A Review of Community Smoke Exposure from Wildﬁre Compared to Prescribed Fire in the United States

: Prescribed ﬁre, intentionally ignited low-intensity ﬁres, and managed wildﬁres—wildﬁres that are allowed to burn for land management beneﬁt—could be used as a land management tool to create forests that are resilient to wildland ﬁre. This could lead to fewer large catastrophic wildﬁres in the future. However, we must consider the public health impacts of the smoke that is emitted from wildland and prescribed ﬁre. The objective of this synthesis is to examine the differences in ambient community-level exposures to particulate matter (PM 2.5 ) from smoke in the United States in relation to two smoke exposure scenarios—wildﬁre ﬁre and prescribed ﬁre. A systematic search was conducted to identify scientiﬁc papers to be included in this review. The Web of Science Core Collection and PubMed, for scientiﬁc papers, and Google Scholar were used to identify any grey literature or reports to be included in this review. Sixteen studies that examined particulate matter exposure from smoke were identiﬁed for this synthesis—nine wildland ﬁre studies and seven prescribed ﬁre studies. PM 2.5 concentrations from wildﬁre smoke were found to be signiﬁcantly lower than reported PM 2.5 concentrations from prescribed ﬁre smoke. Wildﬁre studies focused on assessing air quality impacts to communities that were nearby ﬁres and urban centers that were far from wildﬁres. However, the prescribed ﬁre studies used air monitoring methods that focused on characterizing exposures and emissions directly from, and next to, the burns. This review highlights a need for a better understanding of wildﬁre smoke impact over the landscape. It is essential for properly assessing population exposure to smoke from different ﬁre types.


Introduction
Wildfire has long been an important ecological process of our natural world, only requiring three ingredients-fuel, oxygen, and heat [1]. Prior to European settlement, many forests in the United States were historically shaped by wildfires [2]. Native Americans historically used wildfire as a vegetation management tool to increase density of edible plants, provide material for basketry, and control insects and plant diseases [3]. Historically, in the Western US, frequent fires of low severity burned on the forest floor and resulted in coniferous forests that are more vulnerable to the effects of fire [4]. In California, ecological health, but this relation is confounded by smoke exposure [27]. Understanding relative risk from fire management actions is essential to informed protection of public health.
The objective of this synthesis is to examine the differences in ambient community-level exposures from smoke in the United States from two smoke exposure scenarios-wildfire and prescribed fire.
Several key questions will be addressed: (1) What are the PM 2.5 concentration differences between prescribed fire and wildfire smoke exposures? (2) How do PM 2.5 concentrations from each exposure scenario compare to the National Ambient Air Quality Standards (NAAQS)? (3) How long are communities exposed to PM 2.5 during each exposure scenario? This synthesis will provide public health practitioners, air quality regulators, and natural resource managers with more information on the exposure differences of smoke exposure from wildfire compared with prescribed fire. Ultimately, this information can be used to understand and quantify the health risks associated with smoke exposure from wildfire compared with prescribed fire.

Materials and Methods
A systematic search was conducted to identify scientific papers from peer-reviewed journals to be included in this review. The systematic search followed the Guidelines for Systematic Review and Evidence Synthesis in Environmental Management [28].
The Web of Science Core Collection and PubMed, for scientific papers, and Google Scholar were used to identify any grey literature or reports to be included in this review. The search strategy used the following search terms-wildfire, wildland fire, prescribed fire, grass fire, peat fire, prescribed managed fire, prescribed natural fire and smoke, exposure assessment, air quality. For each search that was performed, we recorded the search date, search terms that were used, database that was searched, and titles that were returned from the search.
The synthesis was restricted to scientific papers that met the following inclusion criteria: (1) studies that were conducted in the United States and (2) reported PM 2.5 concentrations during specific wildfire or prescribed fire events. Studies were appraised for the quality of the methods used for air monitoring or modeling used for concentration estimation. Studies that reported only PM 2.5 occupational exposures during a wildfire or prescribed fire event were not included.
The systematic search resulted in 271 journal articles from PubMed, with 229 unique titles, and 2023 journal articles from Web of Science, with 1093 unique titles ( Figure 1). Once merged, there were 1449 unique scientific journal articles. Next, we reviewed the journal titles and selected 79 relevant articles. During the title review, reasons for articles to be excluded included: (1) were not conducted in the United States; (2) indicated a focus on developing models to estimate PM 2.5 emissions, source apportionment, or plumes; (3) conducted an occupational exposure study; (4) measured other air contaminants; (5) indicated that they were conducted in a laboratory. Of the selected articles, we reviewed their abstracts for extractable information that was relevant to the synthesis objectives. Based on the information provided in the abstracts, such as study methods and results, we selected the article to be further reviewed by reading the full article (N = 34). Sixteen peer-reviewed scientific journal articles met the study criteria and were included in this synthesis.
From each selected journal article, information was extracted and inputted into a table for comparison and analysis (Table 1). Extracted data from each article included: information on the wildfire or prescribed fire event name and date range, reported concentration mean and range, number of reported days that exceeded the NAAQS 24-h standard (PM 2.5 concentration ≥ 35 µg m −3 ) [29], number of days sampled, the data source of the reported concentrations, and what type of average concentration average or sampling time was used for each study.

