Research Project
Cross-Modal Attention for Dust and Sandstorm Forecasting in the Sahel using Surface and Atmospheric Features
My research on combining surface and atmospheric data to forecast dust and sandstorms in the Sahel. SahelWatch is the app for this work.
Research question
How can surface and atmospheric features be combined to forecast dust and sandstorms in the Sahel 24–48 hours before occurrence?
Motivation & study area
The work focuses on the Sahel, using atmospheric and surface data to support earlier warnings of dust and sandstorms.
Methodology
The research investigates cross-modal attention for combining surface and atmospheric features. The existing implementation lists Python, TensorFlow, LSTM and geospatial data.
Software implementation
SahelWatch is the app for this research. It provides a web interface for the dust and sandstorm forecasting work.