For the evaluation of state-of-the-art flood services and their machine learning enhancements, a global approach will be followed focusing on river delta areas. A few national case studies will also be studied in more details, to better understand the needs of local decision-makers through surveys and interviews to drive the advancements.
To demonstrate the value of the ML-enhanced forecasts and projections, Mozambique will be the main case study, as it is one of the most disaster-prone countries in the world and at high risk of compound floods, due to tropical cyclones making landfall within its 2500-km densely populated coastline. Better predicting compound flooding is essential to support humanitarian action and disaster management, to increase the resilience of vulnerable communities, as for example seen during Cyclone Idai and Kenneth in 2019.