PRACTICERESEARCH OF ARTICLE FORESTRY J. For.J. For.115( 116(1):76–86●):000–000 ence smoke impacts. Finally, we discuss how http://dx.doi.org/10.5849/jof.16-042 this quantitative framework can help man- Sidebar 1. Glossary of Technical Terminology Copyright © 2017 Society of American Foresters agers and regulators implement proactive Air resource advisors—Personnel with specialized training who are assigned to coordi- -~--- nate air quality monitoring and smoke concentration and dispersion modeling on wild- fire management strategies that align public health and forest restoration objectives. We land fires. highlight examples from California’s Si- Hazard mapping system (HMS) smoke product—A map produced by the National fire & fuels management erra Nevada, but many of these challenges Oceanic and Atmospheric administration that indicates the location and density of sig- are common to other parts of the world nificant smoke plumes. where increasingly large and severe wild- Prescribed fire—A wildland fire originating from a planned ignition that is intentionally fires are impacting large populations (Wil- managed to achieve resource objectives. liams 2013). Push/pull fire tactics—Tactics for speeding or slowing fire spread, which can be used to Aligning Smoke Management with Ecological influence the rate of smoke production. Socioecological Context Resource objective wildfire—A wildland fire originating from an unplanned ignition The last century of fire suppression in that is intentionally managed to achieve resource objectives such as reducing fire danger or and Public Health Goals California has caused past and current gen- forest restoration rather than being targeted for suppression. Such fires have also been erations to expect less frequent fire and labeled “resource benefit wildfire,” “wildland fire use,” or “managed wildfire” (see Hunter Jonathan W. Long, Leland W. Tarnay, and Malcolm P. North smoke than was integral to seasonally dry et al. 2014). forest ecosystems (Stephens et al. 2007). It Wildland fire—Any nonstructure fire that occurs in vegetation or natural fuels. has also contributed to less frequent, larger, Suppression wildfire—An unplanned wildland fire where the objective is to put the fire Past and current forest management affects wildland fire smoke impacts on downwind human populations. est managers and air quality regulators can and more severe wildfires (Stephens and out. However, mismatches between the scale of benefits and risks make it difficult to proactively manage wildland jointly pursue to minimize public health im- Sugihara 2006, Miller and Safford 2012, fires to promote both ecological and public health. Building on recent literature and advances in modeling smoke pacts while restoring more natural fire re- Fule´et al. 2014). These extreme fires not and health effects, we outline a framework to more directly quantify and compare smoke impacts based on gimes. First, we consider factors that make only directly threaten lives and property but more than 50,000 people and another 5% sults from the ability of present generations emissions, dispersion, and the size and vulnerability of downwind populations across time and space. We apply this challenge so difficult, by examining how they also produce enormous quantities of living in smaller urban clusters (US Census to pass accumulated fuels that result from the framework in a case study to demonstrate how different kinds of fires in California’s Central Sierra Nevada the social and ecological context of fire in- smoke that pose significant health risks, es- Bureau 2015). Many of those urban areas fire suppression onto future generations have resulted in very different smoke impacts. Our results indicate that the 257,314-acre Rim Fire of 2013 volves scale mismatches that create disincen- pecially when affecting large urban areas are situated in valleys or basins that have (North et al. 2012, Lueck and Yoder 2015). probably resulted in 7 million person-days of smoke impact across California and Nevada, which was greater tives for ecologically beneficial use of wild- (Strand et al. 2012, Moeltner et al. 2013, poor air quality due to human activities as In the shorter term, current policies have than 5 times the impact per burned unit area than two earlier wildfires, Grouse and Harden of 2009, that were land fire. Second, we outline a more direct Schweizer and Cisneros 2016). Important well as natural conditions that often trap permitted regulators to curtail fires inten- intentionally managed for resource objectives within the same airshed. The framework and results suggest framework for quantifying smoke impacts health risks include increased mortality and pollutants (Ngo et al. 2010, Nakayama tionally managed for resource objectives in strategies and tactics for undertaking larger-scale burns that can minimize smoke impacts, restore forest to human populations through monitoring respiratory morbidities associated with fine Wong et al. 2011). For example, the four response to nuisance complaints by a few ecosystems, and reduce the potential for more hazardous wildfire and smoke events. and analysis of daily emissions, the location particulate matter (less than 2.5 microns, metropolitan areas in the United States with individuals, despite the potential for such the highest levels of particle pollution are fires to have long-term collective benefits and density of resulting smoke plumes, and known as PM2.5) (Liu et al. 2015). Wildfires Keywords: air quality, California, USA, particulate matter, socioecological systems, wildland fire all located in California’s Central Valley (Engel 2013). Because the impact and like- the size and vulnerability of populations are an unpredictable component of PM2.5 within such plumes. We show how the pollution in California, contributing only (American Lung Association 2015). Because lihood of smoke increase the longer that fire framework can link tools that forest manag- 17,068 tons in 2005, 529,821 tons in 2008, many urban populations already experience is kept out of the system, extensive fire sup- ider use of ecologically benefi- Tensions reflect long-standing division of ers, air regulators, and public health experts and 53,487 tons in 2011. The wildfire emis- poor air quality during the summer, they are pression can result in a vicious cycle that be- cial fire has been suggested as a responsibilities between forest managers particularly vulnerable to health impacts comes more and more costly to escape until are already using, but that are often not sys- sions in 2008 represented 68% of all PM2.5 critical tool for increasing the concerned with restoring forests and air reg- from wildfires (Delfino et al. 2009, Cisneros the system fails, as represented by extreme W tematically applied to resource objective emissions in the state, and they caused nota- pace and scale of forest restoration and for ulators concerned with protecting public ble public health impacts (Wegesser et al. et al. 2014). wildfires (Calkin et al. 2015). Many mem- avoiding extreme wildfires that pose threats health (Sneeuwjagt et al. 2013). A recent up- wildfires and prescribed fires. Through a 2009, Preisler et al. 2015). Such impacts are An important temporal mismatch re- bers of the public agree with researchers and to human health (North et al. 2015, date to wildfire-smoke policy proposed by case study, we examine how smoke impacts likely to worsen in the future, because higher Schweizer and Cisneros 2016). A range of the US Environmental Protection Agency from intentional use of fire are likely to differ temperatures and accumulated fuels are ex- constraints, however, have hindered the use (US EPA) recognized the need to restore and from an extreme wildfire targeted for sup- Management and Policy Implications pected to favor very large fires in fire-prone of fire to meet resource objectives, whether maintain more frequent fire regimes through pression. We focus on the area affected by regions such as California (Barbero et al. through prescribed burning or managing intentional use of fire, while asserting that the Rim Fire, which escaped from a campfire Forest managers and air quality regulators could better align the dual objectives of restoring forests and 2015). Hurteau et al. (2014) found that un- wildfires for resource objectives (hereafter protecting human health remained the on Aug. 17, 2013. to become the largest fire protecting human health by using a common framework to quantitatively manage and evaluate the smoke der a business-as-usual climate scenario, this called “resource objective wildfires”). Air agency’s “highest priority” (Office of the in the history of the Sierra Nevada. Because impacts of different kinds of fires on downwind populations. Some smoke management policies discourage escalation in fire potential is likely to in- quality policy and regulations have been par- Federal Register 2015). Therefore, address- this area also had a recent history of pre- managing large fires for resource objectives and risk shunting inevitable emissions into even larger and crease wildfire emissions in California by ticularly significant constraints on inten- ing both forest restoration and air quality scribed burns and resource objective wild- long-lasting wildfires that expose sizeable human populations to unhealthy concentrations of particulate 50% by the end of this century unless agen- tional use of fire (Carroll et al. 2007, Quinn- objectives remains a central challenge. fires, it afforded a distinctive opportunity to matter. Managing wildland fires under favorable conditions can help to restore forests, reduce hazardous cies take a more proactive approach to fire Davidson and Varner 2012, Engel 2013). This article explores strategies that for- consider how working with fire can influ- fuels, and mitigate potentially harmful smoke impacts on downwind populations by decreasing daily use. emissions and taking advantage of favorable dispersion. Our framework supports smoke management An important spatial mismatch results strategies including the following: incentivizing reduction of human exposure to hazardous smoke levels from the fact that large wildfires can create Received August 10, 2016; accepted December 12, 2016; published online January 19, 2017. smoke impacts on distant urban popula- over space and time rather than area burned; pacing fire spread based on airshed capacity to disperse Affiliations: Jonathan W. Long ([email protected]), USDA Forest Service Pacific Northwest Research Station, Davis, CA. Leland W. Tarnay ([email protected]), tions. The risk to urban populations from the resulting emissions; and communicating with the public to reduce the exposure of downwind USDA Forest Service Region 5. Malcolm P. North ([email protected]), USDA Forest Service Pacific Southwest Research Station. regional-scale smoke impacts has increased populations and the benefits of managing wildland fires for resource objectives. Advancing these strategies as California became the most urbanized will depend on coordinated efforts by forest and fire managers, air quality regulators, and air resource Acknowledgments: This article was informed by data and experience by one of the authors as an Air Resource Specialist for Yosemite National Park, as well as support advisors with specialized skills in evaluating and communicating fire impacts on downwind populations. from the Park’s fire management and resource management staff. We thank T. Procter, H. Preisler, L. Heath, and A. Bytnerowicz for thoughtful reviews and insights state in the United States, with 90% of its that contributed to this article. The USDA Forest Service Pacific Southwest Research Station and Region 5 Air Quality Program provided support for this research. population residing within cities that have
76 Journal of Forestry • January 2018 Journal of Forestry • MONTH 2017 1 2 Journal of Forestry • MONTH 2017 RESEARCH ARTICLE J. For. 115(●):000–000 ence smoke impacts. Finally, we discuss how http://dx.doi.org/10.5849/jof.16-042 this quantitative framework can help man- Sidebar 1. Glossary of Technical Terminology Copyright © 2017 Society of American Foresters agers and regulators implement proactive Air resource advisors—Personnel with specialized training who are assigned to coordi- fire management strategies that align public nate air quality monitoring and smoke concentration and dispersion modeling on wild- health and forest restoration objectives. We land fires. highlight examples from California’s Si- Hazard mapping system (HMS) smoke product—A map produced by the National fire & fuels management erra Nevada, but many of these challenges Oceanic and Atmospheric administration that indicates the location and density of sig- are common to other parts of the world nificant smoke plumes. where increasingly large and severe wild- Prescribed fire—A wildland fire originating from a planned ignition that is intentionally fires are impacting large populations (Wil- managed to achieve resource objectives. liams 2013). Push/pull fire tactics—Tactics for speeding or slowing fire spread, which can be used to Aligning Smoke Management with Ecological influence the rate of smoke production. Socioecological Context Resource objective wildfire—A wildland fire originating from an unplanned ignition The last century of fire suppression in that is intentionally managed to achieve resource objectives such as reducing fire danger or and Public Health Goals California has caused past and current gen- forest restoration rather than being targeted for suppression. Such fires have also been erations to expect less frequent fire and labeled “resource benefit wildfire,” “wildland fire use,” or “managed wildfire” (see Hunter Jonathan W. Long, Leland W. Tarnay, and Malcolm P. North smoke than was integral to seasonally dry et al. 2014). forest ecosystems (Stephens et al. 2007). It Wildland fire—Any nonstructure fire that occurs in vegetation or natural fuels. has also contributed to less frequent, larger, Suppression wildfire—An unplanned wildland fire where the objective is to put the fire Past and current forest management affects wildland fire smoke impacts on downwind human populations. est managers and air quality regulators can and more severe wildfires (Stephens and out. However, mismatches between the scale of benefits and risks make it difficult to proactively manage wildland jointly pursue to minimize public health im- Sugihara 2006, Miller and Safford 2012, fires to promote both ecological and public health. Building on recent literature and advances in modeling smoke pacts while restoring more natural fire re- Fule´et al. 2014). These extreme fires not and health effects, we outline a framework to more directly quantify and compare smoke impacts based on gimes. First, we consider factors that make only directly threaten lives and property but more than 50,000 people and another 5% sults from the ability of present generations emissions, dispersion, and the size and vulnerability of downwind populations across time and space. We apply this challenge so difficult, by examining how they also produce enormous quantities of living in smaller urban clusters (US Census to pass accumulated fuels that result from the framework in a case study to demonstrate how different kinds of fires in California’s Central Sierra Nevada the social and ecological context of fire in- smoke that pose significant health risks, es- Bureau 2015). Many of those urban areas fire suppression onto future generations have resulted in very different smoke impacts. Our results indicate that the 257,314-acre Rim Fire of 2013 volves scale mismatches that create disincen- pecially when affecting large urban areas are situated in valleys or basins that have (North et al. 2012, Lueck and Yoder 2015). probably resulted in 7 million person-days of smoke impact across California and Nevada, which was greater tives for ecologically beneficial use of wild- (Strand et al. 2012, Moeltner et al. 2013, poor air quality due to human activities as In the shorter term, current policies have than 5 times the impact per burned unit area than two earlier wildfires, Grouse and Harden of 2009, that were land fire. Second, we outline a more direct Schweizer and Cisneros 2016). Important well as natural conditions that often trap permitted regulators to curtail fires inten- intentionally managed for resource objectives within the same airshed. The framework and results suggest framework for quantifying smoke impacts health risks include increased mortality and pollutants (Ngo et al. 2010, Nakayama tionally managed for resource objectives in strategies and tactics for undertaking larger-scale burns that can minimize smoke impacts, restore forest to human populations through monitoring respiratory morbidities associated with fine Wong et al. 2011). For example, the four response to nuisance complaints by a few ecosystems, and reduce the potential for more hazardous wildfire and smoke events. and analysis of daily emissions, the location particulate matter (less than 2.5 microns, metropolitan areas in the United States with individuals, despite the potential for such the highest levels of particle pollution are fires to have long-term collective benefits and density of resulting smoke plumes, and known as PM2.5) (Liu et al. 2015). Wildfires Keywords: air quality, California, USA, particulate matter, socioecological systems, wildland fire all located in California’s Central Valley (Engel 2013). Because the impact and like- the size and vulnerability of populations are an unpredictable component of PM2.5 within such plumes. We show how the pollution in California, contributing only (American Lung Association 2015). Because lihood of smoke increase the longer that fire framework can link tools that forest manag- 17,068 tons in 2005, 529,821 tons in 2008, many urban populations already experience is kept out of the system, extensive fire sup- ider use of ecologically benefi- Tensions reflect long-standing division of ers, air regulators, and public health experts and 53,487 tons in 2011. The wildfire emis- poor air quality during the summer, they are pression can result in a vicious cycle that be- cial fire has been suggested as a responsibilities between forest managers particularly vulnerable to