Results
The systematic review identified 16 studies that characterized exposures to PM 2.5 from wildfire and prescribed fire events (Table 1). Generally, studies directly measured PM 2.5 concentrations with existing air monitoring networks or temporary monitoring stations placed in communities that were deployed specifically for fire events. Although there were studies that attempted to model concentrations of PM 2.5 from wildfire or prescribed fire smoke, they did not report PM 2.5 concentrations associated with a specific fire event and did not meet the inclusion criteria.
The systematic search identified nine scientific studies that examined exposure to PM 2.5 from wildfire smoke. The studies covered a wide geographic area and were focused on wildfires that occurred in California, Montana, the Pacific Northwest, and Canada that impacted major cities in the United States. The selected papers reported PM 2.5 concentrations from several large wildfires (region-wide events), occurring at one period or during specific wildfire events.  [31,37]. On average, PM 2.5 concentrations from wildfires were sampled and reported for 30 days; events ranged from 2 to 77 days. During wildfire events, the number of days that exceeded the NAAQS ranged from 2 to 47 days and averaged 11 days. The PM 2.5 concentrations from the Tripod Fire smoke in Eastern Washington resulted in 47 days that were above the NAAQS [34].
Seven scientific studies were identified that measured exposure to PM 2.5 at prescribed fires in Arizona, Georgia and South Carolina. Six studies used air monitoring equipment to measure PM 2.5 concentrations, while one study Hu et al. (2008) [43] simulated PM 2.5 concentrations using fire and atmospheric conditions from a specific prescribed fire event. Almost all sampled prescribed fires were performed as broadcast burns, where fire was applied directly across a predetermined area and was confined to that space. One sampled prescribed fire was conducted as a pile burn operation, where only piles of cut vegetation are ignited and burned [44]. Naeher et al. (2006) and Achtemeier et al. (2006) [41,42] reported PM 2.5 concentrations from the same prescribed fire event where researchers examined the effects of mechanical chipping on smoke measurements. The size of the prescribed fires ranged from 1 to 1200 ha, with the largest event being two adjacent prescribed fires in the Southeast United States, outside of Atlanta (Hu et al., 2008) [43].
Generally, the prescribed fire air sampling occurred during the burn operation and monitors were placed inside or next to the fire perimeter. For example, Robinson et al. (2011) [44] placed monitors next to the fire perimeter on Day 1 of sampling and inside the fire perimeter on Day 2 to capture emissions during the smolder phase of the fire. Naeher et al. (2006) and Achtemeier et al. (2006) [41,42] also placed monitors inside the prescribed fire and along the fire perimeter on the downwind side of the prescribed fire burn unit. Pearce et al. (2012) [45] measured concentrations using a grid of 18 monitors that were placed 10-12 km on the downwind side of the prescribed fire burn unit. Hu et al. (2008) [43] was the only study to report PM 2.5 concentrations from a prescribed fire in an urban center-Atlanta, Georgia-which was 80 km from the prescribed fire.
Reported mean concentration of PM 2.5 from the selected studies ranged from 37.8 µg m −3 , in Atlanta, Georgia, to 3000 µg m −3 at a prescribed fire in Arizona [43,44]. Additionally, the same prescribed fire in Arizona during the flaming phase produced the highest maximum PM 2.5 concentration of 8357 µg m −3 [44]. Only Hu et al. (2008) [43] examined the impacts of a prescribed fire on NAAQS exceedances and reported that one day exceeded the NAAQS (24 h mean = 37.8 µg m −3 ) during the prescribed fire event. Unlike the wildfire studies that generally used a consistent averaging time (24 h), prescribed fire studies averaged concentration over many different time periods. Averaging times ranged from 1.5-2 h samples to a four-day total average.

