Google Research Open-Sources Flood Model, Extending Forecast Horizon by 6 Days

3 articles · Updated · Google Research · Jun 3

Google Research released its hydrology modeling framework on GitHub under an Apache 2.0 license, letting meteorological and hydrological agencies train and adapt AI flood forecasts in their own workflows.
The package includes the original 2024 benchmarked model and a newer ME-LSTM system now used in Flood Hub; Google said the upgraded version extends reliable forecasts by six days in gauged basins and one day in ungauged basins.
Python- and PyTorch-based tools let forecasters combine global weather inputs with local river, watershed and operational data, while retaining control of their own datasets and incorporating local or Indigenous knowledge.
Czech Hydrometeorological Institute testing found the AI forecasts comparable to traditional locally calibrated models, and CHMI built an adapter for Delft-FEWS, offering a template for wider operational adoption.
Google framed the release as a lower-cost way to broaden early-warning capacity, especially for resource-constrained regions, with the World Meteorological Organization backing open, interoperable hydrological tools.

Sources

Google Research5h ago

Google Research Open-Sources Hydrology Framework for Global AI Flood Forecasting

hess.copernicus.org23h ago

HESS - Comprehensive Global Assessment of 24 Gridded Precipitation Datasets Across 18 428 Catchments Using Hydrological Modeling

mdpi.com5h ago

Flood Inundation Area Prediction Under Climate Change Scenarios by Integrating Hydrological and Hydraulic Models with a Hybrid Deep Learning Framework

4 Sources

What is the real price for developing nations to use Google's 'free' flood forecasting tool effectively?

As Google's AI models flood forecasts, will they amplify local expertise or wash it away?

Google Open-Sources Global Flood Forecasting Model v2: AI-Powered Early Warning for Billions

Overview

In March 2024, Google Research released its advanced Global Flood Forecasting Model (v2) and extensive datasets, marking a major step in leveraging technology for a more resilient future. By open-sourcing these tools, Google democratizes access to cutting-edge predictive technology, empowering communities worldwide to better anticipate and respond to floods. This move accelerates research and development across the scientific community and fosters collaborative innovation against climate-related disasters. Driven by a commitment to move beyond documenting past events, Google aims to build systems that prevent future crises and ensure vulnerable populations have the information needed to prepare and stay safe.

...