the oil accumulated against the seaward borders of the reef flats where it remained for the duration of the low tides. By early June, a band of substratum 2 to 3m wide at the reef edge was nearly barren of the normal assemblage of sessile invertebrates and algae, which contributed to the lower abundances of organisms in the quantitative sampling described below.

Table 2.
Leaf area indices and percent defoliated trees in three mangrove transects.

As part of a larger-scale monitoring program, we have surveyed the average of organisms at the seaward edge of the reef 16 times in the period between March 1983 and December 1984. These surveys were repeated to compare changes before and after the oil spill. The surveys consisted of 10 transects, 9 to 22 m long (average length, 18 m), which were perpendicular to the reef edge and spaced randomly with- 20 m intervals. Using point sample methods, we determined the spatial coverage of organisms on the substrate. (The method used random points in sets of 5 per 0.5 m interval.) The number of points sampled per survey ranged between 1,510 and 1,560.

By the first week of June 1986, visual inspection of the reef edge showed that it was being colonized by a thin, transparent mat of algae. In a microscopic examination of a systematic collection of these algae, the mat was mainly Cladophora sp., Enteromorpha sp., and (mostly epiphytic) diatoms. By the last week of June, the mat had become a thicker assemblage of Cladophora, Centroceras, and diatoms, overgrowing both the vacated substrata and the sessile organisms that had survived the oil spill. At this time, this algal mat covered more than 54 percent of the hard substratum-more than 4 times the average abundance, and almost twice the maximum abundance, measured in pre-spill surveys (Table 3).

The two common species of zoanthids on the reef flat, Palythoa caribaeortan and Zoanthus sociatus were less abundant after the oil spill. The reef flat population of Palythoa caribaeorum was concentrated in the area where the oil accumulated at low tide. Before the oil spill, its coverage ranged between I and 2 percent averaged over the whole transects, or 10 to 12 percent in a band 2 m wide at the seaward fringe of the reef flat. In June 1986, the overall spatial coverage of this colonial cnidarian was about 0.12 percent, or less than one tenth the average pre-spill abundance. By September 1986, the coverage had increased to 0.25 percent, still less than at any time before the spill. The population of Zoanthus sociatus extended landward of the Palythoa population, but was not entirely outside the zone of direct oiling. In the post-spill censuses, the abundances of Zoanthus were also below the minimum recorded before the spill (Table 3). Within the areas of the surveys, most of the Porites spp. (scleractinian corals) were found in the same habitats as Palythoa. Averaged over the whole length of the transects, the percent coverage of Porites spp. in aggregate was less than 1 percent of what it had been in the pre-spill surveys, but the corals were always present. No Porites were found in the post-spill surveys.

The abundance of crustose coralline algae, the main reef builders at the reef crest,¹⁰ also decreased in the aftermath of the oil spill. Like the Palythoa and Porites, the peak abundance of these algae was at the seaward border of the reef flat, forming an average cover of more than 25 percent before the spill; this corresponds to 8-9 percent cover averaged over the whole length of the transects. In June 1986, the overall cover was measured as 2.24 percent cover overall, which increased by September 1986 to 5.7 percent cover. Both post-spill measurements were lower than any made in the 16 surveys before the oil spill.

In June 1986, the percent coverage of the calcareous green alga Halimeda opuntia was lower than average, but within the range of values recorded in the pre-spill surveys; however, in September 1986 the coverage was about half that measured in June, which was the lowest abundance ever recorded (Table 3). Besides the possibility of delayed mortality, this may have been the result of the time it takes for this tough alga to slough off the substratum after dying. Overgrowth by fleshy red algae also may have reduced the actual or apparent coverage of Halimeda optunia.

The fleshy red alga Laurencia papillosa was the predominant alga at the edge of the reef, where it formed extensive mats ranging between 22 to 62 percent average overall coverage, and 1 to 4 cm in average overall thickness, depending on the season. In June 1986, the overall cover was 19.9 percent, and the thickness was 0.77 cm, each measure lower than in any previous survey. However, by September 1986, the coverage of the mat had increased to 54 percent and the thickness to 2.24 an, both of which were higher than the pre-spill averages. The coverage of a similar fleshy red alga Acanthophora spicifera was also higher than average in the post-spill surveys. Before the oil spill, the counts of sessile species that could be recognized in the field ranged between 13 and 25 total species per survey and averaged 23 species per survey. In June 1986, the count was 14 species; in September 1986, 19 species (Table 3).

