Abstract
Pollution from wildfires constitutes a growing source of poor air quality globally. To protect health, governments largely rely on citizens to limit their own wildfire smoke exposures, but the effectiveness of this strategy is hard to observe. Using data from private pollution sensors, cell phones, social media posts and internet search activity, we find that during large wildfire smoke events, individuals in wealthy locations increasingly search for information about air quality and health protection, stay at home more and are unhappier. Residents of lower-income neighbourhoods exhibit similar patterns in searches for air quality information but not for health protection, spend less time at home and have more muted sentiment responses. During smoke events, indoor particulate matter (PM2.5) concentrations often remain 3–4× above health-based guidelines and vary by 20× between neighbouring households. Our results suggest that policy reliance on self-protection to mitigate smoke health risks will have modest and unequal benefits.
This is a preview of subscription content, access via your institution
Access options
Access Nature and 54 other Nature Portfolio journals
Get Nature+, our best-value online-access subscription
$32.99 / 30 days
cancel any time
Subscribe to this journal
Receive 12 digital issues and online access to articles
$119.00 per year
only $9.92 per issue
Buy this article
- Purchase on SpringerLink
- Instant access to full article PDF
Prices may be subject to local taxes which are calculated during checkout




Similar content being viewed by others
Data availability
The data to replicate all the results in the main text and supplementary material are available at https://github.com/echolab-stanford/wildfire-exposure-behavior-public.
Code availability
The code to replicate all the results in the main text and supplementary material is available at https://github.com/echolab-stanford/wildfire-exposure-behavior-public.
References
Landrigan, P. J. et al. The Lancet Commission on pollution and health. Lancet 391, 462–512 (2018).
Carleton, T. A. & Hsiang, S. M. Social and economic impacts of climate. Science 353, aad9837 (2016).
Aizer, A., Currie, J., Simon, P. & Vivier, P. Do low levels of blood lead reduce children’s future test scores? Am. Econ. J. Appl. Econ. 10, 307–41 (2018).
Deryugina, T., Miller, N., Molitor, D. & Reif, J. Geographic and socioeconomic heterogeneity in the benefits of reducing air pollution in the United States. Environ. Energy Policy Econ. 2, 157–189 (2021).
Burke, M., Hsiang, S. M. & Miguel, E. Global non-linear effect of temperature on economic production. Nature 527, 235–239 (2015).
Grönqvist, H., Nilsson, J. P. & Robling, P.-O. Understanding how low levels of early lead exposure affect children’s life trajectories. J. Polit. Econ. 128, 3376–3433 (2020).
US Department of Health and Human Services Theory at a Glance: A Guide for Health Promotion Practice (National Cancer Institute, 2005).
Greenstone, M. & Jack, B. K. Envirodevonomics: a research agenda for an emerging field. J. Econ. Lit. 53, 5–42 (2015).
Abatzoglou, J. T. & Williams, A. P. Impact of anthropogenic climate change on wildfire across western US forests. Proc. Natl Acad. Sci. USA 113, 11770–11775 (2016).
Burke, M. et al. The changing risk and burden of wildfire in the United States. Proc. Natl Acad. Sci. USA 118, e2011048118 (2021).
Hurteau, M. D., Westerling, A. L., Wiedinmyer, C. & Bryant, B. P. Projected effects of climate and development on California wildfire emissions through 2100. Environ. Sci. Technol. 48, 2298–2304 (2014).
Liu, J. C. et al. Particulate air pollution from wildfires in the western US under climate change. Climatic Change 138, 655–666 (2016).
Goss, M. et al. Climate change is increasing the likelihood of extreme autumn wildfire conditions across California. Environ. Res. Lett. 15, 094016 (2020).
Reid, C. E. et al. Critical review of health impacts of wildfire smoke exposure. Environ. Health Perspect. 124, 1334–1343 (2016).
Cascio, W. E. Wildland fire smoke and human health. Sci. Total Environ. 624, 586–595 (2018).
Xu, R. et al. Wildfires, global climate change, and human health. N. Engl. J. Med. 383, 2173–2181 (2020).
