One of the major environmental impacts of oil exploration in the Ecuadorian Amazon region is the contamination of the aquatic environment. In addition to occasional major oil spills, there are several minor sources. This study investigates aquatic toxicity due to seepage from production pits and oiled road run-off. Water and sediment samples were collected and acute toxicity was determined in Daphnia magna and Hyalella azteca bioassays. Because exposure to ultra-violet radiation may enhance the toxicity of several polycyclic aromatic hydrocarbons in crude oil, toxicity was studied before and after exposure to natural sunlight in the field using a microbiotest or solar-simulating ultra-violet light in the laboratory using cultured organisms.
Total petroleum hydrocarbon concentrations of analyzed sediments ranged from 4.9 to 6980 mg kg−1 dry weight (dw). Immobility and lethality ranged from 0 to 100% in Daphnia magna and from 0 to 40% in Hyalella azteca. The micro-biotest results were generally confirmed by the tests on cultured Daphnia. A river exposed to oiled road run-off and a pond used for drinking water, 100-m from a production pit, had the highest total petroleum hydrocarbon concentrations. These samples were toxic compared to reference sites. The response increased upon light exposure, suggesting polycyclic aromatic hydrocarbon availability. Minor water phototoxicity was also observed in the drinking water pond (having ten times the total petroleum hydrocarbon concentration nationally allowed in drinking water). Thus, the seepage from unlined production pits and the practice to use oil for dust control on roads in the area pose environmental and human health hazards.
So far little research has been devoted to effect studies of crude oil pollution in tropical freshwater environments. Tropical rainforests probably have optimal conditions for degradation processes, but ambient factors in these environments such as pH, light and temperature could also influence aquatic toxicity. Crude oil has been extracted in the Ecuadorian Amazon for more than three decades, and there is intensive debate about the environmental and social impact of this production (Fundación Natura, 1991; Kimerling, 1993). One of the environmental impacts considered to be severe is the aquatic contamination; high total petroleum hydrocarbon (TPH) concentrations were recently found in four rivers in the region, suspected to cause an increased cancer rate (Sebastián et al., 2001). Aquatic toxicity has, to the author's knowledge, not been studied.
Contamination from oil production activities includes several sources. During exploration processes, drilling mud is deposited in pits. Once the actual production begins, sources of contamination include liquid wastes from the wells, flow lines and separation facilities. The production pits (often of size 100× 100 m or larger) are usually unlined and contaminants may eventually leak into the surrounding environment. There have also been at least 30 major oil spills from the 520 km long Trans-Ecuadorian Pipeline, which runs across the Andes to the coast (Kimerling, 1993). In addition to oil pollution from the industrial activities, roads in the area are frequently oiled for dust control.
The combined exposure of an organism to certain contaminants and solar light can enhance or induce toxic effects. Photosensitization occurs inside the organism, necessitating accumulation of the substance before ultra-violet light exposure (Ankley et al., 1994; Wernersson and Dave, 1998). Environmental concerns are primarily focused on a group of polycyclic aromatic hydrocarbons (PAHs), in addition to their mutagenicity and carcinogenicity. Phototoxicity of PAHs has been observed in several plant and animal species, including human cells (Holst and Giesy, 1989; Mauthe et al., 1995; Krylov et al., 1997; Weinstein et al., 1997; Hatch and Burton, 1998). The photo-induced toxicity of water extracts of petroleum products is probably related to PAHs (Pelletier et al., 1997; Wernersson, 2003). PAHs accumulate on particles and sediments, which tends to protect them from biodegrading processes. Biodegradation is often slower for aromatic compounds than aliphatic, and monocyclic compounds disappear faster than the larger and more complex aromatic compounds (Green and Trett, 1989). Photo-active hydrocarbons may be released into the water when the sediment is disturbed (Davenport and Spacie, 1991). In situ tests on zooplankton in a river with PAH-contaminated sediments revealed increased toxicity if they were simultaneously exposed to sunlight, although this effect was reduced during turbid conditions (Ireland et al., 1996). In addition, mortality in benthic invertebrates exposed to PAH-contaminated sediments may increase upon co-exposure to UV irradiation (Ankley et al., 1994) or natural sunlight in situ (Monson et al., 1995).