The abundances of sea urchins at the reef edge also declined immediately after the spill. As part of a larger monitoring program, censuses of all species of sea urchins are made approximately once per month at the reef edge. Echinometra lucunter and E. viridis are the predominant species in this zone. In nine years of surveys, these urchins were generally most abundant between March and June, when E. lucunter may reach population densities of 1,000-2,000 individuals per 20 m² according to Cubit, et al.⁵ As the oil began coming ashore at Punta Galeta in early May 1986, the population densities of E. lucunter and E. viridis had reached 308 and 56 urchins per 20 m², respectively. By the end of May, the densities were 54 and 8 urchins per 20 m², respectively. The numbers of urchins increased in the following months (Table 4), but were less than the seasonal average for previous years⁵.

Table 3.
Comparisons of spatial coverage, algal thickness, and species counts at the seaward edge of the Galeta reef flat, before and after the oil spill

In addition to the studies reported above, we are continuing surveys of spatial coverage of algae, sea grasses, and sessile invertebrates on the whole reef flat, together with various types of population surveys of sea urchins, stomatopods (mantis shrimp), and gastropods.

Subtidal reefs. As of August 1986, numerous colonies of shallow water corals were dead or dying in depths of 1 to 2 m in the heavily oiled areas. In quadrat sampling, the proportion of dead or dying colonies averaged between 17 and 30 percent on oiled reefs, and was 0 percent on unoiled reefs (see Table 5). Surveys of corals are continuing and include repetitions of surveys that were first made along the coast before the oil spill. These will provide data about coral abundances before and after the spill in areas inside and outside the oiled region.

The oil spill has affected several different tropical marine communities, each of which is dominated by a group of organisms, such as corals, mangrove trees, and algae, that provide the primary local structure of their habitat. The relevant questions are: what will happen to these communities following disturbance by oil, and how long (if ever) will it take them to recover? Answers will depend on the species affected as well as on the nature of the disturbance. For example, corals are very long-lived and rates of recruitment and recovery in disturbed areas may require decades.8 In contrast, mangrove roots are relatively short-lived habitats whose inhabitants routinely recruit, grow, and mature within a few years,¹³ so that recovery from disturbance should be more rapid than for corals unless the roots themselves are too severely altered or destroyed, or widespread mortality of the epibionts has eliminated the source of recruits.

Chronic, long-term, contamination also may affect biological changes in the aftermath of this spill. The heavily oiled mangroves and the coral fill beneath the oil refinery may leak oil to adjacent environments for years to come. At the end of September 1986, five months after the initial oil spill, black oil slick was still present around the mangroves in Bahía Cativá, and translucent oil sheen was present daily at Punta Galeta.

Table 4.
Population densities of Echinometra lacunter and E. viridis In censuses at the seaward edge of the Punta Galeta reef flat – Densities are In number of animals per 20 m2. Heavy oiling of Punta Galeta began on May 8, 1986.

Table 5.
Percentage of the colonies of the coral SidrAu sa-sidcrsa showing recent partial mortality of tissue In habitats exposed 5 to different amounts of oil–The number of conls observed is shown in parentheua.

¹Arbitrary ranking based on observations from over flights, photographs, visits to sites, early path of the oil, and distance from the refinery

²The criterion for recent mortality of tissue was growth of a microbial film over obviously decaying coral tissue

In following the course of the spill, it was apparent that the spatial pattern of oil contamination and the types of organisms affected were very much dependent on the peculiar weather at the time of the spill. The usual rains and shifting winds for the season when the spill occurred would have carried more oil out of the bays and along the coast; instead, aseasonal northerly winds held the oil relatively near the refinery. Higher water levels, such as those normal for November through February, would have protected organisms at the seaward edge of the reef, but would have allowed the oil to penetrate more deeply into the basin habitats of the mangrove forests, causing mangrove death over a greater area. Because of the season, we have seen few oiled birds to date. Most of the migratory shorebirds and ospreys that forage around the reefs had departed northwards a few weeks before the spill. Few swimming and diving birds are resident in this area; however, as of September 1986, considerable oil remained in habitats that are winter feeding grounds for North American shorebirds, which may cause chronic oiling of wading and swimming species when they return. Thus, applying the information gained from this oil spill to others must take into account the circumstances of weather and season.

If mortality follows stress and defoliation of the red mangroves, the impact may be much wider than loss of the trees themselves. The thickets of prop roots of this species serve as breeding and nursery areas for many marine species, and as substrata and shelter for a diverse group of others,^1,4,12,13 including economically important species of fish, mollusks, and crustaceans. Sediment now retained by the root masses could be released if the roots decompose, which is a potential threat to nearby corals and other organisms that are intolerant of siltation.

Although mangroves are ecologically important and ranked among the coastal environments most sensitive to oiling: there has been little research involving the effects of oil on mangrove root communities. Cairns and Buikema2 define mangroves as one of the research areas for which the fewest data on oil effects exist.