health impacts comes more and more costly to escape until are already using, but that are often not sys- sions in 2008 represented 68% of all PM2.5 critical tool for increasing the concerned with restoring forests and air reg- from wildfires (Delfino et al. 2009, Cisneros the system fails, as represented by extreme W tematically applied to resource objective emissions in the state, and they caused nota- pace and scale of forest restoration and for ulators concerned with protecting public ble public health impacts (Wegesser et al. et al. 2014). wildfires (Calkin et al. 2015). Many mem- avoiding extreme wildfires that pose threats health (Sneeuwjagt et al. 2013). A recent up- wildfires and prescribed fires. Through a 2009, Preisler et al. 2015). Such impacts are An important temporal mismatch re- bers of the public agree with researchers and to human health (North et al. 2015, date to wildfire-smoke policy proposed by case study, we examine how smoke impacts likely to worsen in the future, because higher Schweizer and Cisneros 2016). A range of the US Environmental Protection Agency from intentional use of fire are likely to differ temperatures and accumulated fuels are ex- constraints, however, have hindered the use (US EPA) recognized the need to restore and from an extreme wildfire targeted for sup- Management and Policy Implications pected to favor very large fires in fire-prone of fire to meet resource objectives, whether maintain more frequent fire regimes through pression. We focus on the area affected by regions such as California (Barbero et al. through prescribed burning or managing intentional use of fire, while asserting that the Rim Fire, which escaped from a campfire Forest managers and air quality regulators could better align the dual objectives of restoring forests and 2015). Hurteau et al. (2014) found that un- wildfires for resource objectives (hereafter protecting human health remained the on Aug. 17, 2013. to become the largest fire protecting human health by using a common framework to quantitatively manage and evaluate the smoke der a business-as-usual climate scenario, this called “resource objective wildfires”). Air agency’s “highest priority” (Office of the in the history of the Sierra Nevada. Because impacts of different kinds of fires on downwind populations. Some smoke management policies discourage escalation in fire potential is likely to in- quality policy and regulations have been par- Federal Register 2015). Therefore, address- this area also had a recent history of pre- managing large fires for resource objectives and risk shunting inevitable emissions into even larger and crease wildfire emissions in California by ticularly significant constraints on inten- ing both forest restoration and air quality scribed burns and resource objective wild- long-lasting wildfires that expose sizeable human populations to unhealthy concentrations of particulate 50% by the end of this century unless agen- tional use of fire (Carroll et al. 2007, Quinn- objectives remains a central challenge. fires, it afforded a distinctive opportunity to matter. Managing wildland fires under favorable conditions can help to restore forests, reduce hazardous cies take a more proactive approach to fire Davidson and Varner 2012, Engel 2013). This article explores strategies that for- consider how working with fire can influ- fuels, and mitigate potentially harmful smoke impacts on downwind populations by decreasing daily use. emissions and taking advantage of favorable dispersion. Our framework supports smoke management An important spatial mismatch results strategies including the following: incentivizing reduction of human exposure to hazardous smoke levels from the fact that large wildfires can create Received August 10, 2016; accepted December 12, 2016; published online January 19, 2017. smoke impacts on distant urban popula- over space and time rather than area burned; pacing fire spread based on airshed capacity to disperse Affiliations: Jonathan W. Long ([email protected]), USDA Forest Service Pacific Northwest Research Station, Davis, CA. Leland W. Tarnay ([email protected]), tions. The risk to urban populations from the resulting emissions; and communicating with the public to reduce the exposure of downwind USDA Forest Service Region 5. Malcolm P. North ([email protected]), USDA Forest Service Pacific Southwest Research Station. regional-scale smoke impacts has increased populations and the benefits of managing wildland fires for resource objectives. Advancing these strategies as California became the most urbanized will depend on coordinated efforts by forest and fire managers, air quality regulators, and air resource Acknowledgments: This article was informed by data and experience by one of the authors as an Air Resource Specialist for Yosemite National Park, as well as support advisors with specialized skills in evaluating and communicating fire impacts on downwind populations. from the Park’s fire management and resource management staff. We thank T. Procter, H. Preisler, L. Heath, and A. Bytnerowicz for thoughtful reviews and insights state in the United States, with 90% of its that contributed to this article. The USDA Forest Service Pacific Southwest Research Station and Region 5 Air Quality Program provided support for this research. population residing within cities that have
Journal of Forestry • MONTH 2017 1 2 Journal of Forestry • MONTH 2017 Journal of Forestry • January 2018 77 Smoke Impact = / (Emissions x Dispersion x Population Vulnerability) One approach for estimating the smoke are particularly important for evaluating plex, but they would help decisionmakers tory database (Yosemite National Park integrated over time and space impact in economic terms has been to mul- smoke impacts. The National Oceanic and and the public evaluate potential tradeoffs. 2012), operational notes, and smoke man- tiply the number of person-days of impact Atmospheric Administration (NOAA) be- agement plans for prescribed fires on the by individuals’ willingness to pay to avoid gan providing maps of smoke plume density Method for Applying the Framework Stanislaus National Forest and Yosemite exposure. For example, Richardson et al. based on satellite observations through their to Selected Fires National Park (Kent van Wagtendonk, Taro (2012) found that individuals in several Hazard Mapping System (HMS) in Novem- We applied the framework retrospec- Pusina, and Mike Beasley, National Park southern California cities were willing to pay ber 2006 (Ruminski et al. 2007). Predic- tively to compare differences in smoke im- Service, pers. comm., July 19, 2015). In this pacts between resource objective wildfires $84.42 (in 2012 dollars) to avoid a day of tions of surface PM2.5 levels are limited by analysis, we assumed that daily emissions are wildfire smoke impact resulting from the difficulty in determining how much smoke and full-suppression wildfires within the San based only on areas of new fire growth for a Station Fire of 2009, during which ground- has reached the surface, but these data pro- Joaquin River watershed in California’s Si- given day because there is no established level conditions exceeded the 24-hour stan- vide an objective means of defining a poten- erra Nevada, the Sierras that burned be- method for allocating emissions across mul- dard for 1–3 days. Jones et al. (2016) found tial area of influence for a particular fire. tween 2002 and 2013, including 10 re- tiple days. We used FOFEM within a geo- a per capita willingness to pay of $130 (in Our framework can also be applied source objective wildfires (totaling 20,494 graphic information system (GIS) to calcu- 2014 dollars) to avoid wildfire smoke health prospectively to fires by using BlueSky, acres), 17 prescribed fires (totaling 6,636 late emissions for each daily fire progression 2 3 effects in Albuquerque during the Wallow HYSPLIT, CALPUFF, or other modeling acres), 4 small wildfires (totaling 12,025 acres), polygon based on inputs of fuel loads and Fire of 2011. tools to spatially and temporally forecast and the exceptionally large Rim Fire fuel moisture for each vegetation type, fol- Figure 1. Framework for evaluating smoke impacts using source emissions, dispersal and smoke concentrations downstream from (257,314 acres). This comparison requires lowing the approach by Clinton et al. downwind populations, integrated over time and large areas of fire influence. A Quantitative Framework for fires using daily emission and dispersion es- several caveats: (1) the use of resource objec- (2003). This spatially explicit method ac- Evaluating Smoke Impacts timates (Goodrick et al. 2013). Research has tive wildfires has varied historically, reflect- counts for area of the different vegetation ing continuing evolution in policy and prac- We outline a quantitative framework supported this approach, as Fusina et al. types within the polygons for each days managers who have argued that tolerating burned toward more direct measures of (2007) found that plume trajectories gener- tice; (2) some resource objective fires are spread, for each fire, then incorporating (Figure 1) to evaluate smoke impacts that short-term impacts of prescribed fires may ated in BlueSky agree well with satellite ob- categorized as suppression wildfires when crosswalks assigned from the California Gap smoke impact may afford opportunities to can be applied using existing tools and data avoid more harmful impacts (Olsen et al. servations and ground-level PM concen- they no longer meet desired objectives; (3) Analysis Project (GAP) land cover map (Da- better align public health and ecological res- sets to improve understanding of potential 2.5 2014). However, those arguments have not trations. Furthermore, Yao et al. (2013) the Rim Fire burned through and into sev- vis et al. 1998), to estimate total daily emis- toration. smoke impacts across large areas over time. been well aligned with regulatory systems found significant associations between BlueSky- eral of the previous fires; and (4) the limited sions from the FOFEM output. Because we Some fire managers routinely conduct the designed to protect public health, visibility, Measuring Smoke Impacts on forecasted and HMS smoke plumes and re- availability of smoke monitoring data, par- lack systematic and detailed fuel plots, we first two steps, quantifying source emissions and safety, which vary across and within spiratory health outcomes such as asthma- ticularly before 2007, requires a focus on assigned “light” fuel loads for areas that had Public Health and downwind concentrations, but they states (Engel 2013). related physician visits. modeled emissions. experienced fires in the previous 12 years, Smoke and wildfires can impact public have traditionally compiled summaries over Within California, decisions regarding The third step in the framework is to We evaluated the smoke impacts of two while assigning “typical” loads otherwise health in ways other than particulate pollu- the total size and duration of the burn, particular fires are generally made on a case- tion, including ozone pollution, increased quantify the size and vulnerability of the af- relatively large resource objective wildfires: (see Supplemental Table S1 ). Lydersen et rather than tracking the daily variation of the Grouse Fire (3,042 acres) and the by-case basis by local air districts. These de- stress during and after wildfires, and strains fected populations within the areas of smoke al. (2014) found that this approach yielded emissions relative to dispersion. For both Harden Fire (1,653 acres), which were ig- cisions often rely on area burned as a proxy on medical services and communication sys- influence for the duration of the fire. One reasonably accurate consumption estimates prospective planning and retrospective eval- nited by lightning on May 31, 2009, and for reporting and evaluating potential smoke tems (Fowler 2003, Kumagai et al. 2004, approach is to use census data to estimate when applied to large areas within the Sierra uation of fire outcomes, it is important to June 8, 2009, respectively, and then allowed impacts from fires and for setting thresholds Finlay et al. 2012). Despite these broader populations underlying mapped smoke plumes Nevada. We modeled prescribed burns as for providing additional documentation for quantitatively account for the smoke im- weighted by smoke density and the expected to burn for multiple objectives. The Grouse surface fires but allowed wildfires and re- considerations, public health regulations for permits (California Air Resources Board pacts to public health in space and time. likelihood and intensity of impact. Finer Fire is significant as the first instance in source objective wildfires to burn into tree smoke typically focus on a 24-hour average 2015). For example, local regulators have Source emissions can be estimated based analyses of such data can consider demo- which fire managers in Yosemite National crowns. We estimated fuel moistures using of PM . Values that exceed 35 g/m3 are sometimes limited burns to less than 50 2.5 on fuel loads using tools such as the First graphic variables that often indicate vulner- Park explicitly designed the fire contingency operational notes from individual fires and considered unhealthy for sensitive groups, acres per day when air quality conditions Order Fire Effects Model (FOFEM) (Lutes ability, such as age, income, race, education, plan to incorporate thresholds for PM representative values from remote automatic which include pregnant women, young chil- 1 2.5 were considered marginal for dispersing 2014) or the BlueSky Playground. Emis- and health, although fine-scale data on the monitored at sensitive sites. The same inci- weather stations used in Yosemite’s fire dren, elderly individuals, smokers, and peo- smoke (Carolyn Ballard, Fire Management sions from fires may vary by an order of mag- incidence of respiratory disease are usually dent team simultaneously managed the management program. Officer, Sierra National Forest, pers. comm., ple with chronic respiratory problems such nitude, depending on the type of vegetation, lacking (Gaither et al. 2015). Harden Fire for resource objectives within For the second step in the framework, April 19, 2016). Because suitable burn win- as asthma (Delfino et al. 2009, Kochi et al. the fuel loading associated with that vegeta- The framework can be extended to in- an area bounded by containment lines and we adapted the methodology used by Pre- dows are often limited to only a few consec- 2010, Moeltner et al. 2013). Although this tion type (Leenhouts 1998), and the inten- clude monetary metrics by multiplying the rock, while focusing active resources on the isler et al. (2015), who related HMS smoke standard is an important quantitative sity and the severity of the fire activity. For utive days, such constraints can make it in- estimates of person-days of exposure by esti- Grouse Fire. These two fires illustrate the maps to monitored surface PM2.5 values for feasible to burn the hundreds or thousands threshold, higher levels of PM2.5 are associ- example, a forest where fuels have accumu- mated health costs or other valuations. The potential for managing multiple fires to a range of fires in the Central Sierra Nevada. of acres needed for landscape restoration. In ated with broader and greater health risks, so lated unchecked for decades will have much US EPA has developed a tool, BenMAP- achieve resource objectives for an extended They determined that plumes of HMS addition, some air pollution control districts incorporating concentration-response rela- heavier fuel loads and expected emissions CE,4 for estimating economic impacts of air period. Furthermore, when combined, these smoke within 6.2 miles of a monitoring site in California have imposed fees for each unit tionships will improve estimates of smoke than otherwise similar areas subjected to pollution based on population data, mod- resource objective fires were the largest in the were associated with significant increases in impact. Furthermore, exceedances of the of area burned for prescribed fires and re- frequent fires. Consequently, area burned, eled or monitored air quality data, concen- study area for which the HMS smoke den- daily average values of PM2.5, with a likeli- source objective wildfires, although not for 24-hour standard may not signify an actual while easy to measure, may not correlate well tration-response functions, and valuation sity maps were available. hood of 36.5% (95% confidence limit [CL] wildfires targeted for suppression (Sneeuw- health impact on a particular day, as fore- with smoke production. measures (Jones et al. 2016). Another exten- To estimate emissions in the first step in was 27–47%) of exceeding the “norm” on jagt et al. 2013). These fees are used to offset warning of smoke can help people take steps The second step in the framework is to sion of the framework would be to iteratively the framework, we developed maps of daily days with dense smoke and 17% (95% CL the costs of reviewing smoke plans and mon- to reduce their exposure (Rappold et al. evaluate population exposure to pollutants model regional airshed conditions to project fire spread for fires of all types in the Tuol- was 12–23.5%) on days with medium-den- itoring air quality during burns; however, 2014). A broad accounting of smoke im- within observed or modeled smoke plumes. the benefits of using fire at large scales over umne watershed since 2002 using records of sity smoke. The norm was the expected level they also provide a disincentive for wildland pacts also should consider other demo- Because standards to protect public health long periods. Such efforts would be com- fire progression from the Yosemite fire his- on each day for each site when smoke fires that are managed for resource benefits. graphic and institutional factors that influ- are often based on 24-hour average PM2.5 For reasons we discuss throughout the arti- ence vulnerability and adaptive capacity values and available dispersion varies sub- cle, shifting from an emphasis on area (Cross 2001, Trainor et al. 2009). stantially from day to day, daily emissions Supplementary data are available with this article at http://dx.doi.org/10.5849/JOF-2016-042R2.
