Improving the accuracy of the oil spill model

Oil spill modeling is a crucial tool in marine environmental management, enabling early prediction and mitigation of pollution impacts on coastal ecosystems. This study presents a high-resolution oil spill simulation framework developed for the Hai Phong coastal region in northern Vietnam, where risks from maritime activities are significant. The proposed approach utilizes a Lagrangian particle tracking method to simulate the transport and dispersion of oil particles driven by wind, wave, and tidal forces. A hydrodynamic model underpins the simulation, with its input parameters systematically calibrated to improve accuracy. Model calibration was guided by widely accepted error assessment metrics, including Root Mean Square Error, Nash–Sutcliffe Efficiency, and Percent Bias. These indicators were used to iteratively refine the simulation until a satisfactory match with empirical observations was achieved. Unlike traditional models that often rely on static inputs,this study incorporates dynamic environmental forcing and applies adaptive calibration strategies, allowing for reliable forecasts even when the spill source is uncertain or data availability is limited. The model demonstrates strong predictive capacity during the critical early hours following an oil spill, highlighting its potential for integration into rapid-response systems. Overall, the framework contributes to advancing oil spill preparedness and response strategies in data-scarce coastal regions, supporting both ecological protection and maritime risk management.