Systemic Drivers of Electric-Grid-Caused Catastrophic Wildfires: Implications for Resilience in the United States

Abstract

Wildfires are projected to increase in severity and frequency due to climate change, and the electric grid is both a cause of wildfires and is vulnerable to wildfires. Equipment from the electric grid accounts for 10% of fires burned in California and 3% of fires nationally. Recent catastrophic wildfires, such as the Lahaina Fire, Camp Fire, Marshall Fire, and Smokehouse Creek fires, were all started by electrical equipment and show how devastating these events can be because they threaten lives and structures. Vegetation structure, weather and winds, climate and vegetation response, land use, and human activities all impact the likelihood of severe wildfires. We explore the relationship between the built environment, electric grid infrastructure specifically, and its role in causing catastrophic wildfires to find lessons learned for increasing resilience. Electric grid utility companies currently employ multiple methods to mitigate fire, including (1) early detection, (2) grid hardening, (3) vegetation management, and (4) pre-emptive shutoffs. Utility companies need to consider the conditions for wildfire and the impact that each mitigation strategy has on drivers of wildfire behavior, as a single solution will not be adequate. Utility companies need to work with stakeholders to develop a holistic strategy to reduce ignition likelihood and spread likelihood to reduce catastrophic wildfires and improve resiliency.

1. Introduction

Wildfires are one of the greatest challenges of the Anthropocene. Human behavior and the built environment can create the spark for wildfires, which in turn threaten watershed health and the lives and homes of people living within the area. The increasing number of wildfires linked to the electric grid has become a significant concern in the United States [1] and globally [2]. The infrastructure that makes up the electric grid provides a crucial power supply to communities across the United States. It is vulnerable to damage from natural hazards, such as wildfires, and can also cause catastrophic wildfires. Electrical equipment can provide the initial spark that can spread into a wildfire through flashover events [3]. The combination of extreme drought conditions, high wind conditions, and high fuel loading from a century of wildfire suppression have created the perfect storm where one spark can lead to devastating consequences for lives, property, energy security, and ecosystems. In recent years, fire seasons have gotten longer [4] and have increased in magnitude and acres burned [5]. Wildfires are expected to increase in frequency and intensity with climate change [6,7]. In 2000–2015, the western United States experienced on average nine additional days per year of high fire potential. While conditions that allow for wildfires to occur are increasing, there has been an increase in electric-grid-caused wildfires. While wildfire is a natural process and necessary to certain ecological processes, extreme fire behavior can impact water quality form flooding after a fire, and other environmental processes. The electric grid infrastructure can serve as a catalyst for extreme fire behavior, when weather and biophysical conditions facilitate extreme fire behavior.

The intricate relationship between electric grids and wildfires necessitates a comprehensive review of how infrastructure failures contribute to these disasters. High winds can cause power lines to arc or vegetation to come into contact with electrical equipment, sparking fires. Once ignited, the combination of dry conditions, abundant fuel, and rapid spread can lead to large-scale blazes [8]. These catastrophic wildfires present a challenge to environmental safety, utility management, and community livelihoods. The impacts on communities are profound, resulting in loss of life, property destruction, widespread evacuations, and prolonged power outages that can cripple regional economies [9,10,11,12]. The financial impacts of these wildfires on utilities are also substantial, with costs ranging from billions of dollars due to infrastructure repairs, loss of service, and increased insurance premiums. For instance, the Camp Fire in 2018 led to Pacific Gas and Electric (PG&E) (Oakland, CA, USA) declaring bankruptcy due to liabilities exceeding $30 billion [13].

At least 10% of the total wildfires that happened in California from 2016 to 2020 were caused by power lines [14], accounting for 19% of acres burned. These incidents demonstrate the vulnerability of the power lines to causing ignitions and exacerbating wildfire damages. Adequate planning, regular vegetation management, and undergrounding lines are essential to reduce wildfire risks significantly [15].

Given these severe consequences, there is a critical need to understand the vulnerabilities associated with electric grids to develop effective mitigation and planning tools. Advanced technologies, such as machine learning for predictive analytics and enhanced vegetation management strategies, are essential for preventing and controlling wildfires [16,17].

This review aims to comprehensively explore the conditions for electric grid ignition with wildfire behavior spread conditions to clearly outline the unique circumstances of electric-grid-caused fires and contributing factors. This paper addresses the impacts of human disruptions on Earth’s natural systems through four case studies of electric grid-caused fires. Each case study explores the biophysical drivers and conditions of the ignition start, emphasizing the need for effective preventative measures. We explore mitigation recommendations for securing a more reliable and resilient electrical grid to improve human and environmental resilience to wildfire.

