Abstract
This study examines drought dynamics in the arid and semi-arid lowlands of Ethiopia, focusing on selected districts in the Afar and Somali regions using remote sensing and climate-based drought indicators. The results show ongoing environmental stress, limited ecological recovery, and noticeable spatial and temporal variations in drought severity across the study areas. A comparative evaluation of machine learning and deep learning models showed that the CNN-LSTM hybrid model performed best in capturing complex drought patterns and improving prediction accuracy. The study highlights the value of combining spatial and temporal data for drought forecasting and provides an important foundation for developing stronger early warning systems in highly climate-vulnerable regions.
Bio
Dr Elefelious Getachew is an Associate Professor of Software Engineering with over 15 years of experience across industry, academia, and research. He holds degrees from Addis Ababa University and Vrije Universiteit Amsterdam, and has been a visiting researcher at Virginia Tech and a postdoctoral researcher at the University of Milan. His work has received awards from Facebook, IDRC/Carleton University, Research ICT Africa, AAU, and the JP Morgan Faculty Research Award. He is currently Lead Researcher for AI for Societal Challenges at the Center of AI and Robotics, Institute of Advanced Science and Technology, AAU.