FIRE INTELLIGENCE
USING CAMERAS AND ALLIED TECHNOLOGIES
BY RICK MCRAE
In wildfire management, operational officers constantly make decisions that reflect what the fire is doing, or what it might do. The decisions are based on intelligence products, although, unlike military personnel, fire managers rarely use the term intelligence products. The products use sets of information created from what has become in recent years a flood of data.
Fire services tend to be much better at data than at intel. But new intelligence products that better meet the needs of wildland fire managers are being used to monitor the landscape. Several product developers participated in the trade show at the IAWF’s 8th International Fire Behaviour and Fuels Conference in Hobart, Tasmania, in April, and I spoke with them about their technology.
HOW IT WORKS
Imagine there is a suggestion of a new fire ignition on fictional place we’ll call Bloggs Peak, near another fire that is in Day 3 of suppression operations. The duty officer may ask “Is there a new, confirmed fire near Bloggs Peak?” The incident controller might ask “Is there a new fire west of Charlie sector that could require new planning?” The operations officer might ask “Is my action plan still valid if there is a new fire nearby?” The fire behaviour analyst might ask “If there is fire on Bloggs Peak could it form vorticity-driven lateral spread (VLS) this shift?” Each question is asked to support the making of a decision that may become mission-criticalmission critical.
This diversity of potential questions is founded on access to a range of datasets. Data may be accessed from fire towers operators, cameras, aircraft, drones, air and ground observers, satellites, weather stations and radar, and weather services. The relevant datasets are processed, often (hopefully) following endorsed doctrine. The resulting information products are further analysed into intel products to support decision making.
The growth in technology in these processes may, if done correctly, bring a set of benefits including faster processing, different processing using artificial intelligence, a better record of the decision-making processes (for legal matters or for after action reviews, for example), and support of service improvement for future incidents.
My own research has shown how high-resolution fire cameras permitted the first ever detailed parameterisation of a major burn-over event caused by vorticity-driven lateral spread (see Understanding a Critical Fireground Safety Threat, page 28, and read research paper here: https://doi.org/10.3390/ fire9020079).
For some years now, a consortium linked to the University of California, San Diego, has run a data network in the hills and ridges of California. The High Performance Wireless Research and Education Network (HPWREN) provides a data backbone for use by a range of environmental monitoring projects and includes fire cameras.
Each site has four wide-angle, colour cameras , and each provides an image every minute, with metadata visible in the images. The photos are made available in near-real time and as a video every six hours.
The global research collaboration looking at fire thunderstorms (known as pyroCbs) has made good use of HPWREN data, and other new science on extreme wildfires has been validated using the system’s imagery.
WHAT’S HAPPENING
Across Australia, jurisdictions are progressing a range of fire camera initiatives, from simple enhancements to existing tower infrastructure with single cameras, through to fully integrated innovative systems.
The Australia and New Zealand Council for fire and emergency services (AFAC) supports information sharing, coordination, and sector-wide collaboration rather than prescribing or endorsing specific solutions. AFAC’s executive director for national capability, Sandy Whight, indicated in Hobart that national forums such as conferences and collaboration groups provide valuable opportunities to share insights, lessons learned, and emerging practices. A key focus is strengthening coordination across jurisdictions, including consideration of project development and funding challenges to maximise collective capability and avoid duplication of effort.
An example project is the work of a consortium of technology companies – AEM, FTS, and ESS Earth Sciences. ESS has decades of experience setting up portable automatic weather stations (PAWS) for fire services, land managers, and researchers. The company’s meteorology sales manager, Tim Cookes, said there can be challenges adding new capabilities and technology to the PAWS platforms, which are well suited to remote sites but therefore carry limits on communications networks and power supply.
However, Cookes said ESS is advancing the storage and downloads needs that are required if a camera is added to the system.
It is worth noting that field-deployed cameras are engineered to specifications for wavelength bands, fields of view, resolution, time to get each image to the user (latency), and repeat time. All this technology can lead to new hardware needs, which creates a delicate balancing act when cameras are in remote areas.