2. Review of Catastrophic Wildfires Caused by Electric Equipment

Wildfires started by electric grid equipment can be catastrophic, destroying thousands of structures and taking lives. Utilities whose equipment has been found to start the fires are often liable for damages. Utility equipment provides the spark and ignition of the fire, but it is the environmental conditions around the power equipment that lead to rapid fire spread. Vegetation conditions are a huge driver of wildfire rate of spread, with low live and dead fuel moisture and high density of combustible fuels from dead trees contributing to high rates of spread. Climatic conditions also impact the wildfire rate of spread, with high winds fanning the fire and drought and unusual heat contributing to the condition of the fuel. Dense surface fuels also play a part in rapidly spreading wildfires, and in Figure 1, they are shown to be present across all large electric-grid-caused wildfires.

Among the most recent grid-caused fires (Figure 1), high winds, low relative humidity, and high temperatures were expectedly common drivers. The presence of dried-out grasses created fuel bed continuity. There are also key differences in each of these destructive fires that suggest fire weather will become less predictable and more severe amidst a changing climate. Both the 2021 Marshall Fire in Colorado and the 2018 Camp Fire in California (each the most destructive fire in their respective state’s histories) saw relatively strong precipitation during the winter and spring, generating fuel that was later dried out by the heat and extremely dry conditions in the latter half of the year. Wet springs can increase the volume of surface vegetation, but when weather abruptly shifts between extremes, these can dry out and act as fine fuels that increase fire intensity.

Extreme weather, such as record high temperatures in February combined with high winds, drove the Smokehouse Creek Fire to become the largest in the state’s history. During the Marshall Fire in 2021, extremely high winds with gusts up to 115 mph resulted in a catastrophic urban fire.

We filter to grid-caused fires at least 1 acre large because these fires are more consequential and more reliably corroborated across different historic fire databases. Grid-caused wildfires occur more frequently and burn more acres each year (Figure 2). The InFORM Fire Occurrence Data Records tabulate 950,000 wildland fire incidents from 1992 to the present, representing an official record of all known fire occurrences in the U.S. These data were carefully aligned (Appendix A) with the USFS Spatial Wildfire Occurrence Dataset, which represents a cleaned, error-checked, and standardized set of wildfire occurrence records from 1992–2020 [18]. We find a spike in the frequency of grid-caused fires after 2019, with a corresponding increase in burned acreage.

3. Electric Grid Wildfire Ignition Causes

On a national scale, grid-caused fires are rare, accounting for 3% of wildfires nationally, as shown in Figure 3a. As shown in Figure 3b, 61% of grid-caused fires were less than one acre, and 93% were less than 100 acres. The <1% of grid-caused fires over 10 k acres comprise 79% of the total burned acreage since 1992. Grid-caused fires burned 4.15 million acres, representing 4.3% of acreage burned due to fires with a recorded cause and 2.4% of all acres burned since 1992. A total of 98% of the 5072 grid-caused fires from the National Wildfire Coordinating Group dataset were due to electric transmission/distribution systems, with the remaining 2% attributed to power generation causes.

Power grid components can ignite a wildfire due to device malfunctioning, downed energized power lines coming into contact with vegetation, trees, and branches coming into contact with energized lines, lines slapping together and generating a spark, and dense smoke creating an arc. Electrical arcing occurs when electrical current jumps across a gap between conductors. When insulators (which prevent electrical current from flowing to unintended areas) are damaged, the electricity can arc to metal structures or the ground, causing sparks that may ignite fires. In hot weather conditions, power lines can also sag due to the expansion of metal conductors, increasing the likelihood of contact with trees. High winds can cause damage to infrastructures associated with the electric grid by causing power lines to sway or come into contact with vegetation or from crossarms breaking, leading to a downed conductor line.

Utility poles, overhead transmission lines, and associated grid components often go across forested regions, and their interaction with existing vegetation poses a significant fire hazard to the surrounding areas. Deferred maintenance of grid components such as transformers and switchgear can increase the likelihood of equipment causing a spark. Corrosion from the elements can cause faulty connectors and weakened bolts, which can also lead to downed lines.

Reclosers and transformers mounted on poles can explode or overheat, igniting the pole itself or surrounding vegetation. Utility poles are often made of wood, and they are pressure treated with preservatives (e.g., Pentachlorophenol, or penta; Chromated Copper Arsenate, or CCA; Copper Naphthenate, or CuN; 4,5-Dichloro-2-N-Octyl-4-Isothiazolin-3-One, or DCOI; Creosote; and Ammoniacal Copper Zinc Arsenate, or ACZA) to increase the longevity and reduce the likelihoods of decaying or insect infestation. They can catch fire if an insulator fails, allowing electrical current to arc directly onto the pole.

