- Email: [email protected]
Pollution risk assessment of oil spill accidents in Garorim Bay of Korea
MPB-07143; No of Pages 7 Marine Pollution Bulletin xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Marine Pollution Bulletin journal homepage: www.elsevier.com/locate/marpolbul
Pollution risk assessment of oil spill accidents in Garorim Bay of Korea Moonjin Lee a,b, Jung-Yeul Jung b,c,⁎ a b c
Maritime Safety Research Division, Korea Research Institute of Ships and Ocean Engineering, KIOST, Daejeon 305-343, Republic of Korea Ocean Science and Technology (OST) School, KMOU/KIOST Joint Program, Busan 606-791, Republic of Korea Technology Center for Offshore Plant Industries, Korea Research Institute of Ships and Ocean Engineering, KIOST, Daejeon 305-343, Republic of Korea
a r t i c l e
i n f o
Article history: Received 6 February 2015 Received in revised form 21 August 2015 Accepted 23 August 2015 Available online xxxx Keywords: Aquaculture farm Coastline Impact probability Impact time Oil spill accident Pollution risk assessment
a b s t r a c t This study presents a model to assess the oil spill risk in Garorim Bay in Korea, where large-scale oil spill accidents frequently occur. The oil spill risk assessment is carried out by using two factors: 1) The impact probability of the oil spill, and 2) the first impact time of the oil that has been spilt. The risk assessment is conducted for environmentally sensitive areas, such as the coastline and aquaculture farms in the Garorim Bay area. Finally, Garorim Bay is divided into six subareas, and the risks of each subarea are compared with one another to identify the subarea that is most vulnerable to an oil spill accident. These results represent an objective and comprehensive oil spill risk level for a specific region. The prediction of the oil spill spread is based on real-time sea conditions and can be improved by integrating our results, especially when sea conditions are rapidly changing. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction It is expected for the industrialization caused by economic growth to proceed more actively in coastal areas than in inland areas (FAO, 1998). Recently, the frequency of oil spill accidents has increased as maritime transportation activity has increased, damaging not only the economics of coastal communities but also marine ecosystems. South Korea is not an exception and has suffered from oil spill accidents in the past. In 1995, the oil spill accident of Sea Prince occurred near Sori Island, Yeo-Chun, Jeon-Nam Province while seeking refuge from typhoon “Faye” (Cho, 2007). Approximately 5000 tons of Arabian crude oil were spilt, ~ 230 km of coastline was contaminated, and the clean up effort lasted ~ 5 months. In 1995, the oil spill accident of Yuil No. 1 occurred near Nam-Hyeong-Jae Island while being towed toward the port of Busan (Shim et al., 2001). More than 2000 tons of oil were spilt. Also the Honam Sapphire spilt over 1000 tons of crude oil near Yeo-Cheon harbor in 1995 where Sea Prince's collision had happened 4 months before. The Hebei Spirit oil spill was Korea's worst oil spill, and it began on 7 December 2007 (Lee et al., 2009). The Hebei Spirit collided with a crane barge owned by Samsung Heavy Industry near the port of Daesan in Taean County on the Yellow Sea. As a result of the collision, ~10,800 tons of oil were spilt. The port of Daesan is located at the entrance of Garorim Bay, which is the area for which the pollution risk assessment is performed in this study. ⁎ Corresponding author at: Maritime Safety Research Division, Korea Research Institute of Ships and Ocean Engineering, KIOST, Daejeon 305-343, Republic of Korea. E-mail address: [email protected] (J.-Y. Jung).
The following response strategies are essential to minimize the pollution damage caused by oil spills: initial response, effective prevention methods including oil boom and recovery, and rapid aid transport. Effective prevention methods require for adequate response strategies to be established in advance of an oil spill accident, and a risk assessment map should include the location and importance of resources that are to be protected, the areas with a high risk of oil spill accidents, the impact probability of spilt oil, and the first impact time of spilt oil in order to establish an effective response strategy. An accurate estimation of the trajectory of the spilt oil is very important to effectively undertake a response to oil spill, and many studies have focused on developing oil spill trajectory modeling and simulation systems (Reed et al., 1999; French-McCay, 2004; Castanedo et al., 2006; Delgado et al., 2006; Guo et al., 2009; Alves et al., 2014; Lan et al., 2015; Melaku Canu et al., 2015; Nixon and Michel, 2015). Since spilt oil is transported by external environmental forces such as the current, waves and wind, the trajectory of the spilt oil is usually estimated using real-time meteorological and oceanic conditions. However, in the case of quickly-changing sea conditions, there is a high probability of an error in the trajectory of the spilt oil that is estimated using realtime observations of the conditions at sea. Although the trajectory estimation of the spilt oil comes with individual uncertainty, the collection and analysis of data from past oil spill accidents can provide statistical guidance to reducing the damage caused by oil pollution. In other words, an oil spill estimation system should be accompanied with a comprehensive risk assessment in order to reduce the negative impact of the spill (Skognes and Johansen, 2004; Elshorbagy and Elhakeem, 2008; Grifoll et al., 2010; Olita et al., 2012). However, most results that
http://dx.doi.org/10.1016/j.marpolbul.2015.08.037 0025-326X/© 2015 Elsevier Ltd. All rights reserved.
Please cite this article as: Lee, M., Jung, J.-Y., Pollution risk assessment of oil spill accidents in Garorim Bay of Korea, Marine Pollution Bulletin (2015), http://dx.doi.org/10.1016/j.marpolbul.2015.08.037
2
M. Lee, J.-Y. Jung / Marine Pollution Bulletin xxx (2015) xxx–xxx
provide an oil spill trajectory modeled using real-time conditions at sea have been limited to specific oil spill accidents in Korea, which means that a comprehensive risk assessment is not yet available. The configuration of our preliminary study (Lee and Kim, 2009) is used to develop a model to assess the pollution risk of oil spill accidents. The purpose of this study is thus to offer a comprehensive and objective risk assessment model based on statistical methods in order to address the shortcomings of oil spill trajectory systems, especially in the case where sea conditions change quickly, to provide assistance in establishing a response strategy in advance. In this study, the coastline and aquaculture farms in the Garorim Bay area are considered as the risk assessment target area. The history of past spill accidents around Garorim Bay was reviewed to select the site where oil spill accidents have most frequently occurred. We assess the pollution risk of an oil spill in two aspects, 1) the impact probability of the spilt oil, and 2) the first impact time of the spilt oil. To make a quantitative assessment of the pollution risk, the oil spill trajectories are simulated 80 different times. In other words, many oil spill trajectories are statistically simulated based on past sea conditions to analyze and assess the oil spill hazard on coastlines and aquaculture farms. These trajectories are then collected and analyzed to produce a comprehensive and objective risk assessment. The Garorim Bay area was divided into six subareas, and the subarea most vulnerable to oil spills was then determined. 2. Assessment of oil pollution risk 2.1. Risk indicators of oil pollution The risk of oil pollution can be evaluated according to the impact probability and the first impact time of the spilt oil. The impact probability is a very important information when deciding which area to protect from spilt oil. This information can thus be used to effectively allocate resources and plan a response for an oil spill. The first impact time of the spilt oil can provide information on the response priorities.
