© 2006 International Council for the Exploration of the Sea
Dispersant effectiveness on oil spills – impact of salinity
a Department of Civil and Environmental Engineering, University of Cincinnati Cincinnati, OH 45221-0071, USA
b Ecosystems Research Division, National Exposure Research Laboratory US EPA, Athens, GA 30605, USA
*Correspondence to G. A. Sorial: Department of Civil and Environmental Engineering, University of Cincinnati, PO Box 210071, Cincinnati, OH 45221-0071, USA; tel: +1 513 5562987; fax: +1 513 5562599. e-mail: george.sorial{at}uc.edu.
When a dispersant is applied to an oil slick, its effectiveness in dispersing the spilled oil depends on factors such as oil properties, wave-mixing energy, temperature, and salinity of the water. Estuaries represent water with varying salinity, so in this study, three salinity values in the range 10–34 psu were investigated, representing potential salinity concentrations found in typical estuaries. Three oils were chosen to represent light refined oil, light crude oil, and medium crude oil. Each was tested at three weathering levels to represent maximum, medium, and zero weathering. Two dispersants were chosen for evaluation. A modified trypsinizing flask termed a baffled flask was used to conduct the experimental runs. A full factorial experiment was conducted for each oil. The interactions between the effects of salinity and three environmental factors, temperature, oil weathering, and mixing energy, on dispersion effectiveness were investigated. Each experiment was replicated four times in order to evaluate the accuracy of the test. Statistical analyses of the experimental data were performed for each of the three oils independently for each dispersant treatment (two dispersants and oil controls). A linear regression model representing the main factors (salinity, temperature, oil weathering, flask speed) and second-order interactions among the factors was fitted to the experimental data. Salinity played an important role in determining the significance of temperature and mixing energy on dispersant effectiveness for almost all the oil–dispersant combinations. The impact of salinity at different weathering was only significant for light crude oil with dispersant A.
Keywords: baffled-flask test, dispersant, dispersant effectiveness, environmental factors, oil remediation, oil spill
Received 29 September 2005; accepted 3 April 2006.
 |
Introduction |
|---|
 Top Introduction Material and methodsResults and discussionSummary and conclusionsReferences |
|---|
In the event of unintentional releases of oil into coastal waters, oil from slicks can have deleterious impacts on biota in exposed ecosystems. Effects will depend in large part on the ultimate location of the oil as well as on its chemical composition at the time of interaction with the biota (NRC, 1985). Oil slicks usually spread rapidly to a large area because of the action of gravitational and viscous forces, so quick response has to be initiated (Hoult, 1972). Four cleanup strategies that frequently receive consideration include (i) mechanical cleanup or recovery, (ii) burning, (iii) bioremediation, and (iv) treatment with chemical dispersants (NRC, 1989).
Chemical dispersants are made of surfactants that are usually sprayed onto oil slicks to remove oil from the surface and disperse it into the water column, significantly reducing the impact on shorelines and habitats (Lessard and Demarco, 2000). The essential components in dispersant formulations are surfactants, which contain both oil-compatible (lipophilic) and water-compatible (hydrophilic) groups. Following successful application of a chemical dispersant formulation to an oil slick on water, the surfactant molecules reside at oil–water interfaces and reduce the interfacial surface tension. In the presence of mixing energy (provided by wave or wind action), this might result in dispersion of the oil as small droplets into the underlying water column. Such dispersion leads to dilution of the oil in the water and increased oil–water interfacial surface area, which might eventually favour microbial degradation of the oil.
Studies have shown that the salinity of receiving waters can impact dispersion of oil by chemical dispersants (Clayton et al., 1993). Specifically, the intent of dispersant formulations for marine use is to provide maximum dispersion at normal seawater salinity. Mackay et al. (1984) note that higher salinity increases the effectiveness of dispersants by deterring migration of surfactant molecules into the water phase, equivalent to a salting-out effect for the surfactant from the saline medium. Such a situation will tend to promote association of surfactant molecules with oil at oil–water interfaces, which is important for lowering oil–water interfacial surface tensions in the oil–dispersant mixture. In general, increasing salinity will decrease the solubility of dispersants in water, resulting in more surfactant being available to interact and mix with the oil. Experimental studies have demonstrated a general increase in dispersant effectiveness with increasing salinity. Clayton et al. (1993) and Byford et al. (1983) performed tests to determine the effect of salinity on dispersant effectiveness under low temperatures and high-energy conditions, using the Labofina-rotating flask test method. Those tests were conducted with seven dispersants and two types of crude oil. Test conditions were intended to simulate those commonly found in the Arctic environment. Results indicated an overall increase in dispersion with increasing salinity in five of the dispersants tested. Clayton et al. (1993) and Fingas (1991) studied the role of salinity on the effectiveness of three dispersants on three types of crude oil, using the swirling flask test method. These tests also showed an increase in dispersion with an increase in salinity from 0 to 45 psu. However, there could have existed maximum dispersion effectiveness or no effect when the salinity was investigated (Blondina et al., 1999; Moles et al., 2002). It depends on the dispersant–oil combination and the mixing energy.
