How police can use AI to run faster, safer wildfire evacuations

Key takeaways

• Wildfire evacuations require more police training than most agencies provide: Many departments train heavily for SWAT and tactical ops but spend minimal time on evacuation readiness.
• Integrating AI into evacuation planning can save lives: Combining realistic scenario-based training with AI predictions strengthens officer decision-making under pressure.
• Free alerting systems like IPAWS and RAVE improve public safety communication: Law enforcement can quickly notify residents and share live evacuation maps during disasters.
• Traffic modeling software can prevent deadly evacuation delays: Data-driven models show when to issue orders, where to deploy officers and how to avoid road gridlock.
• AI mapping tools give law enforcement real-time wildfire evacuation routes: Platforms like Perimeter and Ladris help agencies predict congestion, estimate evacuation times and choose the safest routes.

By Lieutenant Troy Morton

Imagine a deadly wildfire tearing through a city or county. The next day, the police chief or sheriff must stand before their community, amid the smoky air and scent of destruction, and speak with unwavering certainty: “We were prepared, and we did everything in our power to protect the lives of our people.” If you can’t say that today, then it’s time for change — time to innovate and transform how you approach disaster readiness.

Wildfire risk to communities and police response

Wildfires are the single most significant risk to life in many cities and counties, in some cases more deadly than even the most horrific active shooter events. This has been shown through the recent wildfires in California. The Tubbs Fire in Napa and Sonoma counties in 2017, the Camp Fire in Butte County in 2018 and, most recently, the Palisades and Eaton fires earlier this year in Southern California are just a few examples of how deadly these wildfires can be. In those four fires alone, there were over 130 fatalities. [1]

| RELATED: Eight hours inside the Pacific Palisades wildfire

According to a 2025 USA Today article, the 10-year average number of acres burned by wildfire each year since 2014 has roughly doubled. [2] Author Dinah Voyles Pulver further explains that six of the fires since 2017 have ranked in the top 10 for structures burned in the state’s history, and three fires were in the top five most deadly dating back to 1933. This is staggering information law enforcement should consider when setting budget and training priorities. Sheriffs’ offices and police departments alike are the sole agencies responsible for carrying out evacuations in California.

Wildfire fatalities should be something all agencies consider, no matter where they are. The size, intensity, destruction and frequency of these wildfires continue to increase, placing lives at risk and putting law enforcement in the spotlight when evacuations are required. With the profound responsibility of preserving lives during evacuations, law enforcement must adapt and develop. Fortunately, modern technologies hold the key to a more effective response. The potential of artificial intelligence (AI) to revolutionize law enforcement response is immense, offering hope and optimism in the face of these challenges.

Evacuation challenges for law enforcement

Public safety agencies employ a variety of emergency alert systems and public-facing mapping technologies, each with its own procedures for deployment during wildfires. In some jurisdictions, the local office of emergency services (OES) or office of emergency management (OEM) handles alerting residents while that jurisdiction’s law enforcement agencies carry out the evacuations. Having two different entities involved in evacuations can be complex. The need for coordination and training cannot be overstated.

In other jurisdictions, the OES or OEM is a division of the county sheriff’s office that handles both alerts and evacuations. How advanced and capable the alerting platforms are that agencies use differs as well. Furthermore, most law enforcement agencies lack traffic modeling and other specialized software to estimate evacuation durations, identify optimal routes, predict congestion points and analyze other critical factors. Programs like these would be extremely valuable to officers or deputies when conducting evacuations.

Why wildfires demand more police training

Policing has POST training standards and agency policy mandates in many areas relevant to professional policing. However, there are often no wildfire or disaster training and readiness requirements. Incidents of violence and criminal acts are undoubtedly high risk and require priority; however, wildfires can pose an even greater danger to life. When comparing the training hours and funding allocated to SWAT, bomb squads and other tactical operations to those dedicated to evacuation and disaster response, the difference is astounding. For instance, many SWAT teams in the Sacramento area train two to three times a month throughout the year, compared to approximately one brief training annually before fire season for evacuations. That practice does not align with other high-risk responsibilities law enforcement carries.

Communities often do not advocate for law enforcement reforms or improvements in disaster management until it is too late. This lack of clamor for change, combined with budgetary constraints and unfunded mandates, leaves many jurisdictions without the resources and training needed to operate at a high level in evacuation and disaster response, even though there are well-established state and federal mandates for specific response capabilities.

Executives who lead agencies in counties or cities with a high percentage of wildland-urban interface (WUI) areas should be especially alarmed. These are areas where homes or businesses are adjacent to or intermixed with wildland vegetation. According to a 2025 Guardian article, more than 80% of new California homes are in high fire-risk areas: “In 1990, about 13% of new homes were built in places at high risk of fire. By 2020, that number had more than doubled to 31%.” [3] Fortunately, the guidelines to respond effectively to these disasters are well established.

