Retrospective analysis of burn windows for fire and fuels ...very rare. Interannual variation was...

ORIGINAL RESEARCH Open Access

Retrospective analysis of burn windows forfire and fuels management an examplefrom the Lake Tahoe Basin California USARandy Striplin12 Stephanie A McAfee23 Hugh D Safford14 and Michael J Papa5

Abstract

Background In fire-adapted ecosystems of the western USA prescribed fire is an essential restoration and fuelreduction tool There is general concern that as the fire season lengthens the window for conducting prescribedburns will contract unless management changes are made This could occur because a number of conditions mustbe met before prescribed fire can be used in the field and those are most common during the spring and autumnwhen the need for fire suppression response has been historically less To assess patterns of potential prescribedburning feasibility this study evaluated three conditions (1) permission to burn as granted by air quality regulators(2) weather within burn plan prescription and (3) availability of operational and contingency resources Our 21-yearanalysis (1999 to 2019) combines three independent datasets for a daily comparison of when prescribed fires couldhave been implemented (henceforth burn windows) in the Lake Tahoe Basin (LTB) and analyzes seasonalityinterannual variability and trends

Results Burn windows were most frequent during spring followed by autumn with the fewest burn windowsduring the summer and winter Burn windows lasting multiple days occurred infrequently Two- to three-day burnwindows did not often occur more than twice per month over the study period and longer burn windows werevery rare Interannual variation was considerable Finally an abrupt increase in burn windows was detected in 2008This was determined to be related to a methodological change by air quality regulators and not to any changes inclimate or resource availability

Conclusions While this case study focuses on the LTB the analysis was performed with readily available data andcould be applied easily to other land management units demonstrating a valuable method for planning andprioritizing fire and fuels management activities This type of tool can also identify areas for research For example ifthere were unused burn windows during the winter and early springmdashor they were projected to increasemdashresearch into the ecological impacts of winter and spring burning may allow managers to more confidently adaptto changing climate Moreover this analysis demonstrated that modest and reasonable regulatory changes canincrease opportunities for prescribed burning

Keywords burn window climate change mixed conifer forest prescribed fire western USA

copy The Author(s) 2020 Open Access This article is licensed under a Creative Commons Attribution 40 International Licensewhich permits use sharing adaptation distribution and reproduction in any medium or format as long as you giveappropriate credit to the original author(s) and the source provide a link to the Creative Commons licence and indicate ifchanges were made The images or other third party material in this article are included in the articles Creative Commonslicence unless indicated otherwise in a credit line to the material If material is not included in the articles Creative Commonslicence and your intended use is not permitted by statutory regulation or exceeds the permitted use you will need to obtainpermission directly from the copyright holder To view a copy of this licence visit httpcreativecommonsorglicensesby40

Correspondence randystriplinusdagov1USDA Forest Service Pacific Southwest Region 1323 Club Drive VallejoCalifornia 94592 USA2Department of Geography University of Nevada Reno 1664 North VirginiaStreet Reno Nevada 89557 USAFull list of author information is available at the end of the article

Fire EcologyStriplin et al Fire Ecology (2020) 1613 httpsdoiorg101186s42408-020-00071-3

Resumen

Antecedentes En ecosistemas adaptados al fuego en el oeste de los EEUU las quemas prescriptas son unaherramienta esencial para la restauracioacuten y reduccioacuten de combustibles Existe un consenso general sobre que si laestacioacuten de fuegos se prolonga la ventana de prescripcioacuten para conducir quemas prescriptas se contrae a menosque se hagan cambios en el manejo Esto puede ocurrir dado que cierto nuacutemero de condiciones deben sercumplidas antes de que las quemas puedan ser aplicadas en el campo y eacutestas se dan maacutes comuacutenmente enprimavera y otontildeo cuando las necesidades de supresioacuten son histoacutericamente menores Para determinar el patroacuten deposibilidades de aplicacioacuten de quemas este estudio evaluoacute tres condiciones (1) el permiso de quema otorgado porreguladores de la calidad del aire (2) el tiempo meteoroloacutegico dentro del plan de prescripcioacuten y (3) ladisponibilidad de recursos de contingencia y operacionales Nuestros anaacutelisis de 21 antildeos (1999 a 2019) combinarontres conjuntos de datos independientes para una comparacioacuten diaria sobre cuando las quemas deberiacutean haber sidoimplementadas (ventana de quemas de aquiacute en maacutes) en la cuenca del lago Tahoe (LTB) y analizoacute la estacionalidadla variabilidad interanual y las tendencias

Resultados Las ventanas de quema fueron maacutes frecuentes durante la primavera seguidas del otontildeo con la menorcantidad de ventanas de quemas durante el verano y el invierno Las ventanas de quemas que duraban varios diacuteasocurriacutean de manera infrecuente Las ventanas que duraban dos a tres diacuteas no ocurriacutean maacutes que dos veces por mesen todo el periacuteodo de estudio y ventanas que iban maacutes allaacute de este periacuteodo fueron realmente raras La variacioacuteninteranual tambieacuten fue considerable Finalmente un abrupto incremento de ventanas de quemas fue detectado en2008 Esto fue determinado que estuvo relacionado a un cambio en la metodologiacutea usada por los reguladores decambios en la calidad del aire y no por cambios meteoroloacutegicos o en la disponibilidad de recursos

Conclusiones Aunque este estudio de caso se enfoca en LTB el anaacutelisis fue realizado con datos raacutepidamentedisponibles que pueden faacutecilmente ser aplicados a otras unidades de manejo demostrando ser un meacutetodo valiosopara planificar y priorizar quemas y actividades de manejo de combustibles Este tipo de herramienta puedeidentificar tambieacuten aacutereas para la investigacioacuten Por ejemplo si hubiese ventanas de quemas no usadas durante elinvierno y la primavera tempranamdasho si fuese proyectado su incrementomdashla investigacioacuten sobre los impactosecoloacutegicos de las quemas de invierno o primavera podriacutean permitir a los gestores de recursos adaptarse con maacutesconfianza al cambio climaacutetico Este anaacutelisis demostroacute ademaacutes que cambios regulatorios modestos y razonablespueden incrementar las oportunidades para realizar quemas prescriptas

AbbreviationsCARB California Air Resource BoardCOOP Cooperative Observer ProgramFFP5 Fire Family Plus v 5FR Fire regimeLTB Lake Tahoe BasinPL Preparedness levelNOPS Northern California Geographic AreaNWS National Weather ServiceONCC Northern California Geographic Area

