Investigating environmental determinants and spatiotemporal dynamics of highly pathogenic avian influenza H5N1 outbreaks in India through machine learning

Avian Influenza (AI), caused by highly pathogenic strains of influenza viruses, poses a significant threat to poultry populations and public health worldwide. This study offers a comprehensive evaluation of the spatial and temporal dynamics of HPAI outbreaks in India, employing a multidisciplinary approach that integrates geospatial analysis, machine learning modelling, remote sensing, and environmental risk factor assessment. The resulting incidence map provides a powerful visual representation, offering valuable insights into the distribution and concentration of HPAI (Highly Pathogenic Avian Influenza) outbreaks across the country. The study identifies a peak in HPAI outbreaks during the winter and spring seasons. Critical environmental variables such as air temperature, enhanced vegetation index (EVI), leaf area index (LAI), potential evapotranspiration (PET), rain precipitation rate, specific humidity, and wind speed are identified as significant predictors of HPAI risk. Ensemble technology was adopted by integrating high-performing random forest (RF) and classification tree (CT) models for HPAI risk assessment. The risk map generated and estimated basic reproduction numbers (R₀) indicate, the southern and north-eastern regions of India are vulnerable to HPAI. The findings presented provide a holistic perspective essential for effective surveillance, strategic planning for resource allocation and policy development for disease management, aimed at safeguarding both avian and human populations from the looming threat of influenza outbreaks.