HydroNetNZ: An AI-driven mobile app for impact-based flood forecasts and warnings at household level in New Zealand

'As climate extremes intensify, traditional flood forecasting is no longer enough. Communities don’t just need to know when it will rain. They need to know what that rain will do. Impact-Based Forecasts and Warnings (IBFW) focus on “what the weather will do” rather than just “what the weather will be.” This project aims to enable IBFW for floods in New Zealand by addressing key practical challenges in implementing such systems. A major barrier is the computational intensity of conventional flood hazard (inundation) modelling methods used to generate detailed maps of flood extent and depth. These flood maps are essential for understanding local flood risk, and are typically based on the dynamics of flood drivers such as river flow, sea level, and the topography of the affected area. In this research, we develop AI-based models that generate rapid, high-resolution flood maps in a fraction of the time required by traditional approaches. We also develop a supporting IBFW platform and mobile app to disseminate localised forecasts at the household level and collect crowdsourced flood report, creating a feedback loop that enhances model accuracy and strengthens community resilience.