A number of other factors, such as mixing energy, oil weathering, and temperature also influence dispersant effectiveness. Emulsions of oil droplets in water begin to form when sea energy is sufficient (Fingas, 1991; Fingas et al., 1995). After addition of the dispersants, mixing energy is further required to disperse the oil droplets formed. Clayton et al. (1993) reported that applications of dispersants reduce the interfacial tension between oil and water, resulting in the formation of oil droplets. Experimental studies performed by a number of scientists have indicated that the size of the oil droplets is inversely related to the amount of mixing energy input into test vessels. For example, Clayton et al. (1993) and Fingas et al. (1993) conducted experiments which indicated that the mixing energy reduces the size of the oil droplets. However, the available database for droplet size distribution is very limited. The chemical composition and physical properties of crude oil also determine the behaviour of the oil, and the way its properties will change when the oil is spilled at sea (Kristiansen et al., 1997). Weathering increases the viscosity of the oil through evaporation of the lighter components. Oil viscosity is perceived as a major factor affecting the dispersibility of oil (Canevari et al., 2001). As the oil weathers and the viscosity increases, the effectiveness of the chemical dispersant declines (Daling, 1989). Lower water temperature increases the viscosity of both the oil and the dispersant. A higher water temperature usually increases the solubility of dispersants in water, and also affects the spilled oil temperature. Hence, an increase in temperature reduces oil viscosity and leads to improvement of dispersion. Mackay and Szeto (1981), Byford et al. (1983), Lentinen and Vesala (1984), and Fingas (1991) conducted studies which indicated an increase in dispersion efficiency with increasing temperature. However, there have been conflicting results in the trend of dispersant effectiveness with either increasing or decreasing water temperature. For example, the results of studies performed by Byford et al. (1983) differed from those conducted by Fingas (1991).
To assess the impact of dispersant usage on oil spills, US EPA is currently developing and evaluating models (Weaver, 2004). Because of the complexity of chemical and physical interactions between spilled oils, dispersants, and the sea, an empirical approach to the interaction between the dispersant and the oil slick may provide a useful or practical approach for including dispersants in these models. The overall objective of this research was therefore to create a set of empirical data on three oils and two dispersants, by studying the variation in the effectiveness of the dispersants caused by changes in the salinity of seawater, temperature, oil type, oil weathering, dispersant type, and rotation speed. Recently, the US EPA (Sorial et al., 2004a, b) developed an improved dispersant testing protocol, called the baffled-flask test (BFT), which was the basis of the experiments conducted in the present study. The specific objectives were (i) to conduct a factorial experimental setup with four replicates to determine which of the factors salinity, temperature, oil type, oil weathering, dispersant type, and rotation speed of the BFT are related to the effectiveness of a dispersant used in oil spill response, and (ii) to determine empirical relationships between the amount of oil dispersed and the variables studied.
|
Material and methods |
|---|
|
TopIntroductionMaterial and methodsResults and discussionSummary and conclusionsReferences |
|---|
Modified 150-ml glass baffled trypsinizing flasks with screw caps at the top and Teflon stopcocks near the bottom were used in all experiments (Chandrasekar et al., 2005). An orbital shaker (Lab-Line Instruments Inc, Melrose Park, IL) with a variable speed control unit (40–400 rpm) and an orbital diameter of 0.75 inches (2 cm) was used to provide turbulence to solutions in test flasks. The shaker has a control speed dial to provide a reading of rpm within the instrument. The accuracy is within ±10%, as determined by the manufacturer. A Brinkmann Eppendorf repeater plus pipettor (Fisher Scientific, Pittsburgh, PA), capable of dispensing 4 µl of dispersant and 100 µl of oil with an accuracy of 0.3% and a precision of 0.25%, was used with 100-µl and 5-ml syringe tip attachments. Glassware consisting of graduated cylinders, 125-ml separatory funnels with Teflon stopcocks, pipettes, 50-ml crimp-style amber glass vials, and 50, 100, and 1000-µl gas-tight syringes were also used. A UV mini-1240 UV–VIS Spectrophotometer (UV–VIS spec) (Shimadzu Scientific Instruments, Inc, Wood Dale, IL) was used in all experiments to measure the dispersed oil concentration after extraction.
