Historically, fighting fires that erupt in the wilderness or natural, undeveloped areas involved detection by human eyes—professionals stationed at lookout towers—and surveillance via manned aircraft. Once a wildfire was detected, putting it out entailed hand tools, helicopters and planes dropping flame retardant and water, or fire trucks in more accessible areas. While these methods still largely comprise how wildfires are fought, new conditions due to climate change have necessitated more innovative technologies to protect lives and property.
Several scientific studies have confirmed that climate change is causing longer and more severe fire seasons in many parts of the world, including the U.S., Australia, the Amazon rainforest, Siberia, Canada, Portugal and Greece, among other places. Even formerly ice-covered Greenland has been getting hit with unprecedented wildfires in recent years. A group of 57 peer-reviewed studies since 2013 have shown that climate change is linked to increased wildfire activity through warmer temperatures, dryer conditions and stronger winds. The data indicates that fire seasons are now longer on 25 percent of the Earth’s surface where there’s vegetation.
According to David Pierce, a climate scientist with Scripps Institution of Oceanography, extremely hot days lead to the largest proportional increase in wildfire activity because of the drying effect such extreme weather has on a landscape. So, while winters have generally become wetter, the other three seasons are hotter and dryer. These hotter conditions also spur more lighting strikes, a prime ignition source, which can strike dried out or dead vegetation, thereby increasing the amount of fuel feeding wildfires. Another climate change trend—earlier snowmelt—is also linked to heightened wildfire seasons.
As the planet’s average temperature continues to climb, the threat posed by wildfires on ecosystems and human development will only worsen. Like a taste of what’s to come, the past three years have brought some of the most devastating wildfires in recorded history, taxing emergency response infrastructure around the world. In light of this, in April 2019, the first Wildfire Technology Innovation Summit convened at Sacramento State University in fire-stricken California. The program showcased many new technologies that are in various stages of deployment to help curb the devastating nature of wildfires. However, the event took place before the onset of the COVID-19 pandemic, which has delivered a blow to disaster relief capabilities.
In California, where unprecedented wildfires have engulfed the northern part of the state, fire-fighting crews are shorthanded due to inmate firefighters being released from prison early to help reduce the spread of COVID-19. For residents who have to decide whether to stay and defend their properties or evacuate to a shelter, the pandemic has made crowded shelter conditions riskier. Such complex conditions make it imperative to harness advanced technologies to help with preventing, detecting, and extinguishing wildfires and facilitating relief efforts for impacted communities.
“We’re going to see for the first time the approach and application of technologies that were previously not available to the public safety community,” said Jeff Johnson, fire chief and CEO of Western Fire Chiefs Association, speaking at the summit. “The first is the creation of a national public safety broadband network.…This one innovation puts more tools at our disposal on a go-forward basis than we’ve ever had.”
The network has taken shape as FirstNet, the First Responder Network Authority, and is a public-private partnership between the federal government and AT&T that was launched nationwide in 2018, and in 2019 achieved the distinction of having over 750,000 connections and performing faster than any commercial network. Beyond significantly increasing network coverage for first responders in remote locations or subterranean structures, FirstNet also has an app that allows personnel to view active events and incidents within a 50-mile radius.
Team Awareness Kit
As with the public safety broadband network, other technological innovations that have been developed for different purposes are being harnessed to fight wildfires as the need for them grows . The Team Awareness Kit (TAK) is an Android smartphone geospatial infrastructure app developed by the U.S. Air Force. Although originally designed for military purposes, it’s been impactful for first responders, including during several major hurricanes. More recently it was deployed during the Grizzly Creek Fire that’s being battled in Colorado. TAK facilitates situational awareness, navigation and data sharing. Among other uses, it can map new wildfire growth, pinpoint the location of firefighters, and deliver real-time intelligence on wildfire behavior. During wildfires such as the Grizzly Creek Fire, crew leaders have TAK installed on devices out in the field, enabling a high level of on-the-ground intelligence to aid critical decision-making.