Coordination CenterRAWS Remote automated weather stationUSDA United States Department of AgricultureWRCC Western Regional Climate CenterWUI Wildlandndashurban interfaceWY Water year

BackgroundHuman management has greatly altered forests in thewestern USA since Euro-Americans arrived in the mid1800s In forests that historically experienced frequentmostly low-severity fire (ie Fire Regime 1 forests [FR1]

Hardy et al 2001) logging and fire exclusion havecaused major changes including loss of the large treecomponent increases in stand density and surface andladder fuels as well as compositional shifts towardshade-tolerant and fire-intolerant tree species (Saffordand Stevens 2017) In the last two to three decadeswarming and drying have interacted with these forestchanges to drive an increase in the occurrence of wild-fires that are burning at unprecedented scales and sever-ities killing large areas of canopy trees and increasinglythreatening human life and property (Miller et al 2009Abatzoglou and Williams 2016 Safford and Stevens2017 Holden et al 2018 van Wagtendonk et al 2018)

Reducing risks to communities and natural resourcesis a top priority for land managers As a result theUnited States Department of Agriculture (USDA) ForestService continues to invest heavily in community risk re-duction and has recently emphasized increasing the paceand scale of ecological restoration (USDA Forest ServicePacific Southwest Region 2011 USDA Forest Service2012 Agee et al 2016) In FR1 forests frequent thinningand other types of fuel reduction followed by prescribed

Striplin et al Fire Ecology (2020) 1613 Page 2 of 16

fire are usually the most effective fuels management andforest restoration tools (Agee and Skinner 2005 Stephensand Moghaddas 2005 North et al 2009 aillant andStephens 2009 McIver et al 2013) Some studies have alsofound that prescribed fire alone reduces surface and lad-der fuels and is successful in mitigating the risk of crownfire under extreme weather conditions (Kilgore and Sando1975 Stephens et al 2012)Despite the clear need for management to reduce

wildfire risk a recent analysis found that in much of thewestern USA use of prescribed fire has declined sincethe late 1990s (Kolden 2019) This reduction has oc-curred despite an acknowledged ldquobacklogrdquo in forest man-agement (North et al 2012 Vaillant and Reinhardt2017) Although Kolden (2019) attributes some of thereticence around prescribed fire use in the western USAto societal concerns there are clear practical constraintsto its use as well (Quinn-Davidson and Varner 2012)Prescribed fire is effective at reducing wildfire threats

but there are risks associated with it so the practice isstrictly regulated A variety of conditions need to be metprior to prescribed burning on federal lands (NationalWildfire Coordinating Group 2017) Weather conditions(forecasted and observed) must be within prescriptivecriteria established in the prescribed fire implementationplan Prescribed fire implementation plans (burn plans)establish a set of environmental conditions (the prescrip-tion) under which the burn has a high likelihood ofmeeting project objectives (National Wildfire Coordinat-ing Group 2017) Operational resources (personnel andequipment) for burn implementation and the contin-gency plan must be available and burn permits must beobtained from the jurisdictional air quality regulatorsWeather conditions that meet burn plan prescriptionssufficient resources and permissible burn days for airquality must occur together on the day or days of theburn before it can proceed Fire and resource managersknow through experience that the coincidence of theseevents is limiting and can constrain their ability to meetfuels and restoration objectives Although studies haveevaluated seasonal patterns in the weather conditionssuitable for the use of prescribed fire (eg Yurkoniset al 2019) there is currently no quantitative methodfor assessing the frequency with which all of these limi-tations on prescribed fire coincideDescribing burn window occurrences their trends

and the variables constraining them will increase thelikelihood of success in meeting restoration and fuel re-duction objectives Information about past patterns andtrends in burn windows is important for projecting areasonable treatment area given project objectives Stud-ies suggest that the majority of natural burning historic-ally occurred during the summer wildfire seasonsometimes extending into autumn (Taylor 2004 Beaty

and Taylor 2008) However during the wildfire seasonweather is warmer and drier fire suppression resourcesare often committed to active wildfires and stable calmatmospheric conditions are not as conducive to smokedispersion so prescribed burning is often discouragedWith all of the limitations described above and the needto increase prescribed burning both in-fire-season andout-of-fire-season burning may be necessary Conse-quently information on burn-window likelihood iscritical for managers intending to restore a natural fireregime or simply address a backlog in prescribedburningUnderstanding when and where weather and fuel con-

ditions are within prescription and where in the pre-scriptive range they fall informs managers about whenand how to burn as well as when to plan burns withspecific intended objectives For instance burning at themoister end of the prescription will consume less fueland produce less severe effects than burning at the drierend of the prescription (Knapp et al 2005 Knapp andKeeley 2006 Schwilk et al 2006) Either of those out-comes or any range in between may be optimal forachieving desired objectives Knowledge of seasonal pat-terns in weather conditions can also inform the types oftreatment that are most plausible For example if burnwindows are likely in the spring when soil and fuel mois-tures are higher and less common in the summer or au-tumn managers may choose to target areas with heavyfuels in the spring where lower levels of fuel consump-tion and patchy burns might be desired Drier conditionsin the autumn might be reserved for burns intended tomaximize fuel consumptionInvestigation of historical burn-window occurrences

and their drivers can improve planning and budgeting Ifburn windows are most frequent in spring and autumnbut the workforce tour of duty is scheduled to ramp upin late spring and wind down in early autumn sufficientresources to implement prescribed fire projects may notbe available The current Forest Service paradigm is touse fire suppression resources to implement prescribedfire projects During spring autumn and winter fewerseasonal fire personnel are available to conduct burnseven though weather and atmospheric conditions areoften optimal for implementing prescribed fire Under-standing when regional fire activity and fire suppressionresource limitations inhibit capacity can provide an in-centive to develop innovative staffing solutions such asstaggering seasonal crew start and end dates to allow foradditional staffing in the spring and autumn or formingdedicated prescribed fire crews as California Depart-ment of Forestry and Fire Protection recently did (Stateof California 2020) This information will allow man-agers to scale and better schedule their workforce forsuccess Furthermore knowing when air quality effects

are most likely can inform community outreach and en-hance collaboration and cooperation with air quality reg-ulators Finally analysis of historical and projectedfuture burn-window occurrences may provide insightinto research needed for long-range planning For in-stance research into the ecological implications of win-ter and spring burns (eg Knapp et al 2005 Knapp andKeeley 2006) is warranted if multiple-day burn periodsare or become more likely during spring or winter whenfires have historically been uncommonHere we assess how interactions between weather

conditions air quality regulations and resource availabil-ity influence prescribed fire burn windows in the LakeTahoe Basin (LTB) California USA Our analysis identi-fies the daily co-occurrence of all three conditions from1999 through 2019 identifying burn-window patterns toassist managers in planning and implementing pre-scribed fires In terms of fire management we believethat the LTB serves as a reasonable proxy for otherinhabited parts of the forested West with similar fueland forest conditions but perhaps somewhat higher an-thropogenic ignition densities