The synthetic sea salt "Instant Ocean" (Aquarium Systems, Mentor, OH) was used for all experiments at concentrations (salinity) of 10, 20, and 34 psu (the salinity of ocean water ranges between 32 and 37 psu). The synthetic seawater was prepared by adding sufficient water to 10, 20, or 34 g of salt to make 1 l of solution. Three types of oil samples provided by US EPA, South Louisiana Crude Oil (SLC), Prudhoe Bay Crude Oil (PBC), and Number 2 Fuel Oil (2FO), were used in the study. Of the three oils studied, SLC and PBC are light and medium weight EPA/API standard reference crude oils, respectively, and 2FO is light refined oil. Two dispersants that Venosa et al. (2002) found to be particularly effective dispersants were used in this study, here referred to as dispersants A and B. The solvent dichloromethane (DCM, pesticide quality) was used to extract all sample water and oil–standard water samples.
Weathering of oils
The three oils, SLC, PBC, and 2FO, were used in the study at three levels of weathering. They were weathered by bubbling nitrogen gas at a low flow rate at a room temperature of 22 ± 1°C to achieve volume losses of 0%, 10% (after 3 days), and 20% (after 7 days) for SLC and PBC, and 0%, 3.8% (after 3 days), and 7.6% (after 7 days) for 2FO.
Oil standard procedure
Standard solutions of oil for calibrating the UV–VIS spec were prepared with the specific reference oils and dispersant used for a particular set of experimental test runs. For control treatments with no dispersant, i.e. oil control experiments, only oil was used to make the standard solution. Initially, oil alone stock standard was prepared. The density of 2 ml of the specific reference oil with 18 ml DCM added was measured with a 1-ml gas-tight syringe, and the concentration of the oil solution was determined. Specific volumes of 20, 50, 100, 150, 200, or 300 µl of SLC–DCM stock, or 11, 20, 50, 100, 125, or 150 µl of PBC–DCM stock, or 150, 200, 400, 600, 800, or 1000 µl of 2FO–DCM stock were added to 30 ml of synthetic seawater in a separatory funnel, and extracted three times with DCM. The final DCM volume for the combined extracts was adjusted to 20 ml with DCM. The extracts were transferred to a 50-ml crimp-style glass vial with a Teflon/aluminium seal, wrapped with parafilm, mixed by inverting many times, and stored in a refrigerator at 4 ± 2°C until time of analysis (a maximum of 5 days). For treatments with oil plus dispersant, oil plus dispersant stock standard was first prepared. The density of 2 ml specific reference oil, 80 µl of the dispersant, and 18 ml of DCM was measured with a 1-ml gas-tight syringe, and the concentration was then determined. These stock solutions were used to prepare standard solutions as mentioned above.
Dispersant effectiveness procedure
A volume of 120 ml of synthetic seawater equilibrated at the desired temperature was added to the test flask, followed by the sequential addition of oil, and finally the dispersant. Then, 100 µl of oil was dispensed directly onto the surface of the synthetic seawater with an Eppendorf repeater pipettor with a 5-ml syringe tip attachment. The dispersant was then dispensed onto the centre of the oil slick with a 100-µl syringe tip attachment set to dispense 4 µl, giving a dispersant-to-oil ratio (DOR) of 1:25. The flask was placed on an orbital shaker and mixed for 10 min at the desired rotation speed, at the end of which it was removed from the shaker and allowed to remain stationary for another 10 min. At the end of the settling period, the first 2 ml of sample was drained from the stopcock and discarded, then 30 ml of sample was collected in a 50 ml measuring cylinder. The 30 ml sample was then transferred to a 125 ml separatory funnel, and extracted three times with fresh 5 ml DCM. The extract was then adjusted to a final volume of 20 ml and transferred to a 50-ml crimp-style glass vial with a Teflon/aluminium seal. These vials were stored at 4 ± 2°C until analysis (a maximum of 5 days). The oil standard procedure and test procedures were conducted according to the procedures given by Sorial et al. (2004a, b).
Sample analysis
The experimental sample extracts and the standard solutions were removed from the refrigerator and allowed to equilibrate at laboratory temperature. First, a blank solution (DCM) was introduced to zero the spectrophotometer. Then, the standard solutions were analysed in order of increasing concentration, and the absorbance values were noted at wavelengths of 340, 370, and 400 nm to generate a calibration curve (see below). After this, the experimental samples were analysed.