Zhou, X. et al. Excess of COVID-19 cases and deaths due to fine particulate matter exposure during the 2020 wildfires in the United States. Sci. Adv. 7, eabi8789 (2021).
Heft-Neal, S., Driscoll, A., Yang, W., Shaw, G. & Burke, M. Associations between wildfire smoke exposure during pregnancy and risk of preterm birth in California. Environ. Res. 203, 111872 (2021).
Santana, F. N., Gonzalez, D. J. & Wong-Parodi, G. Psychological factors and social processes influencing wildfire smoke protective behavior: insights from a case study in Northern California. Clim. Risk Manage. 34, 100351 (2021).
Rappold, A. et al. Smoke Sense initiative leverages citizen science to address the growing wildfire-related public health problem. GeoHealth 3, 443–457 (2019).
Reid, C. E. et al. Differential respiratory health effects from the 2008 northern California wildfires: a spatiotemporal approach. Environ. Res. 150, 227–235 (2016).
Kondo, M. C. et al. Meta-analysis of heterogeneity in the effects of wildfire smoke exposure on respiratory health in North America. Int. J. Environ. Res. Public Health 16, 960 (2019).
Wen, J. & Burke, M. Wildfire smoke exposure worsens learning outcomes. Preprint at EarthArXiv https://doi.org/10.31223/X52H06 (2021).
Wildfire Smoke: A Guide for Public Health Officials, 2019 Revision (US Environmental Protection Agency, 2019).
Pellert, M., Metzler, H., Matzenberger, M. & Garcia, D. Validating daily social media macroscopes of emotions. Preprint at arXiv https://doi.org/10.48550/arXiv.2108.07646 (2021).
Baylis, P. Temperature and temperament: evidence from Twitter. J. Public Econ. 184, 104161 (2020).
Baylis, P. et al. Weather impacts expressed sentiment. PLoS ONE 13, e0195750 (2018).
Hutto, C. & Gilbert, E. VADER: a parsimonious rule-based model for sentiment analysis of social media text. Proc. Int. AAAI Conf. Web Soc. Media 8, 216–225 (2014).
Wang, Z., Ye, X. & Tsou, M.-H. Spatial, temporal, and content analysis of Twitter for wildfire hazards. Nat. Hazards 83, 523–540 (2016).
Sachdeva, S., McCaffrey, S. & Locke, D. Social media approaches to modeling wildfire smoke dispersion: spatiotemporal and social scientific investigations. Inform. Commun. Soc. 20, 1146–1161 (2017).
Choi, H. & Varian, H. Predicting the present with Google Trends. Econ. Rec. 88, 2–9 (2012).
Goel, S., Hofman, J. M., Lahaie, S., Pennock, D. M. & Watts, D. J. Predicting consumer behavior with web search. Proc. Natl Acad. Sci. USA 107, 17486–17490 (2010).
Liang, Y. et al. Wildfire smoke impacts on indoor air quality assessed using crowdsourced data in California. Proc. Natl Acad. Sci. USA 118, e2106478118 (2021).
Miller, K. A. et al. Estimating ambient-origin PM2.5 exposure for epidemiology: observations, prediction, and validation using personal sampling in the Multi-Ethnic Study of Atherosclerosis. J. Expo. Sci. Environ. Epidemiol. 29, 227–237 (2019).
Shrestha, P. M. et al. Impact of outdoor air pollution on indoor air quality in low-income homes during wildfire seasons. Int. J. Environ. Res. Public Health 16, 3535 (2019).
Uejio, C. et al. Summer indoor heat exposure and respiratory and cardiovascular distress calls in New York City, NY, US. Indoor Air 26, 594–604 (2016).
Ferguson, L. et al. Exposure to indoor air pollution across socio-economic groups in high-income countries: a scoping review of the literature and a modelling methodology. Environ. Int. 143, 105748 (2020).
Bi, J., Wallace, L. A., Sarnat, J. A. & Liu, Y. Characterizing outdoor infiltration and indoor contribution of PM2.5 with citizen-based low-cost monitoring data. Environ. Pollut. 276, 116763 (2021).
Allen, R. W. et al. Modeling the residential infiltration of outdoor PM2.5 in the Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA Air). Environ. Health Perspect. 120, 824–830 (2012).
