By Robert Hutchens

– Executive Summary –

The next significant advance in wildland firefighting tactics may be using swarms of unmanned aerial vehicles (UAVs) to perform fire attacks on wildland fires before they burn out of control. Swarms are autonomous teams of UAVs deployed under specific parameters to complete a mission. Swarms of UAVs are permitted to make autonomous and cooperative decisions within the mission parameters while under supervisory control by a single operator. This ability contrasts with current UAV operations, which require a pilot for each aircraft. The utilization of UAV swarms for fire suppression operations has the potential to increase safety for firefighters and the public, provide early detection and suppression of wildland fires, and free up traditional aerial assets for deployment on critical fires.

A.              Motivation for Research

Wildland fires pose a clear and present danger to American homeland security because of the impact on natural resources, decreased economic stability for people who make their living from the forest products industry, loss of property and homes, and potential loss of life. The danger that large-scale wildland fires represent is so severe that fire managers must attempt to suppress them using all available tactics, including direct and indirect fire-attack methods. These techniques require placing firefighters in harm’s way to slow or stop the fire progression. Between 2010 and 2019, 134 firefighters were killed in the line of duty while battling wildland fires.[1] The National Interagency Fire Center reports that over the past 10 years, “there were an average of 62,693 wildfires annually and an average of 7.5 million acres impacted annually.”[2] However, the availability of resources to combat these fires remained the same over those years.[3] With trends of increased global warming and longer fire seasons, it is critical to foster innovation in the arena of wildland firefighting to ensure maximal impact by firefighting forces while simultaneously increasing the safety of firefighters.

1.               Augmenting Firefighting Tactics with Swarm Technology

The concept of using single UAVs to assist in intelligence collection on wildland fires is familiar territory. For most large wildland fires, single UAVs may be used to observe the location and movement of the fire or monitor a fire that has escaped control lines. The intelligence produced by UAVs becomes a critical factor in fire prediction and operational planning for the upcoming operational periods. However, UAVs have not been deployed for direct fire-suppression activities, nor have swarms of UAVs been deployed on wildland fires. Swarm technology is in its infancy and will need additional development to be a viable option to apply to suppression operations.

The term “UAV swarm” refers to multiple UAVs operated by one controller. A controlled swarm is a team of UAVs programmed for a specific mission, operated by a single controller. A semi-autonomous swarm is a team of UAVs assisted by a controller to launch and recover. The semi-autonomous swarm has specific mission parameters whereby it can identify a mission, decide which members will complete the mission, complete the mission, and recover to a base station while deconflicting among the members to ensure successful completion.[4]

The notion of using single UAVs for direct fire attacks has neither been significantly explored nor studied due to the necessity of placing a large suppression payload of water or retardant on the fire. However, with the potential of UAV swarm technology that allows many smaller payloads to be placed on station and the advancement of heavy-lift UAV systems that carry as much as 100 pounds, swarms of many UAVs may be effective in suppressing fires. Using a UAV swarm rather than one traditional aircraft could be the next dramatic advance in wildfire suppression. The technology to perform this job is being developed, yet it is neither practical nor scalable at this time.

2.               Wildland Fires Are Becoming More Prevalent and Destructive

Wildland fires must be viewed with a wide-angle lens as a homeland security issue. Many courses of action can be undertaken concurrently to decrease wildland fires’ frequency, severity, and intensity. It will be necessary not to focus on one solution as a panacea to the wildland fire problem but rather to embrace innovation and technology to meet the challenge of decreasing these destructive fires.