Tripping commands in transmission lines can be caused by phase-locked loops (PLL) and DC-side dynamics for utility-scale PV power plants [19]. In [20], the causes of transmission line failures are divided into two segments: (1) the elastic extension of either the conductors or surrounding objects (such as tree limbs) causing electrical contact and arcing, and (2) failures under high strain conditions affecting other grid components (conductors, poles) or surrounding objects (trees). In [21], it is shown that the steel core mechanical strength of aluminum conductors steel-reinforced (ACSR) power conductors can rapidly fall if they obtain a temperature of more than 500 °C. This results in a permanent elongation of the line and decreases the clearance from the underlying vegetation.

One of the largest recent wildfires in Texas, known as the “Smokehouse Creek Fire”, happened in 2024 and burned more than one million acres. Downed power lines with broken utility poles were identified as the key cause of this wildfire, and at least 11,000 people were left without electricity [22].

A transformer is used to change the voltage levels in the utility grid and contain oil or other cooling fluids. They generate a lot of heat during operation, and if this heat is not dissipated properly, it can cause the insulation and oil inside the transformer to break down, leading to a fire [23]. They can also fail due to age, overloading, or internal faults. If overheating or a short circuit occurs, the equipment may rupture, potentially releasing hot oil that could ignite, leading to fires or explosions that may spread to nearby areas [24,25]. The intense heat from a transformer failure can cause molten metal or other materials to fly off, which can ignite vegetation, nearby structures, or other equipment. Though the number of transformers that catch fire varies by location and environment, a 2010 study claimed that at least 730 transformers explode in the U.S. annually [26]. In addition to that, 2.4% to 4% of all transformers can be expected to cause a fire during the average 40-year service life [27].

Other substation and grid components can also pose ignition risks due to malfunctions and various types of faults. Switchgear is commonly located at electric grid substations and serves to control, protect, and isolate electrical equipment. It comprises a variety of equipment, including breakers, fuses, current transformers (CT), potential transformers (PT), and lightning arrestors. It can cause electrical faults that lead to short circuits or sparks during faulted conditions and ignite nearby materials. Arcing faults are the most severe in such equipment and can lead to intense heat generation.

Raptors and other wildlife present near the transmission line zone can become electrocuted and increase the risk of ignition from the grid equipment. The number of wildfires caused by bird species was similar to species killed by electrocution [28]. According to the U.S. Fish and Wildlife Service’s National Forensics Laboratory (Ashland, OH, USA) analysis of 417 electrocuted raptors from the year 2000 to 2015, nearly 80% were bald or golden eagles [29].

4. Wildfire Behavior Drivers

Wildfires behave differently because of the multitude of factors that govern the physics of combustion and the pre-condition of the landscape to be able to spread a fire. Wildfire behavior, such as the rate of spread, flame length, intensity, and direction of the fire, is influenced by environmental factors such as weather conditions, topography, and the type and condition of available fuels [30,31].

Weather conditions such as wind speed, direction, air temperature, and relative humidity play a part in how quickly vegetation dries and is likely to burn, as well as how far flames and embers travel [32]. Two primary meteorological factors that significantly affect wildfires are wind patterns and temperature combined with humidity [33]. Wind is key for both the fire triangle and fire behavior triangle as it brings a fresh oxygen supply for the fire combustion process, moves debris, and can “push” the fire forward, spreading the fire to more fuel sources. Strong winds can increase a fire’s spread rate and enable it to jump natural barriers like rivers or roads. Winds also influence flame length and intensity, potentially escalating a fire’s destructiveness. Winds can vary greatly due to larger-scale weather systems or local topography [34]. Winds are also a primary reason for downed conductor lines and flashover events, with lines coming into contact with one another. High winds also accelerate the ignition caused by campfires, foreign objects, or other human activities by faster spreads of heat and flames. Moreover, they can lift burning embers and carry them over long distances to ignite fire far from the key ignition point and expose new fuel sources to an existing fire. Additionally, the upward wind flow can preheat the surrounding vegetation well ahead of the nearest fire arrival inside the hilly tracts.

Temperature and humidity affect fuel moisture content, a critical determinant of flammability. Higher temperatures reduce fuel moisture, making vegetation more susceptible to ignition and combustion, while higher humidity can dampen fuels and slow fire spread. Higher temperatures also cause the power lines to expand, making them sag closer to the ground and surrounding vegetation.

Topographic features such as elevation, slope, and aspect are also significant drivers of fire behavior [35]. Higher elevations with cooler temperatures and increased humidity typically slow fire progression, while lower elevations with drier conditions can experience more intense fire activity. Terrain features like slopes and valleys create unique local conditions affecting how fires burn. Steeper slopes generally promote faster fire spread due to the preheating of unburned fuels upslope, which is driven by radiant heat and convective currents. Topography also influences wind patterns and microclimates. Landscape features like valleys and ridges can funnel winds, altering their speed and direction and complicating fire behavior predictions. Additionally, topographical variations can create localized weather patterns, such as thermal or katabatic winds, which significantly impact fire dynamics. Power lines are often situated over mountainous terrain with steep slopes, thick vegetation, and poor access to fire suppression vehicles.

