Prediction of Oil Spill Trajectory on the Ocean Surface Using Mathematical Modeling

Abstract

A simplistic mathematical model based on the law of motion, for predicting oil spill trajectory on the ocean surface utilizing random walk technique is proposed in this paper. Validation of the proposed model is performed by comparing results with the General NOAA Operational Modeling Environment (GNOME) model and the available set of periodic Sentinel -1 Synthetic Aperture Radar (SAR) images of the contingency location (Corsica oil spill incident in the Mediterranean Sea). SAR images are processed for speckle noise removal, dark spot detection, feature extraction and classification of dark spots as oil spills and look-alikes for suitability of comparison. The accuracy of prediction is evaluated using Centroid Skill Score (CSS) metric and is compared with that of the prediction results from MEDSLIK – II model. The results of proposed model are found to be in good agreement with available SAR images. The simulation also showed that using an hourly wind and ocean current data on the study region, more accurate prediction of the trajectory is possible.