Researchers at Gwangju Institute of Science and Technology have developed an AI-based, hybrid system to predict wildfire danger. This brand-new, innovative forecast system combines deep learning algorithms with the currently available forecast models for accurately detecting wildfire conditions and enhanced coverage of existing models.
Wildfire Danger- A Rising Threat to Survival
No doubt, raging wildfires across the world are causing substantial economic damage. Also concerning is the ever-increasing number of fatalities caused by wildfires. If there’s a system that can predict this catastrophic event, it can substantially decrease damage since the available forecast systems aren’t as accurate or effective. Due to the limitations of these systems, it becomes impossible to predict an upcoming wildfire.
Details of the New AI-based Model
According to researchers, an Artificial intelligence (AI)-based model can help predict wildfires better than conventional methods because of the involvement of deep learning, which improves prediction accuracy.
Moreover, AI-based mode can determine where and when the fire will happen. This can also help in resource allocation and improving fire prevention. Applying a deep learning algorithm can enhance wildfire danger predictability in the Western US.
“This study is a big step forward as it demonstrates the potential of such an effort for enhancing fire danger prediction without the need for extra computing power.”
Dr. Rackhun Son – Lead author
The research results from a joint effort between US and South Korean researchers. This hybrid method can provide improved predictions from one week before the fire using finer scales (4kmx4km resolution). This can significantly improve fire management and suppression.
What Makes it Different from other AI-based Methods?
AI methods involving the use of data can infer things considerably well, but it is still difficult to determine how it achieved the inference. Researchers combined AI’s deep learning algorithm with computer models that worked on physical principles. This combination helped them diagnose extreme levels of fire danger.
Hence, benefitting from the AI-based model’s computational edge. These predictions are made after considering geographic features and strong winds. Such as in the Western US, where high canyons and mountains are widespread, this method can help make accurate predictions, which coarser models can find difficult. This model can eliminate the reliance on regional downscaling, which is more expensive, computationally exhausting, and time-consuming.
An Ongoing Research….
Co-author of the research, Prof. Kyo-Sun Lim from Kyungpook National University, Korea, stated that the newly developed AI-based model could make accurate high-resolution fire forecasts in much less time and is relatively more cost-effective than traditional systems. He further noted that in this study, they tested AI for fire danger prediction in the Western US but plan to apply it to weather extremities in other parts of the world.
