Using atmospheric dynamics to predict hazards in Asia and Africa
William Boos, Professor
Earth and Planetary Science
Applications for Fall 2025 are closed for this project.
The coupling of rainfall in Earth's tropics with atmospheric fluid dynamics is a vigorous area of research; it is also enormously consequential for many millions of people working in agriculture in Africa and Asia. In this project we will use theoretical atmospheric dynamics, large satellite datasets, and new AI weather and climate models to predict variations in rainfall and water availability for large populations of farmers in the tropics. Participants in this project will work with an interdisciplinary team of atmospheric scientists and AI experts to develop and deploy operational prediction systems in India and equatorial Africa. Through partnerships with economists and social scientists, these predictions will be communicated directly to tens of millions of farmers. A pilot version of this project successfully supplied skillful forecasts of the start of the rainy season to 38 million farmers in India in 2025, and will be expanded across a larger region of India and Africa in future years.
Role: Core tasks include analysis of large observational datasets, use of AI models and their output, and development of prediction algorithms. Students will learn atmospheric and fluid dynamics, methods for analyzing large datasets, and technology involved in state-of-the-art AI weather and climate prediction.
Qualifications: Familiarity with data analysis in python is required, including in a Linux/Unix environment. Ability to work collaboratively in a fast-paced environment is required. Some knowledge of fluid dynamics is beneficial but not required.
Hours: 9-11 hrs
Related website: https://boos.berkeley.edu
Related website: https://envfluids.github.io/monsoon-operational/#project-overview-section