Wildfire, Smoke, and Power Shutoffs

Publication Date: 2025

Click to View the Research Brief

Abstract

This study investigated how residents perceived and responded to co-occurring and compounding wildfire hazards—wildfire, smoke, and power shutoffs—in Northern California. Using a multi-site, qualitative case study approach, we conducted 45 semi-structured interviews with residents and key informants across three counties with high wildfire risk, extreme smoke exposure, and frequent power shutoffs. Participants perceived wildfire as the most significant threat, followed by smoke and power shutoffs. While participants’ response efficacy—their belief in the effectiveness of an action to mitigate threat—was high for wildfire and power disruption, participants expressed lower response efficacy in managing smoke events, risks associated with generator use, and limited access to clean water among residents reliant on well pumps. These findings highlight the need for public health agencies to address public misconceptions about wildfire smoke, expand outreach on safe generator use, and adopt policies that provide equitable access to air filtration systems and backup power. Furthermore, public health agencies need to better communicate universal smoke risks and support residents’ mental health during prolonged smoke events. We have identified several opportunities for improving generator safety education and providing backup power to vulnerable populations. Finally, the study underscores the importance of further research to document and understand compounding effects in other regions and climate contexts.

Introduction

People across the Western United States (“the West”) are increasingly exposed to wildfire, smoke, and wildfire-related power outages, which occur when utilities shut down the supply of electricity to reduce the chance that downed power lines spark a wildfire. Wildfire risk in the West is driven by a variety of factors, including climate change, dense vegetation resulting from past land management policies, and an increasing number of people living in the wildland urban interface (WUI) (Abatzoglou & Williams, 2016¹; Radeloff et al., 2018²; Wu et al., 2018³). Wildfire threatens the lives, homes, and property of people living in fire-prone regions of the West; while wildfire smoke and power shutoffs—like Public Safety Power Shutoffs—threaten populations living both near and far from actively burning areas (Abatzoglou et al., 2020⁴; Childs et al., 2022⁵; Murphy, 2021⁶). Wildfire smoke and power shutoffs pose serious risks to public health and can disrupt daily life (Heaney et al., 2022⁷; Mildenberger et al., 2022; Reid et al., 2016⁸; Zanocco et al., 2021⁹). Also, power shutoffs may hinder wildfire smoke coping strategies, such as running an air filter, thereby exacerbating health issues for those who are medically reliant on electronic devices (Wong-Parodi, 2022¹⁰). We argue that wildfire risk, smoke, and power shutoffs are not just co-occurring hazards, they are also compounding because they “can amplify [overall] effects, especially if they occur over relatively short time periods and overlap geographically” (Cutter, 2018, p. 19¹¹).

Individuals and communities in the West are faced with the challenge of responding and adapting to co-occurring wildfire hazards and their potentially compounding effects, however little is known about the processes or factors influencing individual responses. The study of compounding climate hazards is growing (AghaKouchak et al., 2020¹²; Cutter, 2018; Raymond et al., 2020¹³), yet most research on responses to wildfire risk, smoke, and power disruption in the West has focused on studies of single hazards. To our knowledge, risk perception and response to co-occurring wildfire hazards and the compounding effects of these hazards have not been investigated empirically. Additional research is needed to deepen our understanding of how individuals respond to the compounding effects of co-occurring wildfire hazards. Such research can contribute to new theories and frameworks explaining the psychosocial factors that underpin risk perception and response and describe the ways that co-occurring hazards may lead to compounding effects, subsequent challenges for adaptation, and even maladaptation—actions or policies intended to reduce vulnerability to hazards but that inadvertently increase it or create new risks (Cutter, 2018). Thus, this study sought to better understand how individuals perceive threats and their ability to respond to co-occurring wildfire hazards, explore what compounding effects (if any) individuals experience from co-occurring wildfire hazards, and describe how individuals respond to the effects of both single hazards and compounding effects.