Wildland fires directly threaten U.S. homeland security due to large fires’ extreme economic, social, and emotional impact. The cost to combat these fires continues to increase in monetary expenditures and human costs, both in the lives of firefighters and citizens who have succumbed to wildland fires. The National Fire Protection Agency estimates that the annual cost to suppress wildland fires is now $1.6 billion annually.[5] Historically, there has been a designated “fire season”; however, large-loss wildland fires happen in all months of the year, causing some western states to abandon the notion of a fire season.[6]

As wildland fires continue to present challenges for firefighters and those living in the urban–wildland interface, exploring more efficient and cost-effective methods of combating these destructive fires becomes critical. One cannot escape the fact that global warming is changing the susceptibility of forests to burn more readily. These tinder-dry conditions have created situations in which multiple lives have been lost because the fire progressed at such a rapid rate. A clear example is the Camp Fire in California in 2018, where 85 people lost their lives, and 18,804 structures were lost, primarily in the first five hours of the fire.[7] A more recent example is the Marshall Fire outside of Denver, Colorado, on December 30, 2021, where two people lost their lives, and 991 structures were lost in just a few hours. This fire was fueled by dry conditions and winds of more than 90 miles per hour.[8] Exacerbating these fires is the fact that there are limited resources to combat them in terms of personnel, aircraft, and funds to hire additional assistance. Scarcity of resources has become a significant problem, especially during the heart of fire season when multiple fires are burning and expanding rapidly. During the last few fire seasons, there were times when there simply were not enough resources to commit to firefighting. Fire managers must make difficult decisions to write off sometimes thousands of acres or hundreds of homes when that happens.

3.               Global Warming and the Threat to the Wildland

What were previously believed to be unprecedented fires and fire behavior have become commonplace seemingly every summer. Global warming significantly impacts the wildland environment, creating hotter and dryer climates that perpetuate fire growth. Current and legacy policies of where, how, and when fires are suppressed have increased the fuel loading of American forests. It will take decades for forest management policies to catch up to the current fuel load in American forests.[9] Global warming and climate change initiatives may take decades to impact forests positively. According to the Insurance Information Institute, approximately 10.1 million acres burned in the United States in 2020, and U.S. homes at high or extreme risk of wildfire totaled 4.5 million.[10]

It is essential to acknowledge the negative impact of global warming and climate change on the wildland environment. The most significant effect global warming has on forests is drought and higher average daily temperatures. These higher temperatures have two significant effects on vegetation. The first is creating lower long-term fuel moistures (taking moisture out of the fuel). The second creates a vapor pressure deficit within the environment (taking the moisture out of the air), making the vegetation more susceptible to fire.[11] The drier the fuel, the more susceptible it is to ignition when exposed to heat. Both climate change effects work in conjunction to dry fuels and then keep them dry.

The second impact of global warming is the erratic weather patterns that climate change fosters, bringing large amounts of rain during the rainy season and hotter and drier conditions in the summer months.[12] The rain propagates the growth of light vegetation in the spring, increasing the overall fuel load. When the hotter, drier summer months arrive, the new growth is susceptible to drying. This vegetation is now “cured” and ready to carry a fire, especially when exposed to wind. The fuel volume increases for the next fire season if the vegetation does not burn that season. The weather patterns that global warming influences have an exponential impact on wildland fires by first increasing the fuel load and then severely drying the areas, making them more susceptible to fire.[13] Forest management policies of allowing heavy undergrowth to accumulate have created conditions within the forest environment that promote fires to burn hotter, faster, and more intensely.[14] The failure to act on global warming initiatives and little prescribed burning have created unhealthy forest ecosystems.

The ability to accurately define what is contributing to global warming and climate change specific to their effects on wildland fire will continue to attract public attention for years to come. The effects of global warming are on display every year with the large wildfires that burn in the western United States and other parts of the globe. These large fires will require fire managers to embrace technology and policies that can potentially limit the growth of future fires. Moritz notes that the national understanding of the role of wildfire in global warming is changing. He states that between 2003 and 2007, the question typically asked during large wildland fires was “Who is to blame here?” Conversely, now the question asked is, “Are these fires due to climate change?”[15] Perhaps recognizing that wildfires are an extreme result of global warming will influence policies that positively impact global climate change.