The objective of the present study was to investigate aquatic contamination, toxicity and phototoxicity due to seepage from production pits and oiled road run-off in the Amazon region of Ecuador (Figure 1).
Materials and methods
Toxicity of water and sediment samples was studied in bioassays on a crustacean zooplankton species (Daphnia magna) and an amphipod (Hyalella azteca). Samples were collected from four sites (13 stations) where crude oil was the main polluting factor, from production pits (three sites) and oiled road run-off (one site). After 24, 48 or 96 h of exposure (accumulation) mobility or lethality was checked. The beakers (with survivors) were then irradiated (1 or 2 h) and the response was checked again. Due to limitations on the use of cultured organisms in the field, D. magna juveniles from ephippiae (DaphtoxkitTM) were used to test effects of natural sunlight in the area of collection. The sediment bioassays were repeated under laboratory conditions using solar-simulating UV-light.
Sampling and site descriptions
Surface water samples were collected in April 1998 from the upper 20 cm either from the shore or by submerging a weighted bottle. Sediments were taken from the upper 5 cm, using a box corer. For both sediment and water samples four replicates from an area 1 to 2 m in diameter were mixed to one composite sample from each station in order to reduce the influence of spatial variability. Samples from rivers were taken approximately 1 m from ashore. River banks are probably the most relevant sampling sites if considering effects of oil pollution, because this is where the largest amounts of oil eventually settle in swift flowing waters (Green and Trett, 1989). Sediment and water samples were transported on ice and stored in darkness at 4°C until analysis.
Stations A and B (‘Sachas’) are both inside a pond used for drinking water, 100-m from a production pit still in use (located above the pond), but station A (depth 20 cm) is located close to incoming water. Diffuse penetration of oily water from beneath was observed at several spots. Samples from station B were taken from a depth of 30 cm, close to the outlet. The vegetation had been cleared around the pond (approximately 10 m in diameter) so there were no shaded parts. Stations C (small drinking water pond, well above the production pit) and D (the river Río Yana Quincha, Figure 1) were chosen as references.
Station E and F are both in the river Río Rumiyaco, 5 m upstream and downstream, respectively, of a pipe (approx. 100 m long) outlet from an oil production pit no longer in use. There was some occasional shade from vegetation.
Station H (‘Dayuma’) is an abandoned production pit now filled with water. Station G is from in-flowing water, I from out-flowing and J a minor stream (3-m width) receiving the water from the pit. It was not possible to take sediment samples from these stations, either due to inaccessibility or to a too thin sediment layer (rock).
Finally, stations K-M are from the river ‘Río Shushufindi,’ north of the town of Shushufindi. Station K is upstream and L downstream an oiled road (a common practice in this region to inhibit dusting). Station M is approximately 500 m downstream (east of the town).
The concentrations of TPH in sediment and water from 5 stations (B, E, F, K, L) was determined within two days after sampling by 1,1,2-trichloro-trifluoro-ethane extraction and infrared analysis (Perkin Elmer 1600) in the Laboratorio de Suelos in Napo-Coca, Ecuador. The TPH parameter was chosen in this study as a rough estimate of the amount of oil present. All other analyses (dry weight (dw), total organic carbon (TOC), S, V, Al, Pb, Fe, Co, Cu, Cr, Mn, Ni, Zn) were performed on sediment samples within 3 wk by AnalyCen Nordic AB, Göteborg, Sweden. The analyses were performed according to Swedish standard methods (SIS, 1981, 1988, 1990, 1993a, b, c, d, 1997).
Within one week after sampling, sediment and water toxicity was tested. Due to lack of an available culture facility, hatched ephippiae (resting eggs) of the crustacean D. magna were used in the first tests of the environmental samples. The ephippiae from a Daphtoxkit FTM (Creasel BVBA, Deinze, Belgium) were incubated in Petri dishes with deionized water reconstituted to have a hardness of 250 mg CaCO3 l− 1 and pH 8.0 (Klüttgen et al., 1994). A 60 W lamp was placed about 15 cm above the dishes to irradiate the eggs continuously during incubation. Temperature was 23 ± 1°C. Ephippiae started hatching after three d and neonates were transferred to beakers with incubation medium daily. Fed juveniles (1–4 d old) were used in the tests.