The National Incident Management System (NIMS) and, in California, the Standardized Emergency Management System (SEMS) map out the process for handling large-scale disasters through the Incident Command System (ICS). All agencies must understand the components of ICS. To receive reimbursement for disaster-related expenses such as staff overtime, agencies are required to certify that they adhered to ICS protocols during the incident.

When a law enforcement official and fire incident commander determine evacuation orders, part of the determination is how long it will take law enforcement to evacuate a particular area. In most cases, officers rely only on their experience and local knowledge of the region, with some assistance from a satellite map to see how many homes are impacted. However, that assessment can end up being little more than an educated guess with minimal, if any, data to substantiate it. Policing should not rely on best guesses when other options are available.

AI mapping and evacuation technology tools

The federal government has maintained an alerting platform for decades. That platform, FEMA’s Integrated Public Alert and Warning System (IPAWS), is available to law enforcement at no cost. In California, the California Office of Emergency Services (Cal OES) provides RAVE, an alerting platform, to jurisdictions free of charge as well. The system is offered to all California law enforcement and OES agencies. [4]

In recent years, technology companies have developed advanced mapping programs for use during evacuations. These mapping programs are made available to the public via a website and display areas under evacuation orders, how many homes are impacted, optimal evacuation routes, shelters and roadblocks. Agencies can send links to those websites through their wireless alerts so the community can click through, navigate to the website and have real-time data.

Bailey Farren, a Tubbs Fire survivor and cofounder of Perimeter, Inc., which makes collaborative mapping and software for emergency response, was quoted on the company website saying, “As the daughter of two first responders, I’ve seen my parents respond to emergencies my whole life. But the 2017 Tubbs Fire took all of us by surprise. During that evacuation, we had no information about where to go or how to get there. Amid the chaos, everyone asked for the same thing: a map of what was going on. We developed Perimeter to be just that. A real-time map that answers the most essential questions for life safety: what’s happening, what should I do, and where should I go?” [5]

East of Sacramento, the El Dorado County Sheriff’s Office of Emergency Services utilizes Perimeter. During the Crozier Fire in 2024, El Dorado County’s website at perimetermap.com saw over 1.23 million site visits during the incident. [6] This number was impressive, especially considering the evacuation only impacted a small rural community within the county and was relatively short in duration.

There are also applications available, and still in development, that can analyze specific areas — for example, a neighborhood that could be placed under an evacuation order. These programs analyze the number of homes, average number of vehicles for that number of homes, optimal evacuation routes and traffic congestion points. This provides data-driven results, showing how long an evacuation will take.

One of these companies is Ladris, which offers an AI platform that can run evacuation and traffic models using multiple injects and problems during the evacuation that the user can add. [7] The company website has a quote that speaks volumes. According to Sergeant Bob Womack, former emergency manager for Truckee, California, “Ladris is the holy grail of what we’ve been trying to get to.” The company says this platform takes “best guess” estimates out and replaces them with accurate, data-driven results. This is important to officers and deputies conducting evacuations. They need to know how long evacuations will take and compare that with how fast fires are moving.

How AI improves evacuation decisions

Traffic modeling software is critical when law enforcement allocates resources. It can provide information so officers can decide, for example, if it’s more important to send every officer into a neighborhood to conduct door-to-door notifications to evacuate or if the traffic backup at a stoplight will pose a greater risk, therefore redirecting officers to conduct manual traffic control. Having the ability to alert every household in person means nothing if the community gets trapped on the roadways and cannot escape due to traffic congestion.

How long evacuations will take impacts how early an evacuation order is placed. If the decision to evacuate an area is made too late, lives can be lost. These decisions affect officer safety as well. If officers or deputies are still in the area conducting notifications, they could find themselves trapped, diverting resources to save them instead of those who were already at risk.

In addition to better awareness technologies, the best way to improve law enforcement response is training.

Realistic evacuation training for officers

In-service evacuation training is essential, just like all the other training for high-risk incidents. AI platforms can assist in creating real-life scenarios. As technology advances, those scenarios will continue to have a bigger impact. They can provide situations involving differing types of incidents that require evacuations and exercise officers’ decision-making, thinking processes and allocation of resources to get the community evacuated to safety. “Check-the-box training is closed-loop, involves a stationary stance and targets, does not account for realistic conditions or stressors and involves no decision-making,” noted Illinois police captain James Nightingale in a 2023 article for the FBI. “Officers who undergo realistic scenario-based training, though, learn to make critical decisions with environmental cues rather than work their brain’s cortex motor pathway.” [8]

Having the capability to show officers how long evacuations can take if errors are made, such as failing to place traffic control officers at intersections, will have a more significant impact than simply briefing training on hypothetical situations with no data to substantiate it. These scenarios, done in coordination with fire professionals or an AI platform that can show how fast a fire would move under certain conditions, could make all the difference between life and death if the steps were taken to manage evacuations effectively. Rather than relying solely on in-person training, though, technology may hold the answer to ensuring every deputy or officer can respond as we wish.

