Tracking the Effects of the Marshall Fire on Pets and People

Publication Date: 2022

Abstract

Colorado’s Marshall Fire destroyed more than 1,000 homes in the Town of Superior, the City of Louisville, and unincorporated Boulder County on December 30, 2021. The fire occurred on a weekday, when many people were at work. It also occurred during the holidays, when many people were traveling. When the fire began spreading rapidly in populated areas, roadblocks and dense smoke prevented people who might have returned home from doing so. An estimated 60% of Colorado households include dogs, cats, birds, and horses. Moreover, most pet-owning households include more than one animal. While a precise count of animal lives lost in the fire is not possible, we estimate that more than 1,000 pets died. Through interviews with pet owners whose animals died, this research examined the circumstances they faced on the day of the fire that prevented them from rescuing their pets. The emotional consequences have been considerable, and we suggest that reducing future mass deaths of animals will require strong networks with trusted neighbors. This study also explored the impact of the fire on the veterinary clinics located within the burn zone.

Introduction

At 3:00 a.m. on December 30, 2021, the National Weather Service released high wind warnings for Colorado’s Boulder County. Forecasters predicted wind gusts up to 90 miles per hour. The winds exceeded the forecast due to atmospheric conditions known as “mountain wave amplification,” in which wind speed increases rapidly as air flows over the mountains and down into lower elevations. On December 30, wind speed reached 115 miles per hour, strong enough to blow a FedEx tractor trailer over on a highway (Lundgren, 2021¹). Months of drought in the region had left abundant dried grass as fuel, and around 10:00 a.m., firefighters responded to the first of three reports of brush fires. They quickly extinguished the first fire. The second, known as the Middle Fork Fire, grew large enough to necessitate temporarily closing an interstate highway. Firefighters soon controlled the Middle Fork Fire, and all resources were directed to the third fire, initially reported at 11:00 a.m. The wind quickly pushed flames and embers across the grassland from west to east, into the Town of Superior, with a population of approximately 45,000 people. The hurricane-force winds then pushed the fire into the densely populated City of Louisville and unincorporated Boulder County. As one climate scientist tweeted, “Even though this fire started in a brushy/wooded area, it clearly ended as an urban firestorm (Swain, 2021²; emphasis added). Known as the Marshall Fire, it quickly became the most destructive fire in state history. It occurred on a weekday, when many residents were at work, and just before the New Year, when many others were away for the holidays. Within six hours, the Marshall Fire destroyed 1,084 homes and damaged 149 more, along with 30 commercial structures (Phillips, 2021³). It displaced over 30,000 residents and burned over 6,000 acres.

Any incident that affects large numbers of people will affect animals, too. According to data from the American Veterinary Medical Association (AVMA), Colorado ranks among the top ten states in terms of percentage of households that own pets (AVMA, 2018⁴). Because of their total dependence on people for daily needs, transportation, and escape from hazardous situations, pets are particularly vulnerable in disaster (Irvine, 2009⁵; DeYoung & Farmer, 2021⁶; Farmer et al., 2016⁷). Although the number of companion animals who lost their lives in the Marshall Fire can never be precisely determined, we make an estimate later in this report.

This quick response research sought to accomplish three goals: (1) to conduct a census of lost, found, missing, and deceased pets; (2) to determine how to reduce the number of pet casualties in future events, and (3) to assess the impact of the disaster on veterinarians, especially those with clinics within the burn zone. These three goals are “nested” in the following way. The first research goal of determining the number of animals affected by the fire also intended to assess the means by which these numbers were tracked and conceptualize ways to reduce those numbers in subsequent disasters. Reducing the number of animals injured or deceased would lessen the burden on veterinary clinics and animal shelters, most of which operate on limited resources and with limited staff even under normal conditions. Understanding the burden placed on veterinarians makes it possible to implement ways to meet their unique needs and lessen the burden.

