Smoke on the horizon: leveling up citizen and social science to motivate health protective responses during wildfires

Increased particulate matter pollution, especially related to wildfire smoke exposure, has been linked to a broad range of health outcomes including higher rates of cardiorespiratory emergency visits, hospitalizations, and even death (Adetona et al., 2016; Black et al., 2017; Dennekamp and Abramson, 2011; Dodd et al., 2018; Haikerwal et al., 2015; Liu et al., 2015; Morgan et al., 2010; Rappold et al., 2011; Reid et al., 2016; Wettstein et al., 2018), kidney (Wyatt et al., 2020) and cognitive (Cleland et al., 2022) impacts. As the risks from large wildfires grow on the landscape with hotter and drier climate conditions, a directly related public health danger looms on the horizon in the form of greater and more frequent wildfire smoke exposure. Smoke exposure extends well beyond fire perimeters and jurisdictional boundaries and results in billions of dollars in health burden (Fann et al., 2018; Jones, 2017; Kochi et al., 2010). Current approaches to mitigation of wildfire smoke health risks are projected to have 'modest and unequal benefits' (Burke et al. 2022). Wildfire smoke exposure has a range of impacts; from those that seem minor or nuisances such as upper respiratory, ear-nose-throat symptoms, or scratchy eyes to headaches, depression, anxiety, and impaired sleep to the most extreme outcomes previously described. The heterogeneity of effects on just respiratory health alone makes it difficult to characterize demographic subgroups that are more susceptible to adverse outcomes (Kondo et al., 2019). Simply stated, no one in the vicinity of a fire; young or old, healthy or in poor health, can completely escape the ramifications of smoke exposure including an increasing public health burden. Frequent wildfire smoke from events near or far have made fire season an everyday life consideration in some communities, while others are just beginning to experience impacts. Significant population growth has occurred in recent decades along the Wildland-Urban Interface (WUI; the zone where human structures and wildland intersect), exacerbating the risk of fire and smoke exposure (Peterson et al., 2021). These trends highlight the need to identify health risk communication approaches that better motivate individuals to change behavior and adopt health risk mitigation strategies. Although recommended health responses focus on reducing the amount and time of smoke exposure, audiences may need additional context to move from intentions to actions, especially if those actions are perceived as difficult, costly, or more of a change than they are initially willing to consider. In this manuscript, we use the Smoke Sense citizen science project as a platform to demonstrate and test how social science concepts can be used to facilitate motivation, engagement, and provide efficacy to users contemplating protective action. The goal is to contribute to efforts to reduce the public health burden by closing the gap between intentions and actions that occurs when experiencing smoke events. Specifically, we draw upon the behaviors and experiences observed among participants who have experienced smoke exposure in previous years. We describe an approach to reduce barriers (friction) that slow or prevent people from implementing protective exposure reducing behaviors. Similar approaches have been used in behavioral medicine, for example to design interventions to improve exercise adherence and outcomes in cardiac rehabilitation programs (Sniehotta et al., 2005) or healthy behaviors that facilitate weight loss maintenance (Asbjornsen et al., 2022; Asbjornsen et al., 2020). Methods proposed here are rooted in established social and behavioral science theory but have not been linked to behavioral patterns during smoke exposure episodes. For each concept, we identify specific factors contributing to the intention-action gap and connect them to a relevant theoretical framework.