Vegetation conditions such as structure, density, height, and composition across the landscape all contribute to the rate of spread and flame length. Surface fuels are another factor that determines wildfire behavior. Characteristics like fuel bed depth, continuity, and moisture dictate how the fire spreads. Vegetation types vary across elevations, influencing the availability and type of fuel for combustion. Different vegetation types contribute varying amounts of fuel and have distinct flammability properties, impacting how wildfires ignite, spread, and burn. Forests, grasslands, and shrublands each present unique challenges due to their fuel loads and combustibility. Grasslands with fine fuels can support fast-moving fires, while dense forests may lead to higher intensity fires. The distribution and connectivity of vegetation patches also affect fuel continuity and fire movement. Physical characteristics such as fuel load, particle size, bulk density, and type further influence fire intensity, burn severity, and spread [36,37,38]. Fuel characteristics are dynamic, and changes can affect fire behavior, such as moisture content fluctuations after precipitation or long-term climate shifts altering fuel bed morphology and composition. Fuel accumulation and decomposition over time also affect fire dynamics.

Vegetation flammability largely depends on its moisture level, determining how easily it ignites and sustains combustion. Higher moisture content in vegetation generally makes it less flammable, requiring more energy for ignition, while dry fuels ignite more readily and contribute to rapid fire spread. There are two primary types of fuel moisture: live fuel moisture and dead fuel moisture. Live fuel moisture refers to the water content within living plants, varying with seasons, plant health, and environmental conditions. Dead fuel moisture pertains to moisture levels in dead plant material like fallen leaves, twigs, and branches, which are highly responsive to atmospheric humidity. Understanding and predicting changes in fuel moisture content are essential for wildfire risk assessments and management strategies.

5. Wildfire Case Studies

We have reviewed several case studies of electric-grid-caused wildfires in depth. Each case study identifies feedbacks between ecology, land use management, and risks that energy infrastructure plays in exacerbating the problem of catastrophic wildfire. Three of the fires occurred in California, where the century of wildfire suppression practices has left the landscape with dense forests and understories [39]. Combined with pine bark beetle infestations and drought, California has experienced high tree mortality rates, creating readily available fuel to burn if there is a spark. Forest fires are not the only catastrophic fires. Grassland fires can also be catastrophic when occurring in high wind conditions, which is the case of the Marshall Fire in Colorado.

5.1. Camp Fire

The Camp Fire cost $16.65 billion in damages and suppression costs and burned 153,336 acres. A total of 85 civilians and 5 firefighters died in the blaze, 52,000 people were evacuated, 18,661 structures were destroyed, and 675 structures were damaged. The fire started from the failure of a badly maintained steel hook holding up a PG&E high voltage line on 8 November 2018, and within an hour, it overtook the town of Paradise. It consumed 70,000 acres in one day. It was fueled by hot, dry, sustained, and gusting to 50 mph “Jarbo Winds”, a katabatic wind off the Great Basin to the east, which interacted with the terrain and picked up speed as it funneled through Feather Canyon. In between the fire ignition location and Paradise, there was a lot of dense fuel, attributed to a century of wildfire exclusion. Fires that start far away from a town that has time to develop and expand, fueled by winds, make the fire very difficult to fight once it is in town [40]. A history of fire exclusion has changed the composition and vertical structure of the forest. Larger trees (>61 cm diameter at breast height) have declined, whereas smaller trees (<30 cm) have increased [41], and at mid-elevation, ponderosa pine and mixed conifer forests in California likely had about 60 trees per acre 100 years ago, as opposed 165 and 170 trees per acre today [42].

Eighty-five percent of structures were destroyed in Paradise. Buildings ignited not from the flaming front but from showers of burning debris carried by the wind. While the flaming front of the fire burned hot and traveled quickly, embers traveled up to six miles and created spot fires. The vegetation in between structures and plots of land created dozens of small fires, making it difficult for firefighters to fight effectively.

The area had experienced a wet spring, increasing the amount of fine surface fuels, both in ground cover and volume. Surface fuel continuity is often what allows the fire to spread post-ignition before it transitions into a canopy fire. Then, the area experienced an unusually dry fall, receiving 1/7th inch when it normally gets 5 inches of autumn rainfall, leaving dense surface fuels parched by more than 200 days without significant precipitation and ready to burn. Wind events right before the ignition decreased humidity from 23% to 10% and 1000 h fuels were estimated to have 5% dead fuel moisture (USFS).

5.2. Dixie Fire

The Dixie Fire burned 963,309 acres in northeastern CA in 2021. The fire was ignited when a Douglas fir fell into the PG&E powerline. Two fuses blew, but one stayed on, keeping the line energized and creating an electrical fault that slowly ignited fuels on the ground. The Dixie Fire was the largest single (i.e., non-complex) wildfire and was intensified because of a megadrought; 2021 was also the hottest summer on CA’s record. Megadroughts are severe and persistent multi-year droughts. Researchers used hydrologic modeling and analyzed tree rings from 1528 trees across the western U.S. to reconstruct summer soil moisture to demonstrate that the southwestern U.S. has been in the second driest megadrought since 800 CE [43].

