The following is an extract from our climate newsletter Fix the Planet.聽Sign up to receive it for free聽in your inbox every month.
On 8 August 2023, after days of windy weather, utility poles toppled into the long grass just outside the town of Lahaina on the Hawaiian island of Maui. A few minutes later, the first flames flickered around the downed power lines. Within minutes a brush fire erupted.
The fire was declared fully contained by firefighters, but a few hours later it was whipped up again by the strong winds. In the hours that followed, the wildfire grew rapidly in size and intensity, ploughing through the historic port town of Lahaina. It killed 102 people, destroyed more than 2200 buildings and caused $6 billion worth of damage. Cars exploded in the heat and people jumped into the sea to escape the blaze. Hawaii鈥檚 governor Josh Green described the fires as 鈥渢he largest natural disaster in Hawaii鈥檚 state history鈥�.
Advertisement
It is a tragedy that could have been prevented with the right technology, according to the US Department of Homeland Security, which is working to develop technology聽that can help fire services spot a smouldering fire before the first flames appear.
Wildfires rising
The rise in wildfires is one of the clearest indications of how fast the climate is changing. Multiple studies have shown that climate change is increasing the frequency, intensity and duration of wildfires around the world.
Wildfires are bad for the climate in two key ways. Firstly, burned trees and vegetation release stored carbon dioxide back into the atmosphere. Wildfires in 2023 generated 2170 megatonnes of carbon emissions, , with blazes in Canada accounting for almost a quarter of the total.
It鈥檚 not just the carbon emissions from burning trees we need to worry about. Wildfires also diminish the natural carbon sinks of forests that we rely on to absorb future emissions. and released earlier this month warned that wildfires are reducing the planet鈥檚 ability to store carbon, threatening global climate targets.
Stopping wildfires before they take hold is not only essential for saving lives and property, it is also crucial for the climate.
Sensing danger
There are already plenty of technologies in play to help firefighters spot wildfires as quickly as possible. Satellites can swing over huge swathes of land on a regular basis, reporting when a blaze is under way. Traditionally this method could only spot the largest fires that were already out of control, but satellites are becoming better at detecting blazes early. Next year, the first Google-backed FireSat satellite and it will be able to detect wildfires covering areas as small as a classroom. Meanwhile, 360-degree cameras can be deployed on monitoring towers and alert emergency services at the first sign of a blaze.
But how do you catch a blaze before it has started? A new breed of early warning system relies on chemical sensors to detect the first whiff of smoke in the air, sending an alert to fire crews while a potential blaze is still smouldering.
N5 Sensors, based in Maryland, is one of a host of technology companies in North America and Europe developing these state-of-the-art wildfire sensors. Rather than trying to spot flames, N5鈥檚 product looks for a 鈥渇ire fingerprint鈥� in the air, says CEO . The sensors collect data on levels of chemicals and particulate matter, as well as temperature, humidity and more. Each sensor 鈥� which is powered by solar energy and fitted to a utility pole or traffic light 鈥� can monitor around 400 hectares of land. Multiple sensors can create a network that builds a real-time digital map of ground conditions across a vulnerable landscape.
Using AI algorithms, the network then compares the measurements to readings from nearby sensors and determines whether a fire has started. It can also identify the potential cause based on the chemical signature, says Motayed 鈥� for example, if it has started from a barbecue, or in wild lands. Wind sensors can also be deployed, to help the algorithm predict which direction a fire might spread in. Cellular networks allow the sensors to ping information back to emergency services, within minutes of the first detection.
鈥淚f you look at individual sensor readings and consider them as raw data points, they鈥檙e unprocessed,鈥� says Motayed. 鈥淏ut then the intelligence comes from the AI, where it鈥檚 taking sensor readings from different sensors within each node, but also comparing that against other nodes in the same network. Then it鈥檚 looking at the temperature, pressure, other environmental readings, and then making that assessment that what it鈥檚 detecting is a true fire.鈥�
Pilot towns
N5鈥檚 sensor was chosen as the technology partner for , which began in 2020 to test early wildfire-detection techniques. After lab tests, 200 sensors were deployed in towns across the US and Canada in 2023, including in Hawaii, Colorado and California. A further 200 are being rolled out this year.
This kind of technology works best not out in the wilderness, but in areas near wildfire-vulnerable towns, says at the Department of Homeland Security. After all, early fire detection is only useful if nearby crews can reach and extinguish the blaze quickly.
鈥淲e鈥檙e really focusing for the wildland fire sensors on trying to protect those vulnerable edges of development that are starting to impinge on the wildland areas, where there鈥檚 lots of fuel and potential,鈥� he says. 鈥淲e鈥檙e looking to catch the first couple of molecules or chemical signatures of a fire, so that authorities can be alerted and we can do early suppression.鈥�
The pilot has already produced results. Last year, the sensors caught two lightning strikes on trees in California and Canada. Such strikes can smoulder for days, impossible to spot with satellites and cameras, before suddenly erupting into flames. The sensors also caught a blaze in Colorado, a controlled burn that reignited after fire crews left the scene. The fire was detected 37 minutes before a 911 caller first reported a blaze 鈥� by which time fire crews were already en route to the situation.
鈥淭hat鈥檚 a huge deal,鈥� says at N5 Sensors. 鈥淚f you ask a firefighter, would you like a 30-minute head start in putting out a fire? The answer is unequivocally yes.鈥�
The evidence that sensors can be a useful tool for fire-prone areas is clear, Cotter believes. The sensors are 1000 times more sensitive than the smoke detector in your home, he points out. 鈥淚t鈥檚 still a research project, but we feel very confident that the results we are getting are excellent,鈥� he says.
The pilot is due to end next year. After that, it will be up to local districts in the US and Canada whether they invest in the technology. Cost is an issue 鈥� equipping an area with N5 sensors costs around $25 per hectare of land. Plus, there鈥檚 the demand on fire crews to investigate alerts 鈥� not all districts will have the resources to do so.
While wildfire sensors could be a useful tool for protecting vulnerable towns and cities, they will not be as useful for preventing huge wildfires in remote areas, such as the Amazon rainforest or Siberian tundra where blazes can cover hundreds of square kilometres and release thousands of tonnes of CO2. Preventing wildfires in these areas in addition to smarter technologies.
As for the Hawaiian island of Maui, left devastated by last year鈥檚 fires, there is a glimmer of hope. In August of this year, 80 wildfire sensors were deployed across Maui and the Hawaiian Islands. The hope is this new technology could help prevent another disaster.