Literature Review

Models of Risk Perception and Response to Climate Hazards

Behavioral scientists across many disciplines are interested in understanding the motivating psychosocial factors of behavioral response to climate hazards (Fischer, 2019¹⁴; Fischer & Denny, 2024¹⁵; Hamilton et al., 2018¹⁶; van Valkengoed & Steg, 2019¹⁷; Wilson et al., 2020¹⁸). Commonly used theoretical models include the Protection Motivation Theory (Rogers & Prentice-Dunn, 1997¹⁹), applied to public health and environmental contexts; the Protective Action Decision Model (PADM), applied in natural hazard and disaster contexts (Lindell & Perry, 2012²⁰); and the Model of Private Proactive Adaptation to Climate Change, applied to actions specific to climate change (Grothmann & Patt, 2005²¹). These models suggest that response is underpinned by threat perceptions—an individual’s assessment of the severity and likelihood of harm from a hazard—and efficacy perceptions, which encompass both self-efficacy (confidence in one’s ability to take protective action) and response efficacy (belief in the effectiveness of the action to mitigate the threat). Additionally, these models highlight the role of social norms, which can be categorized as descriptive norms (perceptions of what most people do), injunctive norms (perceptions of what most people approve or disapprove of), and subjective norms (perceptions of social pressure from significant others to perform or not perform a behavior). Other critical factors include trust in information sources, as well as objective measures of adaptive capacity, such as the availability of time and financial resources (Grothmann & Patt, 2005; Lindell & Perry, 2012; Rogers & Prentice-Dunn, 1997). Our study draws from all three behavioral models discussed and will contribute to this emerging area of scholarship by examining how individuals perceive threat and response efficacy to co-occurring hazards of wildfire, smoke, and power disruption.

Empirical Studies of Behavioral Response and Adaptation to Wildfire Hazards

We also draw from empirical studies investigating response and adaptation to the specific hazards of wildfire, wildfire smoke, and power disruption.

Wildfire Risk

Literature on protective responses related to wildfire risk in forested landscapes is well established. Several studies have examined how perceived risk, efficacy, responsibility, negative affect, and social interactions may influence adaptation (Hamilton et al., 2018). Common behaviors that have been documented in research include structural mitigation (e.g., installing fire-proof roofs and siding, cleaning roofs of debris), reducing flammable vegetation around homes (e.g., cutting back trees and branches, planting “fire-safe” vegetation types, removing dry leaf litter) (Hamilton et al., 2018), and when a fire is actively burning, decisions to evacuate or not (McCaffrey et al., 2018²²). Using the PADM, McCaffrey et al. (2018) found that perceived response efficacy (of evacuating or staying to defend), risk attitudes, and official warnings informed individual responses in areas recently experiencing wildfire. Another study found that WUI residents who experienced a wildfire that affects their community but does not burn their home or property (i.e., a “near-miss” wildfire) reported lower risk perceptions, but greater intentions to take fire risk mitigation action (Larsen et al., 2021²³).

Wildfire Smoke

Exposure to fine particulate matter measuring less than 2.5 micrometers (less than 2.5µm) in diameter (PM 2.5), a key pollutant in wildfire smoke, has been linked to poor short-term health outcomes including coughing, shortness of breath, and headaches (Balmes, 2018²⁴), as well as adverse long-term impacts on respiratory (Reid et al., 2016), reproductive (Heft-Neal et al., 2022²⁵), and cardiovascular health (Cascio, 2018²⁶). Levels of PM 2.5 over 35.5 µg/m3 is rated as “Unhealthy for Sensitive Groups” in the Air Quality Index, what we define as extreme smoke exposure. Children, older adults (65 years or more), and individuals with existing chronic cardiovascular and respiratory disease are particularly susceptible to harm from wildfire smoke (Heaney et al., 2022; Reid et al., 2016).