Literature Review

This research contributes to the literatures on pre-disaster evacuation and on the importance of the human-animal bond. Although these two bodies of literature exist separately, they also overlap. Because evacuation from disasters saves lives and reduces injuries, scholars have long sought to understand how to encourage people to comply with evacuation orders before an event (Perry 1979⁸; Perry et al., 2001⁹). Disaster researchers now recognize that pet ownership influences evacuation decisions. For example, in studying evacuation decision-making during U.S. hurricanes from 1961 through 1989, Baker (1991¹⁰) found that providing for pets ranked highly on the list of reasons why people chose not to evacuate. Whitehead and colleagues (2000¹¹) also found that pet ownership strongly predicted failure to evacuate. Drabek (2001¹²) found that pet ownership influenced choice of shelter once evacuated, with prohibitions on pets leading over half of survey respondents to choose to shelter with friends rather than in public shelters.

In the research specifically on animal welfare and disasters, pet evacuation constitutes a prominent focus (DeYoung & Farmer, 2021; Farmer et al., 2016; Irvine, 2009; Zottarelli, 2010¹³). Heath and colleagues (2001a¹⁴) warned that “pet ownership can be a significant threat to public and animal safety during disasters” (664). If a hotel, emergency shelter, or other location that will accommodate pets cannot be secured in advance, attempting to find one during a disaster can delay evacuation (Whitehead et al., 2000). If pet owners cannot find pet-friendly accommodations, they may choose to ignore evacuation orders. In a survey performed by the American Kennel Club less than a year after Hurricane Katrina, 62% of respondents claimed they would defy orders to evacuate if they could not locate a place that would accommodate their pets (American Kennel Club, 2006¹⁵). Some owners might choose to leave their pets behind or be required—or even forced—to do so (Blendon et al., 2007¹⁶; Petrolia & Bhattacharjee, 2010¹⁷). Leaving animals behind can jeopardize animal health (Heath et al., 2000¹⁸; Kajiwara, 2020¹⁹; Mattes, 2016²⁰). It can also affect human mental health and emotional well-being (Glassey & Wilson, 2011²¹; Heath, 1999²²; Trigg et al., 2015²³; Thompson et al., 2014²⁴). Because most pet owners in the United States consider their pets members of the family, losing a pet can result in significant psychological distress and trauma (Austin, 2013²⁵; Hunt et al., 2008²⁶; Kajiwara, 2020; Lowe et al., 2009²⁷). The experience can also have considerable impact on children (Travis, 2014²⁸).

The failure to evacuate pets can lead owners to attempt to rescue pets left behind. In such cases, owners will reenter evacuated areas before disaster personnel deem the sites safe (Heath & Champion, 1996²⁹; Glassey et al., 2022³⁰; Kajiwara, 2020; Mattes, 2016). Research comparing evacuations following a flood and a hazardous chemical spill found that 80% of unauthorized reentries were to rescue pets (Heath et al., 2000). Pet owners put not only themselves at risk, but also endanger any personnel who might need to subsequently rescue those who reentered the area. Owners’ reasons for leaving pets behind at the time of evacuation include thinking they would not be out of their homes for long. In the chemical spill, for example, emergency managers anticipated that the response would take several hours. Instead, it took over two weeks, reflecting the unpredictability of disaster response. Other reasons include not having a disaster plan, which results in owners not knowing where to take their pets and being unable to transport them.

Some research has associated the failure to evacuate animals with a weak human-animal bond, measured by level of attachment and commitment to a household pet (Trigg et al. 2015; 2016³¹). Studies assess attachment and commitment to animals by indicators of care, such as visits to veterinarians and owning leashes or carriers. These studies find that owners having stronger attachments to their household pets are also more likely to have disaster plans inclusive of pets (Heath et al. 2001a; 2001b³²; 2001c³³; Hesterberg et al., 2012³⁴). Conversely, a weaker standard of care indicates a weaker bond with an animal. People who leave their animals behind are those who keep their dogs primarily outdoors or who have no carriers available to transport their cats.