The toxicity of each water sample was tested in 4 polyethylene cups (200 ml of volume), by adding 50 ml of the sample and 5 organisms to each cup. Standard dilution water for the sediment bioassays was prepared from Millipore water (ISO, 1982). To test sediment toxicity, 16 g of sediment was diluted to 100 ml with dilution water; mixed and equal volumes were distributed to two glass beakers (250 ml volume, 6 cm diameter). The suspensions were allowed to settle for 1 to 2 h prior to addition of ten juveniles into each beaker.
Immobility was first recorded after 24 h of exposure (indoors). Vessels (with test organisms) were moved outdoors and exposed to sunlight (2 h on a clear sky for water samples and 1 h on a clouded sky for sediment samples) as described below. Immobility was again checked, after another hour without sunlight exposure (in preliminary tests D. magna often recovered within an hour after UV exposure). For all bioassays, dilution water and a reference toxicant, K2Cr2O7, were also tested (ISO, 1982).
Sediment bioassays on cultured Daphnia magna
Daphnia magna (clone 1) was cultured in trays with the same water type as used for incubation of the ephippiae and given the same feed. During culturing and testing the temperature was 20 ± 1°C and the photoperiod was 16 h light, 8 h darkness. Within 3 w after sampling, juveniles (1–4 d old) were exposed in glass Petri dishes (10 cm in diameter; 50 ml of volume). The exposure was performed with different sediment:water ratios, making it possible to calculate the 50% effect concentration (EC50), that is, the sediment concentration at which 50% of the examined organisms were immobilized, and to investigate ‘concealed’ phototoxicity in the samples having 100% immobility in the 16% sediment bioassay. Dilution series (32, 16, 8, 4 and 2% wet weight (ww)) were prepared by diluting 32 g of sediment to 100 ml with dilution water, mixing, distributing half of the suspension (50 ml) into one Petri dish and adding 50 ml of dilution water to the rest. This suspension was mixed, and so forth. Two dilution series were prepared from each sediment sample. Suspensions were allowed to settle for 1 to 2 h prior to addition of 20 test organisms to each test beaker. Immobility was first recorded after 24 h of exposure; then again after 48 h. Immediately after checking immobility the second time the dishes were irradiated by fluorescent solar-simulating UV-tubes for 2 h as described below. Immobility was checked a third time, after another hour without UV exposure.
Sediment Hyalella azteca bioassay
Sediment toxicity was tested within 4 w on the amphipod H. azteca. Although not a tropical organism, it is widely distributed geographically, including lakes in the Caribbean (Environment Canada, 1997). Hyalella was cultured in a tray with the same medium as was used for D. magna cultures. To minimize stress, a mesh and some shade were provided. During culturing and testing the temperature was 20 ± 1°C and the photoperiod was 16 h light, 8 h darkness. Ten ml of a 16% sediment suspension was transferred to each of four wells of disposable Nunc® 6 well plates and ten 3rd instar (< 3 mm) were added to each well. Lethality was recorded after 96 h of exposure. Surviving organisms were transferred individually to 3-ml wells of disposable Nunc® 24 well plates containing about 1 ml of the overlying solution. The organisms may partially dig themselves into the sediment, and in order to make sure that they were all exposed to the same light dose, survivors were transferred to vessels without sediment present. The vessels were then irradiated by solar simulating UV light for 2 h and immobility was checked after another hour without light exposure.
UV exposure and measurements
Light measurements were made using an IL 1400A Radiometer/Photometer and a water proof UVA detector with spectral response range 315–390 nm (International Light, Newburyport, MA, USA). During natural sunlight exposure, water samples were exposed during 2 h under a more or less clear sky (2.28–3.38 mW cm−2) and sediment samples during 1 h with cloud cover (501–676 μW cm−2).
