The power of AI in wildfire prediction and prevention
A bird's-eye view of a forest wildfire, highlighting the urgent need for AI in wildfire prediction and prevention across the globe. - REUTERS/Filepic
THE impact of AI in wildfire prediction and prevention.
To address the increasing frequency of wildfires in Türkiye, the World Economic Forum Centre for the Fourth Industrial Revolution, Koç Holding, the Turkish Ministry for Agriculture and Forestry, and Deloitte launched the FireAId initiative in January 2022. This initiative is using Artificial Intelligence (AI) to enhance wildfires prediction and prevention.
Since the launch, the Forum has built a global community driven by action and knowledge exchange, including Microsoft, Google, Hewlett Packard Enterprise, Scale AI, Conservation International, Inter-American Development Bank, NASA, Lockheed Martin, Sempra, Qcentroid, Patrick J. McGovern Foundation, Planet Labs and Miami University. Portugal and South Africa also joined Türkiye in this initiative.
As part of the pilot programme in Türkiye, an interactive wildfire risk map was developed, leveraging AI and machine learning algorithms – with inputs from the global community of experts. This map draws on multiple data sources including historical, meteorological, and geographical data, to develop models based on optimal resource allocation.
The first successful implementation in Türkiye has led to an 80% accuracy rate in predicting wildfires 24 hours before their outbreak. This information gives authorities a crucial window to prepare and respond proactively. The success of this pilot demonstrated how effective AI can be in supporting authorities to save lives, protect properties, preserve the environment, and significantly reduce wildfire damage.
Beyond the pilot program, this global multistakeholder collaboration is facilitating the exchange of use-cases and best practices, with the ultimate goal of strengthening global knowledge and accelerating the development and application of emerging technologies in wildfire prediction and prevention. Alongside AI, the community is exploring the integration of other technological domains, such as drones, which are important in managing wildfires.
This initiative provides a platform for technology companies and innovators to showcase cutting-edge solutions. It encompasses a wide range of contributions, including satellite and earth observation data that are relevant to wildfires, data and AI companies specializing in climate and wildfire modelling, as well as companies directly and indirectly impacted by wildfires within their supply chains and business models.
What’s the challenge with extreme weather events?
Climate change continues to be one of the most pressing concerns of our time, and its effects can be seen in the increasing frequency, severity, and spread of wildfires around the world. According to data collected by the National Interagency Fire Center (NIFC), approximately 3 million acres of land were impacted by fires every year since 1983. When recalculated from the year 2000, this number rises to 7 million acres annually.
Wildfires are a significant threat to forests, rural areas, and communities, with the potential to rapidly destroy homes, infrastructure, and agricultural land. Wildlife is also often lost in large numbers, posing an additional environmental concern.
As well as the destruction of property and natural resources, wildfires can have serious health consequences for those living in affected areas, due to the resulting dangerous air quality.
Our approach to implementing AI in wildfire prediction and prevention.
At the World Economic Forum Annual Meeting 2023 in Davos, the Forum launched The Next Frontier in Fighting Wildfires: FireAId Pilot and Scaling report which shared the FireAId initiative’s success and potential. Koç Holding CEO Levent Çakıroğlu explained the technology’s functionality, and highlighted that the algorithms and source codes would be made publicly available which is key to increasing the FireAId initiative’s capabilities and impact.
As more training data becomes available, algorithms will become more effective and capable of making more complex predictions based on a variety of variables. Previous wildfire-prevention AI technologies struggled to interpret non-linear data models, which this new algorithm interprets daily. Wildfire outbreaks do not always follow the same patterns, and the machine-learning model must be able to work around this. Researchers keep implementing new learning techniques to optimise FireAId technology to deal with the unpredictable nature of wildfires.
A FireAId Global Coalition is being established to build upon the successes of the initiative’s initial implementation, and to increase global awareness towards addressing wildfires by harnessing the power of AI technologies. Prioritizing AI in wildfire prediction and prevention will improve cost efficiency, effectiveness, and capacity of wildfire mitigation efforts at scale. The coalition aims to continue convening multiple experts to further develop AI prediction models to improve wildfire prevention, and enhance intervention strategies that can help save lives, reduce insurable risks, and enable targeted use of resources.
Must-Watch Video
Stay updated with our news