Therefore, it is essential to set priorities due to the limited availability of resources. The impact probability of spilt oil is quantitatively defined as the average ratio of the amount of oil spilt in each specific area to the total amount of oil spilt, which can be expressed using Eq. (1), P¼
N 1X Ai N i¼1 Atotal
ð1Þ
where, P is the impact probability and N is the number of oil spill events, Ai is the amount of oil spilt from the ith event in each specific area and Atotal is the total amount of oil spilt from the ith event. The first impact time of the spilt oil is defined as the time that has elapsed to reach a specific area. The elapsed times can be calculated as follows, t ¼ T−T 0
ð2Þ
where, t is the first impact time, T is the first attached time of oil to the area, and T0 is the time of the oil spill incident. T and T0 are in local standard time. 2.2. Study area (Garorim Bay) 2.2.1. General feature Garorim Bay has been designated as the risk assessment target area because huge petrochemical plants and transporting ships are populated and plied. Garorim Bay is located on the western coast of the Korean Peninsula, as shown in Fig. 1, and it is famous for laver and oyster cultures and has rich spawning grounds for important species. The Daesan industrial complex includes large-scale petrochemical plants and Daesan harbor is located in the northeast coast outside of the Bay. Due to the narrow entrance of Garorim Bay, once the spilt oil is entrained in the Bay area, the oil will remain in the Bay area, and there is a low
Fig. 1. Location of Garorim Bay.
Please cite this article as: Lee, M., Jung, J.-Y., Pollution risk assessment of oil spill accidents in Garorim Bay of Korea, Marine Pollution Bulletin (2015), http://dx.doi.org/10.1016/j.marpolbul.2015.08.037
M. Lee, J.-Y. Jung / Marine Pollution Bulletin xxx (2015) xxx–xxx
3
Table 1 List of serious accidents that occurred around Garorim Bay, from 1991 to 2000 (Lee and Kim, 2009). Ship's category
Date
Ship's name
Total weight (tons)
Location of accident
Cause of accident
Volume of oil spilled (kl)
10 ~ 100 kl
11.19.1995
Hwa-jeon (cargo)
1952
1 mile north off the island of Huwang
Sinking
100 ~ 1000 kl
07.05.2000
Safe Express (cargo)
1250
North of the island of An
Sinking
02.21.1991
Pacific Friend (cargo)
4417
1.7 miles west of Kyonggi
Sinking
12.02.1996 04.11.2000 10.01.1993
Tae-young Jasmine (cargo) 501 (dredge) Frontier Express (tanker)
2483 1458 40,721
6 miles northwest of the island of An Pyung-taek harbor Off shore from a lighthouse in Daesan harbor
Sinking Sinking Collision
B/B 7.9 B/A 2.7 B/C 30 B/A 10 WTI 155 MDO 41 B/C 159 MDO 165 Petroleum naphtha 8320
b1000 kl
expectation for further natural decay by the weathering process. Therefore, the Garorim Bay area is one of the most vulnerable areas to oil spill accidents in Korea. 2.2.2. Site of an oil spill accident For the past 10 years, the most frequent accident sites (1991–2000) have been determined as hypothetical oil spill accident sites. Table 1 shows a list of serious accidents that have occurred near Garorim Bay, and Fig. 2 shows the location of the accidents listed in Table 1. As seen in Fig. 2, most of the accidents have taken place near Daesan harbor. Therefore, this study designates the entry fairway into Daesan harbor as the hypothetical site where an oil spill accident occurs. 2.2.3. Target of risk assessment The Environmental Sensitivity Index (ESI) Map of Korea can be utilized to select the target for the risk assessment for oil pollution. The ESI Map of Korea includes various types of information related to an oil spill response, such as the coastline sensitivity classifications for oil pollution, sensitive biological resources, socio-economic features, and oil spill response resources. The coastline and aquaculture farms in Garorim Bay that are environmentally sensitive are selected as a target for the risk assessment of oil pollution. Fig. 3 shows the target areas split into a small grid to quantify the impact probability.
3.2. Simulation of oil behavior 3.2.1. Oil spill prediction model The spilt oil is transported by advection dominated by currents and winds while the diffusion occurs as a result of turbulence. The volume of spilt oil decreases due to biochemical reactions and sinking. In this study, the sea current is calculated by solving a hydrodynamic equation with the effect of the tides and wind. To simulate the behavior of the spilt oil, a Lagrangian model based on a numerical particle tracking method was adopted. To provide realistic results, turbulent diffusion and weathering of the spilt oil are included in the Lagrangian model, and a detailed description is given in Section 3.2.2. In our spill prediction model, numerical particles are released and tracked to simulate the movement and weathering of spilt oil. The movement of spilt oil is represented by the displacement of the particles in each time step. If a particle is located at (x0, y0) at time t and then moves to a new location (x0 + δx, y0 + δy) in the time interval δt, the displacements δx, δy during a time δt are given respectively by, δx ¼ ðU þ u0 Þδt
ð3Þ
δy ¼ ðV þ v0 Þδt:
ð4Þ
3. Simulation of oil spill incident 3.1. Preparing scenarios of oil spill incidents In order to accurately assess the pollution risk of an oil spill (i.e., the impact probability of spilt oil and the first impact time of spilt oil), the oil trajectories of past oil spill accidents should be collected and analyzed under various conditions. In fact, data on oil spill accidents is limited, and the characteristics of the oil trajectory are also difficult to observe in the real sea. Therefore, various hypothetical oil spill accidents were simulated in order to obtain data for the trajectory of the spilt oil. The time for hypothetical accidents varied through four seasons in order to consider the various ocean conditions for a specific time from 1983 to 2002. The behavior of the spilt oil is simulated and is analyzed for a period of five days. Table 2 shows the simulations for the statistic trajectory analysis with the tidal time for each season, which is almost even, which means that the cases analyzed in this study will show statistical results for various sea conditions. As mentioned in Section 2.1, two of the major risk indicators for the assessment in this study, the impact probability and the first arrival time of spilt oil, are calculated from the trajectories of 200 random oil spill accident scenarios. The number of random simulations according to tidal times and seasons is listed in Table 2. Each simulation represents a five day duration for an oil spill after hypothetical accidents at random times between 1983 and 2002.
Fig. 2. Locations of serious oil spill accidents around Garorim Bay from 1991 to 2000 and hypothetical accident site for pollution risk assessment (Lee and Kim, 2009).
Please cite this article as: Lee, M., Jung, J.-Y., Pollution risk assessment of oil spill accidents in Garorim Bay of Korea, Marine Pollution Bulletin (2015), http://dx.doi.org/10.1016/j.marpolbul.2015.08.037
4
M. Lee, J.-Y. Jung / Marine Pollution Bulletin xxx (2015) xxx–xxx
Fig. 3. Index cells for the estimation of oil spill risk around Garorim Bay (Lee and Kim, 2009).
Fig. 4. Observed trajectories of drifting buoys (10:42 ~ 17:50, Dec. 11, 2002) (Lee and Kim, 2009).
The current velocity (U, V) is predicted by the results of the simulation based on the hydrodynamic model discussed in Section 3.2.2, and (u′, v′) represents the effect of the turbulence. Fractional Brownian motion allows for the turbulent diffusion to be computed (Guo et al., 2009; Bo and Addison, 2010; Osborne et al., 2011). In fractional Brownian motion, the Hurst coefficient in South Korea seas range from 0.45 to 2.46 (Lee, 1996).