78 Journal of Forestry • January 2018 Journal of Forestry • MONTH 2017 3 4 Journal of Forestry • MONTH 2017 One approach for estimating the smoke are particularly important for evaluating plex, but they would help decisionmakers tory database (Yosemite National Park impact in economic terms has been to mul- smoke impacts. The National Oceanic and and the public evaluate potential tradeoffs. 2012), operational notes, and smoke man- tiply the number of person-days of impact Atmospheric Administration (NOAA) be- agement plans for prescribed fires on the by individuals’ willingness to pay to avoid gan providing maps of smoke plume density Method for Applying the Framework Stanislaus National Forest and Yosemite exposure. For example, Richardson et al. based on satellite observations through their to Selected Fires National Park (Kent van Wagtendonk, Taro (2012) found that individuals in several Hazard Mapping System (HMS) in Novem- We applied the framework retrospec- Pusina, and Mike Beasley, National Park southern California cities were willing to pay ber 2006 (Ruminski et al. 2007). Predic- tively to compare differences in smoke im- Service, pers. comm., July 19, 2015). In this pacts between resource objective wildfires $84.42 (in 2012 dollars) to avoid a day of tions of surface PM2.5 levels are limited by analysis, we assumed that daily emissions are wildfire smoke impact resulting from the difficulty in determining how much smoke and full-suppression wildfires within the San based only on areas of new fire growth for a Station Fire of 2009, during which ground- has reached the surface, but these data pro- Joaquin River watershed in California’s Si- given day because there is no established level conditions exceeded the 24-hour stan- vide an objective means of defining a poten- erra Nevada, the Sierras that burned be- method for allocating emissions across mul- dard for 1–3 days. Jones et al. (2016) found tial area of influence for a particular fire. tween 2002 and 2013, including 10 re- tiple days. We used FOFEM within a geo- a per capita willingness to pay of $130 (in Our framework can also be applied source objective wildfires (totaling 20,494 graphic information system (GIS) to calcu- 2014 dollars) to avoid wildfire smoke health prospectively to fires by using BlueSky, acres), 17 prescribed fires (totaling 6,636 late emissions for each daily fire progression 2 3 effects in Albuquerque during the Wallow HYSPLIT, CALPUFF, or other modeling acres), 4 small wildfires (totaling 12,025 acres), polygon based on inputs of fuel loads and Fire of 2011. tools to spatially and temporally forecast and the exceptionally large Rim Fire fuel moisture for each vegetation type, fol- Figure 1. Framework for evaluating smoke impacts using source emissions, dispersal and smoke concentrations downstream from (257,314 acres). This comparison requires lowing the approach by Clinton et al. downwind populations, integrated over time and large areas of fire influence. A Quantitative Framework for fires using daily emission and dispersion es- several caveats: (1) the use of resource objec- (2003). This spatially explicit method ac- Evaluating Smoke Impacts timates (Goodrick et al. 2013). Research has tive wildfires has varied historically, reflect- counts for area of the different vegetation ing continuing evolution in policy and prac- We outline a quantitative framework supported this approach, as Fusina et al. types within the polygons for each days managers who have argued that tolerating burned toward more direct measures of (2007) found that plume trajectories gener- tice; (2) some resource objective fires are spread, for each fire, then incorporating (Figure 1) to evaluate smoke impacts that short-term impacts of prescribed fires may ated in BlueSky agree well with satellite ob- categorized as suppression wildfires when crosswalks assigned from the California Gap smoke impact may afford opportunities to can be applied using existing tools and data avoid more harmful impacts (Olsen et al. servations and ground-level PM concen- they no longer meet desired objectives; (3) Analysis Project (GAP) land cover map (Da- better align public health and ecological res- sets to improve understanding of potential 2.5 2014). However, those arguments have not trations. Furthermore, Yao et al. (2013) the Rim Fire burned through and into sev- vis et al. 1998), to estimate total daily emis- toration. smoke impacts across large areas over time. been well aligned with regulatory systems found significant associations between BlueSky- eral of the previous fires; and (4) the limited sions from the FOFEM output. Because we Some fire managers routinely conduct the designed to protect public health, visibility, Measuring Smoke Impacts on forecasted and HMS smoke plumes and re- availability of smoke monitoring data, par- lack systematic and detailed fuel plots, we first two steps, quantifying source emissions and safety, which vary across and within spiratory health outcomes such as asthma- ticularly before 2007, requires a focus on assigned “light” fuel loads for areas that had Public Health and downwind concentrations, but they states (Engel 2013). related physician visits. modeled emissions. experienced fires in the previous 12 years, Smoke and wildfires can impact public have traditionally compiled summaries over Within California, decisions regarding The third step in the framework is to We evaluated the smoke impacts of two while assigning “typical” loads otherwise health in ways other than particulate pollu- the total size and duration of the burn, particular fires are generally made on a case- tion, including ozone pollution, increased quantify the size and vulnerability of the af- relatively large resource objective wildfires: (see Supplemental Table S1 ). Lydersen et rather than tracking the daily variation of the Grouse Fire (3,042 acres) and the by-case basis by local air districts. These de- stress during and after wildfires, and strains fected populations within the areas of smoke al. (2014) found that this approach yielded emissions relative to dispersion. For both Harden Fire (1,653 acres), which were ig- cisions often rely on area burned as a proxy on medical services and communication sys- influence for the duration of the fire. One reasonably accurate consumption estimates prospective planning and retrospective eval- nited by lightning on May 31, 2009, and for reporting and evaluating potential smoke tems (Fowler 2003, Kumagai et al. 2004, approach is to use census data to estimate when applied to large areas within the Sierra uation of fire outcomes, it is important to June 8, 2009, respectively, and then allowed impacts from fires and for setting thresholds Finlay et al. 2012). Despite these broader populations underlying mapped smoke plumes Nevada. We modeled prescribed burns as for providing additional documentation for quantitatively account for the smoke im- weighted by smoke density and the expected to burn for multiple objectives. The Grouse surface fires but allowed wildfires and re- considerations, public health regulations for permits (California Air Resources Board pacts to public health in space and time. likelihood and intensity of impact. Finer Fire is significant as the first instance in source objective wildfires to burn into tree smoke typically focus on a 24-hour average 2015). For example, local regulators have Source emissions can be estimated based analyses of such data can consider demo- which fire managers in Yosemite National crowns. We estimated fuel moistures using of PM . Values that exceed 35 g/m3 are sometimes limited burns to less than 50 2.5 on fuel loads using tools such as the First graphic variables that often indicate vulner- Park explicitly designed the fire contingency operational notes from individual fires and considered unhealthy for sensitive groups, acres per day when air quality conditions Order Fire Effects Model (FOFEM) (Lutes ability, such as age, income, race, education, plan to incorporate thresholds for PM representative values from remote automatic which include pregnant women, young chil- 1 2.5 were considered marginal for dispersing 2014) or the BlueSky Playground. Emis- and health, although fine-scale data on the monitored at sensitive sites. The same inci- weather stations used in Yosemite’s fire dren, elderly individuals, smokers, and peo- smoke (Carolyn Ballard, Fire Management sions from fires may vary by an order of mag- incidence of respiratory disease are usually dent team simultaneously managed the management program. Officer, Sierra National Forest, pers. comm., ple with chronic respiratory problems such nitude, depending on the type of vegetation, lacking (Gaither et al. 2015). Harden Fire for resource objectives within For the second step in the framework, April 19, 2016). Because suitable burn win- as asthma (Delfino et al. 2009, Kochi et al. the fuel loading associated with that vegeta- The framework can be extended to in- an area bounded by containment lines and we adapted the methodology used by Pre- dows are often limited to only a few consec- 2010, Moeltner et al. 2013). Although this tion type (Leenhouts 1998), and the inten- clude monetary metrics by multiplying the rock, while focusing active resources on the isler et al. (2015), who related HMS smoke standard is an important quantitative sity and the severity of the fire activity. For utive days, such constraints can make it in- estimates of person-days of exposure by esti- Grouse Fire. These two fires illustrate the maps to monitored surface PM2.5 values for feasible to burn the hundreds or thousands threshold, higher levels of PM2.5 are associ- example, a forest where fuels have accumu- mated health costs or other valuations. The potential for managing multiple fires to a range of fires in the Central Sierra Nevada. of acres needed for landscape restoration. In ated with broader and greater health risks, so lated unchecked for decades will have much US EPA has developed a tool, BenMAP- achieve resource objectives for an extended They determined that plumes of HMS addition, some air pollution control districts incorporating concentration-response rela- heavier fuel loads and expected emissions CE,4 for estimating economic impacts of air period. Furthermore, when combined, these smoke within 6.2 miles of a monitoring site in California have imposed fees for each unit tionships will improve estimates of smoke than otherwise similar areas subjected to pollution based on population data, mod- resource objective fires were the largest in the were associated with significant increases in impact. Furthermore, exceedances of the of area burned for prescribed fires and re- frequent fires. Consequently, area burned, eled or monitored air quality data, concen- study area for which the HMS smoke den- daily average values of PM2.5, with a likeli- source objective wildfires, although not for 24-hour standard may not signify an actual while easy to measure, may not correlate well tration-response functions, and valuation sity maps were available. hood of 36.5% (95% confidence limit [CL] wildfires targeted for suppression (Sneeuw- health impact on a particular day, as fore- with smoke production. measures (Jones et al. 2016). Another exten- To estimate emissions in the first step in was 27–47%) of exceeding the “norm” on jagt et al. 2013). These fees are used to offset warning of smoke can help people take steps The second step in the framework is to sion of the framework would be to iteratively the framework, we developed maps of daily days with dense smoke and 17% (95% CL the costs of reviewing smoke plans and mon- to reduce their exposure (Rappold et al. evaluate population exposure to pollutants model regional airshed conditions to project fire spread for fires of all types in the Tuol- was 12–23.5%) on days with medium-den- itoring air quality during burns; however, 2014). A broad accounting of smoke im- within observed or modeled smoke plumes. the benefits of using fire at large scales over umne watershed since 2002 using records of sity smoke. The norm was the expected level they also provide a disincentive for wildland pacts also should consider other demo- Because standards to protect public health long periods. Such efforts would be com- fire progression from the Yosemite fire his- on each day for each site when smoke fires that are managed for resource benefits. graphic and institutional factors that influ- are often based on 24-hour average PM2.5 For reasons we discuss throughout the arti- ence vulnerability and adaptive capacity values and available dispersion varies sub- cle, shifting from an emphasis on area (Cross 2001, Trainor et al. 2009). stantially from day to day, daily emissions Supplementary data are available with this article at http://dx.doi.org/10.5849/JOF-2016-042R2.
