Scaling Challenge: Local Precision vs. Global Reach
Specialized, hyper-local early warning systems have been designed to respond to flash floods caused by rainfall in specific urban environments, with examples in Florida (USA), Barranquilla (Colombia), Manila (Philippines), Nakhon Si Thammarat (Thailand), Mayagüez (Puerto Rico), and Barcelona (Spain). These systems typically rely on networks of physical sensors monitoring variables such as direct and radar-estimated precipitation, water levels, and flow velocity. While highly accurate for their specific locations, they are difficult to scale due to the high cost of hardware deployment, site-specific calibration algorithms, and the need for engineering expertise.
On a large scale, initiatives like WMO Flash Flood Guidance System (FFGS), the European runoff index based on climate science (ERIC) flash flood indicator, and the U.S. National Weather Service (NWS) flash flood warning The system provides extensive coverage through remote sensing and numerical weather models. However, these systems face significant obstacles regarding global implementation. A primary issue is their reliance on high-resolution hydrological maps and radar-based weather forecasts, resources largely unavailable in the Global South. Additionally, reliance on professional hydrologists to interpret complex model data and deliver actionable warnings presents another major challenge.
To achieve near-global reach, our model uses only global weather products (NASA IMERG, NOAA CPC) as well as real-time global weather forecasts ECMWF Integrated Forecast System (IFS) High Resolution (HRES) atmospheric model and this AI-based medium-range global weather forecast model by Google DeepMind. The system currently operates at a spatial resolution of 20×20 kilometers, a constraint primarily driven by the resolution of globally available data sources.
