Main content

Sustainability and climate change - how green tech can stop wildfires with sensors that ‘smell’ fires.

Madeline Bennett Profile picture for user Madeline Bennett February 9, 2024
Dryad’s Silvanet IoT system could help prevent the hundreds of thousands of human and millions of animal deaths lost to forest fires every year.


Wildfires are a major contributor to climate change, and their impact is getting worse. 2021 was one of the worst years for forest fires since 2000, causing nine million hectares of tree cover loss globally. Although this dropped to seven million hectares in 2022, this loss is still much higher than 20 years ago. 

The higher number of trees lost to wildfires means a lot more CO2 Is released into the atmosphere. Wildfires in boreal forests normally produce about 10% of global emissions; in 2021, they produced 23% of the total. 

As well as emissions, wildfires have a more immediate and deadly impact, killing anything in their paths indiscriminately. The bushfires that raged across Australia in 2019/2020 killed or displaced three billion animals, a number so high it’s difficult to visualise. Nearly 17 million animals died in the Pantanal fires in 2020. 

Earlier this month, forest fires raging across Chile's Valparaíso region killed at least 112 people. This time last year, more than 20 people died after wildfires swept through Biobío and Ñuble in Chile, further south from Valparaíso. And then there’s the estimated 340,000 premature deaths attributed to wildfire smoke every year.


Getting early detection of these fires is a crucial part of preventing this level of destruction, and that’s exactly what Dryad Networks is trying to achieve with its Silvanet technology.  Dryad takes a different approach to the two other existing technologies used to assist firefighters and tackle wildfires. The first is satellites, which look down from space to identify fires remotely; the other uses cameras to look for smoke plumes rising above the tree canopy. Carsten Brinkschulte, CEO and Co-Founder of Dryad Networks, adds:

Then there's a new way, which is to put sensors into the forest to embed them under the tree canopy and to detect fires using gas sensors. This is our approach. 

Dryad’s Silvanet suite consists of three hardware components: solar-powered gas sensors attached to trees that can ‘smell’ a fire and can be produced and deployed at low-cost and high-volume; mesh gateways, which are the IoT network infrastructure layer so the sensors can communicate. Again, these are solar-powered and attached to trees; and finally, a border gateway that connects an entire mesh network with tens of thousands of sensors to the internet over the 4G LTE network. 

The fourth component is a cloud platform, providing device management, alert management and running the service that users and firefighters engage with.

According to Brinkschulte, while there are several technological approaches to assist detecting forest fires, the key metric is how early can they alert nearby firefighters, and therefore increase the chances of extinguishing the fire successfully. 

Satellites are very good at predicting where a fire will develop, and assisting with evacuation planning or containment efforts, but early detection is difficult from space. Brinkschulte attests:

Because they're either too far away or if they're geo-stationary or lower orbiting, they're not permanently located over a certain place on earth. They stop by every few hours only, so they're not really early detection. But they're very good at predicting spread.

The use of cameras is currently one of the main ways to detect fires from a technical perspective, replacing the traditional way of people with binoculars looking for smoke plumes with technology and AI to identify plumes in the same way. Brinkschulte notes:

They can do a good job at detecting fires. The issue is that the fires have to be quite far developed, because when a fire starts, it's small, like a cigarette being thrown on the floor, starting as a smouldering fire and developing over a few hours into a real fire.

It’s only when the fires develop that the heat becomes immense enough for the smoke plume to rise above the tree canopy and be seen. In the first one to three hours, a small fire emits very little smoke and heat, and stays hidden under the tree canopy.


Dryad’s approach is to ‘smell’ for a fire, monitoring for the gas signatures emitted with forest fires. Brinkschulte explains: 

We put gas sensors, electronic noses into the forest and bed them in under the tree canopy, attaching them to a tree. When there is a fire developing, that enables fire to be detected at the very early stages, what's called a smouldering phase before there is even an open flame, and certainly way before the heat is so immense that it pushes it up through the canopy.