Journal of Forestry • MONTH 2017 3 4 Journal of Forestry • MONTH 2017 Journal of Forestry • January 2018 79 Prescribed fire Rim (2013) sin and San Joaquin Valley. In contrast, 10000 Resource objective wildfire most of the fires managed for resource ob- 5000 Suppression wildfire Frog/Laurel (2006) Tuolumne (2003) jectives had peak emissions either in the 2000 Kibbie (2003) Harden (2009) 1000 early summer or in the fall, when airshed Grouse (2009) 500 ~/ conditions are often more favorable. 200 ~~:I Modeled emissions from the Rim Fire : ,f~',,'i, ~~... ~: ~, I I ',' ;~ •• 100 r~,,, •1' ...... ·,~ '', ;::••• within each vegetation polygon varied by an 50 141 • • 1 order of magnitude because of the different 1 .J, ' ~ii 20 ,: .)}-.n;;;;;~]'.1(J~(:···::)[~:··l\·\j types of vegetation and amount of accumu- 11 ,; "11~ I ! /',, 11• 10 ?.. lated fuels (Figure 3A). However, daily 5 ..,. ~ II 1' ,uuuo'~ iuuu~'a •'~ l ~ ~ ,,: ~ .... ~,:.: .... ,...... ,.r ...... ,,, ,' ...... ~--- ...... emissions from the Rim Fire (Figure 3B) :. ,; ~ ,' 'i --- 2 e!!!!!!!!!!! .. •••••••••H,l, I 1 ..... j i ... ~·, .'~ouocuuuu~ ouououi were quite high even in areas that had rela- 0.5 (.... I I tively light fuels (e.g., in the center of the Daily Emissions (Log2 of Tons PM2.5) I fire, as shown in Figure 3A), because the fire I I burned through those areas so quickly. The Rim Fire generated a surge in emis- Jul Jun Aug Sep Oct Nov Figure 3. Fuel loading and estimated emissions of PM per unit area within different vegetation polygons (A) and estimated emissions of Time of Year (2002-2013) sions that began on August 19 and contin- 2.5 ued through September 8, after which esti- PM2.5 per day within daily progression polygons (B), within the Rim Fire. Figure 2. Estimated daily emissions of PM , as generated by the FOFEM for different types 2.5 mated daily emissions dropped below 265 of fire within the Tuolumne River watershed, California, between 2002 and 2013. Points on tons (Figure 4B). That rapid growth resulted The enormous spatial impact of the medium-density plumes overlaid substantial poral and spatial scales. Our analyses help to each line represent one day of estimated emissions for a specific fire. Note the log base 2 y-axis; see Figures 4A and 4B for the nonlogarithmic version. in high- and medium-density smoke plumes Rim Fire is illustrated by a map on August populations in California and Nevada, illustrate and begin to quantify many of the over large populations in California and Ne- 31 (Figure 5), showing how plumes ex- amounting to 25,000 person-days (Figure potential benefits of resource objective wild- vada between August 22 and September 10 tended north from the Rim Fire over large 4C). Ground-level monitoring at Yosemite fires compared with those of extreme fires: plumes were excluded. In particular, they es- medium-density smoke plumes. We con- (Figure 4D). We weighted the population urban populations in the Lake Tahoe Basin Village showed relatively modest impacts, 1. Reduced fuels and reduced consumption. timated that the daily 96th percentile value fined our analysis to California and Nevada, estimates by the weights reported by Preisler (55,607), Carson City (55,212), Sparks where PM2.5 levels remained between 30 Where fires burn over the footprints of 3 5 3 of PM2.5 increased by 14 g/m (95% CL because archived reports and AirNow et al. (2015) (36.5 and 17% for high- and (90,264), and Reno (225,221), before shift- and 55 g/m from July 1 through July 6 relatively recent fires, emissions estimates 3 6 was 9–19 g/m ) where there was heavy maps did not indicate air quality impacts in medium-density smoke plumes, respectively), ing southward over Fresno (494,665), Mad- (Figure 4E). Although these levels are con- should typically reflect lighter fuels and 3 smoke and by 5 g/m (95% CL was 3.6– other states despite evidence that Rim Fire to estimate 2.9 million person-days associ- era (61,416), Visalia (124,442), and other sidered “unhealthy for sensitive groups,” reduced crown consumption (Wiedin- 3 6.1 g/m ) where there was medium-den- smoke traveled as far as central Canada (Pe- ated with high smoke and 4.1 million per- cities in the Central Valley. Altogether, me- concentrations in Yosemite Valley due to lo- myer and Hurteau 2010). We accounted sity smoke. At many of the higher elevation terson et al. 2015). son-days associated with medium density dium- and high-density HMS smoke from cal sources combined with regional wildfire for this effect within the 10,385 acres of sites, the normal values were far below the 35 smoke or 7 million total person-days of ex- the Rim Fire on that day covered a large area impacts often reach or exceed such levels (Pre- the Rim Fire’s footprint that had experi- 3 Results of the Case Study 2 g/m , so only heavy smoke plumes were posure to higher than normal levels of PM2.5 (251,691 mi ) with a population of 2.8 mil- isler et al. 2015). enced prescribed fires or resource objec- This comparison cannot prove that likely to cause an exceedance. However, in (Figure 4D). Ground-level monitoring indi- lion people, more than 2 million of whom By specifically evaluating smoke disper- tive wildfires since 2002 by changing managing fire for resource objectives has re- some urban sites, the norm during the fire cated that these large smoke plumes coin- resided below high-density smoke. Within sal, our analysis shows that extremely large “typical” fuel loads to “light,” which re- 3 duced smoke impacts in a particular land- season was already close to the 35 g/m cided with highly polluted days in Reno, Yosemite Village, which had few of its typi- and intense fires like the Rim Fire presented duced estimated emissions in those areas scape compared with those for a suppres- threshold for “unhealthy for sensitive groups.” which occurred on August 23–25 and again cal visitors, levels exceeded the “very un- a disproportionately greater public health by 53% and was consistent with the ap- 3 sion-centered strategy, but it illustrates how 3 In such locations, even the modest 5 g/m on August 28–29, when PM2.5 values ex- healthy” standard of 150 g/m (Figure 5B) hazard than did previous fires managed for proach used by Stephens et al. (2007) increase due to medium density smoke managing fires for resource objectives could ceeded the “unhealthy for all populations” on that day. (see Supplemental Table S1 for all the resource objectives. Specifically, the Rim could cause or worsen exceedances (Cisneros reduce smoke impacts relative to those for standard (55.5 g/m3) (Figure 4F). Such In contrast, the Grouse and Harden parameters used). Because recent reburns Fire burned 55 times more area (257,213 et al. 2014). more extreme fires. An actual experiment to high levels are such a serious health concern Fires burned slowly over the early summer of constituted only 4.0% of the entire Rim acres) than the combined footprint of the We mapped the HMS medium- and compare the effects of managing fire for re- that people are advised to avoid going out- 2009, with very modest emissions until the Fire area, accounting for those reduced Grouse and Harden Fires (4,695 acres), but high-density smoke plumes that appeared to source objectives at a scale large enough to doors. Navarro et al. (2016) reported that last week of June (Figure 4A). Both fires emissions lowered the estimated emis- our analysis suggests that it had at least 275 be connected with the Rim, Grouse, and evaluate major smoke impacts seems impracti- very unhealthy and unhealthy days occurred were managed for resource objectives using sions for the entire fire by only 3.2% (also times greater impact in terms of person- Harden Fires. In particular, we had to de- cal given the enormous consequences, and at 10 air monitoring sites in the central Sier- different tactics. For the Grouse Fire, man- see Supplemental Table S1). Had the en- days, or 5.5 times greater impact relative to marcate some of the smoke plumes during modeling approaches face numerous chal- ras, northern Sierras, and Nevada during the agers used an intensive “push-pull” ap- tire area of the Rim Fire been treated the early days of the Rim Fire that over- lenges in representing the effectiveness of the Rim Fire. They used BlueSky modeling cal- proach of checking the fire’s spread during area burned. Impacts from the Rim Fire are with recent fire, its estimated emissions lapped with the American Fire, which was suppression response and the impacts of low- ibrated with ground-level monitoring to es- periods of limited dispersion and speeding even greater than suggested by those figures would have been reduced by 48% (see fully contained by Aug. 29, 2013. To repre- probability or uncertain extreme events (Hasan timate that the Rim Fire exposed 1.2 million up the fire through aerial ignition during pe- because of the increased hazard associated Supplemental Table S2), with most of sent ground-level impacts over time, we ex- and Foliente 2015). people to smoke across 10 counties in Cali- riods of favorable dispersion (Figure 6), par- with unhealthy levels for the general popu- the reduction due to the change of dead amined the values for particulate matter for Within the study area, daily emissions fornia and Nevada during 37 days of active ticularly during its last week, while slowing lation and high-density smoke plumes, nei- fuels from “typical” to “light.” Similarly, the Rim and Grouse Fires at monitoring from both prescribed burns and resource ob- growth. Multiplying our estimate of person- the fire in wet drainages and when disper- ther of which was reported during the Wiedinmyer and Hurteau (2010) found sites in Yosemite Village in both 2009 and jective wildfires remained well below 500 days of exposure by the willingness to pay to sion conditions were less favorable. Mean- Grouse and Harden resource objective wild- that replacing infrequent wildfires with fires. 2013 and in Reno for 2013 (data were not tons PM2.5, whereas the Rim Fire had 20 avoid days of smoke impact found by Rich- while, the Harden Fire was contained on its lighter prescribed burns would roughly available for 2009)(US EPA 2016). days exceeding that threshold (nearly half of ardson et al. (2012) suggests that the cost of western flank and allowed to spread east- halve fire emissions in dry forests in the For the third step of the analysis, we its entire period of active fire growth) and smoke impact may have been nearly $600 ward toward natural barriers so that re- Potential Benefits of Resource Western United States, although they used ArcMap (version 10.1; ESRI, Red- peaked at nearly 11,000 tons PM2.5/day on million. That sum is comparable to the en- sources could be focused on managing the Objective Wildfires also noted that emissions would be more lands, CA) to sum the populations (based on Aug. 26, 2013 (Figure 2). During the late tire estimate of non-air quality losses in en- more complex Grouse Fire (van Wagten- The framework can also be useful for frequent under a prescribed burning sce- 2010 census data) for each census block summer, air quality is already problematic in vironmental benefits resulting from the Rim donk 2012). Our analysis of HMS maps in- comparing the public health impacts of fire nario. whose centroid was within areas of high- and downwind areas such as the Lake Tahoe Ba- Fire (Batker et al. 2013). dicated that there were only 2 days when management strategies across broader tem- 2. More favorable dispersion and potential for
80 Journal of Forestry • January 2018 Journal of Forestry • MONTH 2017 5 6 Journal of Forestry • MONTH 2017 Prescribed fire Rim (2013) sin and San Joaquin Valley. In contrast, 10000 Resource objective wildfire most of the fires managed for resource ob- A 5000 Suppression wildfire Frog/Laurel (2006) Tuolumne (2003) jectives had peak emissions either in the 2000 Kibbie (2003) Harden (2009) 1000 early summer or in the fall, when airshed 500 Grouse (2009) conditions are often more favorable. 200 Modeled emissions from the Rim Fire