MethodsStudy areaThe LTB is at the crest of the Sierra Nevada (approxi-mately 39degN 120degW) and is shared between Californiaand Nevada (Fig 1) It spans 132 283 hectares includingthe approximately 50 000-hectare Lake Tahoe which sitsat approximately 1900 m elevation surrounded by moun-tain peaks rising to gt3300 m Dry mild summers and coldwet winters are typical The January mean minimumtemperature at the South Lake Tahoe airport is minus88 degCthe July mean maximum temperature is 271 degC andaverage annual precipitation is about 518 cm (WesternRegional Climate Center 2017) Interannual variabilityin precipitation is high and like in much of the SierraNevada is increasing over time (Safford et al 2012a)For instance the National Weather Service CooperativeObserver Program (NWS COOP) weather station inTahoe City California (NWS ID 048758) on the northshore of the lake recorded about half the 30-year aver-age precipitation in water year (WY) 2015 (1 Oct 2014to 30 Sep 2015) and twice the 30-year average precipi-tation in WY 2017 (TahoeClim 2017)Vegetation and fire regimes can be stratified by eleva-

tion into three broad groups Near lake level the lowermontane zone (lt2200 m) is dominated by Jeffrey pine(Pinus jeffreyi Grev amp Balf) white fir (Abies concolor[Gordon amp Glend] Lindl ex Hildebr) incense-cedar(Calocedrus decurrens [Torr] Florin) and sugar pine (Plambertiana Douglas) The upper montane zone (2200to 2500 m) is dominated by red fir (A magnifica AMurray bis) lodgepole pine (P contorta Loudon ssp

murrayana [Grev amp Balf] Critchf) and western whitepine (P monticola Douglas ex D Don) The subalpinezone occurs at elevations greater than 2500 m with redfir western white pine mountain hemlock (Tsuga mer-tensiana [Bong] Carriegravere) and whitebark pine (P albi-caulis Engelm) Montane chaparral stands are scatteredthroughout especially in the lower montane zone and thetransition to the upper montane zone Dominant shrubgenera include manzanita (Arctostaphylos Adansspp)Ceanothus L spp and currants and gooseberries (Ribes Lspp) Historic (pre 1850) fire return intervals averagedabout 10 years in the FR1 lower montane zone 40+ yearsin the upper montane zone (Fire Regime III Schmidtet al 2002) and gt200 years in subalpine forests (FireRegime IV) (Elliott-Fisk et al 1997 Manley et al 2000Barbour et al 2002 Taylor 2004 Nagel and Taylor 2005Beaty and Taylor 2008) Manley et al (2000) estimatedthat during an average year in the pre-settlementperiod between 800 and 3200 ha burned and meanfire size was probably around 200 to 400 ha (Saffordand Stevens 2017) In FR1 forests fire severities weregenerally low to moderate and there was relativelylittle mortality of mature trees (Skinner and Chang1996 Manley et al 2000 Taylor 2004) but fires weremore severe at higher elevations (Mallek et al 2013van Wagtendonk et al 2018) Fires occurred mostlyin the late summer and autumn (Taylor 2004 Beatyand Taylor 2008)The human footprint is substantial in the LTB More

than 75 of the land area inside the LTB is designatedas wildlandndashurban interface (WUI California Depart-ment of Forestry and Fire Protection et al 2014) Theproportion of wildland area near and adjacent to com-munities and infrastructure adds notable complexity toprescribed fire operations In addition to approximately50 000 permanent residents Lake Tahoe receives an es-timated 77 million recreational visitors per year(LTBMU 2015) increasing the need to ensure fire safetyand minimize air quality impacts The LTB is also juris-dictionally complex with a matrix of local private stateand federal lands spanning two states five counties onerural district multiple cities and townships and numer-ous fire protection entities

Burn window analysisBurn windows as defined here are determined by thesimultaneous occurrence of (1) CARB burn days (daysdesignated by California Air Resources Board [CARB] asburn days) (2) days when weather and fuel-moistureconditions fall within burn plan prescription and (3)sufficient operational and contingency resources (repre-sented by Northern California Geographic Area and na-tional preparedness levels) Preparedness levels (PL) aredictated by regional burning conditions fire activity and

Fig 1 Map showing the location of remote automated weather station (RAWS) used in the assessment of patterns of potential prescribed burningfeasibility in the Lake Tahoe Basin USA

resource availability The availability or unavailability offirefighting resources can often impact resources to im-plement prescribed fires Information on operational andcontingency resources is not usually incorporated in thistype of analysis but it provides a critical practicalconstraint

CARB burn dayCARB burn days are days on which prescribed burningis permitted by the state board and burning is authorizedby each air district consistent with Title 17 of theCalifornia Code of Regulations (California 2010) Atmos-pheric conditions related to smoke dispersal and othersources of air pollution (eg wildfires other prescribedfires and agricultural burning) factor into burn day de-terminations Prescribed fire ignitions are generally notpermitted by CARB on days that are not designated asburn days although there are occasional exceptionsoften in consultation with the entity conducting theburnWe downloaded archived CARB burn day data from

httpwwwarbcagovsmphistorhistorhtm for 1 January1999 to 31 December 2019 California Code of Regula-tions Title 17 subchapter 2 (California 2010) designationsinclude burn day marginal day and no-burn day CARBdata for the Lake Tahoe Air Basin also include twoadditional designations amended and fair Amended re-fers to days on which the initial forecast condition waschanged from burn to no-burn or from no-burn to burnFair and marginal are days on which burning conditionsare not ideal but burning preferably over smaller areas orof materials that will produce lower emissions is allowed(D Mims California Air Resources Board MeteorologySection Sacramento California USA personal communi-cation 5 September 2019) Over the archive period therewere 4587 burn days 714 marginal days 6 amended days14 fair days and 2353 no-burn days For simplicity ofanalysis we assumed that prescribed burns could onlyoccur on days designated as burn days and not fair ormarginal days on which it is expected that any burningthat does occur will be limited These conditions only ap-plied to the California side of the LTB The Nevada Div-ision of Environmental Protection regulates smokemanagement on the Nevada side and may request cessa-tion of burning activities but they do not proactively des-ignate days as permissive or non-permissive for ForestService burning activities The Washoe County Air Qual-ity Management Division requires land managers to ac-quire permits for prescribed fires that emit greater than907 kg of particulate matter of 10 micrometers or less(PM10) Here we use the more stringent and objectivelydefined California standards to provide a conservative esti-mate of burn windows for the entire LTB

