As the World Health Organisation defines, “A wildfire is an unplanned fire that burns a natural area such as a forest, grassland or prairie.” They are often caused by human activity or a natural phenomenon like lightning which can be happened unexpectedly at any time or anywhere. According to reports, among 50% of wildfires recorded, the reason for the fire is still unknown. Wildfires influence the entire ecosystem, like the flora, fauna, air, soil and water.

Over centuries, forests around the globe have encountered numerous wildfires. Between 2012 and 2021, there was an average of 61,289 wildfires reported a year and an average of 7.4 million acres influenced annually, as per the statistics of the Congressional Research Service. In the year 2021, 58,968 cases were reported that engulfed around 7.1 million acres. Since the era of advanced technologies like Artificial Intelligence has arrived, is there a sensible way to combat wildfires? California has the answer.

Cal Fire takes an intelligent turn

Californian firefighters are effectively leveraging the power of Artificial Intelligence to spot and combat wildfires. They are feeding video from more than 1000 cameras placed across the state into a machine that is capable of alerting the first responders when to retreat and mobilise.

The Cal Fire explained the potential of the ALERTCalifornia AI program, which was launched last month with an incident that happened recently; a camera spotted a fire that broke out at 3 a.m. local time in the remote, scrubby Cleveland National Forest almost 50 miles east of San Diego. As the people were asleep and darkness was concealing the smoke, the situation could have become worse and spread into a raging wildfire. But with the immediate alert of AI, a fire captain was able to call 60 firefighters, seven engines, two bulldozers, tow water tankers and tow hand crews, which saved the potential menace. Not more than 45 minutes, they were able to extinguish the fire.

How it works

The model was developed by engineers from the University of California, San Diego, using AI from DigitalPath, a firm based in Chico, California. The platform currently relies on about 1,038 cameras which were mounted by several public agencies and power utilities across the state; each of them is capable of rotating 360 degrees on the command of remote operators. The Cal Fire provided various incidents in which the AI alerted the fore captains to the fire even before the 911 call was made ever since the AI program began on July 10.

A professor of Geology and Geophysics at UCSD and the principal investigator of ALERTCalifornia, Mr Neal Driscoll, mentioned that the sample size so far was minimal to draw conclusions. But, Cal Fire believes that the technology will be able to serve as a model for other states and countries worldwide in the future as it is a need underscored by the unusually devastating wildfires, especially in areas like Hawaii, Canada and the Mediterranean this season.

Suzann Leininger, a Cal Fire intelligence specialist in El Cajon, just east of San Diego, said, “It’s 100% applicable throughout anywhere in the world, especially now that we’re experiencing a lot larger and more frequent fire regimes and with climate change.” Leininger is also in charge of training the machine. She analyses the previously recorded video from the camera network of what the AI detects as a fire, and then the machine predicts it was right or wrong with a binary yes or no answer. However, any phenomenon can trigger a false positive, for instance, the presence of clouds, dust, or even a truck with smoky exhaust.

Mr Driscoll said as hundreds of specialists are repeatedly working up and down the state, the AI has already strengthened in accuracy within weeks. Apart from the camera network, the AI platform also collects massive amounts of additional details, such as aerial surveys to compute the vegetation that can fuel future wildfires and map the Earth’s surface below the canopy, he added.

Aeroplanes and drones are also used to collect data in infrared and other wavelengths that cannot be observed by human vision. Besides, the platform can be used in the winter to analyse the atmospheric rivers and snowpack. The UCSD team also observes the data on burn scars and their influence on erosion, sediment dispersal, soil and water quality.  

• Photo by Mike Newbry on Unsplash