IoT in the Line of Fire
by Thomas Søderholm, VP Business Development, Nordic Semiconductor
Of such intensity were the blazes of Australia’s devastating “Black Summer” wildfires of 2019-20, that locals evoked paranormal and mythological language to describe them. Beneath a “mountainous curtain of smoke” emanated an “intense, pulsing orange and yellow furnace-like glow” that ran the entire length of a cliff, recalled one property owner. A reporter recounted how fierce winds “like dragon’s breath” propelled the fire forward at a terrifying pace, scorching in a few hours huge swathes of terrain that should have taken days to cover. It was like an “atomic bomb” had gone off, with plants and animals “vaporized,” recalled another local.
Once a conflagration like this began, there was almost nothing that could hold it back (Figure 1). It’s common wisdom in emergency management circles that quelling a wildfire within the first hour offers the best chance of containment. But it would take a day before this fire was even detected, having been ignited by lightning in a remote forest. That was much too late. The blaze would grow into one of the fiercest of the summer, burning for 74 days, destroying 1.2 million acres, more than 300 homes, and causing untold environmental, health and psychological scarring.
Now, as wildfires intensify all around the world—propelled by the forces of climate change—authorities, researchers, and scientists are looking for a better way to limit the damage. Many are turning to technology, specifically IoT solutions that combine sensors, wireless connectivity, and AI. These deployments promise not only quicker detection of fires, but also improvements right across the wildfire management lifecycle, from prevention through to response and recovery.
Such an intervention is long overdue. Though more intense today, wildfires have plagued the globe for as long as humans can remember, and likely well before that. Using charred fossils, scientists date the earliest evidence of wildfires to approximately 420 million years ago, at a location near Wales in the U.K. That first fire too was likely caused by lightning. Today, natural effects remain a common cause of wildfires, though human activity has raced up the ladder.
Almost 85% of U.S. wildfires are now caused by humans, according to the U.S. National Park Service, citing factors including unattended campfires, camping equipment malfunctions, discarded cigarettes and, worst of all, arson.
An Unvirtuous Cycle
Wildfires and the climate share something of an unhealthy relationship, each bringing out the worst in the other. Climate change, through its effects of longer drought periods and higher temperatures, is dramatically increasing the risk of wildfires. In drier conditions, trees and vegetation become more flammable, increasing the likelihood of ignition and the rate at which fires spread.
Yet, just as the climate increases the prevalence of wildfires, each wildfire makes climate change worse. In 2021, wildfires released 1.76 billion tons of carbon, according to the EU’s Copernicus Atmosphere Monitoring Service—a record amount that was more than double Germany’s total annual CO2 emissions. Forests, instead of acting as a carbon sink, can become a net carbon source.
The destructive impact of wildfires is perhaps most starkly seen in the devastating effects on flora, fauna and the human population. The previously mentioned Black Summer bushfires—as wildfires are known in Australia—saw 186,200 square kilometers burned, 3,500 homes lost and more than one billion animals perishing.
Worryingly, there’s little good news on the horizon. Experts predict extreme wildfires will become more intense and frequent, with wildfire events increasing by 50 percent by the end of the century. Even the Arctic now faces rising wildfire risk, says a recent UN report.
Away with Old Ways
Fire has light heartedly been described as man’s greatest “invention.” But now the management and mitigation of wildfires is in desperate need of human inventiveness to address the escalating scale and impact of the blazes. Thousands of years ago it seems we were better at managing wildfires and today Australia’s indigenous peoples are lending their generations of knowledge to modern firefighters.
Conventional strategies to reduce the impact of wildfires haven’t always worked so well. The earliest approaches relied on watchtowers, but these were not efficient as they relied on human observations, says Nature magazine. More recent detection solutions use cameras mounted on poles, satellite imagery and aircraft to detect plumes of smoke.
Though an improvement on having a human in a tower, these options too are limited by being optically based. The task of spotting smoke from a new fire can be hindered by the height of tree canopies, cloud cover and, ironically, smoke haze, resulting in potential detection delays of several hours. Detecting fires in the so-called smoldering phase, or within the first hour, is critical to having a chance to prevent their uncontrollable spread and the resulting damage.
Sensing a New Path
In the wake of Australia’s devastating Black Summer, an official inquiry reached a powerful conclusion about the country’s response efforts. Yes, the fires “were unlike anything seen… before,” but technology had been underutilized, it said. The inquiry urged authorities “to push our technological and our research capabilities much harder.” It specifically highlighted “remote sensing technology,” which it said could make a marked difference to so-called “big fire-risk seasons” through “enhanced capability for early detection of new ignitions, real-time tracking of the fire edge progression and intensity as it spreads, and better understanding of vegetation and fuel load issues before the fires start.” Since then, the UN too has advocated for remote sensing and real-time alarm systems to be deployed to monitor, detect and control fires.
Happily, IoT solutions are now finding their place in wildfire management to support this need. For instance, IoT sensors placed on trees in forests susceptible to wildfires can provide authorities with accurate early warning signs of a fire by detecting the gases emitted during the smoldering phase. Systems that can also integrate these sensors with long-range connectivity and dedicated GNSS—such as Nordic’s nRF9160 SiP (Figure 2)—can aid fire crews even further by pinpointing the precise location of the fire.
