Enhancing nuclear emergency response through wind data assimilation: a particle filter-based approach combined with terrain-modified Gaussian plume model

This study presents a novel approach for enhancing radionuclide dispersion estimation following nuclear accidents by integrating a Particle Filtering (PF)-based real-time wind speed prediction model with a terrain-modified Gaussian Plume Model (TM-GPM). Using the case study of the Akkuyu Nuclear Power Plant (NPP), equipped with four VVER-1200 pressurized water reactors (PWRs), the proposed framework dynamically adjusts wind inputs and dispersion predictions, validated against measured meteorological data. The PF model was cross-validated using multiple statistical techniques (RMSE, MAE, R², and correlation coefficient), achieving high accuracy (R² = 0.988). The results demonstrate how terrain and precipitation significantly affect the dispersion and deposition of radionuclides. This integrated approach offers improved predictive capability for emergency response planning and public health risk assessment near nuclear reactors.