100 within each vegetation polygon varied by an 50 order of magnitude because of the different 20 types of vegetation and amount of accumu- 10 lated fuels (Figure 3A). However, daily 5 emissions from the Rim Fire (Figure 3B) 2 0.308 - 500 0,00- 0.08 0.00-11 .6 were quite high even in areas that had rela- - 500-3,000 1 0.08- 0.16 11.7-20,2 - 3,000 - 5.000 0.5 tively light fuels (e.g., in the center of the 0 .1 6 - 0.26 20.2-37.1 - 5.000-10,600 Daily Emissions (Log2 of Tons PM2.5) 0.26 - 0.37 37.2 • 52.0 i:ii8S Prescribed fires 2002-2012 fire, as shown in Figure 3A), because the fire .,: v! . 0.37-0.53 52 .0 - 73.6 • Resource objective wildfires 2002-2012 burned through those areas so quickly. 10 5 0 10mlles ~ • Large water bodies "-: The Rim Fire generated a surge in emis- .... . -·~ Rocks and unburnable wet areas ·····-:. Jun Jul Aug Sep Oct Nov sions that began on August 19 and contin- Figure 3. Fuel loading and estimated emissions of PM2.5 per unit area within different vegetation polygons (A) and estimated emissions of Time of Year (2002-2013) -- ued through September 8, after which esti- PM2.5 per day within daily progression polygons (B), within the Rim Fire. Figure 2. Estimated daily emissions of PM , as generated by the FOFEM for different types 2.5 mated daily emissions dropped below 265 of fire within the Tuolumne River watershed, California, between 2002 and 2013. Points on tons (Figure 4B). That rapid growth resulted The enormous spatial impact of the medium-density plumes overlaid substantial poral and spatial scales. Our analyses help to each line represent one day of estimated emissions for a specific fire. Note the log base 2 y-axis; see Figures 4A and 4B for the nonlogarithmic version. in high- and medium-density smoke plumes Rim Fire is illustrated by a map on August populations in California and Nevada, illustrate and begin to quantify many of the over large populations in California and Ne- 31 (Figure 5), showing how plumes ex- amounting to 25,000 person-days (Figure potential benefits of resource objective wild- vada between August 22 and September 10 tended north from the Rim Fire over large 4C). Ground-level monitoring at Yosemite fires compared with those of extreme fires: plumes were excluded. In particular, they es- medium-density smoke plumes. We con- (Figure 4D). We weighted the population urban populations in the Lake Tahoe Basin Village showed relatively modest impacts, 1. Reduced fuels and reduced consumption. timated that the daily 96th percentile value fined our analysis to California and Nevada, estimates by the weights reported by Preisler (55,607), Carson City (55,212), Sparks where PM2.5 levels remained between 30 Where fires burn over the footprints of 3 5 3 of PM2.5 increased by 14 g/m (95% CL because archived reports and AirNow et al. (2015) (36.5 and 17% for high- and (90,264), and Reno (225,221), before shift- and 55 g/m from July 1 through July 6 relatively recent fires, emissions estimates 3 6 was 9–19 g/m ) where there was heavy maps did not indicate air quality impacts in medium-density smoke plumes, respectively), ing southward over Fresno (494,665), Mad- (Figure 4E). Although these levels are con- should typically reflect lighter fuels and 3 smoke and by 5 g/m (95% CL was 3.6– other states despite evidence that Rim Fire to estimate 2.9 million person-days associ- era (61,416), Visalia (124,442), and other sidered “unhealthy for sensitive groups,” reduced crown consumption (Wiedin- 3 6.1 g/m ) where there was medium-den- smoke traveled as far as central Canada (Pe- ated with high smoke and 4.1 million per- cities in the Central Valley. Altogether, me- concentrations in Yosemite Valley due to lo- myer and Hurteau 2010). We accounted sity smoke. At many of the higher elevation terson et al. 2015). son-days associated with medium density dium- and high-density HMS smoke from cal sources combined with regional wildfire for this effect within the 10,385 acres of sites, the normal values were far below the 35 smoke or 7 million total person-days of ex- the Rim Fire on that day covered a large area impacts often reach or exceed such levels (Pre- the Rim Fire’s footprint that had experi- 3 Results of the Case Study 2 g/m , so only heavy smoke plumes were posure to higher than normal levels of PM2.5 (251,691 mi ) with a population of 2.8 mil- isler et al. 2015). enced prescribed fires or resource objec- This comparison cannot prove that likely to cause an exceedance. However, in (Figure 4D). Ground-level monitoring indi- lion people, more than 2 million of whom By specifically evaluating smoke disper- tive wildfires since 2002 by changing managing fire for resource objectives has re- some urban sites, the norm during the fire cated that these large smoke plumes coin- resided below high-density smoke. Within sal, our analysis shows that extremely large “typical” fuel loads to “light,” which re- 3 duced smoke impacts in a particular land- season was already close to the 35 g/m cided with highly polluted days in Reno, Yosemite Village, which had few of its typi- and intense fires like the Rim Fire presented duced estimated emissions in those areas scape compared with those for a suppres- threshold for “unhealthy for sensitive groups.” which occurred on August 23–25 and again cal visitors, levels exceeded the “very un- a disproportionately greater public health by 53% and was consistent with the ap- 3 sion-centered strategy, but it illustrates how 3 In such locations, even the modest 5 g/m on August 28–29, when PM2.5 values ex- healthy” standard of 150 g/m (Figure 5B) hazard than did previous fires managed for proach used by Stephens et al. (2007) increase due to medium density smoke managing fires for resource objectives could ceeded the “unhealthy for all populations” on that day. (see Supplemental Table S1 for all the resource objectives. Specifically, the Rim could cause or worsen exceedances (Cisneros reduce smoke impacts relative to those for standard (55.5 g/m3) (Figure 4F). Such In contrast, the Grouse and Harden parameters used). Because recent reburns Fire burned 55 times more area (257,213 et al. 2014). more extreme fires. An actual experiment to high levels are such a serious health concern Fires burned slowly over the early summer of constituted only 4.0% of the entire Rim acres) than the combined footprint of the We mapped the HMS medium- and compare the effects of managing fire for re- that people are advised to avoid going out- 2009, with very modest emissions until the Fire area, accounting for those reduced Grouse and Harden Fires (4,695 acres), but high-density smoke plumes that appeared to source objectives at a scale large enough to doors. Navarro et al. (2016) reported that last week of June (Figure 4A). Both fires emissions lowered the estimated emis- our analysis suggests that it had at least 275 be connected with the Rim, Grouse, and evaluate major smoke impacts seems impracti- very unhealthy and unhealthy days occurred were managed for resource objectives using sions for the entire fire by only 3.2% (also times greater impact in terms of person- Harden Fires. In particular, we had to de- cal given the enormous consequences, and at 10 air monitoring sites in the central Sier- different tactics. For the Grouse Fire, man- see Supplemental Table S1). Had the en- days, or 5.5 times greater impact relative to marcate some of the smoke plumes during modeling approaches face numerous chal- ras, northern Sierras, and Nevada during the agers used an intensive “push-pull” ap- tire area of the Rim Fire been treated the early days of the Rim Fire that over- lenges in representing the effectiveness of the Rim Fire. They used BlueSky modeling cal- proach of checking the fire’s spread during area burned. Impacts from the Rim Fire are with recent fire, its estimated emissions lapped with the American Fire, which was suppression response and the impacts of low- ibrated with ground-level monitoring to es- periods of limited dispersion and speeding even greater than suggested by those figures would have been reduced by 48% (see fully contained by Aug. 29, 2013. To repre- probability or uncertain extreme events (Hasan timate that the Rim Fire exposed 1.2 million up the fire through aerial ignition during pe- because of the increased hazard associated Supplemental Table S2), with most of sent ground-level impacts over time, we ex- and Foliente 2015). people to smoke across 10 counties in Cali- riods of favorable dispersion (Figure 6), par- with unhealthy levels for the general popu- the reduction due to the change of dead amined the values for particulate matter for Within the study area, daily emissions fornia and Nevada during 37 days of active ticularly during its last week, while slowing lation and high-density smoke plumes, nei- fuels from “typical” to “light.” Similarly, the Rim and Grouse Fires at monitoring from both prescribed burns and resource ob- growth. Multiplying our estimate of person- the fire in wet drainages and when disper- ther of which was reported during the Wiedinmyer and Hurteau (2010) found sites in Yosemite Village in both 2009 and jective wildfires remained well below 500 days of exposure by the willingness to pay to sion conditions were less favorable. Mean- Grouse and Harden resource objective wild- that replacing infrequent wildfires with fires. 2013 and in Reno for 2013 (data were not tons PM2.5, whereas the Rim Fire had 20 avoid days of smoke impact found by Rich- while, the Harden Fire was contained on its lighter prescribed burns would roughly available for 2009)(US EPA 2016). days exceeding that threshold (nearly half of ardson et al. (2012) suggests that the cost of western flank and allowed to spread east- halve fire emissions in dry forests in the For the third step of the analysis, we its entire period of active fire growth) and smoke impact may have been nearly $600 ward toward natural barriers so that re- Potential Benefits of Resource Western United States, although they used ArcMap (version 10.1; ESRI, Red- peaked at nearly 11,000 tons PM2.5/day on million. That sum is comparable to the en- sources could be focused on managing the Objective Wildfires also noted that emissions would be more lands, CA) to sum the populations (based on Aug. 26, 2013 (Figure 2). During the late tire estimate of non-air quality losses in en- more complex Grouse Fire (van Wagten- The framework can also be useful for frequent under a prescribed burning sce- 2010 census data) for each census block summer, air quality is already problematic in vironmental benefits resulting from the Rim donk 2012). Our analysis of HMS maps in- comparing the public health impacts of fire nario. whose centroid was within areas of high- and downwind areas such as the Lake Tahoe Ba- Fire (Batker et al. 2013). dicated that there were only 2 days when management strategies across broader tem- 2. More favorable dispersion and potential for
Journal of Forestry • MONTH 2017 5 6 Journal of Forestry • MONTH 2017 Journal of Forestry • January 2018 81 .., w N (I) 0 3 ::x:, .... :!! fire will become increasingly important N C (I) cu I cu (I) 0 0 I.D G) a. a. 0 ::J ::J V> V> .., .., :!! CD )> 5)
. for reducing the likelihood and extent of _ 0 0 0 8 8 - large-scale, extreme fires like the Rim PM2 s
PM2.SJ Fire (Westerling et al. 2015). 0 0) 0, g ton
s ( s A comprehensive comparison of smoke tons ( impacts under different management strate- en 0) g 0 ssion i gies is challenging to model because of com- em emissions A- J:,. 0 g plex processes (such as capacity to suppress dally -- or safely manage fires), feedbacks (such as daily ...... -- .., 0 0 0 0 0 0 0 0 0 0 0 0 0 0 g "' changes in fuels, fire behavior, and emissions ......