Other research suggests a more complex set of factors at work, however, and the notion of a weak bond with animals has been challenged. For example, Zottarelli (2010) found that residents with lower income were more likely to leave animals behind during an evacuation. Irvine (2006³⁵, 2009) found that pet owners were forced to evacuate without their pets during Hurricane Katrina. New Orleans residents who managed to bring their dogs and cats to the Superdome—a designated evacuation site—were forced to leave them behind when they subsequently evacuated that facility because animals were not permitted on public transportation. Media accounts depict National Guardsmen simply letting dogs and cats run free as their guardians watch helplessly. One of the most famous images from the disaster depicts a little white dog named Snowball being torn from a boy’s arms by a police officer as the boy boarded a bus to leave the Superdome (Foster, 2005³⁶). Video showed the boy so upset that he vomited. The officer separated the dog and boy to uphold the policy that prohibits animals on public transportation. Evacuees reported being told that their animals would be rescued later and some thought they could soon return for their animals themselves. As is now widely known, some residents have never returned. An estimated 200,000 pets were left behind in the aftermath of Hurricane Katrina (Irvine, 2009; Farmer et al., 2016). It is unlikely that this number can be attributed to a weak human-animal bond.

In the Fukushima nuclear disaster that followed an earthquake and tsunami in March 2011, pet owners faced a similar situation. The Fukushima Prefecture government gave residents mixed messages about what to do with their pets (Kajiwara, 2020). Although pre-disaster instructions published in 2007 told pet owners to bring their animals with them, instructions given at the time of the evacuation “did not allow residents to evacuate with their companion animals” (Kajiwara, 2020, p. 112). Moreover, officials initially told evacuees they would be gone for only a few days (Mattes, 2016). Instead, evacuees frequently moved from shelter to shelter as the evacuation zone widened, and some who initially evacuated with their pets had to leave them behind when instructed to move. In some shelters, people were forced to relinquish their pets. The number of dogs and cats left behind in the 20-kilometer radius of power plant No. 1 is estimated to be between 10,000 and 20,000. By analyzing pet-related expenditures such as veterinary care, food, and grooming, Kajiwara (2020) concluded that Japanese pet owners have a strong attachment to their pets. Thus, owners’ decisions regarding their animals at the time of the disaster cannot be attributed to a weak bond.

In the United States, media coverage of the animals left behind, and particularly the footage of Snowball, prompted the introduction of the Pets Evacuation and Transportation Standards Act, or PETS Act, in Congress. The wording of the PETS Act specifies that state and local response plans should “take into account the needs of individuals with household pets and service animals prior to, during, and following a major disaster or emergency.” On October 6, 2006, President George W. Bush signed the Act into law. It requires state and local emergency management agencies to include companion and service animals in their disaster response plans. Although the PETS Act also makes funding from the Federal Emergency Management Agency contingent on compliance, enforcement has been non-existent. The Act has had mixed results. In a study following Hurricane Irene, Hunt and colleagues conclude that “the media coverage around Hurricane Katrina and the subsequent PETS legislation have had positive effects on the evacuation of animals and that general awareness about the importance of evacuating pets has increased significantly” (2012, p. 537³⁷). Yet, in a study following Hurricane Harvey, Glassey’s interviews with individuals leading the animal response reveal that “only a minority [of responders] had specific knowledge of the PETS Act” (2018, p. 3). Decker and colleagues (2012³⁸) similarly surveyed animal care and control agencies in Ohio and found that only one-third were aware of the PETS Act. Although the PETS Act represents a cultural shift in attitudes about the importance of saving animals in disasters, “the implementation of animal emergency planning appears sub-optimal and the integration of animal welfare charities to respond effectively remains fragmented in many areas” (Glassey, 2018, p. 3³⁹).

Although planning for animal care in disasters is essential, the circumstances of disasters often defeat even the best intentions. This study examines the inability to evacuate pets rather than the intentional failure to do so following a rapid-onset disaster. It examines a disaster that defied the parameters of previous studies of pet evacuation.