The state of knowledge for the effects of oil on shallow beds of algae and sea grass flats is worse. The National Research Council¹¹ describes the situation as “totally neglected,” even though these habitats are “highly vulnerable to oiling,” a ranking with which Gundlach and Hayes concur. The beds of sea grasses and algae in warm, shallow water typically exhibit rates of primary productivity per unit area that are among the highest of all ecosystems measured.’ In situ measurements with flow respirometry at Punta Galeta show that such rates of productivity are maintained even through seasons of stressful conditions.” l.ike mangroves, sea grasses and rhizophytic algae are effective in controlling erosion.”

The data base accumulated at the Galeta Marine Laboratory includes extensive information about the marine biota, as well as meteorological and hydrographic factors. Estimating the effect of oil spills often hampered by the lack of any ecological baseline information it allows scientists to make before-and-after comparisons or put their shorter-term findings into longer-term contexts. The data base Galeta provides both a measure of what organisms were in the allow marine environments before the spill, as well as estimates of their natural variations in distribution, abundance, and growth rate data that are necessary to distinguish changes caused by oil from the ranges that occur naturally.

Acknowledgments

We thank S. Churgin, R. Sarmiento, R. Morales, and A. Lasday for providing information ; A. Román, T. Wachter, V. Batista, and. Moss for technical assistance; and C. Hansen for photographic Services. We especially thank R. Engh, L. Therrien, and D. Dickenson for the generous cooperation of the U.S. Naval Security Group activity at Galeta.

References

1. Batista, V. E., 1980. Estudio de las Comunidades que Habitan las Raices del Mangle Rojo Rhizophora mangle L. de Punta Galeta, costa Atlantica de Panama. Bachelor’s thesis. University of Bogota, Faculty of Marine Sciences. 89pp
2. Cairns, J., Jr. and A. L. Buikema, Jr., eds., 1984. Restoration of Habitats Impacted by Oil Spills. Butterworth Publishers, Stoneham, Massachusetts
3. Cubit, J. D., 1985. Possible effects of recent changes in sea level on the biota of a Caribbean reef flat and the predicted effects of rising sea levels. Proceedings of the Fifth International Coral Reef Congress, Tahiti, V3, pp111-118
4. Cubit, J. D., G. Batista de Yee, A. RomSn, V. Batista, 1985. El valor de los manglares y arrecifes en la costa de Coldn. in Agonia de Ia Naturaleza, S. Heckadon Moreno and J. Espinosa Gonzglez, eds. Impretex, S.A., Panamd, Republic of Panama. pp183-199
5. Cubit, J. D., D. M. Windsor, R. C. Thompson, and J. M. Burgett, 1986. Water-level fluctuations, emersion regimes, and variations of echinoid populations on a Caribbean reef-flat. Ertuarine, Coastal and Shelf Science, v22, pp719-737
6. Griffith, P. C., J. D. Cubit, W. H. Adey, and J. N. Norris, in press. Computer-automated flow respirometry: Metabolism measurements on a Caribbean reef flat and in a microcosm. Limnology and Oceanography
7. Gundlach, E. R. and M. O. Hayes, 1978. Vulnerability of coastal environments to oil spill impacts. Marine Technology Society Journal, v12, pp18-27
8. Hughes, T. P. and J. B. C. Jackson, 1980. Do corals lie about their age? Some demographic consequences of partial mortality, fission and fusion. Science, v209, pp713-715
9. Lewis, J. B., 1977. Processes of organic production on coral reefs. Biological Review, v.52, pp305-347
10. Macintyre, 1. G. and P. W. Glynn, 1976. Evolution of modern Caribbean fringing reef, Galeta Point, PanamS. The American Association of Petroleum Geologists Bulletin, v60, pp1054-1071
11. National Research Council, 1985. Oil in the Sea : Inputs, Fates and Effects. Steering Committee for the Petroleum in the Marine Environment Update, Board on Ocean Science and Policy, National Research Council. National Academy Press, Washington, D.C. p412
12. Odum, W. E. and E. J. I-Ieald, 1972. Trophic analysis of an estuarine mangrove community. Bulletin of Marine Science, v22, pp671-738
13. Sutherland, J. P., 1980. Dynamics of the epibenthic community on roots of the mangrove, Rhizophora mangle, at Bahiade Buche, Venezuela. Marine Biology, v58, pp75-84
14. Zieman, J. C., R. Orth, R. C. Philips, G. Thayer, and A. Thorhaug, 1984. The effects of oil on seagrass ecosystems. in Restoration of Habitats Impacted by Oil Spills, J. Cairns, Jr. and A. L. Buikema, eds. Butterworth Publishers, Boston, pp37-64