Days within prescriptionThe burn plan prescription refers to a set of measurablecriteria used to determine whether a prescribed fire maybe ignited The prescription includes a set of weatherand fuel parameters (ranges of permissible wind speedsair temperatures humidity fuel moistures etc) withthresholds based on desired fire behavior and effectsTypical LTB burn plan prescription criteria include (1)minimum relative humidity between 20 and 50 (2)10-hour fuel moisture (10-hour fuels are woody mate-rials between 064 and 254 cm diameter) between 7and 20 and (3) maximum wind speeds at 61 m abovethe ground lt112 m sminus2 On an actual burn forecastedweather and on-site measurements determine if condi-tions are within prescription Continuous data were notavailable from all past or likely prescribed burnlocations so we estimated days in prescription by com-paring weather and fuel moisture data from remoteautomated weather stations (RAWS) with the prescrip-tion criteria outlined above and categorizing days thatmet all conditions to be in prescription Specifically weconsidered days to be in prescription if the lowesthourly relative humidity measurement in that 24-hourperiod was between 20 and 50 lowest hourly 10-hour fuel moisture was between 7 and 20 (Nelsonmethod calculated by Fire Family Plus v5 Nelson2000) and highest hourly maximum 61 m windspeeds were lt112 m sminus2) Estimates from RAWS maynot be fully representative of sites where burns willbe conducted Estes et al (2012) reported that RAWS10-hour fuel moisture estimates at one location werebiased low when fuel moistures were over 20 Thiscould have led us to overestimate the available daysin prescription during the spring if findings fromtheir location (350 km northwest of the LTB usingan older 10-hour fuel moisture calculation) hold truein the LTB Based on local experience however webelieve our methods provided a reasonable estimateof the frequency of days within prescriptionNo weather station inside the LTB had continuous

hourly data for all the variables needed for this analysisover the full study period so we combined informationfrom two stations to provide quasi-complete local wea-ther data over the 21-year period The Baron RAWS(3885degN 12002degW elevation 1904 m NWS ID 042616)is currently used by the Forest Service for most fire-related purposes in the LTB but its record extends backonly to 21 July 2011 The Markleeville RAWS (3869degN11977degW elevation 1677 m NWS ID 042802) is 29 kmsoutheast of the Baron RAWS and outside of the LTBbut it has all necessary data over the full study periodAlthough the burn window analysis could have beenperformed using the Markleeville RAWS managers inthe LTB prefer the Baron RAWS Thus we used the

Markleeville RAWS to model Baron RAWS for theperiod 1 January 1999 through 31 July 2011 using linearregression (see below) In order to assess the similarityof weather observations between the two stations weconducted a seasonal Pearson correlation analysis (Zar1999 377) of temperature and humidity Table 1 listsPearson correlation coefficients of seasonal temperatureand relative humidity observations for the period ofoverlap between these RAWSData from the two stations were downloaded from the

Western Regional Climate Center (WRCC httpwwwrawsdrieduindexhtml) We used Fire Family Plus v5(FFP5 Bradshaw and McCormick 2000) to perform qual-ity control and summarize data During quality controlwe noted suspect wind gust speeds (greater than 45 m sminus2)at Baron RAWS from June through 4 October 2016 As aresult we excluded Baron RAWS wind gust speed datafrom 1 June 2016 to 4 October 2016 when the wind sen-sor was replaced The four-month gap in wind speeds wasfilled by regression No outliers or errors (other than a fewmissing hourly records) were noted in other variables dur-ing quality control We extracted daily minimum relativehumidity maximum wind gust speeds and 10-hour fuelmoistures (Nelson 2000) calculated by FFP5 from thehourly data prescription analysis The Nelson (2000)

defaults 10-hour fuel moisture to 25 when there is snowcover at the RAWS site To maintain a normal data distri-bution for the regression we calculated 10-hour fuelmoistures with no snow coverWe used adjusted data from the Markleeville RAWS

to estimate daily minimum relative humidity 10-hourfuel moisture and maximum wind gust speed at BaronRAWS prior to August 2011 Data from the period avail-able at both RAWS were divided into training (1 January2012 to 31 December 2015) and validation (1 January2016 to 31 Dec 2019) periods For each variable we de-veloped a linear regression model between Baron andMarkleeville RAWS over the training period and testedit over the validation period Model fit and validationstatistics are shown in Table 2 We then used the regres-sions to estimate Baron RAWS data for 1 January 1999through 31 July 2011 from daily Markleeville RAWSdata To evaluate the consistency of local patterns weapplied burn-window analysis separately to the Baronand Markleeville RAWS stations and to three othernearby RAWS Dog Valley Stampede and Little Valley(see Fig 1 for station locations) This analysis waslimited to 2012 to 2019 the full period of overlapResults for these additional stations are shown inAdditional files 1 to 3

Table 2 Model results for the regression of Markleeville remoteautomated weather station (RAWS) variables to estimate BaronRAWS prior to August 2011 when Baron RAWS becameoperational as part of our study assessing the patterns ofpotential prescribed burning feasibility in the Lake Tahoe BasinUSA from 1999 to 2019 Baron RAWS relative humidity () 10-hour fuel moisture () and wind gust speeds (km hrminus1) wereestimated by simple regression of these variables fromMarkleeville RAWS in order to obtain weather data from 1January 1999 to 31 July 2011 Baron RAWS recordedobservations were used from 1 August 2011 to December 2019to provide a complete dataset for the entire study period (1999to 2019) These data were used to determine if each day waswithin burn plan prescription criteria Baron RAWS is locatedwithin the Lake Tahoe Basin in Meyers California approximately10 km south of Lake Tahoe Markleeville RAWS is sited inMarkleeville California approximately 35 km south southeast ofLake Tahoe Model Adj R2train and Adj R2valid are coefficients ofdetermination for training (1 January 2012 to 31 December2015) and validation (1 January 2016 to 31 December 2019)respectively Validation R is the Pearson correlation coefficientbetween the modeled and the observed validation data (P le0001 for all regressions)