Limitations and considerations for AI use

AI technology can never be solely relied upon to guide one’s response to an evolving threat. Technology can fail, and officers or deputies must be prepared to manage incidents without it. However, it can be an extremely valuable tool to help make informed decisions, and using platforms to train, reinforce and assess the capabilities of first responders is a necessity.

As the culture and trust of AI change, more industries may be willing to rely on it for essential tasks. The communities law enforcement serves would also need education about the new technology to ease concerns that AI would be trusted to save their lives. In jurisdictions where wildfires are a grave concern, management of those incidents is likely already a priority. If not, it is up to the agency to lobby its city council or board of supervisors to give law enforcement the tools they need to keep the community safe during disasters. [9] Clearly, the status quo is not acceptable, especially knowing the options are available for our use.

Steps to prepare police for wildfire evacuations

Law enforcement must be proactive and prepared before another catastrophic wildfire occurs. Departments should conduct comprehensive reviews of their current policies and procedures regarding evacuation operations and preparedness. Implementing policies that require mandatory training hours on evacuation operations is essential to ensure all officers and deputies meet a minimum standard of preparedness.

Although the law enforcement community may struggle at times with the logistical challenges of emergency management, their roles as first responders naturally align with the principles of the Incident Command System and National Incident Management System. [10] Officers and deputies also need to be familiar with ICS, local variants like SEMS and NIMS to integrate seamlessly into the emergency management structure. Any agency that is not familiar with or utilizing the no-cost alerting platforms IPAWS and RAVE should acquire those programs and ensure staff are competent in their use.

Lastly, agencies should consider purchasing programs such as public-facing mapping platforms and AI-supported traffic modeling software so they can best prepare their agency for response.

| WATCH: How police and fire collaborate during disaster response. The recent California wildfires underscore the critical role of collaboration, cross-training and technology in strengthening police and fire response to large-scale emergencies.

References

1. Deadliest wildfires in California history. Western Fire Chiefs Association. 2025.
2. Pulver DV. “A new wildfire paradigm”: Why California fires are growing more extensive and more destructive. USA Today. January 2025.
3. Chalabi M. More than 80% of new California properties are in high fire-risk areas. The Guardian. February 2025.
4. 9-1-1 operations manual: Chapter XII – Next-generation 9-1-1 alert & warning system (NG 9-1-1 AWS). State of California, Governor’s Office of Emergency Services. 2022.
5. We’re revolutionizing public safety to save lives. Perimeter.
6. Van Riviere. Personal communication. Perimeter COO. August 14, 2024.
7. Dynamic evacuation response capabilities. Ladris.
8. Nightingale J. Focus on training: Human performance-based recommendations. FBI.gov. May 2023.
9. Abid SK, Sulaiman N, Chan SW, et al. Toward an integrated disaster management approach: How artificial intelligence can boost disaster management. MDPI. November 2021.
10. Weber J. Law enforcement’s role in responding to disasters: Aiding EM. American Military University. June 2024.

About the author

Troy Morton is a lieutenant with the El Dorado County Sheriff’s Office in Northern California, with over 22 years of law enforcement experience. Throughout his career, he has served in a variety of roles, including corrections, patrol, crime scene investigations, detective, civil/coroner sergeant, SWAT sniper team leader and patrol operations lieutenant. Currently Morton leads the County Office of Emergency Services, where he has served for the past 4½ years, overseeing all aspects of emergency and disaster management across El Dorado County. During that time, the county has navigated multiple large-scale wildfires, including the Caldor Fire, which is in the top 20 most destructive and largest wildfires in California’s history. Morton holds a bachelor’s degree in criminal justice management from Union Institute & University.

This article is based on research conducted as a part of the CA POST Command College. It is a futures study of a particular emerging issue of relevance to law enforcement. Its purpose is not to predict the future; rather, to project a variety of possible scenarios useful for planning and action in anticipation of the emerging landscape facing policing organizations.

The article was created using the futures forecasting process of Command College and its outcomes. Managing the future means influencing it — creating, constraining and adapting to emerging trends and events in a way that optimizes the opportunities and minimizes the threats of relevance to the profession.