California’s Sierra Nevada also experienced an unprecedented insect-induced tree mortality event from the bark beetle. When a fire burns through an area impacted by beetle kill, the wildfire will be more severe. In [44], found that the basal area of the tree killed by fire and canopy torching was higher where wildfire spread into beetle-kill areas. Drought, climate change, and western bark beetle infestations have additive effects due to compromised tree defenses from the drought and bark beetle population dynamics, where more larvae survive the winter due to warming. Each degree (°C) increase in temperature may increase the number of ponderosa pine killed by upwards of 35–40% °C⁻¹ [45].

The 2021 Dixie Fire in California is, to date, the largest single-source fire (not part of a complex) in the state’s history. A hazard tree came into contact with energized conductors and created a high-impedance fault. Aside from high winds and dry weather, fire growth was catastrophic as much of Plumas County fell into Exceptional Drought (D4), increasing tree mortality and, thus, dead fuels. The 2019 Aerial Detection Surveys in California (the most recent available prior to the Dixie Fire) showed that following the severe drought from 2012–2016, tree mortality was exceptionally high in Plumas National Forest at 149k acres (13% of Plumas NF).

5.3. Caldor Fire

The Caldor Fire burned 221,835 acres, destroyed 1003 structures, and 50,000 residents were evacuated. It could have burned much more. The US Forest Service had already treated 3292 acres within the Caldor Fire perimeter in the South Lake Tahoe area with two types of treatments: prescribed burn, which removes surface fuels, woody litter, and duff layers; and mechanical thinning, which removes trees. Low-intensity fires, such as those caused by a prescribed burn or cultural burning, reduce fuel loading across the landscape, slowing large wildfires and diminishing their severity. Data from over 800 wildfires show that when a wildfire encounters a treated area, fire intensity is reduced by 86% [46]. The mechanical treatment spaced trees out, reducing basal density and separating trees farther than a typical flame length away, preventing fuel continuity and slowing the fire [47].

Flame lengths within treated areas of the fire perimeter were 20 ft or less as opposed to 100 to 150 feet in untreated areas, according to the Forest Service. Canopy fires are incredibly hard to suppress; the heat generated from the fire can make retardant and water drops ineffective. When the fire spread into the treated areas, it transitioned from a canopy fire to a surface fire, spreading along the forest floor. This slowed the rate of spread as well as decreased the heat, or intensity, of the fire, allowing firefighters to control the burn more effectively. According to USFS Fire Management Officer Kyle Jacobson, the treatments enabled over 600 homes in the community to be spared.

The Caples Ecological Restoration Project area is an example of when treated landscapes become wildfire resilient. The Caple’s Creek restoration area had been treated with prescribed burn before the Caldor Fire. The Caldor Fire started to burn into it and then went around it, as demonstrated by the island in Figure 4 for the Caldor Fire.

5.4. Marshall Fire

The Marshall Fire started on 30 December 2021, a time of year when there is frequently snow on the ground. Above-average precipitation caused grasslands to have 3–4 ft tall grass, followed by dry and unseasonably warm temperatures, setting the stage for the area just south of Boulder, CO, USA. The area is not heavily forested, defined mostly by grasslands and isolated tree islands. The spark that started the Marshall Fire was actually two separate ignitions, one from an Xcel Energy unmoored power line and the other by embers from a trash burn on private property. The fire was pushed by 50 mph winds, gusting to 100 mph, from a mountain wave wind event, which occurs when winds flow downhill along the terrain, accelerating toward the base, and is strongest where foothills meet the plains. Mountain wave wind events occur when large-scale winds are perpendicular to large mountain ranges, as is the case in the Rocky Mountain front range in CO (National Weather Service). The Severe Fire Danger Index peaked at the 99th percentile in response to the mountain wave wind event, indicating fuel moisture was very low at the time of ignition. The fuel bed would have consisted of senesced (dormant) or dead grass material built up from the season’s growth. Fine fuels are categorized into the 1-h fuel class, meaning they respond quickly to weather events that zap the grasses of moisture, making the fuel burn more easily [48].

The fire caused $2 billion in damages, burned over 6000 acres, destroyed 1084 residential structures and 7 commercial buildings, and damaged 149 residential structures and 30 commercial buildings. While whole neighborhoods were damaged, destruction of structures was sporadic from house to house, mainly driven by ember transport. This fire highlights that grassland-dominated landscapes can cause destructive fires and how a changing climate is extending the fire season. Alignment of high wind events and an initial spark can be disastrous.