As well as detecting gases like carbon dioxide and oxygen, sensors could also be deployed to detect environmental indicators such as humidity and temperature. This is where IoT edge devices with embedded machine learning (ML) capabilities bring significant value. By not only collecting the data from sensors about air quality and other variables such as humidity, temperature and wind conditions, but also analyzing that data using ML capabilities, tomorrow’s IoT edge-computing devices will be able send back to emergency management teams valuable insights and predictions.
They will make available information about things like the chances of fire breaking out and how quickly, and where an outbreak is likely to spread. As a signatory to the UN’s SDGs, Nordic is particularly excited about the beneficial impact its solutions could have on the challenges posed by wildfires.
Other similar solutions are also being seen. Following catastrophic wildfires in 2021, Vodafone deployed a network of sensors into what it dubbed a “smart forest” in Sardinia, Italy. The sensors were networked to communicate both with each other, and back to a gateway at the edge of the forest, where data about the fires could be sent to a cloud-based center.
Another promising Australian IoT deployment resulting from the recent bushfires proposes to deploy “sensor node pairs,” in which temperature, humidity, plant stem water content and wind data is collected from tree-mounted sensors and combined with soil moisture data collected from ground sensors to assess fire risk. The sensor nodes use Bluetooth® LE to communicate with each other, first responders’ smartphones and cloud gateways.
The value of these solutions extends beyond detection. A critical challenge for authorities during wildfires is the need to rapidly identify residents in danger and evacuate them—a task complicated by the spread of the fire itself, which can render properties inaccessible. Using IoT deployments, fire services can better pinpoint the location and trajectory of burning blazes and broadcast emergency alerts via text message to individuals within a certain radius. Safety for firefighters can also be enhanced. In Australia, technologists are exploring the use of wearable sensors for firefighters, which could improve safety outcomes by improving firefighters’ real-time awareness of factors such as air quality and their colleagues’ whereabouts.
A common experience for industries that deploy IoT is realizing the opportunity to access valuable insights that go beyond the discrete problem the IoT was first introduced to solve, because of the wealth of data these solutions generate. Along these lines, temperature sensors deployed in forests are not only helping with prevention but are now fueling new research findings and insights about the overall behavior of wildfires that can be observed when these forests undergo controlled “hazard reduction” burns.
These insights illuminate how natural fires move and shift and how they are impacted by factors such as wind and ground vegetation—invaluable inputs for formulating strategies for preventing and responding to wildfires. And data produced by sensors after a fire passes also helps conservation agencies monitor forest recovery, animal migration and disease spread, supporting more focused restoration efforts.
Nascent deployments of IoT into forests have illuminated several of the key challenges and considerations. Due to the vastness and natural remoteness of many forests, sensors need to be placed in hard-to-reach areas for extended periods. This makes solutions that support extended battery life critical. Given the task they are performing, sensors must also be heat and weather resistant, and need to be positioned in ways that minimize the chance of being dislodged by curious wildlife.
Physically deploying sensors is also made tricky by the sprawl and density of most forests. Scientists have been exploring how to aerially drop low-cost sensors over vast forest areas, after which they can be remotely configured into a network. Using the same methods, once deployed, a network could be reconfigured should any sensors fail.
Connectivity is another key challenge. Where cellular connectivity is available, it’s an effective and robust solution. Technologies such as NB-IoT are ideal, as they are optimized for solutions involving robust, compact battery powered sensors that must run for long periods and need to send small data volumes across many kilometers—the precise scenario for most wildfire deployments.
When cellular IoT isn’t an option due to lack of infrastructure in remote regions, other IoT connectivity alternatives such as DECT NR+ private mesh networks, LoRaWAN, or satellite IoT are options.
A Worthy Investment
After Australia’s Black Summer fires, its official inquiry pointedly concluded that “investment in disaster prevention is more cost-effective than the huge cost of emergency response and rebuilding.” Put simply, prevention is better than cure—a slogan that resonates loudly for wildfire management as it does any realm of risk management.
On this basis alone, the business case for introducing IoT solutions into the forest stacks up. Even more so, after analysis by Switzerland-based Distrelec Group revealed the cost of fully deploying sensors to cover the forest density of Spain and garner their preventative benefits was just 0.0083% of the amount of money spent tackling forest fires. The group drew similar conclusions for a range of European nations.
With compelling figures like these and beneficial use cases like those outlined earlier, IoT is now also being considered for an even wider set of emergency scenarios. As it promises for wildfires, solutions involving sensors, wireless connectivity data analytics and AI might improve preparedness and response outcomes for a range of natural disasters including hurricanes, earthquakes, floods, tsunamis and monsoons.
Sadly, like wildfires, many of these events are set to increase in prevalence in the years ahead due to humans’ impact on the climate. But with careful deployment of IoT technology and analysis of the data it will provide, perhaps at least we can have greater warning of occurrence. Seeing the dragon cross the horizon is still scary but much better than feeling the heat of its breath.