Supercomputing and Machine Learning against wildfires

With wildfires becoming more intense and challenging to predict, advancements in technologies to fight them are also advancing.

Thanks to the collaboration agreement with the Centro de Supercomputación de Extremadura of Fundación COMPUTAEX, we have implemented new functionalities to boost the performance of our GIS platforms for weather monitoring and wildfire management, utilizing the LUSITANIA supercomputer.

Harnessing Supercomputing for Enhanced Weather Data

To enhance the Canadian wildfire risk indices featured on our Weather Map Site platform, we have created a new methodology to calculate a specialized wildfire risk index aimed at identifying the likelihood of large-scale wildfires.

This index is specifically adapted to Spain’s geographic and climatic characteristics, where wildfires can be more intense than in other parts of continental Europe. . It allows for quick and visual analysis of areas where wildfires, if they occur, are likely to become large-scale emergencies (> 500 Ha).

With the support of LUSITANIA’s advanced computing power, we also have been able to calculate climatologies for the 1991-2020 period, essential for studying anomalies in key parameters such as temperature, humidity, precipitation, and wind.

This allows firefighting resources to assess if the current season’s weather deviates from historical maximum, minimum, and average values. Such data is crucial for planning logistics and developing action strategies with clear references based on past scenarios.

Machine Learning Models for Smarter Resource Allocation

As part of enhancing emergency response, supercomputing has also played a key role in training machine learning models.

Aimed at improving resource allocation for firefighting, these models enable emergency managers to determine the ideal quantity and type of resources required to address the fire effectively.

The result is that deployed teams can operate with the assurance of full support, while maintaining a balanced response and reserving resources for other concurrent emergencies.

Eager to test this innovative weather data? Register on Weather Map Site and get started.