This means fires can be detected within minutes after someone starts even the smallest fire, a key aspect in increasing the probability of firefighters extinguishing a fire when it's still small, he adds: 

The smaller the fire is, the easier it is to put out. If the trees themselves are on fire, and you've got what's called a crown fire, where the fire spreads from one tree to the next, it becomes really dangerous and spreads extremely quickly. In particular, if there's wind. So early detection, or as we call it ultra-early detection, is increasing the chances of preventing fires at all.

This was the case recently in Lebanon, where the Silvanet system successfully identified and raised the alarm about an unauthorized fire at a customer site.

With 139,000 hectares of forest, equivalent to roughly 13% of its land area, Lebanon’s forests are increasingly at risk. The country loses 1,500 hectares of forested areas to fires every year, and long periods of drought are exacerbating the risk of fires in higher-altitude regions that previously experienced fewer wildfires. 

Against this backdrop, telecoms operator Mada rolled out the Silvanet system at a pilot site in central Lebanon as part of its CSR commitments. Last January, the company deployed 91 sensors and 2 gateways, covering a 78-hectare area of forest.

In December, Mada reported that the technology had detected a small illegal fire swiftly, allowing for a prompt response.  The gas sensor had detected a change in air composition; subsequent gas scans identified hydrogen, carbon monoxide and volatile organic compounds. When Silvanet's AI analyzed the patterns, it predicted a 70% probability of smoke, triggering an alert through the mesh network to the customer.

Mada now plans to extend the pilot to a full-scale deployment, safeguarding Mount Lebanon. 


For new customers in investing in the system, Dryad offers a network and deployment planning component as part of its cloud platform. Customers can plan a deployment in the area they want to protect and position the gateways and sensors on a map, says Brinkschulte:

Then we have a mobile application, which guides the forest workers to the location to install them. We try to make it as easy as possible for our customers, which are often not very technical, to install the system. Maybe we're not at the Apple user experience level yet, but that's certainly what we're trying to get to.

Dryad’s main customer base falls into three segments: private forestry, keen to protect their product; municipalities with an objective for public safety; and utilities, power lines and railroad operators looking to prevent fires due to liability issues. 

While Dryad is offering something different to satellites and cameras, all three approaches have their place, according to Brinkschulte, who notes that every technological approach to the same problem has positives and negatives. 

In the end, we envisage, and we see many customers applying all three or at least two of the solutions. There is no silver bullet that solves this problem perfectly.

The best approach is to deploy all three - satellites, cameras, sensors - they're very complementary. The problem is so big, it justifies applying multiple technologies to have an ideal solution where the advantages of one solution cancel out the disadvantages of another.

Looking ahead, Dryad is aiming to prevent 3.9m hectares of forest from burning by 2030, preventing 1.7 billion tonnes of CO2 emissions. To achieve this ambitious goal, it will need to deploy more than a hundred million sensors by 2030. Brinkschulte acknowledges:

There's a relatively small amount deployed at the moment, but this year our goal is to step up production for up to 300,000 units, because we really need to make good on the 2030 target. If we succeed in deploying that amount of sensors, we believe we can reliably or relatively reliably predict there would be substantial fires being detected and prevented, and with that, the emissions avoided.

That means spreading its technology in large volume across large areas. Dryad currently has 50 customers across the world, mostly in Southern Europe, Greece, Spain and Portugal, but also the US and Canada, Chile and South Korea. To date, the firm has manufactured 30,000 sensors and has shipped 15,000, most of which are now deployed. While getting from 30,000 to 300,000 and then a million is a lot of sensors to produce and sell over the next six years, as Brinkschulte notes, it only started to sell Silvanet last year. 

My take

The need for this type of technology is clear and urgent. The more temperatures rise, the more droughts; the more droughts, the higher the impact of wildfires, their severity, spread speed and damage. Coupled with more CO2 emitted by wildfires, further accelerating climate change. 

Dryad’s tech offers a new way to prevent wildfires with relatively little effort but a huge potential impact for reducing CO2 emissions, slowing down climate change and saving billions of lives. Let’s hope this example of green tech in action achieves its ambitions.

A grey colored placeholder image