A Comprehensive Review of Dust Storm Detection and Prediction Techniques: Leveraging Satellite Data, Ground Observations, and Machine Learning

Dust storms and sandstorms are meteorological phenomena that threaten human health, agriculture, transportation, and the environment. These storms, capable of transporting fine particulate matter over long distances, contribute to soil erosion, reduced air quality, and respiratory illnesses. Despite their widespread impact, accurately detecting and predicting dust storms remains a complex task due to the variability of environmental conditions and the limitations of existing data and models. This survey paper provides a comprehensive review of current dust storm detection and prediction research, focusing on numerical and machine learning approaches such as Convolutional Neural Networks (CNNs) and Support Vector Machines (SVMs). By analyzing various published techniques, this paper identifies gaps in existing methods and emphasizes the need for improved detection accuracy and early prediction systems. Key datasets such as MODIS, GOES-16, and ground-based measurements are examined, alongside the integration of hybrid data sources, to enhance detection capabilities. The survey also highlights the importance of developing robust, real-time prediction models to mitigate dust storms’ environmental and public health risks. In synthesizing the existing literature, this paper offers insights into the strengths and limitations of current methodologies while proposing future research directions aimed at improving machine learning applications for dust storm detection and prediction.