Calculation procedure
The area under the absorbance vs. wavelength curve for the standards and experimental samples between wavelengths 340 and 400 nm was calculated using the trapezoidal rule, according to the following equation:
|
| (1) |
The dispersant performance (i.e. the percentage of oil dispersed, or the effectiveness), based on the ratio of oil dispersed in the test system to the total oil added to the system, was determined from
|
| (2) |
oil is the density of the test oil (g l–1), Voil is the volume of oil added to the test flask (100 µl, i.e. 10–4 l), the total oil dispersed is the mass of oil x 120 ml/30 ml, the mass of oil (g) is the concentration of oil x VDCM, where VDCM is the final volume of the DCM-extract of water sample (0.020 l), and the concentration of oil (g l–1) is the area determined by Equation (1) divided by the slope of the calibration curve.
Factorial experimental design
The main aim was to determine the environmental factors that are related to the effectiveness of a dispersant used in oil remediation. The response variable for the experiment was the percentage effectiveness of the dispersant. The factors and levels of each of the factors are the following: salinity (10, 20, and 34 psu), weathering (0, 10, and 20% for SLC and PBC; 0, 3.8, and 7.6% for 2FO), dispersant (A or B), temperature (5°C, 22°C, and 35°C), and flask speed (150, 200, and 250 rpm). A complete factorial experiment was conducted with these levels for each factor. The total number of experimental samples prepared for each oil was 648 (3 salinities x 3 weathering levels x 3 temperatures x 3 flask speeds x 4 replicates x 2 dispersants). The factorial experiment was also performed for each of the three oils separately, i.e. with no dispersant added. The total number of oil control experimental samples prepared for each oil was 324 (3 salinities x 3 weathering levels x 3 temperatures x 3 flask speeds x 4 replicates).
|
Results and discussion |
|---|
|
TopIntroductionMaterial and methodsResults and discussionSummary and conclusionsReferences |
|---|
Statistical analysis was performed separately on each of the nine oil–dispersant combinations, i.e. three oils, with dispersants (A or B) and the oils alone. The results were analysed using analysis of variance, with
= 0.05. The highest order interaction in all cases was assumed to be non-significant, and its degrees of freedom were used for error determination. A significant interaction means that the effect of one input parameter varies at differing levels of another input parameter. The t-test from the REG (regression) procedure was used to test the level of significance for each factor studied. The REG procedure is a general-purpose procedure that performs linear regression analysis (SAS Institute, 2000). The condition for significance as determined by statistical analysis was that the probability of a run being greater than the corresponding Student's t-test value should be <0.0001 (Sorial et al., 2004b). Using this procedure, significant factors were determined for each oil–dispersant combination (see Table 1). More explanation is provided in Empirical relationships.
|
A four-replicate study was also conducted for all experiments to determine the precision of the experimental results for the range of variables studied. The precision objectives were determined using the relative standard deviation (RSD) for percentage effectiveness, based on four-replicate flasks. The acceptance criterion was based upon RSD < 15% (Venosa et al., 2002). The RSD was calculated as standard deviation x 100/average effectiveness.
The effect of salinity at different mixing energies
The mixing energy was provided in the form of revolutions per minute (rpm). Figure 1 shows the results for percentage effectiveness of dispersant A for the three oils in their unweathered condition (0% weathering) and at room temperature (22 ± 1°C). Clearly, the percentage effectiveness for a given oil at 0% weathering and given salinity increased as the speed of the orbital shaker increased from 150 to 250 rpm. This trend in dispersant effectiveness with increase in flask speed was true for oil + dispersant B experiments as well as for oil control experiments (results not shown).
|
Figure 1 also shows that the dispersant effectiveness increased with an increase in salinity from 10 to 34 psu for SLC at a given flask speed. For SLC, the RSD values for dispersant effectiveness among the three salinities at the three temperatures and three weathering levels studied were 7.6, 6.7, and 5.6 at 150, 200, and 250 rpm flask speed, respectively. Therefore, in the case of SLC, the impact of salinity on dispersant effectiveness is nearly the same at the three levels of mixing, and this behaviour is also evident in Figure 1.
For PBC, Figure 1 shows that the impact of salinity is more pronounced at the intermediate speed (200 rpm) than at the other speeds. This behaviour was further confirmed for the other temperatures and weathering conditions studied, by calculating the RSD values of dispersant effectiveness at differing salinity. The RSD values were 4.01, 15.1, and 2.6 at 150, 200, and 250 rpm flask speed, respectively. Comparing the RSD value of 15.1 with the other values obtained could imply a significant impact of salinity at 200 rpm. However, the RSD was very close to 15%, and hence no significance role of salinity could be confirmed.
