This article has undergone a thorough review in accordance with Science X's editorial process and policies. Our editors have identified the following notable characteristics while ensuring the content remains credible: A recent study published in Nature reveals that an artificial intelligence (AI) model has the potential to enhance the accuracy of flood forecasting. The research indicates that this system can match or surpass existing leading methods, offering earlier warnings for significant flooding events. Due to human-induced climate change, certain regions have experienced a rise in the frequency of floods. However, current forecasting methods heavily rely on stream gauges, monitoring stations situated along rivers, which are not evenly distributed worldwide. As a result, forecasting ungauged rivers becomes more challenging, particularly impacting developing countries. To address this issue, Grey Nearing and their team developed an AI model.

This model was trained using data from 5, 680 existing gauges, aiming to predict daily streamflow in ungauged watersheds over a 7-day forecast period. Subsequently, the AI model underwent testing against GloFAS (the Global Flood Awareness System), the globally recognized software for flood prediction in short-term and long-term scenarios. Remarkably, the AI model provided flood predictions five days in advance that were just as reliable as, if not superior to, the current system's same-day forecasts. Furthermore, when predicting extreme weather events with a return window of five years, the AI model demonstrated comparable or better accuracy compared to GloFAS forecasts for events with a one-year return window. These findings strongly suggest that the AI model has the potential to offer flood warnings for both minor and severe events in ungauged basins, providing a longer lead time than previous methods. Consequently, access to reliable flood forecasting in developing regions could be significantly improved. For more information, please refer to the original study: Grey Nearing et al, "Global prediction of extreme floods in ungauged watersheds, " Nature (2024). DOI: 10. 1038/s41586-024-07145-1.