The tidal current is the most important component in coastal environments in South Korea, and the maximum tidal height is 10 m. In this study, real-time tides are simulated by combining the harmonic constants of the tidal currents and the variations in astronomical arguments (Foreman, 1978). The harmonic constants are computed through a harmonic analysis of the results of the hydrodynamic model. In addition to deterministic tidal currents, time-varying wind-driven currents play an important role in the coastal region. Due to the complexity of coastal geometry and the bottom topography, the response of the current to winds differs from place to place and is not instantaneous but delayed. The strength and delay time of the wind-driven current are dependent on the strength and duration of the wind. Therefore, it is necessary to define the impulse response of the currents to the
3.2.2. Simulation of oceanic currents The current that mainly governs the spread of the oil spill is calculated by solving a hydrodynamic equation that includes the effect of the tide and the wind. The basic equations that are used for this model are a continuity equation and two-dimensional, depth-averaged hydrodynamic equations of motion given by ∂ζ ∂U ðD þ ζ Þ ∂V ðD þ ζ Þ þ þ ¼0 ∂t ∂x ∂y
ð5Þ
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ∂U ∂U ∂U ∂ζ kU U 2 þ V 2 ¼0 þU þV þg −f V þ Dþζ ∂t ∂x ∂y ∂x
ð6Þ
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ∂V ∂V ∂V ∂ζ kV U 2 þ V 2 ¼ 0; þU þV þg þ fU þ Dþζ ∂t ∂x ∂y ∂y
ð7Þ
where U and V are depth average current components in the x and y directions, respectively, g is the gravitational acceleration, ζ is the elevation from the mean sea level, D is the water depth from the mean sea level, f is the Coriolis parameter (f = 2Ωsinϕ, ϕ = 37°N), and k is the bottom friction coefficient (k = 0.003).
Table 2 Simulation cases for the statistical trajectory analysis. Tidal time
Season
Flood High Ebb Low Spring Summer Fall Winter flow water flow water No. of computation 51
50
49
50
50
50
50
50
Fig. 5. Simulated trajectories of drifting buoys (10:42 ~ 17:50, Dec. 11, 2002) (Lee and Kim, 2009).
Please cite this article as: Lee, M., Jung, J.-Y., Pollution risk assessment of oil spill accidents in Garorim Bay of Korea, Marine Pollution Bulletin (2015), http://dx.doi.org/10.1016/j.marpolbul.2015.08.037
M. Lee, J.-Y. Jung / Marine Pollution Bulletin xxx (2015) xxx–xxx
5
Fig. 6. Impact probability of spilt oil in the coastline of Garorim Bay.
Fig. 8. First impact time of spilt oil to the coastline of Garorim Bay.
recent history of winds in any desired place in order to predict the winddriven currents. In the present study, we have predicted wind-driven currents using a kernel coefficient of the response functions between the winds and wind-driven currents.
(ADIOS) can compute the weathering effect, including the evaporation, emulsification, etc., according to oil types by using a database of oil properties. 3.3. Verifying the simulated oil behavior
3.2.3. Weathering of oil The amount of spilt oil that decays due to the weathering effect depends on the properties of the oil. In our oil spill prediction model, the number of numerical particles decreases according to the rate of decay of specific oil types since it reacts physically and biochemically in water. The rate of decay of oil is computed using the NOAA oil spill response tool database, and the Automated Data Inquiry for Oil Spills
In order to verify our oil spill prediction model, the trajectories of the drifters are experimentally observed near the entrance of Garorim Bay and are compared to the results of the simulation from our model. Fig. 4 represents the trajectories observed for 3 drifters that were simultaneously released at the same starting position. The trajectories of the drifters were tracked for a period of 7 h and 8 min from 10:42 to
Fig. 7. Impact probability of spilt oil in the aquaculture farms of Garorim Bay.
Fig. 9. First impact time of spilt oil for aquaculture farms in Garorim Bay.
Please cite this article as: Lee, M., Jung, J.-Y., Pollution risk assessment of oil spill accidents in Garorim Bay of Korea, Marine Pollution Bulletin (2015), http://dx.doi.org/10.1016/j.marpolbul.2015.08.037
6
M. Lee, J.-Y. Jung / Marine Pollution Bulletin xxx (2015) xxx–xxx
the Bay is relatively safe. The estimated maximum area of contaminated aquaculture farms is ~ 524 ha. 4.2. Estimated first impact time of spilt oil Another subpart of the risk assessment is to assess the first impact time of spilt oil. Figs. 8 and 9 show the first impact time of the split oil in the winter for the coastline and aquaculture farms, respectively. As shown in Fig. 8, the estimated first impact times for Nanji island and Daesan harbor are ~4 h and ~6 h, respectively. For aquaculture farms inside Garorim Bay, the first impact time is in the range of 6 to 11 h, as shown in Fig. 9. 4.3. Evaluation of the oil spill pollution risk in subareas
Fig. 10. Subarea of Garorim Bay (Lee and Kim, 2009).
17:50 on December 12, 2002. The drifters with vanes with an ‘X’ shape were designed to follow currents at 15 m of depth. In field experiments, the positions of drifters were automatically reported via the Differential Global Positioning System (DGPS) and a wireless modem that was equipped in the drifters. The oil spill prediction model computes the trajectories of the numerical particles to simulate the trajectories observed for a drifter under the same conditions as the field experiment described above. Fig. 5 shows the simulated trajectories of particles, which are very similar to the trajectories observed for the three drifters in Fig. 4.
4. Pollution risk of Garorim Bay 4.1. Evaluation of impact probability of oil spill The impact probabilities of spilt oil on the coastline and aquaculture farms in the winter are shown in Figs. 6 and 7, respectively. In the winter, the impact probability of spilt oil is the highest due to strong wind and tides. In Fig. 6, the impact probability in the coastline is as high as 20% in the Daesan harbor region, and most of the harbor region is exposed to an oil spill risk. On the other hand, most of the Bay coastline region is not exposed to an oil spill, except for the Jigok area where the impact probability is less than 12%. The estimated length of the contaminated coastline is ~ 45 km, and regarding the impact probability in aquaculture farms, the area between the entrance of the Bay and Jigok coast is exposed to a risk of contamination with an impact probability of less than 20%, as shown in Fig. 7. However, the south-west region of
The impact probability and the first impact time of the spilt oil are integrated to compute the risk in the subarea. The entire Garorim Bay area is divided into 6 subareas, as shown in Fig. 10, and Table 3 summarizes the maximum oil spill contamination damage in each subarea of Garorim Bay among 80 different simulations. In Table 3, subarea II (where Daesan harbor is located) shows the largest contaminated coastline length of ~ 14.6 km and the shortest first impact time of 0.2 h among the 6 subareas. Subarea VI will be the largest damaged aquaculture farm area of ~ 315.9 ha, and subarea II shows the shortest first impact time of 1.4 h. 5. Summary The pollution risk from oil spills is dependent on the onset time of an accident, the hydrodynamic conditions at sea, and the wind. Since the hydrodynamic conditions and wind conditions are rapidly changing, there is a high probability for error for oil spill estimation systems based on real-time sea and wind conditions. To eliminate uncertainty and to improve the accuracy of the estimation system, a comprehensive risk assessment analysis should accompany the oil spill estimation system. This study thus combines past data of the hydrodynamic and wind conditions as well as area-specific characteristics to create a statistical model for risk assessment. Many random oil spill trajectories are simulated by using an oil spill estimation model with past sea data, and the results of the simulation are statistically analyzed to predict the possible contamination hazard caused by oil spill accidents, providing a sound and generalized risk assessment method. By applying this method, the pollution risk in Garorim Bay in South Korea is assessed in two aspects, namely the impact probability and the first impact time of the spilt oil. It should be noted that this study is based on the results of hypothetical simulations based on past sea conditions and past oil spill accident statistics, not on data from a specific real oil spill accident, and so this work should not be used to assess a specific oil spill accident. Nevertheless, our results can complement the estimation of trajectories of possible spill accidents and could be used to formulate oil spill contamination response and strategies. Our results also help improve our understanding of the possible pollution risk of spill accidents.