In the case of 2FO, the RSD values for dispersant effectiveness at different salinity for the three temperatures and weathering conditions were 8.4, 5.9, and 1.4 at 150, 200, and 250 rpm flask speed, respectively. As the RSD values are <15%, it is concluded that salinity played no significant role.
The effect of salinity at different temperatures
Figure 2 shows the results for percentage effectiveness of dispersant A for the three oils at their maximum weathering level (20% for SLC and PBC, 7.8% for 2FO) and at maximum flask speed (250 rpm). In the case of SLC at 10 and 34 psu, the percentage effectiveness increased with increase in temperature from 5 ± 1°C to 22 ± 1°C, but decreased at 35 ± 1°C. For example, the dispersant effectiveness values at 5 ± 1°C, 22 ± 1°C, and 35 ± 1°C, were 73.3, 81.5, and 77.7%, respectively, at 10 psu, and 84.1, 92.8, and 90.7%, respectively, at 34 psu. However, the results obtained at 20 psu show that the percentage effectiveness increased with a rise in temperature. For example, the dispersant effectiveness values at 5 ± 1°C, 22 ± 1°C, and 35 ± 1°C were 77.1, 80.9, and 84.7%, respectively. Figure 2a also shows that dispersant effectiveness increased with an increase in salinity from 10 to 34 psu at 5 ± 1°C and 35 ± 1°C. Also, for SLC, the RSD values for dispersant effectiveness at different salinity for the weathering and speed conditions studied were 6.9, 7.9, and 7.7 at 5 ± 1°C, 22 ± 1°C, and 35 ± 1°C, respectively. As the RSD values are <15%, we conclude that there was no impact of salinity on dispersant effectiveness. This conclusion is also evident from Table 1.
|
For PBC, for all three salinities, the percentage effectiveness first increased with a rise in temperature from 5 ± 1°C to 22 ± 1°C, then decreased at 35 ± 1°C. This means that PBC, a medium crude oil, resists dispersion even at high temperature and a high salinity of 34 psu. This behaviour may be due to weathering of the oil during the test at 35°C. For PBC, the RSD values for dispersant effectiveness at the different salinities were 5.1, 2.3, and 5.8 at 5 ± 1°C, 22 ± 1°C, and 35 ± 1°C, respectively. For PBC, the impact of salinity at the three flask speeds and weathering conditions was less pronounced than for SLC, a finding also justified in Table 1.
In the case of 2FO at 10 and 34 psu, the dispersant percentage effectiveness increased with a rise in temperature from 5 ± 1°C to 22 ± 1°C, but it decreased at 35 ± 1°C. However, the results obtained at 20 psu show that the percentage effectiveness increased with a rise in temperature. For 2FO, the RSD values for dispersant effectiveness at different salinity were 5.6, 15.3, and 10.9 at 5 ± 1°C, 22 ± 1°C, and 35 ± 1°C, respectively. Hence, the significance of salinity on dispersant effectiveness is more pronounced at the higher temperatures of 22 ± 1°C and 35 ± 1°C than at 5 ± 1°C. However, the RSD values are less than or close to 15%, and Table 1 indicates clearly that salinity was not a significant factor.
The effect of salinity at different weathering
The effect of salinity at the three weathering conditions studied is presented in Figures 3 and 4, in order to observe the behaviour at the lowest and highest temperatures studied. Figure 3 shows the results for percentage effectiveness of dispersant A on all three oils, at a flask speed of 200 rpm, and at a temperature of 5 ± 1°C. In general, for any oil at a given salinity, as the degree of weathering of the oil increased, the dispersant effectiveness decreased. For example, for SLC at 200 rpm flask speed and 10 psu, the values of dispersant effectiveness at 0%, 10%, and 20% weathering were 72%, 61.5%, and 60.9%, respectively. This is true for oil + dispersant B experiments as well as oil control experiments conducted (results not shown). From Figure 3, we also conclude that dispersant effectiveness increased with increase in salinity from 10 to 34 psu for all three oils at given weathering. For SLC, the RSD values for dispersant effectiveness at different salinity were 3.6, 7.6, and 6.7 at 0%, 10%, and 20% weathering, respectively. This implies that salinity has a greater impact on dispersant effectiveness at the higher weathering levels of 10% and 20%. However, the RSD values are <15%. For PBC, the RSD values for dispersant effectiveness among different salinities were 6.1, 7.0, and 6.5 at 0%, 10%, and 20% weathering, respectively, showing a similar behaviour to that of SLC. For 2FO, the RSD values for dispersant effectiveness at different salinity were 7.7, 5.0, and 6.0 at 0%, 3.8%, and 7.6% weathering, respectively. In the case of 2FO, the impact of salinity on dispersant effectiveness was more pronounced at 0% weathering than at the other weathering conditions, but again the RSD values are <15%.