6. Prevention

There are a multitude of approaches to preventing wildfires. Mitigation efforts focus on two areas: ignition prevention and fire spread prevention. Tools such as AI-enabled cameras installed along electric grid assets can provide early warning that a fire has started. Wildfire suppression professionals have a greater chance of success in controlling a fire at the early stages of a fire, making early detection a critical component to preventing large, catastrophic fires.

6.1. Early Detection

Early detection of wildfires through the use of AI-enabled cameras [49] can improve the success of wildfire suppression efforts. If winds are high, it is critical to control a wildfire within the first few minutes of ignition, and early detection enhances the success of identifying and responding to the fire. Ongoing tactical information about the fire perimeter helps inform evacuation orders, keeping communities safe, as well as inform operations tactics in wildfire suppression. Satellites with thermal bands such as MODIS, GOES, and ECOSTRESS can all be used to monitor heat signatures from wildfires. There are several new commercial thermal satellite offerings, such as OroraTech (Munich, Germany), that provide thermal data at a finer resolution (100 m).

Early warning of fire activity can also be used to alter maintenance protocols or de-energize lines when fire weather is extreme. Data from multi-spectral satellite imagery, LiDAR, and weather models can inform how likely a fire is to spread if there is an ignition. Advances in using machine learning algorithms to detect vegetation conditions such as live and dead fuel moisture.

6.2. Preventive Measures

Electric utilities can utilize weather monitoring systems to detect conditions conducive to wildfires and pre-emptively shut down power in high-risk areas. Public Safety Power Shutoffs (PSPS) help mitigate the risk of power lines igniting wildfires under hazardous conditions. Additionally, the installation of remote-controlled sectionalizing devices can isolate portions of circuits during PSPS events, therefore reducing the number of affected customers. PSPS has proven to be an effective measure to reduce the risk of wildfires. SCE (Rosemead, CA, USA) reports that in 2018–2022, they found approximately 90 incidents of wind-related damage on lines that could potentially have caused an ignition but were prevented from doing so because the line was de-energized [50] (SCE WMP 2023).

Uncleared vegetation near substations or power lines also poses a hazard. Therefore, it is a key parameter in the decision process of activating any utility activities as a precaution to prevent a future fire from happening and an existing fire from spreading rapidly. For transmission lines, utilities must follow minimum vegetation clearance distances based on the voltage rating of the line as laid out by the North American Electric Reliability Corporation [51]. California Public Utilities Commission (CPUC) General Order 95: Rule 35 [https://docs.cpuc.ca.gov/published/Graphics/13352.PDF (accessed on 13 November 2024)]. specifies minimum clearances for lines in California. Communication and electric circuits, energized at 750 volts or less, including their service drops, should be kept clear of limbs and foliage. Based on the field knowledge, if any circuit energized at 750 volts or less shows strain from tree contact, the condition shall be corrected by slacking or rearranging the line, trimming the tree, or placing mechanical protection on the conductor(s). The distance for the trolley contact and supply conductors may be reduced for conductors of less than 60,000 volts when protected from abrasion and grounding by contact with a tree. For the supply conductors that range between 22.5 kV and 105 kV, the minimum clearance shall be 18 inches. This minimum clearance is flexible, up to 5%, in case of heavy line loading in certain regions. It is also suggested to clear all flammable vegetation from within a minimum of 30 feet around any home and other structures and make sure the maximum mature tree height is 10 feet away from the closest power line.

PSPS is another preventive action, and initiating a PSPS depends on a number of parameters such as fire risk scores, humidity, fuel moisture, wind speed thresholds, etc.

Table 1. Wind speed thresholds to activate PSPS for major utilities in California.

Utility	Sustained Wind Speed (km/h)	Wind Gusts (km/h)
PG&E	>30	48–464
SCE	49–64	74–93
SCE	59–78	74–99

shows the wind speed thresholds for the major utilities in California to activate a PSPS event [52,53]. Wind speed thresholds are separated depending on whether the winds are sustained for more than 2 min or the maximum gust at any reading.

6.3. Vegetation Management

The management of vegetation under powerlines is an essential concern in modern infrastructure planning and maintenance, crucial for ensuring the safety and reliability of power supply systems. In recent years, increasing incidences of power outages and fires linked to unmanaged vegetation have elevated the significance of this practice, urging utility providers and policymakers alike to explore effective solutions [54,55].

The historical development of vegetation management practices under powerlines reflects a gradual evolution driven by increasing demand for reliable energy supply and growing awareness of environmental concerns. In the early stages, vegetation management was predominantly reactive, responding to outages caused by overgrown trees and plants impeding powerline functionality. The primary approach was manual or mechanical clearing, often regarded as sufficient without considering broader ecological impacts [56]. As urbanization and energy demands accelerated in the 20th century, the need for systematic and preventive vegetation management became more apparent. Navigating challenges such as labor intensity, cost, and limited efficiency of manual methods led to the exploration of alternative techniques.