Krebs, B., Burney, J., Zivin, J. G. & Neidell, M. Using crowd-sourced data to assess the temporal and spatial relationship between indoor and outdoor particulate matter. Environ. Sci. Technol. 55, 6107–6115 (2021).
Lu, J. G. Air pollution: a systematic review of its psychological, economic, and social effects. Curr. Opin. Psychol. 32, 52–65 (2020).
Rappold, A. G. et al. Cardio-respiratory outcomes associated with exposure to wildfire smoke are modified by measures of community health. Environ. Health 11, 71 (2012).
Brulle, R. J. & Pellow, D. N. Environmental justice: human health and environmental inequalities. Annu. Rev. Public Health 27, 103–124 (2006).
Hajat, A. et al. Air pollution and individual and neighborhood socioeconomic status: evidence from the Multi-Ethnic Study of Atherosclerosis (MESA). Environ. Health Perspect. 121, 1325–1333 (2013).
Zheng, S., Wang, J., Sun, C., Zhang, X. & Kahn, M. E. Air pollution lowers Chinese urbanites’ expressed happiness on social media. Nat. Hum. Behav. 3, 237–243 (2019).
deSouza, P. & Kinney, P. L. On the distribution of low-cost PM2.5 sensors in the US: demographic and air quality associations. J. Expo. Sci. Environ. Epidemiol. 31, 514–524 (2021).
Sun, C., Kahn, M. E. & Zheng, S. Self-protection investment exacerbates air pollution exposure inequality in urban China. Ecol. Econ. 131, 468–474 (2017).
Currie, J., Voorheis, J. & Walker, R. What caused racial disparities in particulate exposure to fall? New evidence from the Clean Air Act and satellite-based measures of air quality. Preprint at National Bureau of Economic Research https://doi.org/10.3386/w26659 (2020).
Di, Q. et al. An ensemble-based model of PM2.5 concentration across the contiguous United States with high spatiotemporal resolution. Environ. Int. 130, 104909 (2019).
Reid, C. E., Considine, E. M., Maestas, M. M. & Li, G. Daily PM2.5 concentration estimates by county, zip code, and census tract in 11 western states 2008–2018. Sci. Data 8, 112 (2021).
O’Dell, K., Ford, B., Fischer, E. V. & Pierce, J. R. Contribution of wildland-fire smoke to US PM2.5 and its influence on recent trends. Environ. Sci. Technol. 53, 1797–1804 (2019).
Grainger, C. & Schreiber, A. Discrimination in ambient air pollution monitoring? AEA Pap. Proc. 109, 277–282 (2019).
Fowlie, M., Rubin, E. & Walker, R. Bringing satellite-based air quality estimates down to earth. AEA Pap. Proc. 109, 283–288 (2019).
Wooldridge, J. M. Introductory Econometrics: A Modern Approach (Cengage Learning, 2015).
Massicotte, P. & Eddelbuettel, D. gtrendsR, R package version 1.4.8.9000 https://github.com/PMassicotte/gtrendsR (2021).
Characteristics of People by Language Spoken at Home 2019: 2015–2019 American Community Survey 5-Year Estimates (US Census Bureau, 2020).
Squire, R. F. Measuring and correcting sampling bias in SafeGraph patterns for more accurate demographic analysis. Safegraph https://www.safegraph.com/blog/measuring-and-correcting-sampling-bias-for-accurate-demographic-analysis (2019).
Weill, J. A., Stigler, M., Deschenes, O. & Springborn, M. R. Social distancing responses to COVID-19 emergency declarations strongly differentiated by income. Proc. Natl Acad. Sci. USA 117, 19658–19660 (2020).
Valdez, D., Ten Thij, M., Bathina, K., Rutter, L. A. & Bollen, J. Social media insights into US mental health during the COVID-19 pandemic: longitudinal analysis of Twitter data. J. Med. Internet Res. 22, e21418 (2020).
Morawska, L. et al. How can airborne transmission of COVID-19 indoors be minimised? Environ. Int. 142, 105832 (2020).
Meager, R. Understanding the average impact of microcredit expansions: a Bayesian hierarchical analysis of seven randomized experiments. Am. Econ. J. Appl. Econ. 11, 57–91 (2019).