Research Design

As others have noted, disaster research poses distinct logistical and methodological challenges (Thompson et al., 2017⁴⁰). These include the sudden and unanticipated onset of events that prevents the design of research in advance, the community disruption that hinders representative sampling, and the time constraints that can limit the generalizability of research findings (Mileti, 1987⁴¹; Stallings, 2007⁴²). Despite these challenges, research on disasters nevertheless employs methods routinely used in social scientific studies, albeit in different contexts (Stallings, 2007; Tierney, 2019⁴³).

Research Questions

Our research questions were:

1. What prevented pet owners from evacuating their pets to safety?
2. What, if anything, might have helped them save their pets?
3. What challenges did veterinarians face in the aftermath of the fire?

As mentioned above, findings related to research question #3 are not reported here.

Data, Methods, and Procedures

We collected data from several sources frequently used in qualitative analyses. The first comprises traditional editorial and social media coverage that appeared from the onset of the event through January 15, 2022. We chose this date because, at two weeks post-event, coverage diminished and then stopped. Along with news coverage collected through a comprehensive search of all accessible media venues, this component of the dataset includes evacuation orders and emergency information for the City of Louisville, the Town of Superior, and Boulder County collected from Twitter and the Everbridge Public Warning platform from the morning of December 30, 2021 through their cessation on January 1, 2022. It also includes 11 pages of transcripts of Marshall Fire Briefings held by the Boulder County Sheriff's Office and publicly available on YouTube.

The second source of data includes pet owners’ posts and responses in the Marshall Fire Lost and Found Pets Facebook group, established the day of the fire and collected with permission of the group moderator through May 2022. We chose this as an endpoint for monitoring the group because posts about lost and found animals had stopped. This group provided a tentative count of lost and found animals and the platform for recruiting owners for interviews, which constitutes the third source of data. We conducted semi-structured interviews with eight owners whose pets died (or were thought to have died) in the fire. Interviews ranged in length from 30 to 60 minutes. All interviews with pet owners occurred within two weeks of the event. Owners of lost pets were openly recruited from the Marshall Fire Lost and Found Pets Facebook group through an announcement about the research, posted with permission of the group’s moderator. We began with a request for interviewees to describe, in their own words, how events unfolded for them the day of the fire. This allowed interviewees to choose starting points and salient directions for the narratives of their experience.

We also conducted interviews of approximately 30 minutes in length with two key staff members of the animal shelter with primary receiving responsibility for found or deceased animals. Interviews with shelter staff focused on numbers of animals received, their status on intake (e.g., healthy/injured/dead on arrival), their subsequent veterinary treatment, and reclaim status (e.g., owner identified and notified; owner unknown). In addition, virtual (Zoom) community discussion groups were offered to veterinary professionals on multiple weekday/weekend and times to allow scheduling flexibility and increase participation. The discussion groups ranged from 60-90 minutes and included veterinarians from affected clinics within the burn zone, veterinary professionals from volunteer response groups, and other members of the veterinary community include pet loss grief counselors. Between two and five veterinarians participated in the virtual sessions. We also established an informal communication outlet through the Slack app to facilitate information exchange within the veterinary professional community.

Sampling and Participants

For interviews, owners willing to participate clicked a link that led them to a scheduling app and consent form. This constitutes a form of convenience sampling, a method used in previous studies involving the collection of ephemeral or perishable data (Quarantelli,1997⁴⁴). Of the eight owners who responded, all but one was female. All were white, consistent with the majority of the population within the affected communities. Veterinarians and shelter staff members were also white and female.

Data Analysis

We used the media coverage and emergency notifications records to create a timeline of the event (see Appendix). We used established techniques of qualitative content analysis to examine the Facebook posts. Similarly, we used qualitative data analysis techniques, including coding and memoing, to analyze interview transcripts.

Ethical Considerations, Researcher Positionality, Reciprocity, and Other Considerations

This research received approval from the Institutional Review Board of the University of Colorado Boulder under protocols #22-0070 and #22-0024. All interviews were conducted either by phone or by Zoom. Participants did not receive compensation. Despite enduring significant loss, of property as well as pets, owners volunteered to speak at length about their pets and their grief in hopes that doing so would help others in future disasters.