Variable Adj R2train Adj R2valid Validation R

Relative humidity () 0782 0795 089

10-hour fuel moisture () 0691 0753 087

Wind gust speed (km hrminus1) 0725 0700 084

Table 1 Pearson correlation coefficients between Baron andMarkleeville remote automated weather stations (RAWS) for theperiod 2011 to 2019 for daily values of each variable bymeteorological season as part of our study assessing thepatterns of potential prescribed burning feasibility in the LakeTahoe Basin USA from 1999 to 2019 Baron RAWS is located inthe Lake Tahoe Basin study area and is sited in MeyersCalifornia approximately 10 km south of Lake Tahoe but it onlyincluded data beginning in mid 2011 Markleeville RAWS is sitedin Markleeville California approximately 35 km south southeastof Lake Tahoe and included complete quality data covering thestudy period 1999 to 2019 Seasonal correlations were measuredto assess the appropriateness of using Markleeville RAWS toestimate Baron RAWS variables by regression in order to obtainweather data for the entire 1999 to 2019 study period Variablestested were those to assess basic climatological site similarityand include average minimum and maximum temperature(Tavg Tmax Tmin respectively) and average minimum andmaximum relative humidity (RHavg RHmin RHmax respectively)Meteorological seasons were winter (1 December to 28 or 29February) spring (1 March to 31 May) summer (1 July to 31August) and autumn (1 September to 30 November)

Season Tavg Tmin Tmax RHavg RHmin RHmax

Winter 094 092 087 078 079 063

Spring 096 090 096 081 088 055

Summer 090 078 096 065 087 042

Autumn 096 088 097 078 085 053

Availability of firefighting resourcesPreparedness level (PL) is a daily index that ranks thecommitment level of fire suppression and incident man-agement resources for a geographic area from 1 (low) to5 (high) PL3 is not a threshold for prescribed fire imple-mentation set by Forest Service policy we used it in thisstudy as a surrogate indicator of operational and contin-gency resource availability to add a ldquoreasonable and feas-iblerdquo element to the analysis although it may not be aperfect proxy of crew availability Here we assumed thatprescribed burning was feasible at PL1 and PL2 bothwithin and outside of the usual fire season At PL3 en-vironmental conditions are such that there is high po-tential for fires greater than 40 hectares to occur withseveral fires less than 40 hectares active in the geo-graphic area The USDA Forest Service et al (2016) de-scribes PL3 as

Mobilization of agency and interagency resources isoccurring within the geographic area but minimalmobilization is occurring between or outside of thegeographic area Current and short-term forecastedfire danger is moving from medium to high or veryhigh Local Units implementing prescribed fire oper-ations are starting to compete for interagency contin-gency resources

The Northern California Geographic Area Coordin-ation Center (ONCC) begins preparedness planning forthe Northern California Geographic Area (NOPS) by 1May and continues through at least 15 October (USDAForest Service et al 2016) Review of PLs revealed gapsin the NOPS data (primarily in the non-fire seasonmonths in 2004 to 2008) but national PL data arecomplete National PL and the existing NOPS PL arevery similar so we used NOPS PL preferentially in the

analysis with national PLs used in those instances whenNOPS PLs are missing

Burn-window occuranceWe determined burn windows by assessing when CARBburn days days meeting burn plan prescription criteriaand NOPS PL lt 3 occurred simultaneously Burn windowswere summarized to identify (1) how often each day ofthe year met each criterion individually and all criteriasimultaneously (2) the seasonal frequency of single-dayand multi-day burn windows and (3) interannual variabil-ity in burn windows All analyses (except trend analysiswhich was performed in R [R Core Team 2016]) were per-formed using spreadsheet tools to facilitate wider use ofthese methods in management settings We initiallyassessed changes in annual burn-window frequency usinglinear regression in R (R Core Team 2016) Because resid-uals often were not normally distributed we tested fortrends with the Mann-Kendall trend test (Mann 1945Kendall 1975 Gilbert 1987) using the Kendall packagein the R program (McLeod 2011) When trends were iden-tified in the number of burn windows we performed trendanalysis on the individual variables (CARB burn days dayswith PL lt 3 and days in prescription) to identify the vari-able or variables driving the trend

ResultsBurn windows were especially rare during peak fire sea-son (July to September) and also December throughJanuary (Figs 2 and 3) Less than one-third of days inOctober and in November were burn windows Januaryand December each had 20 likelihood of burn windows(Fig 3) They were most common from February to Mayand from October through November but the daily like-lihood rarely exceeded 50 in spring and 40 in autumn

0010203040506070809

1

01 J

an08

Jan

15 J

an22

Jan

29 J

an05

Feb

12 F

eb19

Feb

26 F

eb04

Mar

11 M

ar18

Mar

25 M

ar01

Apr

08 A

pr15

Apr

22 A

pr29

Apr

06 M

ay13

May

20 M

ay27

May

03 J

un10

Jun

17 J

un24

Jun

01 J

ul08

Jul

15 J

ul22

Jul

29 J

ul05

Aug

12 A

ug19

Aug

26 A

ug02

Sep

09 S

ep16

Sep

23 S

ep30

Sep

07 O

ct14

Oct

21 O

ct28

Oct

04 N

ov11

Nov

18 N

ov25

Nov

02 D

ec09

Dec

16 D

ec23

Dec

30 D

ec

9991syadfo

noitroporP

-201

9

Day of year

Burn window In prescription Burn day PL lt 3

Fig 2 Burn window and individual burn window component frequency by day of year for the Lake Tahoe Basin USA from 1999 to 2019 Burnwindows were composed of days with co-occurrence of permission to burn by the air quality regulators sufficient resources needed for implementation andweather within burn plan prescription criteria The gray shaded area represents days when all three criteria were met (burn windows) The red line representsdays that met burn plan prescription criteria The blue line indicates California Air Resources (CARB) permissible burn days The black line represents days whenthe Northern California Geographic Area preparedness level (PL) was less than 3

(Fig 2) Burn-window frequency ranged from a high of44 in April and May to a low of 7 in August (Fig 3)Nearby stations showed similar seasonal patterns inburn-window occurrence although the absolute fre-quency of burn windows differed from station to stationwith Dog Valley and Baron RAWS having the most fre-quent burn days and Markleeville RAWS the fewest(Additional files 1 and 2)Over the 21-year analysis period consecutive multi-

day burn windows were uncommon and burn win-dows longer than four consecutive days were veryrare Burn windows lasting two to three days weremost common from February through June and Octo-ber through November yet there were still on aver-age two or fewer two- to three-day burn windowsper year in these months (Fig 4) Slightly longer(four- to five-day) burn windows were most commonin April May October and November but these oc-curred on average less than once per year (Fig 4)