The river waters were turbid; the visual depth was about 30 cm. When the UVA intensity at the surface was 800 μW cm− 2, the UVA intensity at 10 cm depth was about 160 μW cm−2 (20% ambient) and at 30 cm depth about 40 μW cm−2 (5%). The water was clearer in the ponds. At site H, the corresponding intensity was about 260 μ W cm−2 (33%) at 10 cm. On a day with clear weather the surface UV-A intensity at noon was usually > 3 mW cm− 2.
In the laboratory, two fluorescent UV-A-340-tubes (Q-panel Company, Cleveland, Ohio) were placed 40 cm above the bench used for exposure and 12 cm apart. The tubes simulate sunlight in the critical short wavelength UV region between 365 nm and the solar cut-off at 295 nm and with peak emission at 340 nm. The light intensity on the Petri dishes and Nunc® well plates was 0.37 ± 0.02 mW cm−2 during 2 h of exposure. The well plates and Petri dishes were exposed without lids.
To determine whether the increase in toxicity was significant after light irradiation, χ2 tests were performed (Excel 97, Microsoft Corporation), using lethality/survival data before irradiation as predicted values and corresponding data after irradiation as observed values. Immobility of D. magna after 48 h was used for determination of the EC50. A low EC50 corresponds to high toxicity and vice versa. If applicable (where there were at least two concentration levels with partial kills), the probit or the moving average method was used for determination of EC50s (Peltier and Weber, 1985). Spearman rank correlation coefficients between data were calculated using the statistical software SAS and the CORR SPEARMAN procedure.
The concentrations of some elements (S, V, Al, Pb, Fe, Cu, Cr, Mn, Ni and Zn) as well as TOC and TPH of the analyzed sediment samples are shown in Table 1. For comparison, sediment quality criteria have been added when available. For those compounds where sediment criteria were available, the concentrations were always below ‘probable effect levels’ and only Ni and Cu was occasionally above ‘threshold effect levels.’ The highest concentrations of Al, Co, Cu and Zn were found in sediment L, and the highest concentrations of Pb and Ni were found in samples A and B, respectively. The concentrations of TPH in analyzed water samples ranged between 0.05 and 0.12 mg l− 1, which are all above the nationally admitted concentration in drinking water (0.01 mg l− 1), but still not as high as some of the concentrations found by Sebastián et al. (2001).
Immobility after 24 h ranged from 0 to 20% and after subsequent exposure to natural sunlight increased 5 to 40% in the water bioassays (data not shown). Several floating (but still not immobilized) organisms were found in sample B (the drinking water pond below the production pit in use), where the highest increase in toxicity due to light was observed (10% before and 40% after irradiation). This water sample also had the highest TPH concentration (0.120 mg l− 1).
Immobility and lethality of the test organisms in 16% sediment is presented in Figure 2. Toxicity data before irradiation (i.e., after 24 h for ephippial D. magna, 48 h for cultured D. magna and 96 h for H. azteca) are shown in Figure 2a. The toxicity to juveniles from ephippial D. magna was always higher than for cultured organisms but the response was Spearman correlated (correlation coefficients 0.90 and 0.77 before and after UV respectively, p < 0.05) and toxicity peaks were observed in the same samples (B, F, L, M). The sensitivity to the dichromate reference of the cultured and hatched organisms was about the same (average EC50: 0.95 and 0.85 mg l−1, respectively). Hyalella was generally less sensitive than D. magna. There was no correlation between the two species, but at least two of the peaks are discernible (L, M). The ‘reference stations’ C, D, E and K had low toxicity to both organisms (0 to 25% immobility or lethality). The water samples from station B was more toxic to D. magna than that from station A (closer to inlet of surface water), and sample F was more toxic than E (its reference station). The samples from station L (exposed to road runoff) was more toxic than K (its reference station) to all organisms. Daphnia. magna toxicity (before irradiation with UV) was correlated to Cr concentrations (correlation coefficient 0.71) whereas H. azteca toxicity was correlated to V and Zn concentrations (correlation coefficients 0.75 and 0.76 respectively; p < 0.05).