Table 3 Maximum pollution damage of spilt oil in each subarea of Garorim Bay. Subareas
I II III IV V VI
Coastline
Aquaculture farm
Total length (km)
First impact time (hour)
Maximum contaminated length (km)
Maximum damage percentage (%)
Total area (ha)
First impact time (hour)
Maximum contaminated area (ha)
Maximum damage percentage (%)
74.5 73.4 14.0 18.7 91.6 58.9
5.3 0.2 6.4 3.5 7.5 6.1
13.0 14.6 0.2 7.6 1.1 8.3
17.4 19.9 1.4 40.6 1.2 14.1
234.9 67.5 70.2 91.8 847.8 588.6
4.8 1.4 – 4.5 6.3 5.9
37.8 32.4 – 91.8 45.9 315.9
16.1 48.0 – 100.0 5.4 53.7
Please cite this article as: Lee, M., Jung, J.-Y., Pollution risk assessment of oil spill accidents in Garorim Bay of Korea, Marine Pollution Bulletin (2015), http://dx.doi.org/10.1016/j.marpolbul.2015.08.037
M. Lee, J.-Y. Jung / Marine Pollution Bulletin xxx (2015) xxx–xxx
Acknowledgments This research was a part of the project titled “Development of Management Technology for HNS Accident,” funded by the Ministry of Oceans and Fisheries (20150340), Korea.
References Alves, T.M., Kokinou, E., Zodiatis, G., 2014. A three-step model to assess shoreline and offshore susceptibility to oil spills: the South Aegean (Crete) as an analogue for confined marine basins. Mar. Pollut. Bull. 86, 443–457. Bo, Q., Addison, P.S., 2010. Modelling flow trajectories using fractional Brownian motion. Chaos–Fractals Theories and Applications (IWCFTA), 2010 International Workshop on, pp. 420–424. Castanedo, S., Medina, R., Losada, I.J., Vidal, C., Méndez, F.J., Osorio, A., Juanes, J.A., Puente, A., 2006. The prestige oil spill in Cantabria (Bay of Biscay). Part I: operational forecasting system for quick response, risk assessment, and protection of natural resources. J. Coast. Res. 1474–1489. Cho, D.-O., 2007. The effects of the M/V Sea Prince accident on maritime safety management in Korea. Mar. Policy 31, 730–735. Delgado, L., Kumzerova, E., Martynov, M., 2006. Simulation of oil spill behaviour and response operations in PISCES. Environmental Problems in Coastal Regions VI. WIT Press, Southampton, pp. 279–292. Elshorbagy, W., Elhakeem, A.-B., 2008. Risk assessment maps of oil spill for major desalination plants in the United Arab Emirates. Desalination 228, 200–216. FAO, 1998. Integrated Coastal Area Management and Agriculture, Forestry and Fisheries. Food and Agriculture Organization of the United Nations, Rome. Foreman, M., 1978. Manual for tidal currents analysis and prediction. Pacific Marine Science Report. Institute of Ocean Sciences, Patricia Bay. French-McCay, D.P., 2004. Oil spill impact modeling: development and validation. Environ. Toxicol. Chem. 23, 2441–2456.
7
Grifoll, M., Jordà, G., Borja, Á., Espino, M., 2010. A new risk assessment method for water quality degradation in harbour domains, using hydrodynamic models. Mar. Pollut. Bull. 60, 69–78. Guo, W.J., Wang, Y.X., Xie, M.X., Cui, Y.J., 2009. Modeling oil spill trajectory in coastal waters based on fractional Brownian motion. Mar. Pollut. Bull. 58, 1339–1346. Lan, D., Liang, B., Bao, C., Ma, M., Xu, Y., Yu, C., 2015. Marine oil spill risk mapping for accidental pollution and its application in a coastal city. Mar. Pollut. Bull. 96, 220–225. Lee, M., 1996. Fractal Application in Oceanic Dispersion Modelling (Ph.D. Thesis) Pukyong National University. Lee, M., Kim, H.J., 2009. A study on the pollution risk assessment of oil spill accidents. J. Ocean Eng. Technol. (in Korean) 23, 24–30. Lee, S.M., Ha, M., Kim, E.J., Jeong, W.C., Hur, J., Park, S.G., Kwon, H., Hong, Y.C., Ha, E.H., Lee, J.S., Chung, B.C., Lee, J., Im, H., Choi, Y., Cho, Y.M., Cheong, H.K., 2009. The effects of wearing protective devices among residents and volunteers participating in the cleanup of the Hebei Spirit oil spill. J. Prev. Med. Public Health 42, 89–95. Melaku Canu, D., Solidoro, C., Bandelj, V., Quattrocchi, G., Sorgente, R., Olita, A., Fazioli, L., Cucco, A., 2015. Assessment of oil slick hazard and risk at vulnerable coastal sites. Mar. Pollut. Bull. 94, 84–95. Nixon, Z., Michel, J., 2015. Predictive modeling of subsurface shoreline oil encounter probability from the Exxon Valdez oil spill in Prince William Sound, Alaska. Environ. Sci. Technol. 49, 4354–4361. Olita, A., Cucco, A., Simeone, S., Ribotti, A., Fazioli, L., Sorgente, B., Sorgente, R., 2012. Oil spill hazard and risk assessment for the shorelines of a Mediterranean coastal archipelago. Ocean Coast. Manag. 57, 44–52. Osborne, A.R., Kirwan Jr., A.D., Provenzale, A., Bergamasco, L., 2011. Fractal drifter trajectories in the Kuroshio extension. Tellus A 41, 416–435. Reed, M., Johansen, Ø., Brandvik, P.J., Daling, P., Lewis, A., Fiocco, R., Mackay, D., Prentki, R., 1999. Oil spill modeling towards the close of the 20th century: overview of the state of the art. Spill Sci. Technol. Bull. 5, 3–16. Shim, Y.-T., Lee, J.-G., Kim, U., 2001. Oil Removal Operation — Yuil No. 1 and Osung No. 3. International Oil Spill Conference Proceedings, pp. 1417–1422. Skognes, K., Johansen, Ø., 2004. Statmap—a 3-dimensional model for oil spill risk assessment. Environ. Model. Softw. 19, 727–737.