Six-day or longer burn windows occurred about onceevery two years in May and were even rarer in othermonths (Fig 4) Multi-day burn windows of anylength were rare during the peak fire season (Julythrough September) with just 42 occurrences over 21yearsSummer had infrequent burn windows often zero in

any given year especially in August (Fig 5) August burnwindows occurred in only seven of the 21 years studiedJuly and September each had burn windows in 14 daysthroughout the study period May was the most variablemonth and December the least variable (Fig 5) In themonths from November through May burn windowsoccurred in every year but they were highly variable InMay for example there were only two burn windows in2001 but there were 24 burn windows in both 2010 and2011 Analysis of more stations over a shorter timeframe (2012 to 2019) confirms the high degree of inter-annual variability in burn windows particularly in the

20

3237

44 44

29

167 11

29 3020

0

10

20

30

40

50

60

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Per

cent

day

s (

)

Fig 3 Percentage of all days in each month that were burn windows in the Lake Tahoe Basin USA from 1999 to 2019 Days with simultaneousoccurrence of permission to burn by the air quality regulators sufficient resources needed for implementation and weather within burn planprescription criteria were designated as burn windows Error bars show the standard error of the mean

0

05

1

15

2

25

elpitlum fo reb

mun egarevA

-day

bu

rn w

indo

ws

2-3 days 4-5 days 6 days or more

Fig 4 Average multiple-day burn windows per month in the Lake Tahoe Basin USA for the analysis period 1999 to 2019 based on observed andestimated Baron remote automated weather station data Multiple-day burn windows were consecutive days meeting burn-window criteriaRelative monthly frequency of multiple-day burn-window occurrences is depicted These classes do not include single-day occurrences Each classof consecutive-day periods excludes the lower classes (ie 2- to 3-day periods are not counted in the 4- to 5-day periods etc)

summer but also in the winter and spring (Additionalfile 3)Burn windowsmdashdays on which all three criteria are

metmdashwere far less common than the number of daysmeeting any one criterion (Fig 2) Year-round burn-plan prescription was the most consistently limiting fac-tor except in January and occasionally in July throughOctober when CARB burn days were more limiting(Fig 2) During peak fire season (July to September)weather on any given day of the year was in prescriptionless than 60 of the time In other months weather wasin prescription on any given day up to ~75 of the timebut it was rarely over 65 (Fig 2) CARB burn days oc-curred most frequently during late winter and spring(February to May Fig 2) although they were also rela-tively common in October November and DecemberJanuary had relatively few CARB burn days CARB burndays within the peak fire season (July to September)were also relatively uncommon generally less than 50of the time on any given day of the year NOPS PL wastypically lt3 except from mid July through September(Fig 2) when fire activity in the NOPS geographic areausually peaks and firefighting resources are committedto ongoing incidents While NOPS PL was rarely limit-ing it was the most limiting factor about 25 of thetime during August through mid SeptemberAnnual burn-window frequency (Fig 6) increased

significantly over our analysis (Mann-Kendall τ = 0438

2-sided P = 0006) CARB burn days was the onlyvariable with a significant trend (Mann-Kendall τ =0616 2-sided P le 0001) An abrupt increase in CARBburn day frequency occurred around 2008 (Fig 6) rais-ing the question of whether the trend had a physicalbasis The primary criterion used in burn-day decisionsby CARB is 500-hectopascal (hPa) geopotential heightassuming that air quality is not already low (D Mimspersonal communication 2019) Higher geopotentialheight (ridging) indicates higher pressure and typicallywarmer and drier conditions Conversely lower geopo-tential heights are associated with cooler and oftenstormier conditions A positive trend in burn days wouldimply lower 500 hPa heights (ie less ridging) coolertemperatures and likely more precipitation but cool-season ridging has in fact increased since the middle ofthe twentieth century (Swain et al 2016) A CARB me-teorologist (D Mims personal communication 2017)stated that in 2008 mixing heights and transport windswere given increased weight in burn-day decisions forthe Lake Tahoe Air Basin rather than relying as stronglyon 500 hPa height Thus the positive trend in burn win-dows was not due to shifting meteorological conditionsbut to a regulatory change

DiscussionIn frequent-fire (FR1) forests of the western USA fire isa critically important ecological process that has been

Fig 5 Monthly burn-window frequency by year for the Lake Tahoe Basin USA from 1999 to 2019 based on observed and estimated Baronremote automated weather station data Interannual standard deviations for each month are shown in the upper right-hand corner ofeach graph

greatly reduced by human management leading todegraded ecological conditions Much of the yellow pinendashmixed conifer forest is at increased risk of uncharacteris-tically large high-severity wildfires (Westerling et al 2006Miller et al 2009 Safford and Stevens 2017) Forest res-toration and fuel hazard reduction activities are imple-mented to reduce this risk (Ritchie et al 2007 North et al2009 Safford et al 2012b McIver et al 2013) Althoughthe restoration of fire itself (rather than its replacementthrough surrogates) has been described as a key compo-nent of such restoration and hazard reduction programs(Agee and Skinner 2005 Ritchie et al 2007 North et al2009 Stephens et al 2009 Vaillant and Stephens 2009McIver et al 2013) there are numerous challenges in ap-plying prescribed fire broadly Given these challenges es-tablishing and maintaining a prescribed fire program thatwill meet restoration and hazard reduction objectives re-quires flexibility and an understanding of burn-windowpatterns and inherent uncertaintyOur study shows that the annual frequency of burn

windows in the LTB follows a general pattern with thegreatest likelihood in spring followed by autumn (Figs 2and 3) Summer has the fewest burn windows of anyseason but conditions during some summers may besuitable to meet objectives on small spatial scales (eg2019 Fig 5) Autumn burn windows were somewhatless frequent than spring While burn windows are lessfrequent in autumn than they are in the spring man-agers often plan to conduct more complex prescribedunderstory burns in autumn because (1) the historicalfire season in the Sierra Nevada region was mostly sum-mer through autumn but summer has few burn win-dows and 2) autumn precipitation events can assist with