4.               Utilization of Aerial Assets in the Wildland Environment

With the threat of wildland fire exposing citizens to potential loss of life, high monetary costs for suppression, and the emerging potential from other fields of UAV study, exploring UAV-based fire suppression is warranted at this time. It is essential to understand the role of aircraft in fighting wildland fires. Using firefighting helicopters and fixed-wing airplanes has become an integral part of the current overall fire suppression plan, especially on large, fast-moving wildland fires. In general, fire managers use aircraft to leverage their speed and ability to quickly put out a fire before it expands to a significant fire. Rarely, if ever, are aircraft able to suppress a fire entirely. It ultimately takes the firefighters on the ground to fully contain wildland fires. Essentially, aircraft are used to “buy time” for other firefighting forces to be obtained and deployed.

However, using aircraft is an extremely costly method of fighting fires. The U.S. Forest Service (USFS) establishes contract rates for aircraft by type and ability. During the contracting period between 2018 and 2021, Type 1 helicopters (the most powerful and capable of dropping the most water) were contracted at rates between $4,000 and $8,000 per flight hour, depending on the aircraft type.[16] Costs for fixed-wing retardant-dropping aircraft can be between $7,100 and $13,500 per flight hour, not including the retardant cost.[17] The USFS spent “approximately $607 million on contract aircraft in 2018,” including rotor and fixed-wing aircraft.[18] In an era of tight budgets, one must consider the cost-effectiveness of using aircraft to fight fires. If using UAVs to suppress fires can prove less costly and as effective as—or even more effective than—traditional aerial assets, the utilization of UAVs should be explored in depth.

Employing both fixed- and rotary-wing aircraft to suppress fires has limitations, namely, the aircraft’s inability to fly in heavy smoke conditions, weather events, the darkness of night, limitations on pilot flight hours, a limited and specialized group of people who can pilot these aircraft, and the inaccessibility of aircraft during required maintenance. Flying aircraft in an uncontrolled environment close to an active wildland fire allows for no margin of error, and the results can be catastrophic. A significant number of firefighters have been killed in aircraft accidents while engaged in fire suppression operations. In a retrospective study, Butler, O’Connor, and Lincoln found that from 2000 to 2013, 78 deaths of wildland firefighters, or 26.2 percent, were related to aviation.[19] In such cases, highly skilled firefighters and pilots are lost, and the airframe they are flying is usually destroyed. While fire managers seek to minimize risk for firefighters and aircrews, unfortunately, placing men and women in harm’s way to slow or stop a wildland fire remains an effective tactic.

UAVs and UAV swarms may replace crewed aerial assets for fire identification and direct fire suppression in the coming years. The utilization of UAVs and UAV swarms allows for a higher operational pace due to their ability to fly at night and in multiple conditions that crewed aircraft cannot. Additionally, UAVs can theoretically fly missions for a full 24 hours and are constrained only by required maintenance, rest requirements for pilots, and fire conditions that are so severe they do not allow for flight operations. Committing to UAVs and UAV swarms could potentially reduce the number of firefighter deaths from aerial accidents and keep fires smaller and more manageable.

B.              Research Question 

How can emerging UAV swarm technology be implemented as a method of fire attack in the wildland setting?

C.              Methodology and Research Design

This thesis explores how UAV swarms could be a novel approach to direct fire-attack operations by suppressing or retarding a fire from growing beyond the incipient phase, allowing conventional firefighting forces time to arrive, control, and suppress the fire before it grows to a significant wildland fire. A proof-of-concept methodology was used by participating in actual UAV swarm flight missions with the Advanced Robotics Systems Engineering Laboratory (ARSENL) and the Consortium for Robotics and Unmanned Systems Education and Research groups at the Naval Postgraduate School. These flights informed the validity of scenario studies used to develop missions with practical applications in the wildland setting. Practical testing informed the possibility of deployment and highlighted issues that must be addressed before full deployment. Additionally, we analyzed current utilization models of commercial UAV swarm users to define their applications in relation to their potential integration into wildland firefighting operations.