Approach and platform: Smoke Sense is a citizen science project launched by the U.S. Environmental Protection Agency to investigate and educate about wildfire smoke, health, and protective actions via a smartphone app. The platform delivers current and forecast air quality visualizations, educational content on protective behaviors, and enables users to share smoke experiences, symptoms, and exposure-reducing actions. With input from over 60,000 participants, prior iterations informed gaps in health risk communication and adoption of protective behaviors. Behavioral science pillars and proposed features: The authors propose integrating four social/behavioral science domains to bridge intention–action gaps during smoke events: (1) Social norms: Provide dynamic and descriptive norm cues via an in-app newsfeed and map showing aggregated local behaviors (e.g., number of people using HEPA filters in a ZIP code), and personal “My Data” views to compare individual actions to community patterns. (2) Positive reinforcement: Incorporate immediate and sustained reinforcers, such as acknowledgment messages after report submission, points/stars, badges/levels, goals (e.g., complete 10 reports), map visualization of shared reports, rotating newsfeed, push notifications, and privileges for engaged users. (3) Reducing friction and bundling benefits: Pair commonly checked information (weather) with air quality on the home screen; organize educational content tailored to identities/interests (e.g., caregivers, outdoor athletes, asthmatics); enable easy sharing of AQ information and protective actions; allow tracking of up to four locations relevant to the user. (4) Self-efficacy: Use messaging, community-shared actions, and the “My Data” diary (calendar/history of reports and actions) to foster users’ belief that achievable, incremental behaviors (e.g., closing windows, using HEPA filters, adjusting exertion) can reduce exposure. Preliminary user feedback procedure: New feature concepts were rendered as an interactive wireframe. Sonoma Technology, Inc. conducted mixed feedback sessions with 15 potential users (9 female, 6 male; ages 31–72; 5 >60 years; 2 with asthma; 3 affected by the Cedar Creek Fire in Oakridge, Oregon, Oct 2022). Ten unmoderated participants were recruited via an online platform; the remainder participated in moderated sessions. Feedback focused on motivations and use contexts, informing design refinements. Monitoring and evaluation plan: Success metrics will include rates and patterns of protective behaviors reported, engagement with the “My Data” module, attainment of default reporting goals (e.g., 10 reports), report frequency per day/week/month, time spent on reports, shared content, push notification and newsfeed preferences, downloads and session analytics, and correlations with external events (e.g., megafires, media coverage, partner outreach). Iterative updates will be guided by usage analytics and user feedback, including engaging high-contributing users to prioritize feature development.

- The paper proposes, rather than tests, a behavioral-science–informed citizen science approach to motivate health-protective actions during wildfire smoke events, leveraging social norms, positive reinforcement, friction reduction/bundling, and self-efficacy. - Prior Smoke Sense participation exceeded 60,000 users, revealing gaps in perceived efficacy and adoption of recommended behaviors and motivating the proposed redesign. - Preliminary user feedback (n=15; 9F/6M; age 31–72; 2 with asthma; 3 impacted by Cedar Creek Fire) indicated: (a) motivation increases when the real-world use of contributed data is salient; (b) users in poor air quality conditions want immediate condition severity and actionable guidance; users in good air quality prioritize forecasts; (c) preference for frequent, brief checks; (d) some users value gamification/recognition; (e) preference for a single primary pollutant display (PM2.5) rather than multiple pollutants. - The authors hypothesize that presenting local/dynamic norms (e.g., counts of nearby users taking specific protective actions), providing immediate social/natural rewards, bundling weather and AQ information, and supporting personal progress tracking will increase timely adoption of exposure-reducing behaviors and habit formation.

The manuscript discusses the growing public health burden from increasingly frequent and severe wildfire smoke due to climate-driven drought/heat and WUI growth. It argues that information provision alone is insufficient to drive sustained behavior change. By integrating social science strategies into a citizen science app, the approach aims to create a digital community that normalizes protective actions, enhances self-efficacy, and reduces friction to act. Customization, notifications, and iterative feature development informed by engaged users are intended to maintain engagement. Citizen science is positioned as both a democratizing tool for data collection/education and a means to foster social support networks that enhance preparedness and resilience. Over time, familiarity with air quality patterns and community-shared responses can reduce uncertainty and motivate appropriate protective behaviors.

Environmental disasters, including wildfire smoke episodes, demand multifaceted, timely, and context-aware responses. Leveraging citizen science and crowdsourcing aligns with U.S. open government directives to provide actionable information and engage the public in knowledge building. Smoke Sense serves as a platform to iteratively develop and test health risk communication strategies rooted in epidemiologic, clinical, and behavioral science, crowdsourcing problem formulation from 60,000+ participants. The authors propose a next-step iteration that applies social norming, positive reinforcement, friction reduction/bundling, and self-efficacy to motivate engagement, participation, and protective behaviors during smoke events. The concept can extend to other environmental disasters, informing user-centered communication strategies that build resilience and improve public health responses.

- The paper is primarily a conceptual/proposal piece; proposed features and behavior-change mechanisms have not yet been empirically evaluated within the app for causal impact on protective actions. - Preliminary user feedback was from a small convenience sample (n=15) and may not generalize across demographics or regions. - Many protective behaviors occur in private, making normative feedback and adoption rates harder to validate without broader deployment and measurement. - Data sharing is not applicable to this paper as no new datasets were generated or analyzed; therefore, quantitative evidence of effectiveness is not provided. - Implementation and impact may vary with external factors (e.g., timing/severity of smoke events, media coverage, access to tools like HEPA filters/N95s), potentially limiting generalizability.