In the emerging literature on adaptation to wildfire smoke, studies have shown that individuals recognize the threat of smoke to their health and will engage in protective actions (Hano et al., 2020²⁷; Rappold et al., 2019²⁸; Santana et al., 2021²⁹). Frequently reported response behaviors include staying indoors, searching for and tracking information on air quality, wearing a mask, temporarily leaving areas with high smoke levels, and purchasing an air purifier (Burke et al., 2022³⁰; Hano et al., 2020; Kolbe & Gilchrist, 2009³¹; Marfori et al., 2020³²; Rappold et al., 2019; Santana et al., 2021). Behaviors requiring more planning and investment, such as purchasing a home air sensor (Burke et al., 2022; Holm et al., 2021³³) or moving out of the area (Berlin Rubin & Wong-Parodi, 2022³⁴) were less common. A recent study found a positive association between threat and response efficacy perceptions related to wildfire smoke and behavioral intentions, as well as a correlation between descriptive social norms and intended response (Santana et al., 2024³⁵).

Power Shutoffs

Only a few recent studies have examined how individuals understand and respond to the threat of utility-initiated power shutoffs, known as PSPS, when implemented by the dominant electric utility in Northern California, Pacific Gas & Electric (PG&E). Northern California has been heavily impacted by PSPS implemented by PG&E, which has cited the need to prevent fire ignition in areas of high fire risk, particularly during hot and windy conditions (Murphy, 2021). Overall support for PSPS appears to be linked to wildfire smoke exposure. Wong-Parodi (2022) found that greater levels of worry about wildfire smoke predicted support for PSPS in California, and Zanocco et al., (2023³⁶) found that individuals in areas with poorer air quality during wildfires were more likely to support PSPS as a wildfire mitigation measure. Also, personal experience with PSPS has been found to be positively associated with worry about storing perishable food, communication, completing household tasks and work, caring for family, controlling indoor temperatures, and using life support and medical devices (Mildenberger et al., 2022³⁷; Wong-Parodi, 2020³⁸). In survey-based studies, individuals with higher vulnerability—defined as living in a household (a) with someone suffering from a chronic respiratory or cardiovascular health condition, (b) with at least one child under the age of 5, (c) with at least one adult over the age of 65, and/or (d) with income 30% or less than the annual median income for their area and household size reported greater PSPS concern (Wong-Parodi, 2020), and individuals with existing respiratory illness reported greater threat perceptions (Wong-Parodi, 2022). During PSPS events, one study found that travel patterns did not change, but respondents engaged in more preparedness activities (e.g., preparing electronics and back-up power sources; finding information about events; buying key supplies like gas, food, and water; and updating emergency kits and plans) (Wong et al., 2022³⁹). Similarly, Mildenberger and colleagues (2022) found that most people stayed at home during PSPS events and spent an average of $327 dollars preparing for the outage (buying food, gas, flashlights, batteries). Exposure to PSPS has also been positively associated with intentions to adopt solar or energy storage (Zanocco et al., 2021) and intentions to buy fossil fuel generators (Mildenberger et al., 2022).

Research Questions

Scholars have acknowledged that climate hazards are co-occurring which can lead to compounding effects (AghaKouchak et al., 2020; Cutter, 2018; Raymond et al., 2020), and it is therefore essential to understand how the “perceived complexity” of overlapping hazards affects behavioral response and adaptation (Fischer et al., 2022, 1024⁴⁰). However, most empirical studies have focused on single wildfire hazards, with only a few studies addressing the interactions of wildfire risk, smoke, and power disruption. To our knowledge, no studies have documented perceptions of risk and response to co-occurring wildfire hazards and their compounding effects. Our study fills this gap by contributing to the empirical literature on behavioral response to wildfire, smoke, and PSPS. We asked the following research questions:

• How do individuals perceive threat and response efficacy related to co-occurring wildfire hazards?
• In what ways do individuals respond to co-occurring wildfire hazards?
• What compounding effects (if any) do individuals experience as a result of co-occurring wildfire hazards?