Key techniques for mechanical vegetation management include cutting, mowing, pruning, or trimming, often executed with various types of machinery like chainsaws, mowers, and chippers. The primary advantage of mechanical methods is their ability to quickly clear large areas, providing an immediate solution to prevent vegetation from coming into contact with powerlines. However, these approaches are labor-intensive and can be costly due to the continual need for clearing and maintenance, especially in rapidly growing vegetation areas. Additionally, frequent mechanical disturbance can stimulate regrowth [57], which may paradoxically increase the need for more intensive and frequent management activities over time.

6.4. Infrastructure Upgrades

The deployments of the grid-hardening techniques are intended to make the power grid less susceptible to wildfire, as well as optimize the operation of the power grid while minimizing the impact of potential wildfires. There are multiple grid components that have the potential to ignite fires, either independently or in conjunction with unmanaged vegetation. Utility poles typically support power lines, transformers, and other essential grid equipment. Hence, utilizing appropriate grid-hardening measures can benefit utility planners in both the short and long term. Different risk-spend efficiency metrics are proposed by various utilities to evaluate [58] and identify the grid-hardening strategy suited best for them.

Power lines are among the critical infrastructure components that are directly exposed to fire-prone weather conditions in open environments. Hence, undergrounding lines can significantly help prevent wildfires. Burying the overhead lines will reduce the probability of ignition caused by unmanaged vegetation and ignition in fire-prone weather. It also eliminates ignition risk from high winds, tree branches, and debris, which can spark wildfire. One of the most devastating fires in the history of California is the so-called “Thomas fire”, which occurred in 2017 due to the convergence of power lines during high wind conditions. It burned 281,893 acres and left more than 85,000 customers without power. Similarly, the “2018 Camp Fire” in California was also sparked by a power line and caused an estimated $9.3 billion in residential property damage [59]. These multiple wildfire occurrences during the 2017 and 2018 fire seasons led to the Pacific Gas and Electricity (PG&E) utility filing for bankruptcy and acknowledged involuntary manslaughter charges [60,61]. These wildfire occurrences led to a massive power line undergrounding project for the 10,000-mile plan by PG&E. As of the end of 2023, more than 600 miles of undergrounding have been completed since they announced the undergrounding program in 2021. Though this approach is very costly and takes too long to get implemented, it contributes to the long-term prevention of wildfires.

A covered conductor refers to an extra layer over the power lines that significantly reduces the probability of arcing and sparking from contacting vegetation. Recently, replacing wooden poles with composite poles and fire-resistant materials such as steel and fiberglass have also gained attention to reduce the impact of wildfires. This approach will make the poles more resilient to high temperatures and fire compared to wood, reducing the likelihood of ignition if a fire reaches the grid infrastructure. They can also reduce the risk of ignitions from equipment mounted on the pole. Enhancing the continuous monitoring of power lines can be effective in improving grid resiliency [62]. This includes utilizing drones, helicopters, or advanced sensors to regularly inspect power lines and equipment for damage or wear that could lead to failures. Wildfire smoke can increase the line capacitance and interrupt the normal operation of a line-shunt reactor. A holistic mitigation strategy is proposed by analyzing the waveform from simulation and field measurements [63]. A fast transmission line wildfire risk calculation method is proposed in [64] to provide a holistic understanding of the combined risk and rake mitigation approaches based on that. A normalized risk score is similarly developed to evaluate wildfire-grid risk [65].

To reduce the risk of ignition caused by wildlife, varmint-proofing can be used to isolate underground grid equipment from animals [66]. For overhead systems, devices such as squirrel guards, insulator covers, cut-out covers, raptor-safe designs, and line covers are utilized to prevent wildlife from entering critical zones.

Another grid infrastructure change to reduce wildfire ignitions would be Rapid Earth Fault Current Limiters (REFCL) [67]. This technology has been widely deployed in parts of Europe, Asia, and Australia for wildfire prevention and involves changing the grounding of the system. For resonant grounded systems, an arc suppression coil, or Petersen coil, is used to limit any fault current that happens during a single-line-to-ground fault. Additionally, an inverter can be included to provide a residual current compensator or ground fault neutralizer to reduce the fault current even more. In order to convert existing 4-wire distribution feeders in the United States, this topology requires substation upgrades, re-design of the protection and voltage regulation systems, and modifying customer service transformer connections. In addition to reducing the energy released during a ground fault, the protection system can be much more sensitive with the ability to detect even very high-impedance ground faults.