Findings

Census of Lost Pets

Two factors ruled out making a detailed count of lost animals. The first is the absence of a comprehensive census of animals pre-disaster. Although the City of Louisville requires residents to license dogs, neither the Town of Superior nor unincorporated Boulder County have a licensing requirement. At the time of the fire, 618 dogs were licensed in Louisville. Of course, not all dog owners license their dogs. Neither Louisville, Superior, nor Boulder County requires licenses for cats. The second complicating factor is the absence of a comprehensive reporting system for lost animals. The Marshall Fire Lost and Found Pets Facebook group constitutes the main site for reporting lost pets, and this requires a Facebook account as well as the time and willingness to post a report. Despite these limitations, we can make an informed estimate of the number of lost pets. According to surveys by the American Veterinary Medical Association (2018), an estimated 60% of Colorado households include dogs, cats, birds, and horses. Moreover, most pet-owning households include more than one animal. Thus, of the 1,233 homes destroyed or damaged in the Marshall Fire, estimates put at least one pet in 739 of these households. Estimating the number of households with additional pets raises the potential number of animals affected to 1,182.

Some animals were rescued by first responders, volunteers, or other evacuees who took them to local animal shelters, where those who suffered burns and other injuries received treatment. Table 1 reports the numbers of animals received by the Humane Society of Boulder Valley (HSBV). In addition to taking in found animals, HSBV provided 33 evacuees with up to 200 days of free boarding following evacuation. In addition, 24 strays, mostly dogs, were brought in and all were reunited with their grateful families. Only five animals were brought in dead on arrival. HSBV encouraged pet owners to submit “lost” reports in an effort to match with animals brought in (see Table 2). They reported receiving 107 reports. Of these, 26 were cancelled through reuniting with owners and 24 were cancelled through confirming death. The remaining 57 lost reports were still active one month after the fire.

Table 1. Animals Received at the Humane Society of Boulder Valley

Table 2. Lost Reports Filed at the Humane Society of Boulder Valley

Among the eight residents interviewed for this research, 17 pets were lost and assumed deceased in the fire (10 cats; five dogs; two guinea pigs). In sum, it appears that the majority of animals were not rescued. As mentioned, some owners found the remains of their animals in their devastated property, but a heavy overnight snowfall covered the landscape. Owners still hope for sightings and reunions. Photos of hundreds of cats still appear online, eight months after the fire, tagged with #StillMissing. Dedicated, trained volunteers continue to stock food and water stations in the burn zone and monitor humane traps and wildlife cameras.

Problems Resulting from Rapid Onset and Lack of Emergency Notifications

Although the literature has associated the failure to evacuate pets with a weak human-animal bond, the circumstances surrounding the onset and spread of the Marshall Fire challenge that assertion. The fire’s behavior, coupled with pitfalls in the emergency notification system, rather than lack of care for or attachment to animals, caused the high pet fatalities.

At 11:47 a.m., the Boulder County Sheriff’s Office issued the first mandatory evacuation orders through landlines and cellphones and email addresses of residents who had opted into the warning system. After the fire, many residents said they did not know that they needed to opt into the emergency notification system to receive alerts on their cell phones. Of these 215 notifications sent, only 54 contacts confirmed receipt of the message as instructed. To be marked as confirmed, recipients of the message needed to press 1 if they received a reverse 911 call or click on a link if they received a text or email. Countless people likely received these notifications but did not follow that instruction, so the exact number of messages received is impossible to know (McKinley, 2022⁴⁵; Phillips, 2022⁴⁶; Zellinger, 2022⁴⁷). Only twenty minutes later, fire engulfed the first home in Superior. By 1:00 p.m., the entire town of Superior had been ordered to evacuate. However, only six percent of those notifications were confirmed as received (Zellinger, 2022). Evacuation orders were soon issued for most of the city of Louisville and nearby areas of unincorporated Boulder County.