Research Design

To investigate these questions, we conducted a multi-site, qualitative case study of three counties in Northern California with (a) high wildfire risk, (b) high exposure to extreme wildfire smoke, and (c) high past and potential future exposure to power shutoffs. Northern California represents an extreme case (Yin, 2017⁴¹), as it has been at the epicenter of recent catastrophic wildfires in the West (Wang et al., 2020⁴²) and continues to be a region at risk of future wildland fire, smoke exposure, and power disruption (Goss et al., 2020⁴³). A qualitative, interview-based study allowed us to learn directly from residents of fire-prone, smoke-exposed, and power-disrupted regions about their perceptions of co-occurring wildfire hazards (e.g., perceptions of threat and response efficacy) and reports of past responses to such hazards and their compounding effects. A semi-structured interview protocol was used to allow for the pursuit of new lines of inquiry and follow-up questions, and to provide participants with the opportunity to recount their experiences in their own words (Creswell & Clark, 2017⁴⁴).

Community Engagement

Additionally, we partnered with the Fire Adapted Communities Learning Network (FAC Net) on the research design, participant recruitment, and ongoing and future dissemination of findings. FAC Net supports and connects member organizations focused on wildfire resilience by providing community-oriented resources and tools to reduce wildfire risk (FAC Net, 2023).

Our partnership with FAC Net provided connections with network members and affiliate organizations actively working in our study counties. FAC Net staff introduced us to individuals in these groups, who we then contacted for assistance in identifying key informants and residents for interviews. We also reached out to individuals from Rural Voices for Conservation Coalition and local universities before beginning our field work to share ideas and avoid duplicating efforts. We searched the Social Science Extreme Events Research database and did not identify any other researchers actively working in our three case study communities (Peek et al., 2024⁴⁵). Finally, we met regularly with FAC Net and local community partners to build trust and sustain communication about the project.

Study Sites and Access

To select study counties, we used secondary data on wildfire risk, wildfire smoke exposure, and past PSPS events in Northern California. To identify counties that met the first criteria and had high risk of direct wildfire impacts, we used the California Department of Forestry and Fire Protection (CAL FIRE) Fire Hazard Severity Zone maps, which classify lands based on their fuel loading, slope, fire weather, wind, and other factors (Office of the State Fire Marshal, n.d.⁴⁶). For our study, we selected counties that contained areas rated as high or very high fire hazard severity zones. Although we did not use data on recent fire events explicitly in our criteria, we acknowledge that many communities across Northern California have been affected by fire and evacuations in recent years, and we considered these events in the design of our interview protocol.

For the second criteria on wildfire smoke risk, we identified counties with high levels of extreme smoke exposure using data on daily local-level estimates of wildfire smoke PM 2.5 from 2006-2020 across a 10 by 10 km grid of California (Childs et al., 2022).Using this dataset from Childs and colleagues (2022), we identified the counties with the highest increase in the number of extreme smoke days using a two-step process. First, we calculated the average number and trend in extreme smoke days by comparing the average number of smoke days in the five-year periods 2006-2010 and 2016-2020. Second, we sorted all counties by their increase in extreme smoke days and selected those counties in the top quartile (top 25%) for analysis. Between 2016 and 2020, these selected counties experienced an average increase of 6-20 extreme smoke days per year compared to 2006-2010 (Childs et al., 2022). Finally, to identify PSPS-affected regions, we examined publicly available PG&E spatial data on past PSPS events and high likelihood of future PSPS (Pacific Gas & Electric, n.d.⁴⁷). We selected counties where five or more PSPS had been implemented as part of a wildfire mitigation strategy since 2019.