Please cite this article as: Lee, M., Jung, J.-Y., Pollution risk assessment of oil spill accidents in Garorim Bay of Korea, Marine Pollution Bulletin (2015), http://dx.doi.org/10.1016/j.marpolbul.2015.08.037
Contents lists available at ScienceDirect
Marine Pollution Bulletin journal homepage: www.elsevier.com/locate/marpolbul
Pollution risk assessment of oil spill accidents in Garorim Bay of Korea Moonjin Lee a,b, Jung-Yeul Jung b,c,⁎ a b c
Maritime Safety Research Division, Korea Research Institute of Ships and Ocean Engineering, KIOST, Daejeon 305-343, Republic of Korea Ocean Science and Technology (OST) School, KMOU/KIOST Joint Program, Busan 606-791, Republic of Korea Technology Center for Offshore Plant Industries, Korea Research Institute of Ships and Ocean Engineering, KIOST, Daejeon 305-343, Republic of Korea
a r t i c l e
i n f o
Article history: Received 6 February 2015 Received in revised form 21 August 2015 Accepted 23 August 2015 Available online xxxx Keywords: Aquaculture farm Coastline Impact probability Impact time Oil spill accident Pollution risk assessment
a b s t r a c t This study presents a model to assess the oil spill risk in Garorim Bay in Korea, where large-scale oil spill accidents frequently occur. The oil spill risk assessment is carried out by using two factors: 1) The impact probability of the oil spill, and 2) the first impact time of the oil that has been spilt. The risk assessment is conducted for environmentally sensitive areas, such as the coastline and aquaculture farms in the Garorim Bay area. Finally, Garorim Bay is divided into six subareas, and the risks of each subarea are compared with one another to identify the subarea that is most vulnerable to an oil spill accident. These results represent an objective and comprehensive oil spill risk level for a specific region. The prediction of the oil spill spread is based on real-time sea conditions and can be improved by integrating our results, especially when sea conditions are rapidly changing. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction It is expected for the industrialization caused by economic growth to proceed more actively in coastal areas than in inland areas (FAO, 1998). Recently, the frequency of oil spill accidents has increased as maritime transportation activity has increased, damaging not only the economics of coastal communities but also marine ecosystems. South Korea is not an exception and has suffered from oil spill accidents in the past. In 1995, the oil spill accident of Sea Prince occurred near Sori Island, Yeo-Chun, Jeon-Nam Province while seeking refuge from typhoon “Faye” (Cho, 2007). Approximately 5000 tons of Arabian crude oil were spilt, ~ 230 km of coastline was contaminated, and the clean up effort lasted ~ 5 months. In 1995, the oil spill accident of Yuil No. 1 occurred near Nam-Hyeong-Jae Island while being towed toward the port of Busan (Shim et al., 2001). More than 2000 tons of oil were spilt. Also the Honam Sapphire spilt over 1000 tons of crude oil near Yeo-Cheon harbor in 1995 where Sea Prince's collision had happened 4 months before. The Hebei Spirit oil spill was Korea's worst oil spill, and it began on 7 December 2007 (Lee et al., 2009). The Hebei Spirit collided with a crane barge owned by Samsung Heavy Industry near the port of Daesan in Taean County on the Yellow Sea. As a result of the collision, ~10,800 tons of oil were spilt. The port of Daesan is located at the entrance of Garorim Bay, which is the area for which the pollution risk assessment is performed in this study. ⁎ Corresponding author at: Maritime Safety Research Division, Korea Research Institute of Ships and Ocean Engineering, KIOST, Daejeon 305-343, Republic of Korea. E-mail address: [email protected] (J.-Y. Jung).
The following response strategies are essential to minimize the pollution damage caused by oil spills: initial response, effective prevention methods including oil boom and recovery, and rapid aid transport. Effective prevention methods require for adequate response strategies to be established in advance of an oil spill accident, and a risk assessment map should include the location and importance of resources that are to be protected, the areas with a high risk of oil spill accidents, the impact probability of spilt oil, and the first impact time of spilt oil in order to establish an effective response strategy. An accurate estimation of the trajectory of the spilt oil is very important to effectively undertake a response to oil spill, and many studies have focused on developing oil spill trajectory modeling and simulation systems (Reed et al., 1999; French-McCay, 2004; Castanedo et al., 2006; Delgado et al., 2006; Guo et al., 2009; Alves et al., 2014; Lan et al., 2015; Melaku Canu et al., 2015; Nixon and Michel, 2015). Since spilt oil is transported by external environmental forces such as the current, waves and wind, the trajectory of the spilt oil is usually estimated using real-time meteorological and oceanic conditions. However, in the case of quickly-changing sea conditions, there is a high probability of an error in the trajectory of the spilt oil that is estimated using realtime observations of the conditions at sea. Although the trajectory estimation of the spilt oil comes with individual uncertainty, the collection and analysis of data from past oil spill accidents can provide statistical guidance to reducing the damage caused by oil pollution. In other words, an oil spill estimation system should be accompanied with a comprehensive risk assessment in order to reduce the negative impact of the spill (Skognes and Johansen, 2004; Elshorbagy and Elhakeem, 2008; Grifoll et al., 2010; Olita et al., 2012). However, most results that
http://dx.doi.org/10.1016/j.marpolbul.2015.08.037 0025-326X/© 2015 Elsevier Ltd. All rights reserved.
Please cite this article as: Lee, M., Jung, J.-Y., Pollution risk assessment of oil spill accidents in Garorim Bay of Korea, Marine Pollution Bulletin (2015), http://dx.doi.org/10.1016/j.marpolbul.2015.08.037
2
M. Lee, J.-Y. Jung / Marine Pollution Bulletin xxx (2015) xxx–xxx
provide an oil spill trajectory modeled using real-time conditions at sea have been limited to specific oil spill accidents in Korea, which means that a comprehensive risk assessment is not yet available. The configuration of our preliminary study (Lee and Kim, 2009) is used to develop a model to assess the pollution risk of oil spill accidents. The purpose of this study is thus to offer a comprehensive and objective risk assessment model based on statistical methods in order to address the shortcomings of oil spill trajectory systems, especially in the case where sea conditions change quickly, to provide assistance in establishing a response strategy in advance. In this study, the coastline and aquaculture farms in the Garorim Bay area are considered as the risk assessment target area. The history of past spill accidents around Garorim Bay was reviewed to select the site where oil spill accidents have most frequently occurred. We assess the pollution risk of an oil spill in two aspects, 1) the impact probability of the spilt oil, and 2) the first impact time of the spilt oil. To make a quantitative assessment of the pollution risk, the oil spill trajectories are simulated 80 different times. In other words, many oil spill trajectories are statistically simulated based on past sea conditions to analyze and assess the oil spill hazard on coastlines and aquaculture farms. These trajectories are then collected and analyzed to produce a comprehensive and objective risk assessment. The Garorim Bay area was divided into six subareas, and the subarea most vulnerable to oil spills was then determined. 2. Assessment of oil pollution risk 2.1. Risk indicators of oil pollution The risk of oil pollution can be evaluated according to the impact probability and the first impact time of the spilt oil. The impact probability is a very important information when deciding which area to protect from spilt oil. This information can thus be used to effectively allocate resources and plan a response for an oil spill. The first impact time of the spilt oil can provide information on the response priorities.