controlling prescribed fires reducing the chance of fireescape (Fettig et al 2010) Moreover fuel moisture istypically lower in autumn than in spring so if maximumfuel consumption is the chief objective late-seasonburns will be more effective (Knapp et al 2005) If in-creasing forest heterogeneity or maintaining litter andduff layers are key objectives higher fuel moisture inspring facilitates creating a patchier residual surface-fuelpattern (Knapp et al 2005 Knapp and Keeley 2006)Since burn windows are most prevalent in the springtaking advantage of those opportunities could help tobetter meet fuels and restoration program goalsIn areas with a predominantly late-season fire regime

however many species may not be adapted to early-season burning if the historical regime was one of pre-dominantly summer to early fall fire (Knapp et al 2007)and the ecological impacts of spring fires are not wellunderstood For example Harrington (1993) and Thieset al (2005) found that ponderosa pine (Pinus ponderosaDougl ex Laws) mortality was greater after autumnthan spring burns in Colorado and Oregon USA butSchwilk et al (2006) found no significant difference inoverstory tree mortality between early- and late-seasonburning in the southern Sierra Nevada Fettig et al(2010) measured higher mortality of large trees afterspring burns Few studies have focused on the long- andshort-term effects of spring burning on understory plantand animal species in montane forests Kerns et al(2006) found decreased prevalence of exotic species afterearly-season burns Knapp et al (2007) found lower im-pacts to understory perennial species but impacts ap-peared to be more related to fire intensity than toseason per se

Rsup2 = 03695

Rsup2 = 07103

0

50

100

150

200

250

300

350

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

2018

2019

Num

ber

of d

ays

Burn windows CARB burn days

Fig 6 Burn window and California Air Resources (CARB) burn day annual time series with trend lines and linear regression coefficients of determination for theLake Tahoe Basin USA 1999 to 2019 The solid black line represents the number of days of each year that were burn windows The dashed line represents thenumber of days of each year that CARB designated as burn days Significant increasing trends were detected in burn windows (Mann-Kendall τ = 0438 2-sidedP = 0008) Subsequent trend analyses of the three component variables (burn plan prescriptions CARB burn days preparedness level lt 3) identified CARB burndays as the component variable responsible for the trend (Mann-Kendall τ = 0616 2-sided P le 0001) Coefficient of determination (R2) values in the figure arelinear regression un-adjusted coefficients of determination

Burn windows were also reasonably common in thewinter (Figs 2 and 3) Winter burning can be limiteddue to the occurrence of inversions that trap smoke atlow altitudes degrading air quality Topographic basinsand valleys like the LTB are especially prone to winterinversions under high-pressure conditions when loweratmosphere mixing is attenuated (Blandford et al 2008Wang et al 2015) Regardless large parts of the LTB aresnow covered in most winters Although snow cover wasnot considered in prescription criteria here it inhibitsmost burning (pile burning can occur if piles are ex-posed and accessible) During recent droughts howeversome parts of the LTB were snow free for much of thewinter (eg 2013 2015) Burning could be accomplishedduring burn windows in such drought years If snow-free or low-snow winters become more common in thefuture as some studies suggest (Hayhoe et al 2004Knowles et al 2006 Cayan et al 2008) prescribed burn-ing may become increasingly possible during winter Aswith spring burning the ecological ramifications of win-ter burns are not well understood Research on pre-scribed burns during the winter and spring will help tocharacterize the advantages and disadvantages of burn-ing during seasons or conditions outside current man-agement practice and the historical fire seasonThere were no significant trends in annual burn-

window frequency once the effect of CARBrsquos policy-driven increase in burn days was removed CARB alteredburn day determination criteria in 2008 in response torequests by LTB land managers following the destructiveAngora Fire in 2007 in order to increase fire-hazard re-duction opportunities using prescribed fire (D Mimspersonal communication 2017) Because of data limita-tions we did not examine when prescribed fires wereimplemented over the full period of this analysis but ra-ther those days when prescribed fire could have beenimplemented based on our criteria As a result we donot know if the additional burn days were utilized butthe trend in burn windows associated with a change inburn day criteria demonstrates that reasonable regula-tory changes can increase opportunities to implementprescribed burningHistorical studies indicate that montane forests of the

LTB supported frequent fires before the arrival of Euro-Americans (Taylor 2004 Maxwell et al 2014) with hun-dreds to thousands of hectares burning per year (Manleyet al 2000) This fire frequency and extent are propor-tionate to forest fire regimes throughout much of the Si-erra Nevada Between 2010 and 2018 (2014 data aremissing) burn logs for the LTB Management unit indi-cate that the USDA Forest Service treated about 323 haper year in the LTB utilizing about 51 burn windows ineach year averaging roughly 64 ha per day The fewestburn windows (34) were exploited in 2013 and the most

(81 burn windows) in 2010 Average area burned perburn window ranged from 30 ha per burn window in2012 to 124 ha per burn window in 2018 Prescribedburns averaged about 47 ha in size and individual burnsrarely exceeded 80 ha (although burning adjacent unitscould function as a single larger fire) Thus treated areaswere typically notably smaller than historical fires whichare thought to have averaged about 200 to 400 hectaresin size in this part of the Sierra Nevada (Safford andStevens 2017)On average there were 96 burn windows each year in

the LTB To attain Manley et alrsquos (2000) (probably con-servative) estimate of ~800 hectares burned in an aver-age year before 1850 managers would need to burn anaverage of 85 hectares during each burn window Theycurrently burn at a rate slightly below 7 ha per burnwindow and utilize on average just over half of theavailable burn windows This suggests either that it isnot possible to use all available burn windows androutinely treat 85 ha per burn window with current re-sources and risk tolerance or that there may beadditional constraints on burning that were not consid-ered here Although our analysis suggests that resourcesare usually not a limiting factor (Fig 2) PL is an imper-fect proxy It is designed to assess wildfire readiness andnot the capacity to conduct prescribed burns Becausethe fire season is concentrated during the summermonths the temporary workforce is often reduced dur-ing spring and autumn decreasing resource availabilityfor forest management activities at a time when burnwindows and particularly multi-day burn windows aremore common (Figs 2 3 4)Increasing staffing during the spring and autumn would