From this secondary data analysis, we identified a list of counties which met all three selection criteria. Next, with our partners at FAC Net, we narrowed the list to three counties based on the presence of practitioner networks and active community organizations providing services in the county related to wildfire recovery or risk mitigation. We did so because (a) we wanted to collaborate with these organizations on participant recruitment; (b) support from wildfire practitioner networks allowed us to identify key informants and connect to residents; and (c) involving community partners ensured that our research findings were relevant and actionable for the counties. The three counties we selected for the study were:

1. Lake County has a population of 68,163, according to the 2020 census (U.S. Census Bureau, n.d.-a[^uscbdna]). It is in the north-central area of California surrounding Clear Lake and the county seat is Lakeport. The county contains moderate, high, and very high fire hazard severity zones. The region has experienced five PSPS events, and Lake County was among the counties in the top quartile for increases in extreme smoke days in extreme smoke days. Over the past decade, Lake County has experienced several significant wildfires, including the Rocky Fire (2015), Valley Fire (2015), and Mendocino Complex Fire (2018).
2. Tehama County has population of 65,829, according to the 2020 census (U.S. Census Bureau, n.d.-b[^uscbdnb]). It is in the north-eastern area of California, bisected by the Sacramento River, and the county seat is Red Bluff, located between Sacramento and Redding along Interstate-5. The county contains moderate, high, and very high fire hazard severity zones. The region has experienced nine PSPS events, and Tehama County was among the counties in the top quartile for increases in extreme smoke days. Tehama County has been impacted by recent major wildfires, notably the August Complex Fire (2020), which burned over 1 million acres across multiple counties.
3. Tuolumne County has a population of 55,620, according to the 2020 census (U.S. Census Bureau, n.d.-c[^uscbdnc]). It is in Northern California in the foothills of the Sierra Nevada mountains, and the county seat is Sonora. The county contains moderate, high, and very high fire hazard severity zones. It has experienced five PSPS events, and Tuolumne County was among the counties in the top quartile for increases in extreme smoke days. Tuolumne County experienced the Rim Fire (2013), which burned over 257,000 acres, making it one of the largest wildfires in California’s history at the time.

Interview Data Collection and Analyses

Sampling Strategy

In our three study counties, we worked with local partner organizations to identify interview participants using a combination of purposive and snowball sampling (Creswell & Poth, 2016⁴⁸; Patton, 2015⁴⁹). We conducted semi-structured interviews with three groups:

• Employees in local governments, community support organizations, land management agencies, or other stakeholder groups who had personally experienced all three wildfire hazards and had knowledge about others’ experiences
• County residents who had experienced all three wildfire hazards
• Residents belonging to populations vulnerable to wildfire smoke (e.g., older adults, outdoor workers) or power disruption (e.g., users of electronic medical devices)

The first two groups of interviewees were key informants who assisted us in identifying through snowball sampling residents vulnerable to wildfire smoke or power disruption. Some key informants also belonged to these vulnerable populations.

Interview Sample

From February to June 2023, we conducted a total of 45 interviews in Lake County (n = 17), Tehama County (n = 11), Tuolumne County (n = 17). Table 1 provides details about participant characteristics such as mode of interview, age, gender, and stakeholder affiliation. Due to record snowfall in Northern California in the winter and spring of 2023, it was difficult to reach many participants in person. Thus, most interviews were held virtually. When in person, we met interviewees at their homes or public locations where privacy could be ensured (e.g., a private room at a library or community center). Interviews were audio recorded and transcribed verbatim. All interviews were conducted by the first author.

Table 1. Interview Participant Characteristics

Interview Guide

We developed a semi-structured interview protocol informed by the behavioral models of risk perception and response to climate hazards discussed in the literature review, as well as empirical literature specific to behavioral adaptation to wildfire-related hazards. We designed questions to elicit data about variables related to risk perception and response, and factors known to influence risk perception and response. Interview topics included: (a) length of county residency and place attachment, (b) experiences with single wildfire hazards, (c) experiences with co-occurring wildfire hazards and their compounding effects, (d) threat and response efficacy perceptions of single and co-occurring wildfire hazards, (e) responses and protective actions to co-occurring wildfire hazards and their compounding effects, and (f) knowledge of and participation in wildfire-related programs and efforts providing help to residents at the community-level.