Of the total notifications sent out to residents, how many were actually received? Table 3 lists the success rate of each notification sent out by the Boulder County Sheriff's Department. Of the 24,289 emergency notifications sent out by the Boulder County Sheriff's office during the Marshall fires, only 4,637, or 19%, were confirmed as received. As mentioned, many area residents were at work or out of town when the fire started. In interviews, residents who were at work learned of the fire either through notification or media coverage. By then, roads were congested with evacuees, and some were closed to all but emergency vehicles, making travel slow and some routes inaccessible. One interviewee said that the drive from his workplace in Boulder to his neighborhood in Louisville, which normally took 15-to-20-minutes, took nearly two hours that day. By the time he reached his neighborhood in hopes of rescuing his cat, his street was blocked off and he could not return to see if his home was still standing.

Table 3. Notifications from the Boulder County Emergency Notification System

Accounts of Evacuation and Loss

Some residents who were home at the time evacuated with their animals. However, countless cats who could not be caught or located were left behind. One interviewee had gone to check on her horse, whom she boarded at a stable near the site of the already-contained Middle Fork Fire. While stuck in traffic, she received a panic-stricken phone call from her husband. He and the children—and their five pets, which included two dogs, a cat, and two guinea pigs—were stranded at home without the car. A stranger rescued her husband, both children, and the dog, but the cat could not be found when they left the house. The daughter had initially picked up the two guinea pigs intending to carry them with her, but she set them down to get dressed and they ran off.

Evacuation orders, dense smoke, fast-moving flames, and road closures prevented residents who were not at home from re-entering their neighborhoods to rescue their pets. Two owners were out of state at the time and followed the fire’s progress online and through reports from friends in the area. One of these owners watched the fire on their home security cameras as she and her husband drove back from another state. They had left their two cats in the care of a neighbor, whose home in Superior was one of the first to go up in flames. The security camera stopped recording and the screen said “Loud audible noise” before going black. After the fire, trained bloodhounds brought in by volunteers searched the site but found no scent. The owners also sifted the site hoping to find remains, but found nothing. A Louisville pet owner who was also away from home lost three dogs and a cat when the pet sitter left them behind. Neighbors watched him leave but did not know the pets were in the house. She emphasized how traumatic the loss was for her 14-year-old daughter, who had begun therapy as a consequence. At the time of the interview, she was visiting the site every day to sift, but had found no remains. “Despite losing everything,” she said, “it’s the pets we grieve.” One family was visiting the area from out of state and were staying in an Airbnb in Superior. Understandably, they had not signed up for emergency alerts and were not following local news during what was to be a vacation. They left their two Labrador retrievers in the property for what they thought would be just a few hours. Both dogs perished in the fire; bloodhounds searched the area within a few days and found their remains. Volunteers also found a charred food bowl used by the dogs, a picture of which was circulated widely on social media.

Impact on Veterinary Medical Professionals

The second author had access to consolidated information about the impact of Marshall Fire on the veterinary professional community through her role as the director of the area veterinary medical reserve corps (vMRC). As an organization with connections to the local veterinary community, the vMRC collaborated with area professional organizations, such as the Colorado Association of Certified Veterinary Technicians and the Colorado Veterinary Medical Association, as well as other volunteer and animal care groups. The veterinary medical community sought information about affected animal care, need for medical supplies, or search and rescue support. Unfortunately, communication between disaster response organizations working within the disaster zone and the veterinary medical community was fractured and slow. Veterinary medical professionals were unaware of how to access official information outlets. Consequently, various groups on social media platforms quickly arose to allow communication between area veterinary professionals. Lack of early and consistent communication between government and volunteer groups resulted in multiple self-deployments of veterinary professionals during the period of active disaster as well as in the initial weeks of recovery as lost animals were still being sought.