Vivalt, E. How much can we generalize from impact evaluations? J. Eur. Econ. Assoc. 18, 3045–3089 (2020).
2019 Tiger/Line Shapefiles (US Census Bureau, 2019).
GADM Data (GADM, 2018).
Acknowledgements
We thank the Robert Wood Johnson Foundation and Stanford’s Center for Population Health Sciences for funding (M.B., S.H.-N., J.L. and A.D.), SafeGraph for data access, Stanford University and the Stanford Research Computing Center for computational resources and support, and members of the ECHOLab and seminar participants at Cornell, Columbia, MIT, Stanford, UC Berkeley and UC Santa Barbara for helpful comments. The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.
Author information
Authors and Affiliations
Contributions
All authors contributed to the conception and design of the study. S.H.-N., J.L., A.D., J.W. and M.L.C. constructed the smoke data. J.L. and A.D. constructed the Google search data. P.B. constructed the Twitter dataset. M.S. and J.A.W. constructed the mobility dataset. S.H.-N. constructed the PurpleAir dataset. M.B. and S.H.-N. led the econometric analysis. All authors contributed to analysing the results and writing the paper.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Human Behaviour thanks Priyanka deSouza, Francesca Dominici and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Extended data
Extended Data Fig. 1 Counties included in analyses that use EPA pollution monitors, and correlation in smoke PM2.5 between monitor pairs as a function of distance between monitors.
a. Counties in red are those with EPA pollution monitors from which we construct smoke PM2.5 measures for the behavioral analyses. b Colors depict a heatmap of the 85,102 pairwise correlations, with lighter colors depicting areas with more data and shown in legend at right; solid black line is the median correlation at each distance. Sample is restricted to stations with at least 1000 days of data. Mean width of counties in our data is 55 km, and mean width of metro areas is 228 km. Source for a: US Census Bureau.
Extended Data Fig. 2 Time spent indoors at home in America.
Data are from repeated rounds of the American Time Use Survey. Top panels show data by average income, age, season, and race/ethnicity. Bottom map shows averages by state across survey rounds. Source for e: US Census Bureau.
Extended Data Fig. 3 Effect of smoke PM2.5 on different mobility measures, and heterogeneity by income.
a Percent of mobile phones estimated to be completely at home on a given day at the US county level, 2019–2020. Black lines are regression point estimates from spline fits conditional on fixed effects, with shaded areas showing bootstrapped 95% confidence intervals. Number of observations in each regression is shown in upper left corner of each panel. Histograms at the bottom show the log distribution of smoke PM2.5 exposure in each sample. b Same but for % fully away from home on that day. c-d Effect of smoke PM2.5 on mobility as a function of income. Lines show the marginal effect of a heavy smoke exposure (50ug smoke PM2.5 on that day) on percent of individuals completely at home on that day (c) or completely away from home (d), as a function of median household income in that county. Colors represent models run with either date fixed effects (blue) or state-by-date fixed effects (orange). Dark lines show regression point estimates, shaded area the bootstrapped 95% CI.
Extended Data Fig. 4 Exposure to average and acute smoke PM2.5 at the county level does not differ systematically by income.
Daily smoke PM2.5 exposures by income decile across US counties, 2006-2020. Dots represent daily observations where smoke PM2.5 was non-zero. Plot is truncated at 300ug for clarity; not plotted are 71 days (0.001% of the sample) in which smoke PM2.5 exceeded 300. Statistics at right show the percent of observations across the study period with daily smoke PM2.5 observations above the listed value.
Extended Data Fig. 5 Higher income US census tracts are more likely to have PurpleAir monitors.
Grey bars show the distribution of tract-level median household income across all US census tracts in the contiguous US, red bars the income in tracts with at least one outdoor PurpleAir sensor, and blue bars the income in tracts with at least one indoor PurpleAir sensor. Vertical lines give the median of each distribution.
Extended Data Fig. 6 Infiltration estimates are highly correlated across alternate statistical models and methods of deriving PM2.5 concentrations from Purple Air data.
Correlation between infiltration estimates from statistical models with different lag structures and different PM2.5 concentration estimates (see Supplementary Table 11 for details). 1a is our preferred specification presented in the main results.