Therefore, it is essential to set priorities due to the limited availability of resources. The impact probability of spilt oil is quantitatively defined as the average ratio of the amount of oil spilt in each specific area to the total amount of oil spilt, which can be expressed using Eq. (1), P¼
N 1X Ai N i¼1 Atotal
ð1Þ
where, P is the impact probability and N is the number of oil spill events, Ai is the amount of oil spilt from the ith event in each specific area and Atotal is the total amount of oil spilt from the ith event. The first impact time of the spilt oil is defined as the time that has elapsed to reach a specific area. The elapsed times can be calculated as follows, t ¼ T−T 0
ð2Þ
where, t is the first impact time, T is the first attached time of oil to the area, and T0 is the time of the oil spill incident. T and T0 are in local standard time. 2.2. Study area (Garorim Bay) 2.2.1. General feature Garorim Bay has been designated as the risk assessment target area because huge petrochemical plants and transporting ships are populated and plied. Garorim Bay is located on the western coast of the Korean Peninsula, as shown in Fig. 1, and it is famous for laver and oyster cultures and has rich spawning grounds for important species. The Daesan industrial complex includes large-scale petrochemical plants and Daesan harbor is located in the northeast coast outside of the Bay. Due to the narrow entrance of Garorim Bay, once the spilt oil is entrained in the Bay area, the oil will remain in the Bay area, and there is a low
Fig. 1. Location of Garorim Bay.
Please cite this article as: Lee, M., Jung, J.-Y., Pollution risk assessment of oil spill accidents in Garorim Bay of Korea, Marine Pollution Bulletin (2015), http://dx.doi.org/10.1016/j.marpolbul.2015.08.037
M. Lee, J.-Y. Jung / Marine Pollution Bulletin xxx (2015) xxx–xxx
3
Table 1 List of serious accidents that occurred around Garorim Bay, from 1991 to 2000 (Lee and Kim, 2009). Ship's category
Date
Ship's name
Total weight (tons)
Location of accident
Cause of accident
Volume of oil spilled (kl)
10 ~ 100 kl
11.19.1995
Hwa-jeon (cargo)
1952
1 mile north off the island of Huwang
Sinking
100 ~ 1000 kl
07.05.2000
Safe Express (cargo)
1250
North of the island of An
Sinking
02.21.1991
Pacific Friend (cargo)
4417
1.7 miles west of Kyonggi
Sinking
12.02.1996 04.11.2000 10.01.1993
Tae-young Jasmine (cargo) 501 (dredge) Frontier Express (tanker)
2483 1458 40,721
6 miles northwest of the island of An Pyung-taek harbor Off shore from a lighthouse in Daesan harbor
Sinking Sinking Collision
B/B 7.9 B/A 2.7 B/C 30 B/A 10 WTI 155 MDO 41 B/C 159 MDO 165 Petroleum naphtha 8320
b1000 kl
expectation for further natural decay by the weathering process. Therefore, the Garorim Bay area is one of the most vulnerable areas to oil spill accidents in Korea. 2.2.2. Site of an oil spill accident For the past 10 years, the most frequent accident sites (1991–2000) have been determined as hypothetical oil spill accident sites. Table 1 shows a list of serious accidents that have occurred near Garorim Bay, and Fig. 2 shows the location of the accidents listed in Table 1. As seen in Fig. 2, most of the accidents have taken place near Daesan harbor. Therefore, this study designates the entry fairway into Daesan harbor as the hypothetical site where an oil spill accident occurs. 2.2.3. Target of risk assessment The Environmental Sensitivity Index (ESI) Map of Korea can be utilized to select the target for the risk assessment for oil pollution. The ESI Map of Korea includes various types of information related to an oil spill response, such as the coastline sensitivity classifications for oil pollution, sensitive biological resources, socio-economic features, and oil spill response resources. The coastline and aquaculture farms in Garorim Bay that are environmentally sensitive are selected as a target for the risk assessment of oil pollution. Fig. 3 shows the target areas split into a small grid to quantify the impact probability.
3.2. Simulation of oil behavior 3.2.1. Oil spill prediction model The spilt oil is transported by advection dominated by currents and winds while the diffusion occurs as a result of turbulence. The volume of spilt oil decreases due to biochemical reactions and sinking. In this study, the sea current is calculated by solving a hydrodynamic equation with the effect of the tides and wind. To simulate the behavior of the spilt oil, a Lagrangian model based on a numerical particle tracking method was adopted. To provide realistic results, turbulent diffusion and weathering of the spilt oil are included in the Lagrangian model, and a detailed description is given in Section 3.2.2. In our spill prediction model, numerical particles are released and tracked to simulate the movement and weathering of spilt oil. The movement of spilt oil is represented by the displacement of the particles in each time step. If a particle is located at (x0, y0) at time t and then moves to a new location (x0 + δx, y0 + δy) in the time interval δt, the displacements δx, δy during a time δt are given respectively by, δx ¼ ðU þ u0 Þδt
ð3Þ
δy ¼ ðV þ v0 Þδt:
ð4Þ
3. Simulation of oil spill incident 3.1. Preparing scenarios of oil spill incidents In order to accurately assess the pollution risk of an oil spill (i.e., the impact probability of spilt oil and the first impact time of spilt oil), the oil trajectories of past oil spill accidents should be collected and analyzed under various conditions. In fact, data on oil spill accidents is limited, and the characteristics of the oil trajectory are also difficult to observe in the real sea. Therefore, various hypothetical oil spill accidents were simulated in order to obtain data for the trajectory of the spilt oil. The time for hypothetical accidents varied through four seasons in order to consider the various ocean conditions for a specific time from 1983 to 2002. The behavior of the spilt oil is simulated and is analyzed for a period of five days. Table 2 shows the simulations for the statistic trajectory analysis with the tidal time for each season, which is almost even, which means that the cases analyzed in this study will show statistical results for various sea conditions. As mentioned in Section 2.1, two of the major risk indicators for the assessment in this study, the impact probability and the first arrival time of spilt oil, are calculated from the trajectories of 200 random oil spill accident scenarios. The number of random simulations according to tidal times and seasons is listed in Table 2. Each simulation represents a five day duration for an oil spill after hypothetical accidents at random times between 1983 and 2002.
Fig. 2. Locations of serious oil spill accidents around Garorim Bay from 1991 to 2000 and hypothetical accident site for pollution risk assessment (Lee and Kim, 2009).
Please cite this article as: Lee, M., Jung, J.-Y., Pollution risk assessment of oil spill accidents in Garorim Bay of Korea, Marine Pollution Bulletin (2015), http://dx.doi.org/10.1016/j.marpolbul.2015.08.037
4
M. Lee, J.-Y. Jung / Marine Pollution Bulletin xxx (2015) xxx–xxx
Fig. 3. Index cells for the estimation of oil spill risk around Garorim Bay (Lee and Kim, 2009).
Fig. 4. Observed trajectories of drifting buoys (10:42 ~ 17:50, Dec. 11, 2002) (Lee and Kim, 2009).
The current velocity (U, V) is predicted by the results of the simulation based on the hydrodynamic model discussed in Section 3.2.2, and (u′, v′) represents the effect of the turbulence. Fractional Brownian motion allows for the turbulent diffusion to be computed (Guo et al., 2009; Bo and Addison, 2010; Osborne et al., 2011). In fractional Brownian motion, the Hurst coefficient in South Korea seas range from 0.45 to 2.46 (Lee, 1996).