/ after fires), sensitivity to extreme events, and Estimated I ... , ...... 0 0 . Estimated uncertainty of climate. Despite such chal- 3 3 3 3 3 3 3 3 3 3 3 3 3 1 1 1 1 1 1 1 1 1 1 1 1 1 009 009 009 0 0 013 0 009 lenges, smoke monitoring and smoke 2009 2009 2009 2009 20 2013 20 20 /2 /2 /2 /2 /2 /2 /2 /2 / / / / / 1 212009 7 9 6/2009 6/2009 5/2013 5 8/20 8/200! 1/2 009 4 4 0/2009 8 8/ / / / / /31 / /21 / / 17/20 1 11/2013 13/20 1 23/2013 25/2013 1 / / modeling are essential for guiding future 9 9 9 9 9 7/ 7/ 7 7 6/ 6/4 6 6 / / / /1 /1 9/21/20 9 9 9 9 9 9 9119 7 6116 6/ 6/22/2009 6/24/2009 6/26/2009 6/28/2009 6/30/2009 6/20 6/10/2009 6/12/2009 6/ 5/3 8/22/20 8/24/20 8/26/20 8/28 8/30/2013 8/20 8/ management strategies, given changing 0 o _ 0 0 § g (uglm3) (uglm31 large areas can burn with relatively minor ;;:, ::r C s: .. ., '< 0 0 g smoke impacts. An important strategy is to PM2.5 PM2.5 manage fires at least in part based on moni- ... "' ;;l - tored smoke concentrations at key monitor- average average ing sites, rather than applying predeter- g g hr hr ;;:, ;;:, ::r 82 Journal of Forestry • January 2018 Journal of Forestry • MONTH 2017 7 8 Journal of Forestry • MONTH 2017 fire will become increasingly important for reducing the likelihood and extent of large-scale, extreme fires like the Rim OREGON ID~ Fire (Westerling et al. 2015). A comprehensive comparison of smoke impacts under different management strate- gies is challenging to model because of com- plex processes (such as capacity to suppress or safely manage fires), feedbacks (such as changes in fuels, fire behavior, and emissions after fires), sensitivity to extreme events, and uncertainty of climate. Despite such chal- c... . lenges, smoke monitoring and smoke UTAH modeling are essential for guiding future management strategies, given changing Rim Fire (20 13) climate and fire regimes (McKenzie et al. Harden Fire (2009) 2014). Grouse Fire (2009) Strategies to Support Greater Use of Fire The framework and results of the case study suggest several management approaches to align ecological restoration with reduction in public health impacts through greater use of fire to achieve resource objectives. Managing Fires Based on Direct Smoke Impacts Although large-resource objective fires are challenging to implement and do not en- 2010 Census Block Groups tirely avoid smoke impacts, the Grouse Fire Populalon density (personslmll) and other resource objective fires (Schweizer 0-28 and Cisneros 2014) have demonstrated that 28-259 large areas can burn with relatively minor 259-2580 smoke impacts. An important strategy is to ..2590 -25900 ..>25900 manage fires at least in part based on moni- High density smoke 813112013 tored smoke concentrations at key monitor- Medium danaity smoke 8/31/2013 ing sites, rather than applying predeter- mined area limits or assuming a direct Figure 5. HMS smoke plumes from the Rim Fire on Aug. 31, 2013, a day of extensive heavy correlation with smoke impacts. This ap- smoke impact, overlying population density of census tracts in California and Nevada. proach is directly responsive to human health impacts while providing greater flexibility to take advantage of favorable or concentrations projected using tools 4. Creating anchors that facilitate future fire dispersion and treat more area during a Figure 4. Comparisons between three measures of potential smoke impact during the Grouse and Harden fires of 2009 (left) and the Rim such as the BlueSky framework. When management. In addition, strategic place- given fire. wildfire of 2013 (right), including estimated daily emissions (top); estimated person-days of smoke impact associated with medium- and dispersion is limited, firefighters can slow ment of relatively large resource objective high-density smoke plumes (middle, note that the y-axis scale on the chart of impacts during the Grouse and Harden fires [C] is 250 times progression by herding the fire into bar- fires within fire-prone landscapes can cre- Public Communication to Reduce smaller than the one for the Rim Fire [D]); and 24-hour rolling average PM2.5 values (bottom, shown in relation to associated levels of health riers, roads, and areas of previous fuel re- ate anchors for limiting the spread of haz- Smoke Impact and Support Use concern) for Yosemite Village, California, and Reno, Nevada (data were not available for Reno in 2009). duction. Conversely, they can encourage ardous wildfires and increasing opportu- of Fire fire growth during periods of favorable nities to use future wildfires for resource Smoke exposure can be reduced by dis- dispersion to dilute emissions. They can benefit (North et al. 2015). For example, tilling monitoring and modeling tools into less ozone. As maintenance burns reduce greater lofting and concentrations of source objectives when weather and fire burn perimeters (“black-lining”) under the cumulative effect of decades of man- information that people can use to their fuel levels over time, managers may be smoke reported from extreme fires, behavior conditions are more moderate marginal dispersion conditions, just be- aging wildland fires for resource objec- modify behavior (Olsen et al. 2014, Rap- able to burn more safely earlier in the which often deliver pollution to distant, than under extreme wildfires, their fore a mild front arrives and then ignite tives in Yosemite National Park created pold et al. 2014). The combination of ad- summer and or later in the fall, when dis- large urban populations in lower-eleva- slower fire spread can curb daily emis- larger interior areas aerially once more fa- opportunities to suppress the Rim Fire vance warning and education can also in- persion is often more favorable and tion valleys (Colarco et al. 2004, Peter- sions. In addition, managers can employ vorable dispersion arrives. Using such on its northern and eastern edge (Figure crease public support for managing fire for ozone concentrations are lower (Jaffe son et al. 2015). the push-pull tactics burn described for techniques requires resources and coor- 3A). As fuels continue to accumulate and resource objectives over time (Sneeuwjagt et al. 2013). Fires managed for resource 3. Greater ability to regulate fire spread. Be- the Grouse Fire to regulate daily emis- dination among fire operations, air qual- climate conditions increase the likeli- et al. 2013, Blades et al. 2014). It is impor- objectives are less likely to result in the cause wildfires would be managed for re- sions based on monitored concentrations ity, and meteorology specialists. hood of large fires, such strategic use of tant for outreach efforts to explain how for- Journal of Forestry • MONTH 2017 7 8 Journal of Forestry • MONTH 2017 Journal of Forestry • January 2018 83 large fires in the contiguous United States. Int. FULE´, P.Z., T.W. SWETNAM, P.M. BROWN, D.A. LIU, J.C., G. PEREIRA, S.A. UHL, M.A. BRAVO, portunities for managers. Environ. Manage. J. Wildl. Fire. 24(7):892–899. FALK, D.L. PETERSON, C.D. ALLEN, G.H. AND M.L. BELL. 2015. A systematic review of 54(3):571–582. BATKER, D., Z. CHRISTIN,R.SCHMIDT, AND I. DE APLET, ET AL. 2014. Unsupported inferences of the physical health impacts from non-occupa- PETERSON, D.A., E.J. HYER, J.R. CAMPBELL, LA TORRE. 2013. The economic impact of the high-severity fire in historical dry forests of the tional exposure to wildfire smoke. Environ. M.D. FROMM, J.W. HAIR, C.F. BUTLER, AND 2013 Rim Fire on natural lands. Earth Eco- western United States: Response to Williams Res. 136:120–132. M.A. FENN. 2015. The 2013 Rim Fire: Impli- nomics, Tacoma, WA. 53 p. and Baker. Global Ecol. Biogeogr. 23(7):825– LUECK, D., AND J. YODER. 2015. The economic cations for predicting extreme fire spread, py- BLADES, J.J., S.R. SHOOK, AND T.E. HALL. 2014. 830. foundations of firefighting organizations and roconvection, and smoke emissions. Bull. Am. Smoke management of wildland and pre- FUSINA, L., S. ZHONG,J.KORACIN,T.BROWN, A. institutions. J. For. 113(3):291–297. Meteorol. Soc. 96(2):229–247. scribed fire: Understanding public preferences ESPERANZA,L.TARNAY, AND H. PREISLER. LUTES, D.C. 2014. First Order Fire Effects Model PREISLER, H.K., D. SCHWEIZER,R.CISNEROS, T. and trade-offs. Can. J. For. Res. 44(11):1344– 2007. Validation of BlueSky Smoke Predic- user guide version 6.1. USDA Forest Service PROCTER,M.RUMINSKI, AND L. TARNAY. 1355. tion System using surface and satellite observa- Rocky Mountain Research Station, Fort Col- 2015. A statistical model for determining im- CALIFORNIA AIR RESOURCES BOARD. 2015. Smoke tions during major wildland fire events in lins, CO. 66 p. pact of wildland fires on particulate matter management program technical tools. Available Northern California. P. 403–408 in The fire LYDERSEN, J.M., B.M. COLLINS, C.M. EWELL, (PM2.5) in Central California aided by satellite online at www.arb.ca.gov/smp/techtool/ environment—Innovations, management, and A.L. REINER, J.A. FITES,P.GONZALEZ, D.S. imagery of smoke. Environ. Poll. 205:340– techtool.htm; last accessed Apr. 2, 2015. policy; conference proceedings, 26–30 March SAAH, AND J.J. BATTLES. 2014. Using field data 349. CALKIN, D.E., M.P. THOMPSON, AND M.A. 2007, Destin, FL, Butler, B.W. and W. Cook to assess model predictions of surface and QUINN-DAVIDSON, L., AND J.M. VARNER. 2012. FINNEY. 2015. Negative consequences of posi- (comps.). USDA Forest Service Proc. RMRS- ground fuel consumption by wildfire in conif- Impediments to prescribed fire across agency, tive feedbacks in US wildfire management. P-46CD, Rocky Mountain Research Station, erous forests of California. J. Geophys. Res. Bio- landscape and manager: An example from For. Ecosyst. 2(1):9. Fort Collins, CO. geosci. 119(3):223–235. northern California. Int. J. Wildl. Fire 21(3): CARROLL, M.S., K.A. BLATNER, AND P.J. COHN. GAITHER, C.J., S. GOODRICK, B.E. MURPHY, AND MCKENZIE, D., U. SHANKAR, R.E. KEANE, E.N. 210–218. 2007. Managing fire danger in the forests of N. POUDYAL. 2015. An exploratory spatial STAVROS, W.E. HEILMAN, D.G. FOX, AND A.C. RAPPOLD, A.G., N.L. FANN,J.CROOKS, J. the US inland Northwest: A classic “wicked analysis of social vulnerability and smoke RIEBAU. 2014. Smoke consequences of new HUANG, W.E. CASCIO, R.B. DEVLIN, AND D. problem” in public land policy. J. For. 105(5): plume dispersion in the US South. Forests 6(5): wildfire regimes driven by climate change. DIAZ-SANCHEZ. 2014. Forecast-based inter- 239–244. 1397–1421. Earth’s Future 2(2):35–59. ventions can reduce the health and economic Figure 6. Fire managers pushed the Grouse Fire to spread during periods of favorable burden of wildfires. CISNEROS, R., D. SCHWEIZER,H.PREISLER, D.H. GOODRICK, S.L., G.L. ACHTEMEIER, N.K. LAR- MILLER, J.D., AND H.D. SAFFORD. 2012. Trends Environ. Sci. Technol. smoke dispersion, including times at night. 48(18):10571–10579. BENNETT,G.SHAW, AND A. BYTNEROWICZ. KIN,Y.LIU, AND T.M. STRAND. 2013. Model- in wildfire severity 1984–2010 in the Sierra 2014. Spatial and seasonal patterns of particu- ling smoke transport from wildland fires: A re- Nevada, Modoc Plateau, and southern Cas- RICHARDSON, L.A., P.A. CHAMP, AND J.B. LOO- late matter less than 2.5 microns in the Sierra view. Int. J. Wildl. Fire 22(1):83–94. cades, California, USA. Fire Ecol. 8(3):41–57. MIS. 2012. The hidden cost of wildfires: Eco- est management can mitigate smoke impacts framework to more directly account for re- nomic valuation of health effects of wildfire Nevada Mountains, California. Atmos. Pollut. HASAN, S., AND G. FOLIENTE. 2015. Modeling MOELTNER, K., M.K. KIM,E.ZHU, AND W. from extreme fires not only in rural forest gional-scale smoke impacts from these smoke exposure in Southern California. J. For. Res. 5(4):581–590. infrastructure system interdependencies and YANG. 2013. Wildfire smoke and health im- communities but also in downwind urban events using surface monitoring and satellite Econ. 18(1):14–35. CLINTON, N., J. SCARBOROUGH,Y.TIAN, AND P. socioeconomic impacts of failure in extreme pacts: A closer look at fire attributes and their populations. RUMINSKI, M., S. KONDRAGUNTA,R.DRAXLER, observations of smoke. Managing large fires GONG. 2003. A GIS based emissions estima- events: Emerging R&D challenges. Nat. Haz- marginal effects. J. Environ. Econ. Manage. AND G. ROLPH. 