The implementation of microgrids and other localized energy generation systems (e.g., solar PV, battery storage, vehicle to grid (V2G)) that can operate independently from the main grid is an effective strategy for supplying power to areas that are isolated due to their high wildfire risk. These systems reduce reliance on long-distance transmission lines that cross high-risk wildfire areas. Anzo Borrego Springs microgrid was deployed by San Diego Gas and Electric, covering an area of 60 miles of power lines, and the Laguna wastewater treatment plant microgrid was developed by PG&E in Santa Rosa [68]. Microgrids can limit public safety power shutoffs up to 2–3% of the annual energy demand, whereas eradicating them will significantly increase the energy cost 20. In [69], it is shown that a system with installed microgrids can maintain operation under severe fire-related conditions without scheduled or unplanned outages.

Different solutions that are applicable to improve grid resiliency can also play key roles in grid hardening for wildfire prevention. Distribution network reconfiguration to reduce the power losses (https://www.nature.com/articles/s41598-024-73928-1 (accessed on 19 October 2024), dynamic simulation-based studies to locate the vulnerable areas, advanced algorithms to identify the tipping points that can lead to cascading failure can be very effective to employ grid-hardening initiatives. In addition to these techniques, optimizing the vegetation and PSPS can also reduce the number of affected customers and provide more resilient grid operation.

6.5. Protection and Fault-Clearing Systems

In addition to infrastructure upgrades, new protection systems and fault-clearing technologies can be used to reduce the risk of wildfire ignition. The probability of ignition is direction proportional to the duration of the arc, so detecting and clearing faults quickly provides a significant improvement in wildfire reduction. There are many ways to increase the speed of the protection devices. Fault interrupters can detect a fault and isolate the faulted line to avoid sparking. Advanced fault detection systems such as arc fault detection technologies should be installed to automatically cut power when a fault occurs. Moreover, the utilization of the fault and load transmitter (FLT) and fault and load receiver (FLR) system enables precise fault location through accurate fault indication and load monitoring [70]. This system facilitates faster fault identification and informed switching decisions.

Traveling-wave-based fault locators can be used to enhance fault detection capabilities within a single tower span. Additionally, transformers can be equipped with internal fault devices (IFD) to reduce the ignition risk, while capacitor banks equipped with a neutral current sensor and SCADA communication systems allow real-time data transmission to operation centers, which will further reduce the wildfire risk [71]. The replacement of regular reclosers with advanced technologies, along with upgrading relays, and proper coordination of different relays, switches, and breakers to shorten fault durations and limit energy flow to faults, can mitigate potential ignitions. Feeder and recloser relays capable of detecting high-impedance faults can also be selectively implemented to improve fault detection [72].

Electric utilities employ additional mitigation strategies, including Open Phase Detection (OPD), which can sense power line separations and disconnect power before the line reaches the ground, preventing potential ignitions. High-impedance relays use protective elements to reduce the risk of relatively low-energy fault conditions that are typically undetectable by conventional protection schemes.

Moreover, fast-acting fuses can interrupt electrical current quickly and reduce the risk of ignitions when there is an electrical fault, such as when a tree falls on a power line during high winds. Faster grid protection settings, known as fast-curve settings, on high fire risk areas during elevated fire conditions can result in a quicker reduction in fault energy, decreasing the ignition risk. Upgrading relay hardware to expand the number of circuits with these protection settings and deploying transformers with biodegradable fluid that have higher flash and fire points can further enhance wildfire resilience.

7. Conclusions

Mitigation measures to prevent a fire from igniting and spreading are necessary for public safety. Steps can be taken to increase electric grid resilience to wildfire while also reducing risk exposure to people living in the wildland-urban interface, strengthening sustainability of the built and natural environment interactions. The execution of vegetation management under powerlines involves substantial economic and operational challenges. These challenges stem from the complexity of rectifying immediate hazards while balancing long-term operational and financial efficiency. Vegetation management demands significant resource allocation in terms of manpower, technology, and ongoing maintenance activities, posing a continuous financial burden on utility companies. However, the integration of advanced technologies, such as remote sensing coupled with predictive modeling, offers potential solutions to some of these challenges. Automation and technological innovation can streamline operations over large areas, reducing labor costs and enhancing precision in vegetation assessment and management.

The grid-hardening measures mentioned above are already being implemented by several electric utilities across the United States. Regular asset monitoring using more data-driven techniques along with enhanced grid technologies—such as optimizing grid topology, sectionalizing the grid based on fire weather predictions, advanced power flow control, and dynamic line rating—can help reduce the probability of wildfire ignition. Conducting detailed wind studies, upgrading fuses and switches at critical locations, installing fire-safe arresters, performing protection coordination analysis, deploying automated reclosers and community-based microgrids, and implementing self-healing restoration strategies would also be highly beneficial to mitigate the wildfire impacts. Using advanced detection and machine learning approaches to understand wildfire risk, conditions for ignition and spread is critical for the proactive management of wildfires. It is important that utilities understand and anticipate the feedbacks between weather, ecosystem response, and land management impacts on wildfire behavior. Communication and strategic partnering across stakeholders is critical to reducing both ecosystem and built environment vulnerability to wildfire.