The tidal current is the most important component in coastal environments in South Korea, and the maximum tidal height is 10 m. In this study, real-time tides are simulated by combining the harmonic constants of the tidal currents and the variations in astronomical arguments (Foreman, 1978). The harmonic constants are computed through a harmonic analysis of the results of the hydrodynamic model. In addition to deterministic tidal currents, time-varying wind-driven currents play an important role in the coastal region. Due to the complexity of coastal geometry and the bottom topography, the response of the current to winds differs from place to place and is not instantaneous but delayed. The strength and delay time of the wind-driven current are dependent on the strength and duration of the wind. Therefore, it is necessary to define the impulse response of the currents to the
3.2.2. Simulation of oceanic currents The current that mainly governs the spread of the oil spill is calculated by solving a hydrodynamic equation that includes the effect of the tide and the wind. The basic equations that are used for this model are a continuity equation and two-dimensional, depth-averaged hydrodynamic equations of motion given by ∂ζ ∂U ðD þ ζ Þ ∂V ðD þ ζ Þ þ þ ¼0 ∂t ∂x ∂y
ð5Þ
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ∂U ∂U ∂U ∂ζ kU U 2 þ V 2 ¼0 þU þV þg −f V þ Dþζ ∂t ∂x ∂y ∂x
ð6Þ
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ∂V ∂V ∂V ∂ζ kV U 2 þ V 2 ¼ 0; þU þV þg þ fU þ Dþζ ∂t ∂x ∂y ∂y
ð7Þ
where U and V are depth average current components in the x and y directions, respectively, g is the gravitational acceleration, ζ is the elevation from the mean sea level, D is the water depth from the mean sea level, f is the Coriolis parameter (f = 2Ωsinϕ, ϕ = 37°N), and k is the bottom friction coefficient (k = 0.003).
Table 2 Simulation cases for the statistical trajectory analysis. Tidal time
Season
Flood High Ebb Low Spring Summer Fall Winter flow water flow water No. of computation 51
50
49
50
50
50
50
50
Fig. 5. Simulated trajectories of drifting buoys (10:42 ~ 17:50, Dec. 11, 2002) (Lee and Kim, 2009).
Please cite this article as: Lee, M., Jung, J.-Y., Pollution risk assessment of oil spill accidents in Garorim Bay of Korea, Marine Pollution Bulletin (2015), http://dx.doi.org/10.1016/j.marpolbul.2015.08.037
M. Lee, J.-Y. Jung / Marine Pollution Bulletin xxx (2015) xxx–xxx
5
Fig. 6. Impact probability of spilt oil in the coastline of Garorim Bay.
Fig. 8. First impact time of spilt oil to the coastline of Garorim Bay.
recent history of winds in any desired place in order to predict the winddriven currents. In the present study, we have predicted wind-driven currents using a kernel coefficient of the response functions between the winds and wind-driven currents.
(ADIOS) can compute the weathering effect, including the evaporation, emulsification, etc., according to oil types by using a database of oil properties. 3.3. Verifying the simulated oil behavior
3.2.3. Weathering of oil The amount of spilt oil that decays due to the weathering effect depends on the properties of the oil. In our oil spill prediction model, the number of numerical particles decreases according to the rate of decay of specific oil types since it reacts physically and biochemically in water. The rate of decay of oil is computed using the NOAA oil spill response tool database, and the Automated Data Inquiry for Oil Spills
In order to verify our oil spill prediction model, the trajectories of the drifters are experimentally observed near the entrance of Garorim Bay and are compared to the results of the simulation from our model. Fig. 4 represents the trajectories observed for 3 drifters that were simultaneously released at the same starting position. The trajectories of the drifters were tracked for a period of 7 h and 8 min from 10:42 to
Fig. 7. Impact probability of spilt oil in the aquaculture farms of Garorim Bay.
Fig. 9. First impact time of spilt oil for aquaculture farms in Garorim Bay.
Please cite this article as: Lee, M., Jung, J.-Y., Pollution risk assessment of oil spill accidents in Garorim Bay of Korea, Marine Pollution Bulletin (2015), http://dx.doi.org/10.1016/j.marpolbul.2015.08.037
6
M. Lee, J.-Y. Jung / Marine Pollution Bulletin xxx (2015) xxx–xxx
the Bay is relatively safe. The estimated maximum area of contaminated aquaculture farms is ~ 524 ha. 4.2. Estimated first impact time of spilt oil Another subpart of the risk assessment is to assess the first impact time of spilt oil. Figs. 8 and 9 show the first impact time of the split oil in the winter for the coastline and aquaculture farms, respectively. As shown in Fig. 8, the estimated first impact times for Nanji island and Daesan harbor are ~4 h and ~6 h, respectively. For aquaculture farms inside Garorim Bay, the first impact time is in the range of 6 to 11 h, as shown in Fig. 9. 4.3. Evaluation of the oil spill pollution risk in subareas
Fig. 10. Subarea of Garorim Bay (Lee and Kim, 2009).
17:50 on December 12, 2002. The drifters with vanes with an ‘X’ shape were designed to follow currents at 15 m of depth. In field experiments, the positions of drifters were automatically reported via the Differential Global Positioning System (DGPS) and a wireless modem that was equipped in the drifters. The oil spill prediction model computes the trajectories of the numerical particles to simulate the trajectories observed for a drifter under the same conditions as the field experiment described above. Fig. 5 shows the simulated trajectories of particles, which are very similar to the trajectories observed for the three drifters in Fig. 4.
4. Pollution risk of Garorim Bay 4.1. Evaluation of impact probability of oil spill The impact probabilities of spilt oil on the coastline and aquaculture farms in the winter are shown in Figs. 6 and 7, respectively. In the winter, the impact probability of spilt oil is the highest due to strong wind and tides. In Fig. 6, the impact probability in the coastline is as high as 20% in the Daesan harbor region, and most of the harbor region is exposed to an oil spill risk. On the other hand, most of the Bay coastline region is not exposed to an oil spill, except for the Jigok area where the impact probability is less than 12%. The estimated length of the contaminated coastline is ~ 45 km, and regarding the impact probability in aquaculture farms, the area between the entrance of the Bay and Jigok coast is exposed to a risk of contamination with an impact probability of less than 20%, as shown in Fig. 7. However, the south-west region of
The impact probability and the first impact time of the spilt oil are integrated to compute the risk in the subarea. The entire Garorim Bay area is divided into 6 subareas, as shown in Fig. 10, and Table 3 summarizes the maximum oil spill contamination damage in each subarea of Garorim Bay among 80 different simulations. In Table 3, subarea II (where Daesan harbor is located) shows the largest contaminated coastline length of ~ 14.6 km and the shortest first impact time of 0.2 h among the 6 subareas. Subarea VI will be the largest damaged aquaculture farm area of ~ 315.9 ha, and subarea II shows the shortest first impact time of 1.4 h. 5. Summary The pollution risk from oil spills is dependent on the onset time of an accident, the hydrodynamic conditions at sea, and the wind. Since the hydrodynamic conditions and wind conditions are rapidly changing, there is a high probability for error for oil spill estimation systems based on real-time sea and wind conditions. To eliminate uncertainty and to improve the accuracy of the estimation system, a comprehensive risk assessment analysis should accompany the oil spill estimation system. This study thus combines past data of the hydrodynamic and wind conditions as well as area-specific characteristics to create a statistical model for risk assessment. Many random oil spill trajectories are simulated by using an oil spill estimation model with past sea data, and the results of the simulation are statistically analyzed to predict the possible contamination hazard caused by oil spill accidents, providing a sound and generalized risk assessment method. By applying this method, the pollution risk in Garorim Bay in South Korea is assessed in two aspects, namely the impact probability and the first impact time of the spilt oil. It should be noted that this study is based on the results of hypothetical simulations based on past sea conditions and past oil spill accident statistics, not on data from a specific real oil spill accident, and so this work should not be used to assess a specific oil spill accident. Nevertheless, our results can complement the estimation of trajectories of possible spill accidents and could be used to formulate oil spill contamination response and strategies. Our results also help improve our understanding of the possible pollution risk of spill accidents.
