.png)
Rapid and accurate fire intelligence empowers more effective wildfire mitigation
(Article printed in July 2023 issue of Asia Pacific Fire Magazine
Within an uncanny certainty, the team of Portuguese civil protection authority firefighters assured us there would be wildfires starting at 11am. They were right. Our trusty Cessna 182 was airborne before lunch, and by dinner we had flown over more than a dozen separate fires across several hundred kilometers of scenic, densely vegetated countryside.
Like many regions of the world, Portugal has an increasingly worrisome wildfire problem. It’s not that fire is new to the area or its population. It has, of course, shaped the landscape for millennia. Land use, land management and human activity are the wild cards in today’s equation. Add climate change, more extreme weather occurrences, compounding impacts of land-management shortcomings,
increased pressure in the urban interface and a general lack of adequate retention
measures to the mix, and it’s no surprise that essentially all fire-prone regions around the world are in such a critical state.
As we leapfrog from fire to fire on this warm and windy Tuesday evening, we maintain an elevation of about 4,000ft, an ideal height for the cutting-edge
multispectral imaging payload fixed to the wing of the aircraft. The imaging
system, about the size of a small suitcase, incorporates sensors in long-wave, mid-wave, short-wave, NIR and visible spectrums, synchronously operated by onboard AI software that captures imagery over the fires, stitches it together and creates accurate geolocated maps and fire perimeters that can be transferred to crews and responders on the ground in near real time .
‘This is a sophisticated solution to a complex problem,’ said Greg Denton, a wildfire intelligence expert from Overwatch Imaging. ‘Wildfire is dynamic, unpredictable, dangerous and increasingly more costly to address.’
Between 2011 and 2020, Portugal experienced an average of over 22,000 wildfires annually, as reported by the country’s National Authority for Civil Protection. In 2020 alone, Portugal spent approximately 258 million euros on fire-suppression efforts, accounting for a considerable portion of the national budget. This trend has been consistent over the past decade, with expenditures on firefighting resources continuously increasing due to the growing scale and complexity of wildfires.
‘Rapid and accurate fire intelligence makes an enormous difference in how efficiently and effectively crews can address wildfires,’ Denton continued. ‘If crews can catch fires when they’re small, and focus on the hottest, most critical sections on the front lines, we can increase efficiencies, reduce time to contain and, ultimately, achieve significant long-term cost reduction.’
For a sense of scale, compared to Portugal’s roughly 258 million Euros in 2020, the USA spent nearly $4.5 billion in wildfire suppression and response in 2021.
Overwatch Imaging, a small company located in the west coast state of Oregon, knows how bad fire seasons have become in recent years first hand. The company
was founded on developing airborne imaging technology to collect, analyze and deliver wildfire intelligence for crews and responders to utilize as quickly and effectively as possible. Overwatch systems are integrated into a wide variety of aircraft and are utilized by nearly all of the major agencies and providers across the country to map the nation’s major fires each year.
‘This is an exciting joining of forces between cutting-edge technologies like sensor fusion, artificial intelligence and edge computing, with good old-fashioned, boots-on-the-ground and shovels-in-the dirt hard work,’ Denton said. ‘Translating remote-sensing data into simple visual representations was once a complex task,
involving a professional interpretation and processing, sophisticated software and a significant amount of time. With modern systems like this, we can deliver more accurate and actionable intelligence from the seat of a small aircraft at 5,000ft, in
under five minutes.’
Portugal has taken measures to improve its wildfire prevention, detection and
response capabilities in recent years. This includes enhanced forest-management
practices, investment in firefighting resources, and international cooperation for
sharing knowledge and resources during major wildfire incidents. That’s how a small tech company from Oregon found itself in Lisbon, accompanied by agency leaders enthusiastically impressed with capabilities that could immediately impact their fire missions and their bottom line.
Anecdotally, if the average response and mop-up phase of a typical fire is four or five days, and crews can utilize accurate fire data to increase their efficiency and reduce their mop-up time by one day, that represents a significant saving in resources and expense and contributes to more rapid operational readiness for future response needs.
Back in the Cessna 182, we spot a column of grey and copper smoke ballooning from a suburb dangerously close to Lisbon. As we approach the fire overhead and initialize the sensor, the northern flank reaches a forested island of eucalyptus trees that sends a fiery ball of black smoke into the air. We radio our compatriots on the ground and are told a crew is approaching the fire from the southern flank and will be on-scene within 10 minutes – just enough time for us to capture, process and relay a map from above.
The system is controlled from a simple web-based interface and is largely autonomous once initiated by a pilot or operator. As the aircraft passes over a fire, the system captures the scene below in a steady stream of still images across a scan pattern and simultaneously across different spectral bands. The aircraft covers the fire in a ladder scan pattern, with the number of passes depending on the size of the fire and the speed and altitude of the aircraft. Typically, a few passes or less is all that’s needed. The system’s onboard processing then stitches the images together, runs AI analytics on the multispectral data to indicate hot spots
and fire perimeters, and processes the georeferenced maps into a size and format that can be relayed with low-bandwidth requirements to crews on the ground.
Back on the ground that evening, we learned that the fire flared up for about an
hour but was quickly subdued by sunset, cooler temperatures and calming winds.
Tomorrow’s mission will start at dawn, with a small hand crew looking for hotspots near the perimeter to extinguish before they can reignite and instigate another flare-up.
For this particular crew, that’s a tedious, physical and unprecise job of walking
accessible sections of the fire perimeter, looking for smoke, picking through dirt,
stumps and rocks, and literally feeling for heat. Perhaps in the near future, they’ll simply pull out their phones, pull up a fresh map from their intelligence support and follow coordinates to where their efforts are needed most.
More information on Overwatch Imaging's smart sensors and AI-driven software HERE.