2007. Use of environmental for resource objectives can shift the release of tion system for wildfire and prescribed burn- ards 78(3):2143–2168. 66(3):476–496. satellite imagery for smoke depiction and Engaging Air Resource Advisors ing. P. 14 in EPA 12th Annual Emission Inven- HUNTER, M.E., J.M. INIGUEZ, AND C.A. FARRIS. NAKAYAMA WONG, L.S., H.H. AUNG, M.W. inevitable emissions to conditions that min- transport model initialization. In 16th Annual Air resource advisors, who are com- tory Conference. US Environmental Protection 2014. Historical and current fire management LAME´, T.C. WEGESSER, AND D.W. WILSON. imize large-scale smoke impacts, by control- international emission inventory Conference: monly assigned to large wildfire incident Agency, Washington, DC. practices in two wilderness areas in the southwest- 2011. Fine particulate matter from urban am- ling fire spread based on available dispersion Emission inventories “Integration, analysis, and teams, can also help managers mitigate the COLARCO, P., M. SCHOEBERL,B.DODDRIDGE, L. ern United States: The Saguaro Wilderness Area bient and wildfire sources from California’s and monitored impacts and creating an- communications.” US Environmental Protec- negative health effects of prescribed fires and MARUFU,O.TORRES, AND E. WELTON. 2004. and the Gila-Aldo Leopold Wilderness Complex. San Joaquin Valley initiate differential inflam- tion Agency, Washington, DC. 16 p. chors for containing future hazardous fires. Transport of smoke from Canadian forest fires USDA Forest Service Gen. Tech. Rep. RMRS- matory, oxidative stress, and xenobiotic re- resource objective wildfires. Such trained SCHWEIZER, D., AND R. CISNEROS. 2014. Wild- to the surface near Washington, DC: Injection GTR-325, Rocky Mountain Research Station, sponses in human bronchial epithelial cells. specialists can facilitate public communica- When well supported by firefighting, air land fire management and air quality in the quality monitoring and modeling, and pub- height, entrainment, and optical properties. J. Fort Collins, CO. 38 p. Toxicol. In Vitro 25(8):1895–1905. southern Sierra Nevada: Using the Lion Fire as tion and prepare documents used by air Geophys. Res. Atmos. 109(D6):D06203. HURTEAU, M.D., A.L. WESTERLING,C.WIEDIN- NAVARRO, K.M., R. CISNEROS, S.M. O’NEILL, D. lic communications resources, this approach a case study with a multi-year perspective on quality regulators in permitting burns and CROSS, J.A. 2001. Megacities and small towns: MYER, AND B.P. BRYANT. 2014. Projected ef- SCHWEIZER, N.K. LARKIN, AND J.R. BALMES. PM impacts and fire policy. can overcome existing disincentives for 2.5 J. Environ. considering exemptions for exceedances of Different perspectives on hazard vulnerability. fects of climate and development on California 2016. Air-quality impacts and intake fraction Manage. 144:265–278. air quality standards that might be caused by achieving ecological and public health goals. Global Environ. Change Part B Environ. Haz- wildfire emissions through 2100. Environ. Sci. of PM2.5 during the 2013 Rim megafire. En- SCHWEIZER, D.W., AND R. CISNEROS. 2016. For- such fires. ards 3(2):63–80. Technol. 48(4):2298–2304. viron. Sci. Technol. 50(21):11965–11973. est fire policy: Change conventional thinking Endnotes DAVIS, F.W., D.M. STOMS, A.D. HOLLANDER, JAFFE, D.A., N. WIGDER,N.DOWNEY,G.PFIS- NGO, M.A., K.E. PINKERTON,S.FREELAND, M. of smoke management to prioritize long-term Incentivizing Reduction of Smoke 1. For more information, see http://playground. K.A. THOMAS, P.A. STINE,D.ODION, M.I. TER,A.BOYNARD, AND S.B. REID. 2013. Im- GELLER,W.HAM,S.CLIFF, L.E. HOPKINS, ET air quality and public health. Air Qual. Atmos. airfire.org. BORCHERT, ET AL. 1998. The California Gap pact of wildfires on ozone exceptional events in AL. 2010. Airborne particles in the San Joaquin Health 10(1):33–36. Impacts Rather than Area Burned Analysis Project. Final Rep., Univ. of Califor- the Western U.S. Environ. Sci. Technol. Valley may affect human health. Calif. Agric. Rather than using area burned, public 2. For more information, see www.arl.noaa.gov/ SNEEUWJAGT, R.J., T.S. KLINE, AND S.L. STE- HYSPLIT_info.php. nia, Santa Barbara, Santa Barbara, CA. 255 p. 47(19):11065–11072. 64(1):12–16. PHENS. 2013. Opportunities for improved fire health objectives can be better met by mea- 3. For more information, see src.com/calpuff/ DELFINO, R.J., S. BRUMMEL,J.WU,H.STERN, B. JONES, B.A., J.A. THACHER, J.M. CHERMAK, AND NORTH, M., B.M. COLLINS, AND S.L. STEPHENS. use and management in California: Lessons suring exceedances of health thresholds at calpuff1.htm. OSTRO,M.LIPSETT,A.WINER, ET AL. 2009. R.P. BERRENS. 2016. Wildfire smoke health 2012. Using fire to increase the scale, benefits from Western Australia. Fire Ecol. 9(2):14– monitoring sites or person-days of expected 4. For more information, see www.epa.gov/ The relationship of respiratory and cardiovas- costs: A methods case study for a Southwestern and future maintenance of fuels treatments. J. 25. harm. Replacing unit-area fees on restorative benmap. cular hospital admissions to the southern Cal- US ‘mega-fire.’ J. Environ. Econ. Policy 5(2): For. 110(7):392–401. STEPHENS, S.L., R.E. MARTIN, AND N.E. CLIN- ifornia wildfires of 2003. Occup. Environ. Med. 181–199. NORTH, M.P., S.L. STEPHENS, B.M. COLLINS, burning with flat annual fees, as employed 5. For more information, see inciweb.nwcg. TON. 2007. Prehistoric fire area and emissions 66(3):189–197. KOCHI, I., G.H. DONOVAN, P.A. CHAMP, AND J.K. AGEE,G.APLET, J.F. FRANKLIN, AND P.Z. gov/. from California’s forests, woodlands, shru- by some air districts, would reduce one of ENGEL, K.H. 2013. Perverse incentives: The case J.B. LOOMIS. 2010. The economic cost of ad- FULE´. 2015. Reform forest fire management. 6. For more information, see airnow.gov/. blands, and grasslands. For. Ecol. Manage. the disincentives for intentionally using of wildfire smoke regulation. Ecol. Law Q. verse health effects from wildfire-smoke expo- Science 349(6254):1280–1281. 251(3):205–216. large fires to restore landscapes (Sneeuwjagt 40(3):623–672. sure: A review. Int. J. Wildl. Fire 19(7):803– OFFICE OF THE FEDERAL REGISTER. 2015. Treat- STEPHENS, S.L., AND N.G. SUGIHARA. 2006. Fire et al. 2013). Literature Cited FINLAY, S.E., A. MOFFAT,R.GAZZARD,D.BAKER, 817. ment of data influenced by exceptional events, management and policy since European settle- AMERICAN LUNG ASSOCIATION. 2015. Most polluted AND V. MURRAY. 2012. Health impacts of KUMAGAI, Y., M.S. CARROLL, AND P. COHN. proposed rule. Available online at www. ment. P. 431–443 in Fire in California’s ecosys- Conclusion cities. Available online at www.stateoftheair. wildfires. PLOS Curr. Disasters. Available online at 2004. Coping with interface wildfire as a hu- federalregister.gov/articles/2015/11/20/ tems, Sugihara, N.G., J.W. van Wagtendonk, org/2015/city-rankings/most-polluted- http://currents.plos.org/disasters/article/health- man event: Lessons from the disaster/hazards 2015-29350/treatment-of-data-influenced- The combination of a warming climate J. Fites-Kaufman, K.E. Shaffer, and A.E. Th- cities.html; last accessed Jan. 12, 2016. impacts-of-wildfires/. literature. J. For. 102(6):28–32. by-exceptional-events; last accessed Dec. 11, 2015. ode (eds.). Univ. of California Press, Berkeley, and accumulation of forest fuels ensures a BARBERO, R., J.T. ABATZOGLOU, N.K. LARKIN, FOWLER, C. 2003. Human health impacts of for- LEENHOUTS, B. 1998. Assessment of biomass OLSEN, C.S., D.K. MAZZOTTA,E.TOMAN, AND CA. future with more large fires and smoke in dry C.A. KOLDEN, AND B. STOCKS. 2015. Climate est fires in the southern United States: A liter- burning in the conterminous United States. A.P. FISCHER. 2014. Communicating about STRAND, T.M., N. LARKIN, K.J. CRAIG,S.RAF- western US forests. We have outlined a change presents increased potential for very ature review. J. Ecol. Anthropol. 7:39–59. Conserv. Ecol. 2(1):1. smoke from wildland fire: Challenges and op- FUSE,D.SULLIVAN,R.SOLOMON,M.RORIG, 84 Journal of Forestry • January 2018 Journal of Forestry • MONTH 2017 9 10 Journal of Forestry • MONTH 2017 large fires in the contiguous United States. Int. FULE´, P.Z., T.W. SWETNAM, P.M. BROWN, D.A. LIU, J.C., G. PEREIRA, S.A. UHL, M.A. BRAVO, portunities for managers. Environ. Manage. J. Wildl. Fire. 24(7):892–899. FALK, D.L. PETERSON, C.D. ALLEN, G.H. AND M.L. BELL. 2015. A systematic review of 54(3):571–582. BATKER, D., Z. CHRISTIN,R.SCHMIDT, AND I. DE APLET, ET AL. 2014. Unsupported inferences of the physical health impacts from non-occupa- PETERSON, D.A., E.J. HYER, J.R. CAMPBELL, LA TORRE. 2013. The economic impact of the high-severity fire in historical dry forests of the tional exposure to wildfire smoke. Environ. M.D. FROMM, J.W. HAIR, C.F. BUTLER, AND 2013 Rim Fire on natural lands. Earth Eco- western United States: Response to Williams Res. 136:120–132. M.A. FENN. 2015. The 2013 Rim Fire: Impli- nomics, Tacoma, WA. 53 p. and Baker. Global Ecol. Biogeogr. 23(7):825– LUECK, D., AND J. YODER. 2015. The economic cations for predicting extreme fire spread, py- BLADES, J.J., S.R. SHOOK, AND T.E. HALL. 2014. 830. foundations of firefighting organizations and roconvection, and smoke emissions. Bull. Am. Smoke management of wildland and pre- FUSINA, L., S. ZHONG,J.KORACIN,T.BROWN, A. institutions. J. For. 113(3):291–297. Meteorol. Soc. 96(2):229–247. scribed fire: Understanding public preferences ESPERANZA,L.TARNAY, AND H. PREISLER. LUTES, D.C. 2014. First Order Fire Effects Model PREISLER, H.K., D. SCHWEIZER,R.CISNEROS, T. and trade-offs. Can. J. For. Res. 44(11):1344– 2007. Validation of BlueSky Smoke Predic- user guide version 6.1. USDA Forest Service PROCTER,M.RUMINSKI, AND L. TARNAY. 1355. tion System using surface and satellite observa- Rocky Mountain Research Station, Fort Col- 2015. A statistical model for determining im- CALIFORNIA AIR RESOURCES BOARD. 2015. Smoke tions during major wildland fire events in lins, CO. 66 p. pact of wildland fires on particulate matter management program technical tools. Available Northern California. P. 403–408 in The fire LYDERSEN, J.M., B.M. COLLINS, C.M. EWELL, (PM2.5) in Central California aided by satellite online at www.arb.ca.gov/smp/techtool/ environment—Innovations, management, and A.L. REINER, J.A. FITES,P.GONZALEZ, D.S. imagery of smoke. Environ. Poll. 205:340– techtool.htm; last accessed Apr. 2, 2015. policy; conference proceedings, 26–30 March SAAH, AND J.J. BATTLES. 2014. Using field data 349. CALKIN, D.E., M.P. THOMPSON, AND M.A. 2007, Destin, FL, Butler, B.W. and W. Cook to assess model predictions of surface and QUINN-DAVIDSON, L., AND J.M. VARNER. 2012. FINNEY. 2015. Negative consequences of posi- (comps.). USDA Forest Service Proc. RMRS- ground fuel consumption by wildfire in conif- Impediments to prescribed fire across agency, tive feedbacks in US wildfire management. P-46CD, Rocky Mountain Research Station, erous forests of California. J. Geophys. Res. Bio- landscape and manager: An example from For. Ecosyst. 2(1):9. Fort Collins, CO. geosci. 119(3):223–235. northern California. Int. J. Wildl. Fire 21(3): CARROLL, M.S., K.A. BLATNER, AND P.J. COHN. GAITHER, C.J., S. GOODRICK, B.E. MURPHY, AND MCKENZIE, D., U. SHANKAR, R.E. KEANE, E.N. 210–218. 2007. Managing fire danger in the forests of N. POUDYAL. 2015. An exploratory spatial STAVROS, W.E. HEILMAN, D.G. FOX, AND A.C. RAPPOLD, A.G., N.L. FANN,J.CROOKS, J. the US inland Northwest: A classic “wicked analysis of social vulnerability and smoke RIEBAU. 2014. Smoke consequences of new HUANG, W.E. CASCIO, R.B. DEVLIN, AND D. problem” in public land policy. J. For. 105(5): plume dispersion in the US South. Forests 6(5): wildfire regimes driven by climate change. DIAZ-SANCHEZ. 2014. Forecast-based inter- 239–244. 1397–1421. Earth’s Future 2(2):35–59. ventions can reduce the health and economic Figure 6. Fire managers pushed the Grouse Fire to spread during periods of favorable burden of wildfires. CISNEROS, R., D. SCHWEIZER,H.PREISLER, D.H. GOODRICK, S.L., G.L. ACHTEMEIER, N.K. LAR- MILLER, J.D., AND H.D. SAFFORD. 