|
|
Figure 4 shows the results for percentage effectiveness of dispersant A on all three oils at an intermediate flask speed of 200 rpm and a high temperature (35 ± 1°C). Again, as the degree of weathering of the oil increased, the dispersant effectiveness decreased. For SLC, the RSD values for dispersant effectiveness at different salinity were 6.8, 14.8, and 11.4 at 0%, 10%, and 20% weathering, respectively. This implies that salinity has a greater impact on dispersant effectiveness at weathering levels of 10% and 20% than it does at 0%. For PBC, the RSD values for dispersant effectiveness at different salinity were 11.5, 11.6, and 10.8 at 0%, 10%, and 20% weathering, respectively. For 2FO, the RSD values for dispersant effectiveness at different salinity were 17.7, 16.6, and 13.9 at 0%, 3.8%, and 7.6% weathering, respectively. In the case of 2FO, Figure 4 shows that with increase in salinity from 10 to 20 psu, the dispersant percentage effectiveness increased, but that it decreased at 34 psu at all three levels of weathering, indicating a negative impact of salinity.
General discussion
Table 2 shows the effect of salinity at different mixing energies. The percentage effectiveness values for each oil–dispersant combination and oil control experiment are shown. The range in values of effectiveness is shown for all temperatures and weathering levels studied. The corresponding RSD values were calculated. Overall, for SLC, the RSD range among dispersant effectiveness values was 9–46%, 18–51%, and 26–110%, for experiments with dispersants A, B, and oil controls, respectively. Similarly for PBC, the RSD range was 20–53%, 17–54%, and 24–100%, for experiments with dispersants A, B, and oil controls, respectively. For 2FO, the ranges of RSD among dispersant effectiveness values were 17–64%, 24–54%, and 15–52%, for experiments with dispersants A, B, and oil controls, respectively. If an RSD value > 15% is considered to be significant, then salinity plays an important role in determining the significance of flask speed on dispersant effectiveness for all three oils. This is also evident from Table 1, which lists the significant factors for each of the oil–dispersant combinations. Speed was also a significant factor for SLC and PBC oil control experiments.
|
The effect of salinity at different temperatures is reflected by the results listed in Table 3, which lists the percentage effectiveness values for each oil–dispersant combination and oil control experiment. The range in effectiveness values is shown for all flask speeds and weathering levels studied. The corresponding RSD values for these results were calculated. In the case of SLC, the RSD among dispersant effectiveness values varied between 3% and 23%, 3% and 30%, and 44% and 124%, for experiments with dispersants A, B, and oil controls, respectively. Similarly for PBC, the RSD ranged between 10% and 54%, 1% and 44%, and 3% and 80%, for the same experiments, respectively. For 2FO, the RSD among dispersant effectiveness values ranged between 8% and 65%, 14% and 45%, and 3% and 82%, for the same experiments, respectively. Again, if an RSD value > 15% is considered to be significant, then salinity plays an important role in determining the significance of temperature on dispersant effectiveness, for all three oils except for SLC with dispersant A, which gave comparatively less RSD than the other experiments. This is also evident from Table 1, which shows that temperature is indeed a significant factor for all oil–dispersant combinations except SLC with dispersant A. Temperature was also a significant factor for SLC oil control experiments.
|
The effect of salinity at different degrees of weathering can be seen in Table 4, which lists the percentage effectiveness values for each oil–dispersant combination and oil control experiment. The range in effectiveness values is shown for all flask speeds and temperatures studied. The corresponding RSD values for these results were calculated. Overall, for dispersant A experiments, the RSD among dispersant effectiveness values varied between 3% and 36% for SLC, between 1% and 12% for PBC, and between 1% and 14% for 2FO. For dispersant B experiments, the RSD among dispersant effectiveness values varied between 1% and 6% for SLC, between 1% and 9% for PBC, and between 1% and 9% for 2FO. For oil control experiments, the RSD among dispersant effectiveness values varied between 1% and 15% for SLC, between 1% and 11% for PBC, and between 1% and 14% for 2FO. Based on these RSD values, we conclude that the impact of salinity on weathering was significant for SLC with dispersant A experiments only (also evident from Table 1).