Table 3 Maximum pollution damage of spilt oil in each subarea of Garorim Bay. Subareas
I II III IV V VI
Coastline
Aquaculture farm
Total length (km)
First impact time (hour)
Maximum contaminated length (km)
Maximum damage percentage (%)
Total area (ha)
First impact time (hour)
Maximum contaminated area (ha)
Maximum damage percentage (%)
74.5 73.4 14.0 18.7 91.6 58.9
5.3 0.2 6.4 3.5 7.5 6.1
13.0 14.6 0.2 7.6 1.1 8.3
17.4 19.9 1.4 40.6 1.2 14.1
234.9 67.5 70.2 91.8 847.8 588.6
4.8 1.4 – 4.5 6.3 5.9
37.8 32.4 – 91.8 45.9 315.9
16.1 48.0 – 100.0 5.4 53.7
Please cite this article as: Lee, M., Jung, J.-Y., Pollution risk assessment of oil spill accidents in Garorim Bay of Korea, Marine Pollution Bulletin (2015), http://dx.doi.org/10.1016/j.marpolbul.2015.08.037
M. Lee, J.-Y. Jung / Marine Pollution Bulletin xxx (2015) xxx–xxx
Acknowledgments This research was a part of the project titled “Development of Management Technology for HNS Accident,” funded by the Ministry of Oceans and Fisheries (20150340), Korea.
References Alves, T.M., Kokinou, E., Zodiatis, G., 2014. A three-step model to assess shoreline and offshore susceptibility to oil spills: the South Aegean (Crete) as an analogue for confined marine basins. Mar. Pollut. Bull. 86, 443–457. Bo, Q., Addison, P.S., 2010. Modelling flow trajectories using fractional Brownian motion. Chaos–Fractals Theories and Applications (IWCFTA), 2010 International Workshop on, pp. 420–424. Castanedo, S., Medina, R., Losada, I.J., Vidal, C., Méndez, F.J., Osorio, A., Juanes, J.A., Puente, A., 2006. The prestige oil spill in Cantabria (Bay of Biscay). Part I: operational forecasting system for quick response, risk assessment, and protection of natural resources. J. Coast. Res. 1474–1489. Cho, D.-O., 2007. The effects of the M/V Sea Prince accident on maritime safety management in Korea. Mar. Policy 31, 730–735. Delgado, L., Kumzerova, E., Martynov, M., 2006. Simulation of oil spill behaviour and response operations in PISCES. Environmental Problems in Coastal Regions VI. WIT Press, Southampton, pp. 279–292. Elshorbagy, W., Elhakeem, A.-B., 2008. Risk assessment maps of oil spill for major desalination plants in the United Arab Emirates. Desalination 228, 200–216. FAO, 1998. Integrated Coastal Area Management and Agriculture, Forestry and Fisheries. Food and Agriculture Organization of the United Nations, Rome. Foreman, M., 1978. Manual for tidal currents analysis and prediction. Pacific Marine Science Report. Institute of Ocean Sciences, Patricia Bay. French-McCay, D.P., 2004. Oil spill impact modeling: development and validation. Environ. Toxicol. Chem. 23, 2441–2456.
7
Grifoll, M., Jordà, G., Borja, Á., Espino, M., 2010. A new risk assessment method for water quality degradation in harbour domains, using hydrodynamic models. Mar. Pollut. Bull. 60, 69–78. Guo, W.J., Wang, Y.X., Xie, M.X., Cui, Y.J., 2009. Modeling oil spill trajectory in coastal waters based on fractional Brownian motion. Mar. Pollut. Bull. 58, 1339–1346. Lan, D., Liang, B., Bao, C., Ma, M., Xu, Y., Yu, C., 2015. Marine oil spill risk mapping for accidental pollution and its application in a coastal city. Mar. Pollut. Bull. 96, 220–225. Lee, M., 1996. Fractal Application in Oceanic Dispersion Modelling (Ph.D. Thesis) Pukyong National University. Lee, M., Kim, H.J., 2009. A study on the pollution risk assessment of oil spill accidents. J. Ocean Eng. Technol. (in Korean) 23, 24–30. Lee, S.M., Ha, M., Kim, E.J., Jeong, W.C., Hur, J., Park, S.G., Kwon, H., Hong, Y.C., Ha, E.H., Lee, J.S., Chung, B.C., Lee, J., Im, H., Choi, Y., Cho, Y.M., Cheong, H.K., 2009. The effects of wearing protective devices among residents and volunteers participating in the cleanup of the Hebei Spirit oil spill. J. Prev. Med. Public Health 42, 89–95. Melaku Canu, D., Solidoro, C., Bandelj, V., Quattrocchi, G., Sorgente, R., Olita, A., Fazioli, L., Cucco, A., 2015. Assessment of oil slick hazard and risk at vulnerable coastal sites. Mar. Pollut. Bull. 94, 84–95. Nixon, Z., Michel, J., 2015. Predictive modeling of subsurface shoreline oil encounter probability from the Exxon Valdez oil spill in Prince William Sound, Alaska. Environ. Sci. Technol. 49, 4354–4361. Olita, A., Cucco, A., Simeone, S., Ribotti, A., Fazioli, L., Sorgente, B., Sorgente, R., 2012. Oil spill hazard and risk assessment for the shorelines of a Mediterranean coastal archipelago. Ocean Coast. Manag. 57, 44–52. Osborne, A.R., Kirwan Jr., A.D., Provenzale, A., Bergamasco, L., 2011. Fractal drifter trajectories in the Kuroshio extension. Tellus A 41, 416–435. Reed, M., Johansen, Ø., Brandvik, P.J., Daling, P., Lewis, A., Fiocco, R., Mackay, D., Prentki, R., 1999. Oil spill modeling towards the close of the 20th century: overview of the state of the art. Spill Sci. Technol. Bull. 5, 3–16. Shim, Y.-T., Lee, J.-G., Kim, U., 2001. Oil Removal Operation — Yuil No. 1 and Osung No. 3. International Oil Spill Conference Proceedings, pp. 1417–1422. Skognes, K., Johansen, Ø., 2004. Statmap—a 3-dimensional model for oil spill risk assessment. Environ. Model. Softw. 19, 727–737.
Please cite this article as: Lee, M., Jung, J.-Y., Pollution risk assessment of oil spill accidents in Garorim Bay of Korea, Marine Pollution Bulletin (2015), http://dx.doi.org/10.1016/j.marpolbul.2015.08.037