Milo Smith, a lifelong resident of Southern California, has been driven by a deep-seated motivation to tackle the challenges posed by natural disasters with innovative solutions. With a strong connection to his hometown and years of firsthand experience enduring the smoke haze of wildfire seasons, Milo has dedicated himself to the cause of wildfire prevention and detection. Drawing upon his expertise in machine learning, he has embarked on a journey to address the pressing issues surrounding wildfires and make a meaningful impact on the safety of his community and beyond. After extensive research and analysis of existing tools, Milo presents here his findings along with his own alerting solution, which he made available to the public for free. 

Some of the services mentioned in this blog are for-profit. The Fire Networks and FAC Net do not directly or indirectly endorse any of the listed services; this blog represents the opinions of the author.

 

Getting ready for a wildfire isn’t a last-minute effort; rather, it’s a continuous process, even a lifestyle, that incorporates practical steps to reduce a fire’s risks and adverse effects. Over time, numerous technological advancements have emerged, significantly enhancing the ability of communities to prepare more effectively for wildfires. Among these innovations are several notable technological tools including satellite imagery, sensors, infrared cameras, drones utilized for data collection, machine learning (ML), and artificial intelligence (AI), employed for data post-processing and generating actionable insights.

Various companies, governmental bodies, and nonprofit organizations have harnessed these technologies to develop final products to assist communities in addressing fire-related challenges. In this blog post, I seek to provide an overview of several such solutions and evaluate their potential benefits.

 

AirNow

AirNow is a result of a partnership between the U.S. Environmental Protection Agency, the National Oceanic and Atmospheric Administration (NOAA), the National Park Service, NASA, the Centers for Disease Control, and tribal, state, and local air quality agencies. AirNow’s data provides insights into current fire conditions, including fire locations, smoke plumes, and air quality data from permanent and temporary air quality monitors. This data is readily available through their website, which also offers interactive maps.

Image of the AirNow interface, which shows a map of the US with various colored blobs representing air quality.

AirNow: an interactive map of air quality.

To collect the data, AirNow and its partners employ air sensors and CCTVs. Sensors take inputs such as temperature, humidity, air pressure, and gas levels. AirNow collects this data from PurpleAir sensors, stationary long-term monitoring stations operated by state, local or tribal air quality agencies, and temporary monitors for monitoring smoke events. CCTVs or webcams are also used to capture images, like images of smoke or fire glow.

Understanding the importance of air pollution exposure is crucial, as the deterioration of air quality is a serious concern for human health. This emphasizes the critical necessity of creating robust air quality monitoring networks and effective information dissemination systems, all with the overarching goal to safeguard public health and minimize smoke exposure during or before fire incidents.

 

Watch Duty

Watch Duty functions both as a non-profit organization and as an app sharing its name. Its primary role involves gathering information on wildfires from official sources, volunteer reports, and radio communications. Subsequently, the collected data undergoes a filtration process, and in times of emergencies, the organization issues wildfire alerts to app users.

User interface for WatchDuty, which shows a map of the western US with a flame icon located at fires in various areas.

WatchDuty’s map of active fires

A key data source harnessed by Watch Duty is the information intercepted via radio scanners. When a 911 signal is intercepted, Watch Duty orchestrates team collaboration to cross-check this data with other sources to confirm the presence of wildfires. The end result is accessible to users and includes real-time alerts and comprehensive wildfire maps.

Emergency alerts are crucial in preparing us for impending dangers and ensuring that we take the necessary actions, such as evacuating an area when required. Like the other tools listed, Watchduty is not an official emergency alert system, but it’s a valuable addition to consider.

 

Frontline Wildfire Tracker App

Frontline Wildfire Tracker is an application that offers practical support and information during wildfire seasons. It provides various tools, such as customizable wildfire maps based on the user’s home address, alerts for fire weather watches, lightning events, active wildfires, and evacuation notices. Additionally, users can create groups within the app to share essential notifications during wildfire events and evacuations.

Frontline Wildfire's user interface.

Frontline Wildfire’s app: evacuation assistance

The app employs technology that utilizes satellite imagery to identify wildfires, although a certain time lag is involved. Consequently, the app is primarily designed to gather pertinent information to aid in wildfire preparedness, such as evacuation planning and finding shelter. These fire maps are accessible within the app and also on the website.

 

Augurisk

Augurisk is a platform designed for assessing natural hazard risks. It provides predictive analysis and risk scoring for various events, including wildfires. The free app offers information about the vulnerability of specific U.S. locations to environmental emergencies and crime. Additionally, it conducts risk assessments on a county level, considering factors like climate change and other relevant aspects.

User interface of Augurisk.

Augurisk’s risk assessment for LA County.


The platform’s proprietary data is subsequently processed through
machine learning (ML) prediction algorithms to produce risk scores for specific areas. Augurisk’s data is available in various formats, including reports and benchmarks, which can be obtained through their API. Augurisk can help the public stay well-informed about the potential risks posed by impending wildfire.

Due to the proprietary nature of the platform’s data sources, there may be a bias and DEI issues may not be fully accounted for.

 

FireSafe

FireSafe is a free, real-time wildfire alert bot in California powered by social media. As soon as a fire outbreak occurs in California, the FireSafe bot swiftly captures and analyzes relevant social media posts, identifies their locations, and promptly issues alerts indicating the potential wildfire’s location. These alerts are delivered in real-time, almost without any delay between data acquisition and alert issuance. They also offer extensive coverage and precision down to the coordinates, enabling users to accurately gauge their proximity to the wildfire and take all necessary precautions.

User interface for FireSafe.

FireSafe’s alert message for an ongoing fire.

The FireSafe bot harnesses the capabilities of social media. When a wildfire ignites, individuals frequently share posts regarding the initial indicators they’ve observed, such as smoke, ashes, or visible flames. The FireSafe bot captures these messages, employs its Machine Learning (ML) algorithms to evaluate their relevance to the wildfire, geotags them to pinpoint the wildfire’s exact location, and promptly issues an alert once processing is finished.

Real-time wildfire alerts play a pivotal role because detecting wildfires in their early stages significantly mitigates the adverse consequences of the fire. This timely awareness empowers communities to implement essential measures while enabling first responders to swiftly and efficiently coordinate their efforts.

 

While wildfires are a natural occurrence, advancements in technology aimed at wildfire prevention, detection, and containment play a crucial role in managing these incidents and mitigating their negative impacts. Social media, machine learning, satellite imagery, and sensors each possess their unique strengths and weaknesses, yet collectively they work toward a common objective: ensuring constant preparedness for wildfire risks.

 

****