Figure 1.
Coverage map of satellite data and information on the study area.
Coverage of satellite map and study area around the 2007 Hebei Spirit oil tanker disaster. (A) Coverage map of acquired satellite data from 8th to 25th December, 2007. The smallest green box is the coverage area of KOMPSAT-2 optical sensor, and other boxes are those by SARs as indicated in the legend. ENVISAT ASAR and TerraSAR-X were in descending orbit and RADARSAT-1 in ascending orbit. (B) Bathymetry and positions of Automatic Weather System (AWS) stations and tidal current observation.
Figure 2.
SAR images of oil slicks acquired by different times as listed in Table 1, where (A) ENVISAT-ASAR, (B) RADARSAT-1, (C) TerraSAR-X, and (D) ENVISAT-ASAR. The position of the Hebei Spirit was marked by the white circle in (A).
Table 1.
Parameters of satellite sensors and wind data at the AWS stations closest to the centroid of the slick areas at each satellite data acquisition time.
Figure 3.
Flowchart for the numerical simulation of oil slick movement.
Each parameter represents vector or scalar and explained in detail in the text. The central box represents equations for each step in the simulation.
Figure 4.
Tidal current and wind field around the oil slick area.
Examples of tidal current and wind fields around oil slick area at 10∶40 on December 11, 2007. (A) tidal current data from EFDC without wind-induced current during spring/low tide and (B) wind field data from AWS.
Figure 5.
Comparison of EFDC and KHOA current data.
Left and right columns show the comparisons of current velocity and direction respectively. The solid lines and dots correspond respectively to Environmental Fluid Dynamics Code (EFDC) and Korea Hydrographic and Oceanographic Administration (KHOA) data.
Figure 6.
Average wind and tidal current during simulation period.
Mean wind and tidal current averaged over 30 minutes at the position of Hebei Spirit. (A) represents wind speed (blue line) and direction (green line), and (B) represents tidal current velocity (blue line) and direction (green line). Vertical dashed lines represent the accident time and satellite data acquisition times.
Table 2.
Start and end times of simulation for the cases 1–5.
Figure 7.
Comparison of oil slick areas by satellite images and simulation.
Comparison of oil slick areas extracted from satellite images on the right column and those by simulation on the right column. From top to bottom rows: cases 1–5. See the text for further details.
Figure 8.
Diagram for quantitative evaluation of simulation results.
Cells with light gray color in the simulated binary image indicate those occupied by oil particles that overlap with the cells containing oil particles in the satellite image, and cells with dark color are those occupied by oil particles but do not overlap with the oil cells in the satellite image.
Figure 9.
Matching rates as a function of wind drift factor.
Matching rates of the simulated oil slick areas compared with those of satellite images as a function of different wind drift factor for cases 1–5.
Figure 10.
Scatter diagram of the wind drift factor and wind speed.
Scatter diagram of the wind drift factor and wind speed for the cases 1–5 when the matching rates are optimum. The solid line is the regression line with the coefficient of determination R2.
Table 3.
Mean wind speed, wind drift factor, and matching rate of simulation result compared with satellite images.
Figure 11.
Comparison of matching rates (triangles) of the new model with those (squares) of the conventional model with a fixed wind drift factor of 0.03. The regression line for the new model was defined by equation (3) in the text.