CONAGUA, Google Sign Agreement to Enhance Flood Forecasting

CONAGUA, through Mexico’s National Meteorological Service, and Google signed a four-year technical cooperation agreement to use AI for weather and flood forecasting, improve early warnings and distribute official alerts through Android, Google Search, and Google Maps.

As Mexico faces a growing succession of extreme weather events, the country’s National Water Commission (CONAGUA), through the National Meteorological Service (SMN), and Google are combining meteorological expertise with AI to improve how authorities anticipate and communicate climate risks. The agreement comes as Mexico faces intense activity in the Pacific, including Hurricane Polo, which made landfall in Baja California Sur, and Tropical Storm Rachel, which is moving parallel to the Mexican coast and is expected to strengthen.

Under the cooperation agreement, the SMN will gain early access to Google’s latest AI-based weather models and predictive tools for the next four years. The initiative will focus on identifying extreme hydrometeorological events earlier and strengthening early-warning capabilities for the National Civil Protection System. The collaboration also establishes a framework for the SMN to evaluate and technically validate AI products under Mexico’s specific meteorological, hydrological, and geographic conditions.

AI Adds a New Layer to Flood Forecasting

One of the central components is Google’s Flood Hub, which will operate as an additional source of information within the SMN’s existing systems. The technology can help anticipate sudden floods in urban areas up to 24 hours in advance. Rather than replacing meteorologists or establishing a precise prediction for every street or property, the system provides an additional layer of information that can help specialists identify conditions associated with flood risk.

This distinction is relevant for public-sector organizations responsible for emergency preparedness. AI-generated forecasts must be evaluated alongside existing meteorological and hydrological information before authorities issue official warnings.

The agreement allows the SMN to assess these technologies under Mexican conditions while maintaining control over the information it provides. According to the agreement, both organizations will operate with their own resources, while each party retains ownership, control and sovereignty over its information in accordance with applicable regulations.

The National Meteorological Service's Fabián Vázquez says AI is already changing how atmospheric conditions are analyzed and forecast. “AI is no longer a tool of the future; it is a reality that is transforming the way we understand and forecast the atmosphere,” he notes.

For the SMN, the objective is therefore not to replace existing forecasting capabilities, but to complement them with additional computational models and data that can support analysis, forecasting and early warning processes.

From 15-Day Forecasts to Immediate Alerts 

The second technological component is WeatherNext 3, an AI weather model developed by Google DeepMind. The model incorporates satellite information and meteorological variables to produce forecasts with horizons of up to 15 days. Within the cooperation framework, this capability can support monitoring of extreme events, including hurricanes, while giving meteorologists another source against which to evaluate potential scenarios. The initiative addresses a second challenge beyond forecasting: distributing information quickly enough for people to act on it.

Official alerts issued by CONAGUA and the SMN will be integrated through a standardized Common Alerting Protocol (CAP) feed. This will allow warnings to be distributed automatically across digital platforms, including Android, Google Search and Google Maps. This could enable users to receive an official warning while using services they already access, rather than having to actively visit an institutional website to check weather conditions.

For communities exposed to intense rainfall, flooding, or hurricanes, the additional distribution channel can help reduce the time between an official warning and its arrival to the public.

Sebastián Valverde, Director General, Google Mexico, says the value of the technology lies in its ability to translate AI capabilities into tools that can support public safety. “AI gains its true value when it becomes a tool for protecting and improving people’s well-being,” Valverde says.

Building Mexico’s Climate Resilience With Technology

The timing of the agreement highlights the operational challenge facing Mexican authorities. Hurricane Polo recently made landfall in Baja California Sur as a Category 2 system, bringing the potential for significant rainfall to parts of the peninsula. At the same time, Tropical Storm Rachel is moving parallel to Mexico’s Pacific coast, with heavy rainfall and strong waves expected across parts of Guerrero, Michoacan and Colima.

These events illustrate why earlier detection alone is insufficient. Forecasting systems must connect data, analysis, and public communication so that information can support decisions before extreme conditions materialize. The initiative also creates an opportunity to evaluate how emerging AI models perform against the country’s particular geography and weather patterns.