Significant (p < 0.05; χ2 test) increase in D. magna toxicity occurred after the light exposure (both natural and simulated) of samples A, B and L as shown in Figure 2b. In sample B, H. azteca phototoxicity is also significant. For ephippiae of D. magna, toxicity increased in sample E under natural sunlight, but this was not confirmed in the tests using cultured D. magna. The pH of the overlying test water of sample F and M was below 5 (4.7 and 4.4, respectively) during the bioassays, and this may have contributed to the high toxicity (low EC50s) of these samples. In samples A and B an oily film was visible on the water surface.
The EC50 values (based on sediment: water ratios) before and after UV radiation for cultured D. magna are shown in Table 2. The toxicity ranking of the most toxic samples to D. magna was M > F > B. The highest EC50 ratios (ratio between EC50 before and after UV irradiation) were found for sediment B (3.3) and L (> 2.5). The pH increased gradually with less sediment present. In sample M, no ‘concealed’ phototoxicity was discovered and only a minor toxicity increase could be detected in sample F (EC50 ratio of 1.3).
Acute water toxicity was low at all sites, although the results should be interpreted with caution since, for practical reasons, the bioassays had to be performed in plastic cups (probably causing hydrophobic substances to adsorb to the walls). Furthermore, any water sampling should be repeated at several time intervals in order to make any firm conclusions about water toxicity. Light induced increases in toxicity were especially observed in sample B, coinciding with several floaters, probably due to oil contamination.
The immobility of D. magna in sediments from site A, B and L were somewhat lower in the bioassay on cultured test organisms (performed 3 wk later) and could be due to exposure situation or to degradation and volatilization of hydrocarbons during storage. However, the main peaks were the same, showing that although limited numbers of organisms were available, sediment bioassays using hatched ephippiae of D. magna could be quite useful for establishing preliminary results in field situations like these.
According to the concentrations of V [often used as an indicator of oil pollution (Khalaf et al., 1982)] and TPH, sample L (140 mg kg− 1 V and 494 mg kg− 1 TPH) and B (6980 mg kg− 1 TPH) were probably the samples most contaminated by oil. Site F was also contaminated by hydrocarbons (110 mg kg− 1 TPH), ten times higher than its reference station. The TPH:TOC ratio, possibly representing the bioavailable fraction of TPH in sediments (DiToro et al., 1991), was highest in samples B, F, L, although the highest TOC concentrations were probably also significantly influenced by their TPH fractions (Schreier et al., 1999). The rank order of the TPH concentrations and the D. magna EC50 ratios are the same, suggesting that the phototoxic hydrocarbons (probably the PAH fraction) were correlated to the TPH concentrations and bioavailable.
Even if the sediment samples were diluted using a medium with pH 8, the pH of some of the test solutions decreased below 5 (sediments F and M). At least in part, the low pH could have been responsible for the high D. magna toxicity of these samples. Acidity may be derived from weathered oil (Traulsen et al., 1999), supported by the higher pH at the reference site, but because tropical rainforest soils can be very acidic, it is not possible to exclude a natural origin, especially at station M (Baillie, 1996). Heavy metal concentrations were relatively low according to sediment quality guidelines (Table 1). However, these guidelines were not developed for tropical sediments and the lower pH values likely affected the availability of metals (Erickson et al., 1996). Native invertebrate species would have to tolerate acidity, but it is not certain that they do not respond to the combination with heavy metals.
The organisms used in this study were selected because of their sensitivity to toxicants and phototoxicity and of the availability of historical records. Boese et al. (1997) found that species normally exposed to light may better tolerate fluoranthene phototoxicity. However, phototoxicity has been detected in a diversity of species, temperate and tropical, including Hawaiian tropical reef animals normally exposed to high intensities of UV light (Peachey and Crosby, 1996).
