EMS Annual Meeting Abstracts
Vol. 22, EMS2025-523, 2025, updated on 30 Jun 2025
https://doi.org/10.5194/ems2025-523
EMS Annual Meeting 2025
© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.
Artificial Intelligence and Machine Learning at Basque Meteorology Agency
Santiago Gaztelumendi and José Antonio Aranda
Santiago Gaztelumendi and José Antonio Aranda
  • Basque Meteorology Agency (Euskalmet) , Vitoria-Gasteiz, Basque Country (s-gaztelumendi@euskalmet.eus)

Artificial intelligence (AI) and machine learning (ML) represents an important evolution in computer science and data processing that is quickly transforming a vast array of industries. Driven by the increasing need for digitalization over the last decade, AI/ML are being used in businesses of all types to automate processes, analyze large datasets for insights, improve decision-making, enhance customer experiences, and optimize operations.

The Basque Meteorological Agency (Euskalmet) is the official weather service of the Basque Country, responsible for weather monitoring, forecasting, and civil protection support. It provides weather data, develops climate studies, and offers public information, contributing to emergency preparedness and climate change research. Established in 1990 and reorganized in 2024 as a public entity under the Basque Government, Euskalmet also plays a key role in managing alerts, collecting meteorological data, and communicating weather-related risks to the public.

AI and ML are powerful tools that enhance observation, modeling, forecasting, and analysis capabilities in the geosciences, making meteorological and climate tools and systems more efficient, intelligent, and useful for society. In the case of Euskalmet, AI/ML are primarily used or planned to be used, in relation to their utility in improving forecast accuracy, analyzing large volumes of meteorological and climatological data, detecting extreme weather events and anomalies, downscaling models to generate high-resolution local projections, automating operational tasks such as quality control and satellite image classification, supporting hybrid physical-data modeling approaches, and enabling data-driven decision-making for weather-sensitive sectors. These technologies enhance Euskalmet’s capacity to monitor, predict, and communicate weather and climate information more effectively and efficiently, particularly during severe weather and impact events.

In this work, we will present the strategy we are pursuing at Euskalmet regarding the exploration of opportunities offered by AI/ML. We will provide an overview of the steps we are taking, including the activities already undertaken and those planned, with a focus on our experience as a small-scale meteorological center.

We will also share details of Research, Development and innovation projects and operational implementations we are currently doing in this field, which may serve as inspiration for others following a similar path. Particular attention will be given to key aspects dealing with surveillance, forecasting, modeling, climate, and observation. Finally, we will present future plans and conclusions based not only on our technical experience but also on other relevant aspects such as management, funding, infrastructure, partnerships, and more.

How to cite: Gaztelumendi, S. and Aranda, J. A.: Artificial Intelligence and Machine Learning at Basque Meteorology Agency, EMS Annual Meeting 2025, Ljubljana, Slovenia, 7–12 Sep 2025, EMS2025-523, https://doi.org/10.5194/ems2025-523, 2025.