Search

An oil spill forecast system for offshore China was developed based on Visual C++. The oil spill forecast system includes an ocean environmental forecast model and an oil spill model. The ocean environmental forecast model was designed to include timesaving methods, and comprised a parametrical wind wave forecast model and a sea surface current forecast model. The oil spill model was based on the “particle method” and fulfills the prediction of oil particle behavior by considering the drifting, evaporation and emulsification processes. A specific database was embedded into the oil spill forecast system, which contained fundamental information, such as the properties of oil, reserve of emergency equipment and distribution of marine petroleum platform. The oil spill forecast system was successfully applied as part of an oil spill emergency exercise, and provides an operational service in the Research and Development Center for Offshore Oil Safety and Environmental Technology.

Based on an empirical orthogonal function (EOF) analysis of the monthly NCEP Optimum Interpolation Sea Surface Temperature (OISST) data in the South China Sea (SCS) after removing the climatological mean and trends of SST, over the period of January 1982 to October 2003, the corresponding TCF correlates best with the Dipole Mode Index (DMI), Nino1+2, Nino3.4, Nino3, and Nino4 indices with time lags of 10, 3, 6, 5, and 6 months, respectively. Thus, a statistical hindcasts in the prediction model are based on a canonical correlation analysis (CCA) model using the above indices as predictors spanning from 1993/1994 to 2003/2004 with a 1–12 month lead time after the canonical variants are calculated, using data from the training periods from January 1982 to December1992. The forecast model is successful and steady when the lead times are 1–12 months. The SCS warm event in 1998 was successfully predicted with lead times from 1–12 months irrespective of the strength or time extent. The prediction ability for SSTA is lower during weak ENSO years, in which other local factors should be also considered as local effects play a relatively important role in these years. We designed the two forecast models: one using both DMI and Nino indices and the other using only Nino indices without DMI, and compared the forecast accuracies of the two cases. The spatial distributions of forecast accuracies show different confidence areas. By turning off the DMI, the forecast accuracy is lower in the coastal areas off the Philippines in the SCS, suggesting some teleconnection may occur with the Indian Ocean in this area. The highest forecast accuracies occur when the forecast interval is five months long without using the DMI, while using both of Nino indices and DMI, the highest accuracies occur when the forecast interval time is eight months, suggesting that the Nino indices dominate the interannual variability of SST anomalies in the SCS. Meanwhile the forecast accuracy is evaluated over an independent test period of more than 11 years (1993/94 to October 2004) by comparing the model performance with a simple prediction strategy involving the persistence of sea surface temperature anomalies over a 1–12 month lead time (the persisted prediction). Predictions based on the CCA model show a significant improvement over the persisted prediction, especially with an increased lead time (longer than 3 months). The forecast model performs steadily and the forecast accuracy, i.e., the correlation coefficients between the observed and predicted SSTA in the SCS are about 0.5 in most middle and southern SCS areas, when the thresholds are greater than the 95% confidence level. For all 1 to 12 month lead time forecasts, the root mean square errors have a standard deviation of about 0.2. The seasonal differences in the prediction performance for the 1–12 month lead time are also examined.