# AI revolution in global flood forecasting Google's machine learning models extend flood predictions up to 7 days in advance. - **Source:** [Google AI Blog](https://ai.googleblog.com) - **Published:** 2024-03-20 - **Topic:** [Data & AI](https://brifing.si/c/data-ai) - **Original article:** - **On Brifing:** https://brifing.si/s/ad8a302b-58d5-402a-abdf-457481c68d9b ## Global flood threat Floods are the most common natural disaster, causing over $50 billion in annual damages. Due to climate change, flood frequency has doubled since 2000. 19% of world population (1.5 billion people) lives under substantial flood risk. ## Early warning systems Accurate flood forecasts can save thousands of lives. Google launched its flood initiative in 2017. The system provides real-time alerts via Search, Maps and Android. However, scaling in data-scarce regions remained challenging. ## Machine learning solution Research published in Nature showed ML models improve forecasts even in data-poor areas. AI extended prediction windows to 5 days and brought African/Asian forecasts to European standards. ## Global collaboration Models were evaluated with ECMWF. Flood Hub now offers 7-day river forecasts across 80+ countries. Governments and NGOs use this data for early action to protect vulnerable populations. ## India pilot project The 2018 Ganges-Brahmaputra pilot showed LSTM models outperformed traditional hydrology systems. This success led to expansion across India and Bangladesh. Yale University tested warning system effectiveness. ## Solving data scarcity Low-GDP countries have less river data but higher flood risks. ML trains on global data to predict in ungauged basins. An open-source hydrology dataset was shared in Nature Scientific Data. ## LSTM technology Unlike traditional models, LSTM-based systems process daily inputs and system states for time-series prediction. The two-stage LSTM model analyzes precipitation and watershed data. ## Predicting extreme events Models accurately predict extreme floods not in training data. Uncertainty estimation methods integrate multiple data sources, crucial for infrastructure-poor regions. ## Hope for the future Google's AI flood system provides critical defense against climate change disasters. This globally accessible technology has significant potential to reduce casualties, especially in developing nations. ## Sources covering this story - http://blog.research.google/2024/03/using-ai-to-expand-global-access-to.html — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._