Data Analysis Procedures

Field notes and analytical memos were written throughout the interview data collection process, which documented contextual observation, initial themes, and patterns across interviews. Interview transcripts and memos were coded and analyzed using an abductive approach (Timmermans & Tavory, 2012⁵⁰) (i.e., a combination of deductive and inductive techniques) using Dovetail, an online coding software. Interviews were coded deductively by identifying interview responses corresponding to predetermined and expected topics related to the research questions. Interviews were also coded inductively using thematic analysis, identifying emergent patterns and themes across responses (Charmaz, 2014⁵¹). All coding was conducted by the first author.

Ethical Considerations and Researcher Positionality

The study was approved by the University of Michigan’s Institutional Review Board (protocol ID HUM00230737) on February 21, 2023. Our study adhered to three core principles of ethical research: respect for persons, beneficence/non-maleficence, and justice (National Commission, 1978⁵²) and incorporated the principles of fidelity and gratitude from Browne and Peek’s Ethical Toolkit (2014⁵³).

To ensure respect for participants and fully informed consent, we provided detailed study information (e.g., goals, dissemination plans) during recruitment and at the beginning of interviews. A multi-step consent procedure addressed potential coercion (Rosenstein, 2004⁵⁴), including reading the consent form aloud, inviting questions before interviews began, and providing a copy for participants to keep. Participants received a gift card incentive before interviews to clarify that compensation was not contingent on completion.

To reduce potential distress, we implemented techniques to monitor and respond to participants’ emotional well-being (Browne and Peek, 2014; Collogan et al., 2004⁵⁵). Mid-interview check-ins offered opportunities for breaks or an early end to the interviews, if needed; however, most participants reported low distress and opted to continue. To promote justice, we provided fair incentives to both participants and community members assisting with recruitment.

Aligned with beneficence, our findings aim to advance understanding of behavioral responses while offering actionable recommendations for adapting to wildfire hazards. To avoid overburdening specific populations, we coordinated with other scholars working in Northern California (Barron Ausbrooks et al., 2009⁵⁶). Upholding fidelity and gratitude, we maintain ongoing communication with community partners, sharing findings, and exploring outreach opportunities.

Findings and Discussion

Here we discuss our preliminary findings related to our three research questions. First, we address how individuals living in fire-prone, smoky, and power disrupted regions in Northern California perceived threat and response efficacy related to co-occurring wildfire hazards. We also address how individuals responded to co-occurring wildfire hazards. And finally, we end by discussing compounding effects that individuals faced as a result of co-occurring wildfire hazards.

Threat, Response Efficacy, and Response to Co-Occurring Wildfire Hazards

Wildfire

The majority of participants perceived wildfire to be a very serious threat. In events when fire, smoke, and power shutoffs occurred simultaneously, most participants described primarily being concerned with their own physical safety and the safety of their friends, family, and neighbors from the direct effects of a fire. With respect to response efficacy, most interview participants reported a high-level of efficacy to respond to wildfire, both through actions reducing risk (e.g., creating defensible space around their home) and emergency preparedness (e.g., planning an evacuation route). However, many participants discussed how during severe fire conditions, they did not believe that any amount of risk reduction on their personal property would stop a fire from destroying their home or damaging their property. Relatedly, several participants shared a worry about an insufficient number or quality of evacuation routes in their community. A Tuolumne County resident said:

It could get to the point where you feel scared for your life. Both the Moccasin and the Rim Fire ended up in our area with evacuation orders… And in this area, there’s not a whole lot of convenient evacuation routes

A participant in Lake County described a similar worry, saying he had bought an inflatable boat as a protective measure, because his home is right along Clear Lake:

There’s only one way out of this area. There’s only one road out. So, we bought a boat. It’s inflatable…We have an emergency bag that we have [ready] if we go out on the lake, and all the neighbors have the same thing. These boats are lined up ready to go.

Such responses demonstrate how residents developed creative solutions when faced with limited evacuation options in their communities.

