NERT DADS: A Near-Real-Time Dust Aerosol Detection System

Global climate is in part affected by airborne particles known as dust aerosols, which are abundant in dry deserted areas such as the northwestern region of Africa. Researchers have found that dust aerosols can travel long distances, even across continents, to participate in the life cycle of other ecosystems. However, in spite of dust aerosols being good for nature, elevated concentrations and prolonged exposure to dust aerosols may deteriorate the quality of life for human beings. Motivated by the potential benefit to our communities we propose a system that detects dust aerosols in nearreal time, allowing people to study their own geographical region and prepare for overcoming adverse scenarios caused by major dust events. The proposed system makes use of data science algorithms to estimate the probability of dust aerosols based on carefully chosen multispectral data. During training, our system performs to a 92% of accuracy, and produces a probabilistic view that enables researchers to observe and study the behavior of dust aerosols at a global scale, facilitating the modeling of dust behavior as it relates to global climate.