Google Launches AI Rainfall Tracking System to Improve Flood and Drought Monitoring

Date:

New tools for studying disasters in poorly monitored regions are offered by the technology, which maps global rainfall every 15 minutes.

Access to a global rainfall-tracking AI system has been opened by Google DeepMind, potentially helping scientists monitor floods and droughts in areas lacking weather infrastructure.

An announcement was made by the company on Nov. 13 stating that developers and researchers now have access to historical archives and live data from PARDIG, its AI-driven global rainfall tracking system.

Satellite observations and atmospheric data are combined by the technology to calculate precipitation at 15-minute intervals across a worldwide grid featuring an approximate 10-kilometer resolution.

Access to rainfall intelligence could be expanded by its release across developing nations and isolated areas, where sparse weather radar coverage complicates tracking storms, flooding, and extended dry spells.

Precipitation measurements collected by NASA’s Global Precipitation Measurement satellite are used by PARDIG to learn, according to Google, while the system incorporates extra satellite imagery and atmospheric data to estimate global rainfall.

Estimates over oceans and regions with sparse observation networks can be generated by PARDIG, unlike conventional monitoring systems that rely heavily on ground-based weather stations.

Google’s WeatherNext 3 forecasting model is also supported by the dataset, utilizing PARDIG-generated precipitation data throughout its training process.

Historical rainfall records dating back to 1999 can be accessed by researchers, alongside a separate dataset covering 2016 through the present day.

Both datasets have been released by Google under a Creative Commons license permitting commercial reuse with attribution, potentially enabling businesses to integrate the data into agricultural monitoring, water management, and disaster-risk tools.

Google’s AI Outperforms NASA’s Rainfall Benchmark by Over 30%

More than 30% higher hourly precipitation detection skill was achieved by PARDIG than NASA’s IMERG-Final dataset when measuring light-to-moderate rainfall, according to DeepMind.

The system was evaluated by the company against radar networks across the United States, Europe, and Japan, alongside data from over 36,000 global weather stations.

Improved identification of intense rainfall events was also reported in its research paper, addressing an area where existing global precipitation datasets often struggle to capture accurate readings.

More detailed models of extreme weather events, including flash floods, landslides, and drought conditions, could be developed by researchers with the help of these capabilities.

The July 2025 Texas Hill Country floods were highlighted by Google as one example, comparing PARDIG’s estimates against ground-based radar readings and NASA’s existing precipitation dataset.

Certain limitations are carried by the technology, however, which could restrict its effectiveness during rapidly evolving emergencies.

Near-real-time data is published by Google in six-hour batches, creating an average delay of roughly five hours between observation and availability, even though PARDIG estimates rainfall every 15 minutes.

The dataset cannot be used by emergency services as an immediate replacement for live weather alerts.

Snowfall in mountainous regions can be underestimated by the system, Google also acknowledged, noting that it remains less reliable across certain polar areas.

Meanwhile, PARDIG is still classified as experimental by the company, which advises users to depend on official meteorological agencies when making public-safety decisions.

Marton K.
Marton K.https://thecoingraph.com
Marton is seasoned crypto and finance journalist with over four years of experience. He has contributed to several high-profile outlets.

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Share post:

Subscribe

spot_imgspot_img

Popular

More like this
Related

Vitalik Buterin Backs Faster Path to Quantum-Safe Ethereum Upgrade

Under a newly outlined framework, bulky digital signatures would...

Former FBI Agent Admits Stealing $2.47 Million in Cryptocurrency

A former FBI agent pleaded guilty to stealing nearly...

Greece Introduces 10% Tax on Crypto Gains and 15% Levy on Staking Income

Greece’s parliament approved a sweeping cryptocurrency tax framework that...

Claude Enters Personal Finance as Anthony Pompliano Accuses Anthropic of Copying His Startup

Anthropic recently rolled out a new feature that lets...