New research reveals that ‘foundation models’ trained on vast, general time‑series data may be able to forecast river flows accurately, even in regions with little or no local hydrological records.

The approach could improve flood warnings, drought planning, and water resource management in parts of the world where monitoring data is limited.

The study, published in ‘Machine Learning: Earth’, was conducted by researchers from The University of Texas at Austin and Hydrotify LLC.

Dr Alexander Sun from the University of Texas at Austin and Hydrotify LLC explained, “Reliable water information is essential for communities everywhere, but many regions still lack the long-term records needed to support traditional forecasting methods.”

Lack of flood warnings causes a chain of devastation

In many parts of the world, river gauges are sparse, records are incomplete, and monitoring networks are difficult to maintain.

Without long, reliable datasets, communities often have little warning before floods, limited insight into drought risk and fewer tools to guide water allocation and infrastructure planning.

When flood warnings are absent, the consequences shift from manageable to catastrophic. Without lead time, communities are caught in “flash” conditions, leaving no window to move valuables, fortify homes, or evacuate.

This lack of preparation leads to higher mortality rates and astronomical economic losses, as infrastructure and personal property are destroyed without any attempt at mitigation.

As climate pressures grow, the ability to produce useful flood warnings without relying on extensive local records is becoming increasingly important.

Efficient flood warning systems bring enhanced protections

Flood warning systems are vital for several reasons, including:

• They provide the precious minutes needed to reach higher ground, significantly reducing drowning risks.
• Even an hour’s notice allows residents to move vehicles or install flood gates, potentially saving thousands in damages.
• Systematic alerts allow first responders to pre-position resources and clear evacuation routes before they become impassable.

Overall, an early warning transforms a chaotic survival situation into a coordinated response, acting as the primary shield between a natural event and a human disaster.

AI models show strong performance in monitoring river flows

The research team evaluated several advanced AI models known as time-series foundational models (TSFMs).

Originally trained on time-series data from sectors such as energy, transport, and climate, these TSFMs were tested on a large US river dataset comprising more than 500 basins. One model in particular, called Sundial, performed nearly as well as a long short-term memory (LSTM) model that had been fully trained on decades of river flow records.

The AI models showed their strongest performance in basins dominated by strong seasonal patterns, such as snowmelt‑driven flow.

“Approaches like this show how new AI tools could help close that gap by giving more places access to data-driven flood warning predictions,” Sun commented

“While there is still progress to be made, especially in more complex river systems, this work points to a future where improved flood forecasting is possible even in areas that have been underserved for decades.”

A growing value for real-world water forecasting

The researchers noted that the capacity of TSFMs scales with the size of their training data.

As future generations of TSFMs incorporate more Earth science data, including hydrological and climate records, their value for real-world water forecasting is likely to continuously grow.