The Fire Foresight and Indicium Dynamics companies have tackled this challenge and to developed a purpose designed system that is being deployed for land managers Tasmania Parks and Wildfire Service and Sustainable Timber Tasmania, as well as for large-asset owners around the world. Indicium’s Elizabeth McDermott said the units carry a high-quality camera and Automatic Weather Station, and use AI for intelligence generation, and have a power generators and communications. The AI can access datasets on lightning detection and registered burns. Assessed new ignitions are passed to duty officers to verify.
A recent case study showed the camera system’s value. While two wildfires were being suppressed in eastern Tasmania, the system reported a new ignition in the same area. Without rapid detection, the new fire may have negated suppression actions for the original fires.
SEEING FIRE
It is worth knowing that there are three infra-red frequencies used for fire detection, especially with satellites. The 2.8µm band is close to visible and, similarly, it is easily blocked by smoke and dust. The 3.9µm band is often called the “fire band” because it works so well. It is prone to reflections and scattering.
The 10.8µm band is a thermal band – it indicates the temperature of something. It is useful for smoke plumes.
In Australia, communities tend to be in flat or undulating countryside where it is easy to see smoke or even flames, there is typically a useful road network, and locals are keen to catch a fire before it grows legs. However, there are challenges for fire detection in rugged or remote terrain, starting with the need for access to vantage points for installation and servicing. Good communication is also needed. Then there is efficacy.
Ken Hall, operations director for the Australian Capital Territory Rural Fire Service, said its camera system has multiple applications, including early detection and situational awareness for the incident management teams, call-taking centre, and field crews. ACT also has its emergency services starting to look at rain and storm cells as they develop.
OTHER SYSTEMS
Some fire monitoring capabilities are being developed using concepts from modern military systems. Many military aviation platforms use a range of technologies to achieve sensor integration. For future fires, a camera may see a candidate for new smoke, and the system may assess that a field crew nearby is best placed to investigate it. The crew’s sector leader would get a request to divert the crews to that task, and that request may be both automated and approved by the ops officer. The system may suggest a need to put more crews into the original sector while confirmation of the new smoke is awaited.
In addition to camera monitoring, new drone technology is being developed. Dan Williams from FireChain, an Australian-designed aerial firefighting platform, said water-carrying drones can reduce both the final area burned and the sensitivity to the effects of climate change. Other drones can be used for detection systems and for ignition work.
For a glimpse into the future, University of California San Diego’s ALERTCalifornia research network of more than 1,200 natural-hazard monitoring cameras shows what can be achieved. ALERTCalifornia works alongside HPWREN to run the continuously growing camera network used by emergency managers across the state. By collaborating with firefighting experts, the network incorporates the latest technologies, such as AI detection and near-infra-red cameras, making it an essential emergency response tool in California.
LOOKING AHEAD
As with anything done on the fireground, there are societal, political, and legal consequences should bad outcomes occur. Therefore, several factors must be considered in the design of fire monitoring systems.
Only science endorsed through agency doctrine should be used. New lessons must be endorsed in pre-season planning. Management must stay up to speed with the science.
Society has a mix of welcoming of and fear for artificial intelligence. Decision makers need to remain cognisant of this.
All fire services are increasingly using and providing out-of-area assistance, both interstate and international. Not all crews will be familiar with local systems.
If we collect more high-quality data, then who knows what new learning we might generate. While this benefit does not show up on annual budget sheets, it is of immense value if just one fire crew avoids a dangerous loss of situational awareness and ensuing consequences.
It is essential that field data be made available to the researchers studying fire behaviour. It is often the unexpected observations that are of critical importance.
Rick McRae served as a headquarters technical specialist in what evolved to become the ACT Emergency Services Agency in Canberra from 1989 until his recent retirement. He worked in business planning, arson investigation, multi-hazard risk assessment, as planning officer for major incidents, weather specialist, and as a research scientist focussing on extreme wildfires, and especially pyroCbs. McRae has conducted case studies, described new phenomena, and developed predictive tools. He maintains a website that aims to present operationally useful material on extreme wildfires –https://www.highfirerisk.com.au/. McRae is now an adjunct professor with the Bushfire Research Group at UNSW Canberra.