To illustrate the potential value of UAV swarms for direct fire attacks, this thesis utilized a theoretical fire in the Tillamook State Forest in western Oregon. The fire is “attacked” using both a 50-aircraft UAV swarm and a single-engine, fixed-wing air tanker. The single-engine air tanker (SEAT) is a conventional aircraft traditionally used to attack the fire. A comparative analysis was performed between a UAV swarm and conventional SEAT aircraft attacking this fire. Innoslate 4 (V 4.5.1.0), the Department of Defense Architectural Framework–compliant systems engineering modeling software, was employed to model the flight scenarios for the UAV swarm and the air tanker using accurate flight, refuel, and reload cycle times. For comparison, the flights were constrained by one normal fuel cycle of the SEAT. The fire attack methods were compared through the lenses of the total firefighting product delivered and comparative cost per gallon dropped for both resources. This information led to conclusions about the viability of using UAV swarms as firefighting assets either in place of or as adjunct to conventional aerial firefighting equipment. Finally, current aerial wildland firefighting methods were analyzed quantitatively, focusing explicitly on the cost of using conventional fixed-wing aircraft versus swarm technology against the theoretical fire. Through comparative analysis and practical testing, a theoretical deployment model was built, addressing concerns of stakeholders and the potential of performing UAV swarm operations as a commercial operation.

D.              Research Objectives

• Examine the feasibility of utilizing UAV swarms for wildland firefighting.
• Identify how UAV swarms may be applied to wildland firefighting.
• Identify current UAV swarm applications specific to the military that could be adapted to wildland firefighting.

E.              Chapter Overview

This thesis is heavily influenced by the personal experience of firefighters in the urban–wildland interface. Chapter I discussed the motivation of this research and its criticality. Wildland firefighting is a highly specialized field that brings its own language and jargon, so Chapter II defines background information about wildland firefighting and introduces terms relevant to understanding this thesis. Chapter III addresses academic literature relevant to the thesis topic and the research question. In Chapter IV, methodologies for comparative analysis are described and explained. As there is limited research on swarm technology, Chapter IV delineates given assumptions and constraints within the analysis. It also relies heavily on the Innoslate 4 systems engineering software to graphically illustrate the process of fighting the fire. The results of our experimentation and modeling are presented at the end of the chapter. Chapter V discusses the results of the comparative analysis between a traditional fixed-wing air tanker and a flight of 50 UAVs in a swarm. From this comparative analysis, recommendations were devised for UAV swarm implementation in the future. Chapter VI reviews the conclusions and recommends follow-on research to further develop the topic of UAV swarm use in the wildland environment. The research should enhance the viability of using UAV swarms for direct fire attack. While the practical application of this technology and research may be years in the future, these findings should provide a place for future researchers to start.

---

[1] Department of Homeland Security et al., Firefighter Fatalities in the United States in 2019 (Emmitsburg, MD: U.S. Fire Administration, 2020), 9, https://www.nfpa.org/-/media/Files/News-and-Research/Fire-statistics-and-reports/Emergency-responders/FFF-2020.ashx.

[2] Katie Hoover and Laura A. Hanson, Wildfire Statistics, CRS Report No. IF10244, version 49 (Washington, DC: Congressional Research Service, January 2021), 1, https://crsreports.congress.gov/‌product/pdf/IF/IF10244/49.

[3] Hoover and Hanson, 1.

[4] Moulay A. Akhloufi, Nicolás A. Castro, and Andy Couturier, “UAVs for Wildland Fires,” in Proceedings of SPIE Defense and Security, ed. Michael C. Dudzik and Jennifer C. Ricklin, vol. 10643, Autonomous Systems: Sensors, Vehicles, Security, and the Internet of Everything (Bellingham, WA: International Society for Optics and Photonics, 2018), M9, https://doi.org/10.1117/12.2304834.