Perhaps unsurprisingly, and even during a wildfire event co-occurring with smoke or power disruption, most participants discussed how emergency preparedness and evacuation were top of mind when a fire was actively burning nearby. During past fires, participants recalled prioritizing their response to the fire, while related concerns and responses to smoke and power shutoffs were less salient. Given the immediate life-threatening dangers of wildfire, this finding shows how not all wildfire hazards are, or should be, considered equally; there may be times when the urgent need to address one hazard supersedes responses to others. This speaks to the limited capacity to respond that many individuals may face, regardless of time, money, or other types of support. Based on interview data, behavioral responses are less likely to focus on wildfire smoke in the event of a wildfire evacuation.

Wildfire Smoke

Most participants also perceived wildfire smoke to be a serious threat; however, many expressed more concern for other individuals whom they perceived to be vulnerable to smoke (e.g., those with asthma, older adults) than for themselves. One participant focused on the health of her children, “There were a lot of times where things got canceled because we didn’t want our children to be exercising and inhaling smoke under [those] physical conditions.” Another participant said:

With the smoke in the air, making sure that anyone in my work… with respiratory illnesses… we’ll reserve a block at the hotel, which has a good [air] filtration system, and try to make sure that anyone with issues gets into a room.

Related to concern for others, some participants expressed a belief that they were less susceptible to smoke compared to others, particularly those with firefighting experience or backgrounds in fire emergency response. This finding may be influenced by the limited number of participants we spoke to who had an ongoing respiratory health issue (n=10). Although some groups are more vulnerable to smoke than others, this finding highlights a potential gap in risk perception: the false perception among some residents, that individuals who are not over 65, children, or suffering from asthma are not susceptible to the negative health impacts of smoke. This misconception aligns with broader challenges in hazards risk communication, where groups outside the “high-risk” categories may underestimate their susceptibility. Reported responses to wildfire smoke included wearing a mask, staying indoors, using an air filter, or in some cases, temporarily leaving the area. However, while residents demonstrated knowledge of protective actions, many participants expressed frustration and a low-level of efficacy to respond to increasingly frequent and severe smoke events, citing the uncertainty of the number of days smoke would be present during a smoke event, and the challenges of repeated interruptions to recreational and professional outdoor activities. Many interviewees reported negative impacts on their mental health and well-being from exposure to wildfire smoke. One participant recalled her experience,

I think the worst experience was the Dixie Fire… it was close by, and it was miserable, it was just miserable living in that smoke. You couldn’t get out of it. It was anywhere you went. It was just there. We couldn’t go outside. It was coming in the swamp cooler. Yeah, that was pretty miserable.

In the interviews, many participants focused on the increasing frequency of wildfire smoke events, and their frustration about the uncertainty of how long smoke events might last; it could be hours, days, or even weeks, and not knowing the length of a smoke event made it difficult to plan for the future. This finding demonstrates the need for further study of the mental health impacts of smoke, and how response efficacy may be linked to negative mental health impacts from prolonged smoke events with uncertain timeframes.

Power Shutoffs

Finally, interview participants discussed the inconvenience and challenges posed by power shutoffs, especially PSPS implemented by PG&E in 2019, 2020, and 2021. In response to these shutoffs, many participants reported purchasing gas or propane generators for supplemental power. Only a few interview participants had solar power and back-up batteries, highlighting disparities in energy resilience strategies. Participants generally described high response efficacy in addressing power disruption in their community. However, many participants acknowledged that purchasing a generator or solar panels was an expense not all households could afford, raising concerns about the inequitable impacts of power shutoffs, particularly during hot temperatures and smoke events.

Compounding Effects Resulting From Co-Occurring Wildfire Hazards

Beyond the individual challenges posed by each hazard, interview participants reported several compounding effects resulting from co-occurring wildfire hazards. The first impact cited was increased heat exposure. Several participants discussed the challenges of managing hot temperatures during wildfire season. For example, in many mountain or foothill towns, residents do not have air conditioning and open their windows at night to cool down their home. When there was smoke in the air, however, keeping windows and doors closed to prevent smoke infiltration resulted in a hotter indoor environment, which made it less comfortable for sleep or other daytime activities and increased the potential for heat-related illness.

