Preventing Wildfires, from Grazing to AI

Thousands evacuated, over 150,000 hectares burned across various regions, and the largest-ever activation of the EU’s civil protection mechanism. It’s Spain’s numbers for the month of August, as it battles 20 major wildfires amid one of the country’s worst heat waves. But beyond the horror stories, the situation has put in the spotlight the lack of prevention efforts in the Mediterranean country. Extreme heat fuels wildfires, yes, but now, conversations and anger also focus on the fact that the investment general and local governments make in wildfire prevention, has gone from 364 million euros in 2009, to 176 million in 2022. Prevention has many faces. It can range from education, as 85% of wildfires are triggered by human activity, including power lines, to controlled burns, mechanical clearing or livestock grazing on the grass and bushes that help wildfires spread. And then there's also smart tech. There are different solutions that can detect wildfires in very early stages. One of the main solutions we see are AI-enabled cameras. These solutions are trained to spot smoke from far away. They’re being used around the world, but California stands out. The state has ALERT California, a project developed by the University of San Diego, that helps the state’s fire departments spot and track wildfires across its territory thanks to a network of more than 1,100 AI-enabled cameras. When smoke or a fire is detected, images are sent to a control centre that verifies the issue and sends alerts to emergency centres. The project, with a funding of 20 million dollars, has helped state firefighters Calfire reach their goal of putting 95% of fires out at 10 acres or less. A good return on investment considering the emotional and monetary cost of wildfires. The US is investing heavily in these AI-camera solutions. We’ve seen company Pano AI recently raised 40 million dollars, with their size and client list, which includes utilities like Austin Electric, continuing to grow. And recently, utility PacifiCorp opened their own wildfire intelligence center in Oregon in which they’ll spot and monitor wildfires using an AI camera system by ALERTWest. Back in Europe, this month, Cyprus announced they were deploying a network of 15 AI wildfire detecting cameras from FireWatch AI. Then, there are sensors. UK company Pyri has developed a biodegradable pinecone shaped sensor that only activates with heat. The sensors are dispersed in remote areas and when a wildfire reaches them, heat triggers a reaction inside the sensor that sends an RF signal to alert of the fire. There’s also what is known in the industry as smell sensors, IoT sensors that can detect fire gases or particles and alert of the presence of wildfires. Hawaii deployed N5 Shield sensors after the Lahaina wildfire. Berlin-based Dryad Networks has small smell sensors that create a mesh network that hops and hops into the forest, communicating with each other and their platform when there are fire gases present. Cameras and sensors detect, to stop the fires from growing. But when the fire has grown into a full-blown wildfire, solutions like drones, satellites or predictive AI models are used for situational awareness and planning. So satellites are great for tracking wildfires, I think not so good at detecting them early at least. Cameras are good at detecting fires, over large areas. But sensors are good for what we call ultra early detection of fires. That's where we fit in. And we aim to complement cameras and satellites. With wildfires, it’s all about team work. And we’ve seen that with the international proof of concept BurnMonitor in 2020, which brought together Inria, Moraga Orinda Fire District, Berkeley University, Analog Devices, Planet Labs and Zonehaven to test end-to-end wildfire detection, alerting, and evacuation solutions. The system was able to detect every fire and to coordinate with the different aspects. So the sensors on the ground, the cloud system, the constellation of satellites, and then the fire department itself. As temperatures become hotter each summer and forest areas become drier, there needs to be more investment in prevention. Clearing woods of the dry grass and bushes, and looking for extra help like that of smart tech. The issue is that, on many occasions, the measures come too late. Most projects in which utilities or cities deploy wildfire prevention technology normally come after a wildfire has already occurred. Wouldn't it be better to invest before? One can never be 100% protected against a wildfire, but we can prevent many of them by using the right methods and technologies together, for each case.