[5] Jesse Roman, Angelo Verzoni, and Scott Sutherland, “The Wildfire Crisis: Greetings from the 2020 Wildfire Season,” NFPA Journal, November/December 2020, http://www.nfpa.org/News-and-Research/‌Publications-and-media/NFPA-Journal/2020/November-December-2020/Features/Wildfire.

[6] “2021 Incident Archive,” Cal Fire, accessed January 14, 2022, https://www.fire.ca.gov/incidents/‌2021/.

[7] Alexander Maranghides et al., A Case Study of the Camp Fire—Fire Progression Timeline, NIST Technical Note 2135 (Gaithersburg, MD: National Institute of Standards and Technology, 2021), 3, https://doi.org/10.6028/NIST.TN.2135.

[8] Kyle Cooke, “‘A Horrific Event’: 991 Structures Destroyed, Three Missing in Marshall Fire,” Rocky Mountain PBS, December 30, 2021, https://www.rmpbs.org/blogs/news/superior-louisville-grass-fire-colorado-evacuations/.

[9] Lee E. Frelich and Peter B. Reich, “Will Environmental Changes Reinforce the Impact of Global Warming on the Prairie–Forest Border of Central North America?,” Frontiers in Ecology and the Environment 8, no. 7 (September 2010): 371, https://doi.org/10.1890/080191.

[10] Cal Fire, “Facts + Statistics: Wildfires,” Insurance Information Institute, accessed October 26, 2021, https://www.iii.org/fact-statistic/facts-statistics-wildfires.

[11] Robinson Meyer, “The Most Important Number for the West’s Hideous Fire Season,” Atlantic, September 15, 2020, https://www.theatlantic.com/science/archive/2020/09/most-important-number-for-the-wests-wildfires-california/616359/.

[12] A. Park Williams et al., “Observed Impacts of Anthropogenic Climate Change on Wildfire in California,” Earth’s Future 7, no. 8 (2019): 894, https://doi.org/10.1029/2019EF001210.

[13] “Climate Change Indicators: Wildfires,” Environmental Protection Agency, accessed March 9, 2022, https://www.epa.gov/climate-indicators/climate-change-indicators-wildfires.

[14] Courtney A. Schultz, Matthew P. Thompson, and Sarah M. McCaffrey, “Forest Service Fire Management and the Elusiveness of Change,” Fire Ecology 15, no. 1 (2019): 2, https://doi.org/10.1186/‌s42408-019-0028-x.

[15] Max A. Moritz, “Wildfires Ignite Debate on Global Warming,” Nature 487, no. 7407 (2012): 273, https://doi.org/10.1038/487273a.

[16] “Helicopter Services Hourly Flight Rates, Fuel Consumption, and Weight Reduction Chart,” U.S. Forest Service, February 16, 2019, https://www.fs.usda.gov/sites/default/files/media_wysiwyg/flt_chrt_‌awarded_2018-2021.pdf.

[17] “FAQs–Austin Airtanker Base,” Texas A&M Forest Service, accessed January 13, 2022, https://tfsweb.tamu.edu/uploadedFiles/TFSMain/Preparing_for_Wildfires/Contact_Us(4)/FAQs-updated%20July%2019.pdf.

[18] U.S. Forest Service, Aviation Annual Report 2020 (Washington, DC: U.S. Forest Service, 2021), 7, https://www.fs.usda.gov/sites/default/files/2021-06/CY2020_USFSAviationReport_Final_1.pdf.

[19] Corey R. Butler, Mary B. O’Connor, and Jennifer M. Lincoln, “Aviation-Related Wildland Firefighter Fatalities—United States, 2000–2013,” Morbidity & Mortality Weekly Report 64, no. 29 (2015): 793, https://doi.org/10.15585/mmwr.mm6429a4.