Abstract
As the severity and number of dust storms grow, weather expert and researcher techniques evolve in an attempt to understand the nature behind such events and predict future storms with higher accuracy. Dust storm behaviour varies based on five attributes. These are wind speed, pressure, temperature, humidity and surface type. Storms can affect heavily congested cities and can cause increased difficulty in outdoor activities and everyday operations. However, their prediction using historical data has not been used yet in a fully efficient and comprehensive way. This study examines the process of both predicting and identifying dust storms using historical dust events through the combination of hybrid artificial intelligence networks combining Bayesian networks (BN) with a case-based reasoning (CBR) approach. Their outcomes are then being utilised to ignite appropriate policy making against dust-storm events using rule based system (RBS) techniques.