|
Empirical relationships
A linear regression model was fitted to the experimental data for each of the oil–dispersant combinations, utilizing the REG procedure (SAS Institute, 2000). All factor terms and their interactions were included in the model, regardless of their significance. This is needed to define all interactions and quadratic relationships. The model takes the form
|
| (3) |
|
| (4) |
is the tabulated t-test value at = 0.05 (Montgomery, 1991). A parameter is significant if the probability of a run being greater than the corresponding calculated t-value is <0.0001.
|
Figure 5 shows a comparison of measured and estimated values of dispersant effectiveness on SLC. Each of the plots shows the data cluster along the 1:1 line, indicating a close match between estimated and measured values. Figure 5b, for SLC with dispersant A, and Figure 6c, for SLC with dispersant B, especially show a good match between measured and estimated values owing to the high values of r2. Similarly, Figure 6 shows a comparison of measured and estimated values of dispersant effectiveness on PBC. In the case of PBC, there is tight clustering along the 1:1 line for PBC with dispersants A and B, the r2-values being 91.37% and 98.03%, respectively.
|
|
Figure 7 is a comparison of measured and estimated values of dispersant effectiveness on 2FO. Figure 7b and c, for dispersants A and B, especially shows a tight cluster along the 1:1 line for 2FO, for which the r2-values were as high as 92.09% and 95.88%, respectively.
|
|
Summary and conclusions |
|---|
|
TopIntroductionMaterial and methodsResults and discussionSummary and conclusionsReferences |
|---|
A full factorial experiment with four replicates was conducted to determine the impact of salinity on three environmental factors, the mixing energy, the temperature, and the oil weathering. All experiments were analysed using an analysis of variance with
= 0.05. The REG procedure was used to perform linear regression analysis. Results from the BFT experiments conducted revealed the general observations listed below:
- Dispersion efficiency increased with increase in mixing energy, with no exceptions. The impact of mixing energy is more pronounced than the impact of salinity for the different oil combinations considered.
- Dispersion efficiency does not follow a general trend with increase in temperature, and is different for each oil, depending on the oil properties. The impact of salinity on dispersant effectiveness is more pronounced at higher temperature than at lower temperature, i.e. the significance of salinity on dispersant effectiveness increased with increase in temperature for all three oils. In general, salinity plays an important role in determining the significance of temperature on dispersant effectiveness for all oil–dispersant combinations, except SLC with dispersant A.
- Dispersion efficiency decreased with increase in the level of weathering for only one oil–dispersant combination. The impact of weathering is only significant for SLC with dispersant A.
- In general, dispersion efficiency increased with increase in salinity for most oil–dispersant combinations.
- This research work has successfully created a set of empirical data on three oils and two dispersants that could serve as an input to the oil-spill simulation models being developed by EPA. The empirical correlation for the collected experimental data predicted with good accuracy the effectiveness of the dispersant. The results of this research are expected to be incorporated into EPA's model of oil spills (Weaver, 2004).
|
Acknowledgements |
|---|
This research was supported by the U.S. Environmental Protection Agency (US EPA) under Contract no. 68-C-00-159. Although this work was reviewed by EPA and approved for publication, it may not reflect official agency policy. The comments of the reviewers were much appreciated.
|
References |
|---|
|
TopIntroductionMaterial and methodsResults and discussionSummary and conclusionsReferences |
|---|
-
Blondina G.J., Singer M.M., Lee I., Ouano M.T., Hodgins M., Tjeerdema R.S., Sowby M.L. (1999) Influence of salinity on petroleum accommodation by dispersants. Spill Science and Technology Bulletin 5:127–134.[CrossRef]
Byford D.C., Green P.J., Lewis A. (1983) Factors influencing the performance and selection of low-temperature dispersants. Proceedings of the Sixth Arctic Marine Oil Spill Program, Edmonton(Environmental Protection Services, Environment Canada, Gatineau, Quebec, Canada) pp. 140–150.
Canevari G.P., Calcavecchio P., Becker K.W., Lessard R.R., Fiocco R.J. (2001) Key parameters affecting the dispersion of viscous oil. Proceedings of the International Oil Spill Conference, Tampa, Florida(American Petroleum Institute, Washington, DC) pp. 11–20.
Chandrasekar S., Sorial G.A., Weaver J.W. (2005) Dispersant effectiveness on three oils under various simulated environmental conditions. Environmental Engineering Science 22:324–336.[CrossRef][Web of Science]
Clayton J.R., Payne J.R., Farlow J.S., Sarwar C. (1993) Oil Spill Dispersants Mechanisms of Action and Laboratory Tests(CRC Press, Boca Raton, Florida) 103 pp.