Another compounding effect discussed was the pollution, noise, expense, and fire risk from generators. As mentioned, a common response to increasingly frequent power disruption was to purchase a generator, but generators were also cited as dirty, loud, and expensive to maintain and fuel. Several participants also mentioned risks that could result from generator misuse or poor placement; for example, generators could produce sparks and start a fire, or pollution produced by generators could be dangerous if it infiltrated a home—an example of maladaptation.

Finally, a third effect mentioned was the lack of access to clean water by residents who have well-water during a power shutoff. Participants in Tuolumne County cited how power shutoffs prevented electric pumps from delivering well-water to their homes. In most cases, it was not feasible to hook up a generator to the well-pump, forcing residents to purchase clean drinking water and limit use. These effects show that when responding to co-occurring wildfire hazards, residents must be aware of both the single hazards and these potential compounding effects that could result from the interaction of two or more hazards. More study is needed to better understand the frequency and character of these impacts and the extent to which residents outside of Lake, Tehama, and Tuolumne Counties face the same compounding effects. Further study of these compounding effects could advance models of response to co-occurring climate hazards that could be applied and tested across spatial and temporal scales.

Conclusions

Implications for Policy or Practice

Findings most relevant and actionable to communities include the need for public health agencies to enhance communications about the universal health risks of wildfire smoke, increased support and research on the mental health impacts of wildfire smoke exposure, and guidance and education for residents on safe generator use. To improve the ability of residents to respond to smoke while evacuating safely or seeking temporary shelter, community organizations and/or public health agencies can include masks and/or filtration devices in evacuation kits or “go-bags.”

This also relates to the Centers for Disease Control and Prevention Crisis and Emergency Risk Communication (CERC) model, which emphasizes the importance of tailoring messages to counteract such misconceptions through communication about universal health risks (Centers for Disease Control and Prevention, 2024⁵⁷). If applying CERC principles, public health campaigns would focus on dispelling the notion that only specific groups are at risk and emphasize that wildfire smoke poses serious health risks for all populations, regardless of age or pre-existing conditions. This approach could close gaps in risk perception and encourage protective responses.

Some local initiatives do provide models for addressing disparities in power shutoff impacts. For example, the Self-Generation Incentive Program (SGIP) in California, has aimed to subsidize battery storage systems for lower-income residents and medically vulnerable populations (California Public Utilities Commission, n.d.⁵⁸). Similar programs could be expanded to include generators. Insights from literature on heat waves also show the importance of targeted interventions (Gronlund et al., 2014⁵⁹). Cooling centers and distributed resources can mitigate health risks during heat-related emergencies, but access is often limited for lower-income groups. Thus, local governments and community organizations could prioritize outreach and education on safe generator use, while ensuring equitable access through subsidies or distribution programs tailored to at-risk populations.

Limitations

Our study has limitations, including its relatively small sample size (n=45) and the retrospective nature of our interview protocol, as we asked participants to recall past experiences, which may introduce recall bias. Additionally, we spoke with a limited number of participants who had ongoing respiratory health issues (n=10), which may have influenced findings about smoke risk perceptions.

Future Research Directions

To ensure that research on co-occurring wildfire hazards is meaningful and responsive to the needs of at-risk communities (Norström et al., 2020⁶⁰), we are working closely with local stakeholders and FAC Net to share findings from this study and inform policy-level solutions at the local level. This study will also inform a report and recommendations for communications, programs, and policy interventions aimed at promoting adaptation for our local partners and FAC Net.

We recommend that future research conduct larger studies, such as surveys, to determine if the compounding effects described here are experienced in other communities and to better understand the prevalence of additional impacts resulting from co-occurring wildfire hazards.

Acknowledgments. We would like to thank Steve and Kathy Berman and Western Fire and Forest Initiative for funding and supporting this project. We also would like to acknowledge the participation and assistance by FAC Net and FAC Net member organizations during recruitment, and the participation of individual staff members from some of those organizations. We also thank the interview participants for their time and willingness to share. And finally, we thank Mia Bittner who provided research assistance in support of the qualitative data analyses.

References