Forecasting
As various types of sensors have improved and become less expensive, these devices have been a boon to the forecasting of wildfire events. Such use of sensors is ubiquitous in the general world of weather forecasting, but with the rise of the destructiveness of wildfires, these devices now play a major role in fire forecasting as well. Many factors go into accurately forecasting the likelihood and severity of a wildfire: temperature, humidity, wind speeds, the amount of both live and dead vegetation (fuels), the moisture content of those fuels, and the proximity of power lines and built infrastructure.
“Weather forecasting accuracy continues to improve with improved observations, physical understanding and artificial intelligence/machine learning,” stated Mary Glackin director of Science & Forecast Operations at IBM. “Applications of IoT and AI/ML will improve detection and prediction of fires.”
Glackin spoke of how IBM’s Global High Resolution Atmospheric Forecasting System is being applied to wildfire prediction using “an ensemble of models.” These models come from different countries and are optimized using AI to produce hourly forecasts. However, such forecasts aren’t typical. They’re designed to consider the complex interplay of weather, topography and fuels—to not only predict the likelihood of wildfire, but also the likelihood of whether factors such as wind gusts will exceed the models’ threshold. This approach can help first responders prepare for worst-case scenarios.
Modeling Fires
Advances in drones, infrared cameras, thermal imaging, corona cameras, satellite imagery and LiDAR have created parallel advances in the capability to create models of wildfire behavior. Many of these devices can be powered by renewable energy sources in remote locations that lack grid access. LiDAR sensors can be mounted on helicopters or placed at fixed points to map changes in vegetation over time. Data gathered from a wireless network of these devices can be utilized through machine learning techniques where models of landscapes and built infrastructure are created at high resolutions.
Funding from the National Science Foundation has led to the development of a prototype for such modeling that produces a knowledge base that can be analyzed and shared. The prototype, known as WIFIRE, uses supercomputers at the University of California San Diego to integrate continuously gathered data into powerful models that can predict what a fast-moving wildfire will do next. WIFIRE is currently being tested in the San Diego area, where several destructive fires have burned over the course of this century. The WIFIRE team is hoping that the prototype may become a global cyber infrastructure to manage the threat of fires everywhere.
Davis Sapsis, chief scientist for Cal Fire, said that one of the critical aspects of wildfire modeling is being able to predict where the smoke will go, due to the important feedback loops that exist between smoke and fire behavior. This entails not just modeling surface conditions but also what is taking place in the upper levels of the atmosphere. Accurate wildfire modeling doesn’t require running just a few models, but running scores of them. Cal Fire staff run models for 44,000 possible ignition points in California every day.
San Diego Gas & Electric company has a mobile app that allows personnel in the field to run simulations for a particular ignition point and determine the likelihood of containment if a fire were to erupt. The utility company also has a situational awareness platform that relies on numerous weather stations installed out in the field producing data used to generate more than 10 million virtual wildfires across its service territory. It also uses the Santa Ana Wildfire Threat Index, which assesses wildfire threat levels during very windy conditions and uses over 30 terabytes of data to produce a single graph.
Power Lines Pro
One common way that deadly wildfires ignite is when damaged utility power lines come in contact with vegetation. A line of software, Power Lines Pro (PLP), can reduce the risk of power lines causing or being affected by wildfires. The software has been used in Australia and elsewhere to identify damaged utility infrastructure and properly maintain vegetation. PLP has an analytics system that leverages a cloud-based big data platform, backed by predictive algorithms to simulate wildfires under various conditions, taking into account the blow out of conductors and forces acting upon utility poles. One key aspect of PLP functionality is that it can determine how clearance violations between utility lines and vegetation can increase as wind speeds increase. This can provide a more accurate prediction of asset failure under high-risk conditions and provide a risk assessment for utility companies and firefighters.