Daling P.S. (1989) A study of the chemical dispersibility of fresh and weathered crudes. Eleventh Arctic and Marine Oilspill Program, Vancouver, British Columbia(Environmental Protection Services, Environment Canada, Gatineau, Quebec, Canada) pp. 481–499.
Fingas M.F. (1991) Dispersants: a review of effectiveness measures and laboratory physical studies(Environmental Emergencies Technology Division, Environment Canada, Ottawa, Ontario, Canada).
Fingas M.F., Fieldhouse B., Mullin J.V. (1995) Water-in-oil emulsions: how they are formed and broken. Eighteenth Arctic Marine Oil Spill Program Technical Seminar, Ottawa, Ontario(Environmental Protection Services, Environment Canada, Gatineau, Quebec, Canada) pp. 21–42.
Fingas M.F., Kyle D.A., Tennyson E.J. (1993) Physical and chemical studies on dispersants: the effect of dispersant amount and energy. Proceedings of the Sixteenth Arctic and Marine Oil Spill Program Technical Seminar, Environment Canada, Ottawa, Ontario(Environmental Protection Services, Environment Canada, Gatineau, Quebec, Canada) pp. 861–876.
Hoult D.P. (1972) Oil spreading on the sea. Annual Reviews in Fluid Mechanics 4:341–368.[CrossRef]
Kristiansen T.S., Lewis A., Daling P.S., Hokstad J.N., Singsaas I. (1997) Weathering and dispersion of naphthenic, asphaltenic, and waxy crude oils. Proceedings of the International Oil Spill Conference, Washington, DC(American Petroleum Institute, Washington, DC).
Lentinen C.M. and Vesala A.M. (1984) Effectiveness of oil dispersants at low salinities and low water temperatures. Symposium of Oil Spill Chemical Dispersants – Research Experience and Recommendations, West Palm Beach, Florida(American Petroleum Institute, Washington, DC) pp. 108–121.
Lessard R.R. and Demarco G. (2000) The significance of oil spill dispersants. Spill Science and Technology Bulletin 6:59–68.[CrossRef]
Mackay D., Chau A., Hossain K., Bobra M. (1984) Measurement and prediction of the effectiveness of oil spill chemical dispersants. Oil Spill Chemical Dispersants, Research, Experience and Recommendations, ASTM STP 840(American Society for Testing and Materials, Philadelphia, PA) pp. 38–54.
Mackay D. and Szeto F. (1981) The laboratory determination of dispersant effectiveness – method development and results. Proceedings of the International Oil Spill Conference, Atlanta, GA(American Petroleum Institute, Washington, DC) pp. 331–337.
Moles A., Holland L., Short J. (2002) Effectiveness in the laboratory of Corexit 9527 and 9500 in dispersing fresh, weathered, and emulsion of Alaska North Slope Crude Oil under subArctic conditions. Spill Science and Technology Bulletin 7:241–247.[CrossRef]
Montgomery D.C. (1991) Design and Analysis of Experiments 3rd edn (John Wiley, New York).
NRC. (1985) Oil in Sea: Inputs, Fates, and Effects(National Research Council, National Academy Press, Washington, DC) 601 pp.
NRC. (1989) Using oil spill dispersants on the sea. Report of the Committee on Effectiveness of Oil Spill Dispersants(National Research Council, National Academy Press, Washington, DC) 335 pp.
SAS Institute. (2000) The REG Procedure Overview – SAS/STAT User's Guide http://www.id.unizh.ch/software/unix/statmath/sas/sasdoc/stat/chap55/sec1.htm.
Sorial G.A., Venosa A.D., Koran K.M., Holder E., King D. (2004) Oil spill dispersant effectiveness protocol. 1. Impact of operational variables. ASCE Journal of Environmental Engineering 130:1085–1093.[CrossRef]
Sorial G.A., Venosa A.D., Koran K.M., Holder E., King D. (2004) Oil spill dispersant effectiveness protocol. 2. Performance of the revised protocol. ASCE Journal of Environmental Engineering 130:1073–1084.[CrossRef]
Venosa A.D., King D.W., Sorial G.A. (2002) The baffled flask test for dispersant effectiveness: a round robin evaluation of reproducibility and repeatability. Spill Science and Technology Bulletin 7:299–308.[CrossRef]
Weaver J.W. (2004) Characteristics of Spilled Oils, Fuels, and Petroleum Products: 3a. Simulation of Oil Spills and Dispersants Under Conditions of Uncertainty(United States Environmental Protection Agency, Washington, DC) EPA 600/R-04/120.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||