Extended Data Fig. 7 Understanding variation in household-specific infiltration estimates.
a Posterior estimates of monitor-specific infiltration rates from a Bayesian hierarchical model are very similar to “raw” estimates from our monitor-specific time-series regressions, indicating that true heterogeneity rather than sampling noise is what is driving observed differences in estimated infiltration. b Ability of random forest (RF) or gradient boosted trees (GBT) model to explain variation (r2) in infiltration across monitors remains low; models use predictors in (c). c For each predictor, we calculate the effect on infiltration of moving from the 5th to the 95th percentile of that predictor in the test dataset, holding the other predictors constant at their average value in the test dataset; estimates are shown for RF and GBT models and for four alternate spatial buffers used to construct housing predictors. Housing Index is constructed by averaging standardized values of home value, number of stories, number of baths, number of bedrooms, height, and area. A/C measures the inverse distance weighted proportion of matched CoreLogic houses that have air conditioning. Median Income is the median household income in the Census tract population. Race variables (i.e. all demographic covariates except Hispanic) are measured among the non-Hispanic/Latino population. AI/AN stands for American Indian and Alaska Native. NHPI stands for Native Hawaiian and other Pacific Islander. HDD and CDD stand for heating degree days and cooling degree days, respectively. d Range of household-level infiltration estimates for the full sample and for sub-samples when behavior (ie opening/closing of doors, use of air purifier) is expected to matter less: when it’s raining, nighttime, and periods when it is cold ( < 10∘C) and low PM2.5 ( < 30μg/m3). e-f during periods when behavioral factors are more likely minimized, infiltration varies more strongly with income and housing age.
Extended Data Fig. 8 Monitor-specific infiltration estimates using indoor/outdoor ratios versus regression-based approaches.
I/O estimates are only modestly correlated with our preferred regression-based estimates that measure the marginal effect on indoor PM2.5 concentrations of a unit increase in outdoor concentrations. For each monitor I/O ratio was calculated across all observations with hourly indoor PM2.5 less or equal to outdoor PM2.5. δI/δO was estimated as described in Methods.
Extended Data Fig. 9 Outdoor and indoor PM2.5 concentrations on a smoke day in CA.
Very similar outdoor PM2.5 concentrations during a smoke event on Aug 20th, 2021 over a high-income area of the peninsular Bay Area were associated with widely varying contemporaneous indoor PM2.5 concentrations. © OpenStreetMap contributors.
Extended Data Fig. 10 Variation in indoor PM2.5 across monitors with similar outdoor PM2.5 during the Aug/Sep 2020 smoke event in the Bay Area.
Each dot is average outdoor PM2.5 and average indoor PM2.5 for an individual monitor in the Bay Area over the Aug/Sep 2020 smoke event, with monitors grouped into 5 μg/m3bins based on outdoor exposure. Numbers at top show the number of monitors in each bin (black), and the ratio of maximum to minimum indoor PM2.5 across monitors within each outdoor PM2.5 bin (red). Monitors with average outdoor PM2.5 exposures within 5 μg/m3of each other experienced > 20x differences in indoor PM2.5 exposures.
Supplementary information
Rights and permissions
About this article
Cite this article
Burke, M., Heft-Neal, S., Li, J. et al. Exposures and behavioural responses to wildfire smoke. Nat Hum Behav 6, 1351–1361 (2022). https://doi.org/10.1038/s41562-022-01396-6
Received:
Accepted:
Published:
Issue date:
DOI: https://doi.org/10.1038/s41562-022-01396-6
This article is cited by
-
Socially vulnerable communities face disproportionate exposure and susceptibility to U.S. wildfire and prescribed burn smoke
Communications Earth & Environment (2025)
-
A unified framework integrating psychology and geography
Nature Human Behaviour (2025)
-
Roof renewal disparities widen the equity gap in residential wildfire protection
Nature Communications (2025)
-
Behavioral Responses to Wildfire Smoke: A Case Study in Western Montana
Journal of Community Health (2025)
-
The impact of wildfire smoke exposure on excess mortality and later-life socioeconomic outcomes: the Great Fire of 1910
Cliometrica (2025)