Performance and verification of the oil spill modeling system
for the Sea of Japan
Sergey VARLAMOV 1) and Jong-Hwan YOON2)
1)Institute of Environmental Sciences, RISSHO University,
1700 Magechi, Kumagaya city, Saitama 360-0194, JAPAN
e-mail:
2) Research Institute for Applied Mechanics, Kyushu University,
6-1 Kasuga-koen, Kasuga city, Fukuoka 816-8580, JAPAN
Oil spill fate and drift prediction system is one of practical applications for high performance state-of-the-art atmospheric and ocean circulation models. The oil spill physical properties and drift model is also an important part of this system. The version of the oil spill analysis and prediction system was developed by authors and adopted for the Sea of Japan. Examples of applications for modeling of the Nakhodka oil spill in the Sea of Japan that took place January 1997 were reported at the last 2nd Workshop on Next Generation Climate Models in Toulouse, 2000 (Varlamov et. al., 1999, 2000). During past year the serious modifications of the spill model as well as of the ocean circulation subsystem were done. System is oriented on the operational forecast of accidental oil spills and includes the meteorological data processing subsystem, ocean currents simulation models, oil spill simulation model and the software for results graphical presentation and analysis. The spill model is based on the particles tracking method and the last version includes the simulation of the oil physical properties that changes the modeled oil droplet characteristics during the spill evolution (density, size, viscosity and water content in the oil emulsion). Oil spill prediction system structure is presented and shortly discussed. The meteorological subsystem utilizes the production of the JMA official regional atmospheric prediction system. The ocean circulation model is a part of the system and both the 3D and 2.5D versions were tested. The combination of local wind drift model with the barotropic model of the Japan Sea (2.5D model) is implemented in to the operational version of system (Varlamov et. al., 2000).
April 3, 1997 in the Sea of Japan one more oil spill accident happened, although not so dramatic as that caused by the tanker Nakhodka oil spill. Korean tanker O-Sung No 3 was ruptured in the Tsushima Strait close to the Korean coast and about 229 kiloliters of the heavy Bunker C oil were spilled. Largest part of oil was recovered from the sea by joint Korean and Japanese actions, however the pollution of the Tsushima Island shoreline was reported.
Strong tidal and residual currents characterize the Tsushima Strait area. The ability of the ocean circulation model to reproduce correctly both these components was carefully checked. At the open boundaries at the outer area of the Tsushima Strait the 2.6 Sv mean inflow current was applied and the sea level anomalies relative to the long-term mean values were forced by 16 tidal harmonics. The simulated volume transport through the Western channel of the Tsushima strait was about 1.5 times larger from this value for the Eastern channel. The sea level simulation compared with observations for Hakata and Idzuhara tidal stations also gave good comparison: root mean square errors for hourly values were 12 and 9 cm correspondingly compared with observed root mean square sea level variability 46 cm for Hakata and 45 cm for Idzuhara stations.
At the outflow open boundaries the free gravitational wave conditions were adopted, although the tidal variations could be also applied. Normal sea current anomalies in tidal-forced open boundary points were adapted from internal points with fixed Newton type adaptation parameter. It avoided the possible instability generation in boundary regions.
The O-sung spill simulation demonstrated good comparison of simulated oil spreading with observed one. At the initial stage the wind drift of surface slicks was most important and slick drifted against the mean sea current. About April 6 the highest concentration oil patches changed the drift direction. Fig. 1 demonstrates the position of modeled oil particles and relative oil concentration for the upper 2 m layer. By rectangle the observed position of oil patches for this time is marked. Large difference between the particle’s positions and the concentration field is produced owing to the different size of used particles: largest droplets that drift at the sea surface form the area of highest oil concentration.
System code parallelization and related increase of the system performance will be discussed.
Fig.1 Position of modeled oil particles (left) and relative oil concentration (right) in upper 2 m April 7, 1997. Initial accident point near 34 36N, 128 35E is marked by small circle.
References
1. Varlamov S.M., J.-H. Yoon, N. Hirose, H. Kawamura and K.Shiohara, Simulation of the oil spill processes in the Sea of Japan Sea with regional ocean circulation model. Journal of Marine Science and Technology, 1999, 4 (3), pp. 94-107.
2. Varlamov S.M., J.-H. Yoon, H. Nagaishi & K. Abe, Japan Sea oil spill analysis and quick response system with adaptation of shallow water ocean circulation model. Reports of Research Institute for Applied Mechanics, Kyushu University, 2000, No. 118, pp. 9-22.