The organisms have to be allowed to accumulate toxicants from the sediment before light exposure. In reality, it is also more likely that the organisms are first exposed to contaminated sediment and then to UV light while in the water column or on the sediment surface. Therefore, standard bioassays were first performed (with 24, 48 or 96 h accumulation), and then the organisms were irradiated (1–2 h). Daphnia magna does not burrow into the sediment and was therefore left in the test beakers. Hyalella azteca can ‘hide’ and since the purpose was to compare different stations with each other, the organisms were picked up from the sediment and subsequently exposed in the overlying water in order to expose them to the same light dose. The exposure time and intensity in the sediment bioassays is very short compared to light exposures tolerated by the organisms (Wernersson and Dave, 1998). Even after 2 h of strong sunlight (during the water bioassay) there was no more than 5% immobility in the test beakers containing pure water.
The presence of UV light at oil contaminated sites may both be positive and negative. On one hand, photolysis may constitute a major abiotic degradation pathway of several PAHs and could in the long run decrease aquatic toxicity (Halmann, 1995). However, degradation does not always reduce toxicity: enhanced toxicity of illuminated crude oil extracts has for example been observed in algae, and the toxicity increased at higher temperatures (Gaur and Singh, 1991). In addition, accumulated substances may become activated (phototoxic) by exposure of the organism to UV light. The sediments from sites B and L were continuously exposed to natural light (due to lack of extensive vegetation) and the conditions were probably good for biodegradation (Prado-Jatar et al., 1993). However, the sediments were found to be acutely phototoxic. This could be due to a continuous supply of contaminants, non-sufficient degradation or a high load. At least the first was true in the cases where phototoxicity was observed: the pond at station B is still in use and the road at station L is frequently oiled. Production had ceased at the oil pit close to station F; TPH concentrations were lower and sediments were not phototoxic. Unfortunately, the abandoned pit (stations G to J) was filled with water and could not be examined extensively, but at least no phototoxicity was observed in the water samples. Continuous supplies of photoactive hydrocarbons, little shade and shallow waters would probably, in spite of optimal degradation conditions, pose an additional threat to sensitive aquatic species in the area.
Phototoxicity is probably dependent on the availability of PAHs, some of which are known to be potent carcinogens. If there is some correlation between phototoxic and carcinogenic PAH bioavailability, there is also a long term health concern. Some substances such as benzo(a)pyrene and dibenzo[a,h]anthracene possess both abilities (Delistraty, 1998). The area has a poorly developed infrastructure and rural people collect all their domestic water (for drinking, cooking and bathing) from the rivers or ponds. A recent study observed excess cancer rates in a village in this region (Sebastián et al., 2001) and according to staff at the Laboratorio de Suelos, Colegio Padre Miguel Gamboa in Napo-Coca, several people drinking the water from the pond at site A-B have become ill. Contamination and other environmental impacts due to the oil industry in Ecuador would be reduced with existing techniques, if only the guidelines were followed (Bravo and Martínez, 1997). One corrective measure would be to line the production pits. Considering the extent and frequency of oiling roads in the region, the environmental consequences would be reduced by using another practice to inhibit dust. There are direct human health aspects, since many people, especially children, walk barefoot and thus repeatedly have skin contact with crude oil.
In conclusion, contaminants such as hydrocarbons are leaking into the surrounding water system from the production pits and oiled roads. ‘Normal’ and light induced toxicity of sediment from a drinking water pond, close to and below a production pit in use was observed in all test organisms, and of river sediment exposed to road runoff in D. magna, suggesting PAH availability. These sources pose an environmental (and health) hazard. The micro-biotest combined with natural sunlight irradiation was good at predicting phototoxicity in the laboratory using cultured organisms.
The author would like to thank staff at Laboratorio de Suelos at the Colegio Padre Miguel Gamboa in Napo-Coca, Ecuador and Dr. Ramiro Merino, Tanya Caceres, Ramiro Castro and Yolanda Pastor at the Biology Department, Pontificia Universidad Católica del Ecuador (PUCE) in Quito for invaluable help. I am also very grateful to my travel companion, Spanish translator and field assistant, Astrid Nuñez from the Department of Economics, Göteborg University. The study was funded by the Royal Swedish Academy of Sciences, “Adlerbertska Forskningsfonden” and “Knut och Alice Wallenbergs stiftelse.”