2012. Trends Environ. Sci. Technol. smoke dispersion, including times at night. 48(18):10571–10579. BENNETT,G.SHAW, AND A. BYTNEROWICZ. KIN,Y.LIU, AND T.M. STRAND. 2013. Model- in wildfire severity 1984–2010 in the Sierra 2014. Spatial and seasonal patterns of particu- ling smoke transport from wildland fires: A re- Nevada, Modoc Plateau, and southern Cas- RICHARDSON, L.A., P.A. CHAMP, AND J.B. LOO- late matter less than 2.5 microns in the Sierra view. Int. J. Wildl. Fire 22(1):83–94. cades, California, USA. Fire Ecol. 8(3):41–57. MIS. 2012. The hidden cost of wildfires: Eco- est management can mitigate smoke impacts framework to more directly account for re- nomic valuation of health effects of wildfire Nevada Mountains, California. Atmos. Pollut. HASAN, S., AND G. FOLIENTE. 2015. Modeling MOELTNER, K., M.K. KIM,E.ZHU, AND W. from extreme fires not only in rural forest gional-scale smoke impacts from these smoke exposure in Southern California. J. For. Res. 5(4):581–590. infrastructure system interdependencies and YANG. 2013. Wildfire smoke and health im- communities but also in downwind urban events using surface monitoring and satellite Econ. 18(1):14–35. CLINTON, N., J. SCARBOROUGH,Y.TIAN, AND P. socioeconomic impacts of failure in extreme pacts: A closer look at fire attributes and their populations. RUMINSKI, M., S. KONDRAGUNTA,R.DRAXLER, observations of smoke. Managing large fires GONG. 2003. A GIS based emissions estima- events: Emerging R&D challenges. Nat. Haz- marginal effects. J. Environ. Econ. Manage. AND G. ROLPH. 2007. Use of environmental for resource objectives can shift the release of tion system for wildfire and prescribed burn- ards 78(3):2143–2168. 66(3):476–496. satellite imagery for smoke depiction and Engaging Air Resource Advisors ing. P. 14 in EPA 12th Annual Emission Inven- HUNTER, M.E., J.M. INIGUEZ, AND C.A. FARRIS. NAKAYAMA WONG, L.S., H.H. AUNG, M.W. inevitable emissions to conditions that min- transport model initialization. In 16th Annual Air resource advisors, who are com- tory Conference. US Environmental Protection 2014. Historical and current fire management LAME´, T.C. WEGESSER, AND D.W. WILSON. imize large-scale smoke impacts, by control- international emission inventory Conference: monly assigned to large wildfire incident Agency, Washington, DC. practices in two wilderness areas in the southwest- 2011. Fine particulate matter from urban am- ling fire spread based on available dispersion Emission inventories “Integration, analysis, and teams, can also help managers mitigate the COLARCO, P., M. SCHOEBERL,B.DODDRIDGE, L. ern United States: The Saguaro Wilderness Area bient and wildfire sources from California’s and monitored impacts and creating an- communications.” US Environmental Protec- negative health effects of prescribed fires and MARUFU,O.TORRES, AND E. WELTON. 2004. and the Gila-Aldo Leopold Wilderness Complex. San Joaquin Valley initiate differential inflam- tion Agency, Washington, DC. 16 p. chors for containing future hazardous fires. Transport of smoke from Canadian forest fires USDA Forest Service Gen. Tech. Rep. RMRS- matory, oxidative stress, and xenobiotic re- resource objective wildfires. Such trained SCHWEIZER, D., AND R. CISNEROS. 2014. Wild- to the surface near Washington, DC: Injection GTR-325, Rocky Mountain Research Station, sponses in human bronchial epithelial cells. specialists can facilitate public communica- When well supported by firefighting, air land fire management and air quality in the quality monitoring and modeling, and pub- height, entrainment, and optical properties. J. Fort Collins, CO. 38 p. Toxicol. In Vitro 25(8):1895–1905. southern Sierra Nevada: Using the Lion Fire as tion and prepare documents used by air Geophys. Res. Atmos. 109(D6):D06203. HURTEAU, M.D., A.L. WESTERLING,C.WIEDIN- NAVARRO, K.M., R. CISNEROS, S.M. O’NEILL, D. lic communications resources, this approach a case study with a multi-year perspective on quality regulators in permitting burns and CROSS, J.A. 2001. Megacities and small towns: MYER, AND B.P. BRYANT. 2014. Projected ef- SCHWEIZER, N.K. LARKIN, AND J.R. BALMES. PM impacts and fire policy. can overcome existing disincentives for 2.5 J. Environ. considering exemptions for exceedances of Different perspectives on hazard vulnerability. fects of climate and development on California 2016. Air-quality impacts and intake fraction Manage. 144:265–278. air quality standards that might be caused by achieving ecological and public health goals. Global Environ. Change Part B Environ. Haz- wildfire emissions through 2100. Environ. Sci. of PM2.5 during the 2013 Rim megafire. En- SCHWEIZER, D.W., AND R. CISNEROS. 2016. For- such fires. ards 3(2):63–80. Technol. 48(4):2298–2304. viron. Sci. Technol. 50(21):11965–11973. est fire policy: Change conventional thinking Endnotes DAVIS, F.W., D.M. STOMS, A.D. HOLLANDER, JAFFE, D.A., N. WIGDER,N.DOWNEY,G.PFIS- NGO, M.A., K.E. PINKERTON,S.FREELAND, M. of smoke management to prioritize long-term Incentivizing Reduction of Smoke 1. For more information, see http://playground. K.A. THOMAS, P.A. STINE,D.ODION, M.I. TER,A.BOYNARD, AND S.B. REID. 2013. Im- GELLER,W.HAM,S.CLIFF, L.E. HOPKINS, ET air quality and public health. Air Qual. Atmos. airfire.org. BORCHERT, ET AL. 1998. The California Gap pact of wildfires on ozone exceptional events in AL. 2010. Airborne particles in the San Joaquin Health 10(1):33–36. Impacts Rather than Area Burned Analysis Project. Final Rep., Univ. of Califor- the Western U.S. Environ. Sci. Technol. Valley may affect human health. Calif. Agric. Rather than using area burned, public 2. For more information, see www.arl.noaa.gov/ SNEEUWJAGT, R.J., T.S. KLINE, AND S.L. STE- HYSPLIT_info.php. nia, Santa Barbara, Santa Barbara, CA. 255 p. 47(19):11065–11072. 64(1):12–16. PHENS. 2013. Opportunities for improved fire health objectives can be better met by mea- 3. For more information, see src.com/calpuff/ DELFINO, R.J., S. BRUMMEL,J.WU,H.STERN, B. JONES, B.A., J.A. THACHER, J.M. CHERMAK, AND NORTH, M., B.M. COLLINS, AND S.L. STEPHENS. use and management in California: Lessons suring exceedances of health thresholds at calpuff1.htm. OSTRO,M.LIPSETT,A.WINER, ET AL. 2009. R.P. BERRENS. 2016. Wildfire smoke health 2012. Using fire to increase the scale, benefits from Western Australia. Fire Ecol. 9(2):14– monitoring sites or person-days of expected 4. For more information, see www.epa.gov/ The relationship of respiratory and cardiovas- costs: A methods case study for a Southwestern and future maintenance of fuels treatments. J. 25. harm. Replacing unit-area fees on restorative benmap. cular hospital admissions to the southern Cal- US ‘mega-fire.’ J. Environ. Econ. Policy 5(2): For. 110(7):392–401. STEPHENS, S.L., R.E. MARTIN, AND N.E. CLIN- ifornia wildfires of 2003. Occup. Environ. Med. 181–199. NORTH, M.P., S.L. STEPHENS, B.M. COLLINS, burning with flat annual fees, as employed 5. For more information, see inciweb.nwcg. TON. 2007. Prehistoric fire area and emissions 66(3):189–197. KOCHI, I., G.H. DONOVAN, P.A. CHAMP, AND J.K. AGEE,G.APLET, J.F. FRANKLIN, AND P.Z. gov/. from California’s forests, woodlands, shru- by some air districts, would reduce one of ENGEL, K.H. 2013. Perverse incentives: The case J.B. LOOMIS. 2010. The economic cost of ad- FULE´. 2015. Reform forest fire management. 6. For more information, see airnow.gov/. blands, and grasslands. For. Ecol. Manage. the disincentives for intentionally using of wildfire smoke regulation. Ecol. Law Q. verse health effects from wildfire-smoke expo- Science 349(6254):1280–1281. 251(3):205–216. large fires to restore landscapes (Sneeuwjagt 40(3):623–672. sure: A review. Int. J. Wildl. Fire 19(7):803– OFFICE OF THE FEDERAL REGISTER. 2015. Treat- STEPHENS, S.L., AND N.G. SUGIHARA. 2006. Fire et al. 2013). Literature Cited FINLAY, S.E., A. MOFFAT,R.GAZZARD,D.BAKER, 817. ment of data influenced by exceptional events, management and policy since European settle- AMERICAN LUNG ASSOCIATION. 2015. Most polluted AND V. MURRAY. 2012. Health impacts of KUMAGAI, Y., M.S. CARROLL, AND P. COHN. proposed rule. Available online at www. ment. P. 431–443 in Fire in California’s ecosys- Conclusion cities. Available online at www.stateoftheair. wildfires. PLOS Curr. Disasters. Available online at 2004. Coping with interface wildfire as a hu- federalregister.gov/articles/2015/11/20/ tems, Sugihara, N.G., J.W. van Wagtendonk, org/2015/city-rankings/most-polluted- http://currents.plos.org/disasters/article/health- man event: Lessons from the disaster/hazards 2015-29350/treatment-of-data-influenced- The combination of a warming climate J. Fites-Kaufman, K.E. Shaffer, and A.E. Th- cities.html; last accessed Jan. 12, 2016. impacts-of-wildfires/. literature. J. For. 102(6):28–32. by-exceptional-events; last accessed Dec. 11, 2015. ode (eds.). Univ. of California Press, Berkeley, and accumulation of forest fuels ensures a BARBERO, R., J.T. ABATZOGLOU, N.K. LARKIN, FOWLER, C. 2003. Human health impacts of for- LEENHOUTS, B. 1998. Assessment of biomass OLSEN, C.S., D.K. MAZZOTTA,E.TOMAN, AND CA. future with more large fires and smoke in dry C.A. KOLDEN, AND B. STOCKS. 2015. Climate est fires in the southern United States: A liter- burning in the conterminous United States. A.P. FISCHER. 2014. Communicating about STRAND, T.M., N. LARKIN, K.J. CRAIG,S.RAF- western US forests. We have outlined a change presents increased potential for very ature review. J. Ecol. Anthropol. 7:39–59. Conserv. Ecol. 2(1):1. smoke from wildland fire: Challenges and op- FUSE,D.SULLIVAN,R.SOLOMON,M.RORIG, Journal of Forestry • MONTH 2017 9 10 Journal of Forestry • MONTH 2017 Journal of Forestry • January 2018 85 N. WHEELER, AND D. PRYDEN. 2012. Analyses line at airnowtech.org/; last accessed Jan. 11, Western US. Final Rep. US For. Serv. Cooper- of BlueSky Gateway PM2. 5 predictions dur- 2016. ative Agreement, Washington, DC. ing the 2007 southern and 2008 northern Cal- VAN WAGTENDONK, K. 2012. Fires in previ- WIEDINMYER, C., AND M. HURTEAU. 2010. Pre- ifornia fires. J. Geophys. Res. Atmos. (1984– ously burned areas: Fire severity and vegeta- scribed fire as a means of reducing forest car- 2012) 117(D17):D17301. tion interactions in Yosemite National Park. bon emissions in the Western United States. TRAINOR, S.F., M. CALEF,D.NATCHER, F.S. P. 356–363 in Rethinking protected areas in a Environ. Sci. Technol. 44(6):1926–1932. CHAPIN, A.D. MCGUIRE,O.HUNTINGTON, P. changing world: Proc. of the 2011 George WILLIAMS, J. 2013. Exploring the onset of DUFFY, ET AL. 2009. Vulnerability and adapta- Wright Society biennial conference on parks, high-impact mega-fires through a forest land tion to climate-related fire impacts in rural and protected areas, and cultural sites, Weber, S. (ed.). management prism. For. Ecol. Manage. 294: urban interior Alaska. Polar Res. 28(1):100– The George Wright Society, Hancock, MI. 4–10. 118. WEGESSER, T.C., K.E. PINKERTON, AND J.A. YAO, J., M. BRAUER, AND S.B. HENDERSON. 2013. US CENSUS BUREAU. 2015. 2010 census urban LAST. 2009. California wildfires of 2008: Evaluation of a wildfire smoke forecasting sys- and rural classification and urban area criteria. Coarse and fine particulate matter toxicity. En- tem as a tool for public health protection. En- Available online at www.census.gov/geo/ viron. Health Perspect. 117(6):893–897. viron. Health Perspect. 121(10):1142–1147. reference/ua/urban-rural-2010.html; last ac- WESTERLING, A.L., J. MILOSTAN, AND A.R. KEY- YOSEMITE NATIONAL PARK. 2012. Fire history geo- cessed Apr. 8, 2015. SER. 2015. Modeling potential fire impacts with database for Yosemite National Park. Available US ENVIRONMENTAL PROTECTION AGENCY. landscape vegetation scenarios and changing cli- online at irma.nps.gov/App/Reference/Profile/ 2016. AirNow-Tech Database. Available on- mate for the Sierra Nevada and other areas in the 2188901; last accessed Dec. 14, 2015. 86 Journal of Forestry • January 2018 Journal of Forestry • MONTH 2017 11