The veterinary clinics in the disaster-affected area came through the disaster with few physical losses although staff members and clinic functions were shared among neighboring clinics in the initial weeks after the disaster and the clinics could reopen. The majority of veterinary clinics were operating at or beyond capacity prior to the Marshall Fire. Backlog of non-essential procedures from the COVID19 pandemic as well as the influx of new animals to many households have overwhelmed many clinic resources. With the addition of urgent visits that occurred in the aftermath of Marshall Fire, area clinics were stressed further. Clinic staff members suffered trauma, however, stemming from damage to or loss of their homes, damage or loss to neighborhoods around the clinic, as well as from the loss of their patients and the grief experienced by their clients.

Reporting and data collection on the needs of displaced area pet owners was attempted through clinic reporting but with minimal success. Anecdotal reports shared among area professionals suggest that many clinics provided food, medication replacement, and initial medical care at no cost to the client, but did not track the provided services and supplies effectively enough to report.

Three debrief/after-action review sessions were offered to the veterinary medical community through the vMRC. Virtual discussion sessions offered conversations between colleagues, memorializing the event, and discussing the experience between colleagues. Over the course of the three-event series, 16 individuals from the vMRC team, area clinics and shelters, pet loss counselors attended the sessions.

In the immediate weeks after Marshall Fire, Colorado State University’s Argus Center and local pet loss support resources provided onsite mental health support to area clinics and shelters. Mental Health Partners and the Colorado Spirit team provided longer-term, embedded mental health support during a six-week pilot period exploring the best ways to provide post-disaster support to medical professionals serving the affected communities. Personal stories of experiences during the disaster event, emotions around feeling helpless and unable to assist, and concern for the risk to their own animals while they are at work if another similar rapid-onset disaster were to occur. The weeks following Marshall Fire continued to be dry and carry a high fire risk. Area clinics report loss of staff members due to the anxiety related to the potential for another event.

We obtained sponsorship through the University of Colorado’s Office for Outreach and Engagement to provide snacks and beverages during these sessions and encourage participation and rapport-building between clinic staff and mental health support team members. Multiple anniversary events are in planning.

Conclusions

In its rapid onset in a highly populated area, the Marshall Fire was an unprecedented event in Colorado’s Front Range. Our research provides a powerful reminder of the need to include pets in emergency planning for households. This information is widely available from veterinarians, on various websites, from insurance providers, and other sources. However, even the most comprehensive plans would have been of little use in the Marshall Fire because so many owners were away from home, either at work or traveling. Future research needs to explore pitfalls in the emergency notification system, which prevented many pet owners from returning home in sufficient time to rescue their animals. This research also highlights the need to prepare for events that seem unlikely to happen. For example, few suburban homeowners would feel vulnerable to wildfire. The Marshall Fire showed how quickly a wildfire can become an urban firestorm, as it moved from grassland to suburban neighborhoods in the course of a few hours.

Based on this research, we now serve in an advisory capacity with Animal Help Now, a local animal welfare non-profit. The director of Animal Help Now lost his home in Superior in the fire, but was able to evacuate with his two cats. Animal Help Now has long hosted a wildlife emergency app and is now developing one for pet rescue. The app would connect pet owners with trusted contacts who will have permission to enter homes and rescue animals when owners are away. However, social factors, rather than strictly technological ones, may be the best route to animal rescue. We encourage pet owners to make friends with neighbors who also have pets and make such arrangements informally. The Marshall Fire has taught us that the best way to save the life of a pet is to know one’s neighbors.

Finally, this research contributes to the literature on the importance of the human-animal bond, especially in the context of disaster response. In particular, our research challenges studies associating the likelihood of pet evacuation with levels of attachment and commitment to animals. In the Marshall Fire, as in Hurricane Katrina and the Fukushima nuclear disaster, factors well beyond the control of pet owners prevented many from saving their animals. Physical distance and closed roads, not weak bonds, led to the loss of animals’ lives.

Acknowledgments

The authors thank Riley Coll, Ashley Mendoza-Salmon; Lindsay Singewald, Emily Sickler, and Sophia Valentine for valuable research assistance.

References