Discussion
Due to differences in study objectives and methodology, PM 2.5 concentrations from wildfire smoke were found to be lower than reported PM 2.5 concentrations from prescribed fire smoke. Although the acres burned on wildfires was up to 100 times larger, monitoring location, distance and concentration averaging time was shown to have an impact on the reported PM 2.5 concentrations. Wildfire studies focused on assessing air quality impacts to communities that were close to the fire (for example 12-36 km) and urban centers that were far from the wildfire. However, prescribed fire studies used air monitoring methods that focused on characterizing PM 2.5 exposures and emissions directly from, and next to, the burns site.
Wildfire and prescribed fire smoke exposure, similar to other emissions, is dependent on proximity to the source. Wildfire studies that were examined measured smoke at locations that ranged from 7 to 242.8 km from the wildfires, while prescribed locations ranged from next to the burn perimeter (0 km) and up to 80 km away from the burn. The dependence on proximity and smoke direction was demonstrated by Burley et al. (2016) [36], showing that megafires, such as the Rim and King fires, largely missed their monitoring site due to smoke plume direction, while the smaller and closer Aspen Fire transported more directly and had the highest exposure impacts at Devils Postpile National Monument. Hu et al. (2008) [43] was the only prescribed fire study identified that assessed the air quality impact from PM 2.5 to a large urban area. The 24-h PM 2.5 concentration in an urban area (Atlanta, Ga) that was estimated from this prescribed burn was 37.8 µg m −3 and in the range of the measured wildfire concentrations. In addition, the distance of the burn (80 km) was also similar to the monitor distance for wildfires.
The selected wildfire studies largely reported PM 2.5 mean concentrations that were generally averaged over a 24 h time period. However, the prescribed fire studies reported mean concentrations that were sampled over time periods ranging from 1-96 h. The short duration prescribed fire sampling events resulted in mean concentrations (198.1-3000 µg m −3 ) that were higher than the prescribed fires that reported 22-24 h average PM 2.5 concentrations (37.8-74.01 µg m −3 ). The shorter prescribed fire sampling events captured the periods of higher smoke emissions, while the longer averaging time for wildfire studies resulted in lower mean PM 2.5 concentrations.
Wildfire exposures are often episodic and short-term, but if they happen often, over a course of a fire season over many years, they could be considered long-term exposures. From the studies that were reviewed, the wildfire events that were included occurred over multiple weeks and months, while the prescribed fire events occurred over a few days. The duration of an event is important to consider because the longer exposure durations can lead to higher cumulative exposures to air contaminants [46]. This review highlights the lack of consistent information about exposures to PM 2.5 from fire smoke, especially from prescribed fires. Monitoring for prescribed fire was more focused on capturing the smoke emission directly next to the fire and not downstream from the burn, while wildfire studies either used existing urban sites and/or monitored for sensitive receptors. There were many studies identified during the initial search that have assessed smoke from wildfires or prescribed fires, but there were few studies that directly reported concentrations of PM 2.5 to meet the inclusion criteria. Characterization of PM 2.5 air quality impacts to communities from prescribed fire smoke is needed to better understand how PM 2.5 exposures are different compared to those of wildfires. Prescribed fire exposure studies should be designed to examine emissions directly from the burn but also consider and measure the impacts on downwind communities. Additionally, one could use an area of the United States that is prone to frequent wildfires and estimate exposure through modeling from recent specific wildfires and prescribed fires to examine exposure differences. This approach was suggested by Baker et al. (2016), as it would lead to better model inputs for fire size and emissions, and could be validated against an existing monitoring network [47]. An additional approach that could be used would be a health impact assessment used by Fann et al (2018) [24] to estimate the incidence and economic value of human health impacts attributable to wildfire smoke compared to prescribed fire smoke [24]. Lastly, improved exposure estimates could be used to quantify the risk of adverse health effects from each of these different exposure scenarios [48].

Conclusions
Destructive wildfires have higher rates of biomass consumption and have greater potential to expose more people to smoke than prescribed fires. Naturally ignited fires that are allowed to self-regulate can provide the best scenario for ecosystem health and long-term air quality. Generally, prescribed fire smoke is much more localized, and the smoke plumes tend to stay within the canopy, which absorbs some of the pollutants, reducing smoke exposure. Land managers want to utilize prescribed fire as a land management tool to restore fire-adapted landscapes. Thus, additional work is needed to understand the differences in exposures and public health impacts of smoke of prescribed fire compared to wildfire. One way to do this would be for managers to collaborate with air quality departments (internal to agency or external) to monitor PM 2.5 concentrations in communities near a prescribed fire.
Consistent monitoring strategies for all wildland fires, whether prescribed or naturally occurring, are needed to allow the most robust comparative analysis. Currently, prescribed fire monitoring is often focused on capturing the area of highest impact or characterizing fire emissions, while wildfire monitoring often relies on urban monitors supplemented by temporary monitoring of communities of concern. A better understanding of smoke impact over the landscape and related impacts is essential for properly assessing population exposure to smoke from different fire types.