appear to be a reasonable response particularly because itmight allow for larger burns on days when managers canburn However interannual variability in burn-window fre-quency is high during those seasons (Fig 5) creating chal-lenges for managers who want to take advantage of periodswhen burn windows are frequent yet reduce costs associatedwith keeping crews on payroll when burning opportunitiesdo not occur Exploring relationships between burn-windowpatterns and large-scale climatic drivers (eg El NintildeoSouthern Oscillation) could help better forecast burn-window availability in upcoming seasons and potentially re-duce uncertainty for managers Developing innovative crewstaffing programs may be required to meet these challengesForest Service Region 5 is currently transitioning to a unifiedprogram of work for all national forests in its region entitledOne Region One Program of Work (USDA Forest ServicePacific Southwest Region 2019) This encourages sharing ofcrews personnel with needed skills and resources acrossunits to meet management goals in the face of changing cli-mate declining budgets and shrinking staffs Other optionsinclude interagency crews formed through state local and

federal partnerships that could help ease the financial burdenwhile recognizing fuels reduction and restoration prioritiesand multi-resource management crews that are prescribed-fire qualified but can also be used for other types of workThe recent institution of year-round full-time prescribed fireteams by CAL FIRE some of which are stationed near theLTB may be a catalyst for this sort of collaborative workIf resource availability cannot be increased the other

option is to increase the number of available burn win-dows by introducing greater flexibility in air quality orprescriptive standards Such flexibility was demonstratedby CARB when it changed burn-day determination cri-teria for the LTB in 2008 significantly increasing thenumber of burn windows Since days in prescription areless frequent than other criteria studied here practicesthat relax some prescriptive criteria may be especiallyhelpful One possibility is a matrix approach to prescrip-tions in which parameters offset each other (eg lowdead fuel moisture is offset by high live fuel moisture orlower fuel moisture and humidity are offset by low windspeeds Raybould and Roberts 2006) Permitting higherlevels of tree mortality in prescribed fires would alsoallow greater flexibility in burn prescriptions Currentprescribed fire prescriptions are often designed tominimize overstory mortality However even low-severity burning in wildfires can kill 20 or more of af-fected trees and it has been suggested that prescribedfires should aim to better mimic the impacts of historicalwildfires for example by permitting higher mortalitylevels in canopy trees (Safford et al 2012b)Retrospective analyses like this provide a tool to evalu-

ate multiple concurrent constraints on prescribed burn-ing can also be used to test the effectiveness of staffingand regulatory changes If managers compare availableand actual burn windows and find that they are notexploiting burn windows in the early spring or lateautumn due to resource issues they could plan short ex-tensions to some seasonal hire terms By applying differ-ent prescriptive criteria to the weather data used hereand evaluating how those criteria influence the numberand timing of burn windows managers could identifywhen modest changes to prescription criteria wouldexpand burn windows most conducive to meeting man-agement goals This tool could also be used in collabor-ation with air quality regulators to detect times of yearwhen otherwise multi-day burn windows are truncatedby no-burn days and assess the costs and benefits ofadditional regulatory changes Multi-day burn windowswould allow larger burn projects to be completed

ConclusionsForest managers navigate a complex system of environ-mental policy and regulatory requirements as well asconsider public opinion to plan and implement

prescribed fires (Quinn-Davidson and Varner 2012 Ryanet al 2013 North et al 2015a b Kolden 2019) Weatherand resource limitations like those investigated hereconstrain managersrsquo ability to meet restoration objectiveswith prescribed fire (Quinn-Davidson and Varner 2012North et al 2015b) Given the importance of prescribedfire and the myriad constraints to its implementationmanagers need tools to help reduce uncertainty whenplanning fuels-management programs This study mayassist forest managers in planning and prioritizing pre-scribed fire programs by quantifying constraints andopportunities and identifying areas for management-relevant researchPrescribed fire is an important tool for restoring FR1

forests and reducing fuels loads but its current use onthe ground in the western USA is making a vanishinglysmall contribution to reducing the fire deficit (Northet al 2012 Quinn-Davidson and Varner 2012 Northet al 2015a Kolden 2019) Using methods that are easilyapplicable to other management units operating undersimilar regulatory regimes we showed that (1) burn win-dows occur infrequently (2) multi-day burn windowsare rare and (3) there is high interannual variability inburn window occurrence particularly in the spring andautumn These conditions characterize much of thewestern USA and challenge managers trying to plan effi-cient and effective burning programsConsidering the limitations to prescribed fire imple-

mentation can also help managers and regulators iden-tify modest changesmdashlike those implemented by CARBin the LTBmdashthat can enhance prescribed burning op-portunities Quantitative assessment of prescribed burn-ing opportunities is particularly important now becausethe fire season is growing in length (Westerling et al2006 Jolly et al 2015) and the periods preferred for pre-scribed burning are shifting earlier in the spring andlater in the fall when seasonal staffing is often reducedand the ecological consequences of prescribed fire areless well understood Analyzing historical burn windowpatterns and the factors that constrain them can helpmanagers pinpoint optimal periods in the calendar thatare most likely to provide opportunities to burn safelyefficiently and sustainably

Supplementary informationSupplementary information accompanies this paper at httpsdoiorg101186s42408-020-00071-3

Additional file 1 Percent each day of the year was a burn windowfrom 2012 to 2019 for Baron remote automated weather station (RAWSelevation 1931 m) and four comparable RAWS nearby at similarelevations and forest types but outside the Lake Tahoe Basin USA Burnwindows for our study assessing the patterns of potential prescribedburning feasibility in the Lake Tahoe Basin from 1999 to 2019 weredesignated as days with simultaneous occurrence of weather within burn

plan prescription criteria sufficient resources for implementation andpermission from air quality regulators to burn The general burn-windowfrequency pattern exhibited at Baron RAWS is consistent overall highestfrequencies in spring and autumn lowest during summer MarkleevilleRAWS (elevation 1676 m) and Little Valley RAWS (elevation 1920 m)tended to have higher burn-window frequencies in winter while Stam-pede RAWS (elevation 1891 m) tended to have the lowest Dog ValleyRAWS (elevation 1821 m) had highest frequencies in March and AprilThese burn-window frequencies reflect differences in the weather-generated prescription variables (relative humidity 10-hour fuel moistureand wind gust speeds)

Additional file 2 Percent days for each month that met burn-windowcriteria from 2012 to 2019 for Baron remote automated weather station(RAWS) and four comparable RAWS nearby but outside the Lake TahoeBasin USA Burn windows for our study assessing the patterns of poten-tial prescribed burning feasibility in the Lake Tahoe Basin from 1999 to2019 were composed of days with co-occurrence of permission to burnby the air quality regulators sufficient resources needed for implementa-tion and weather within burn plan prescription criteria Monthly burn-window frequencies for each RAWS are shown for comparison The gen-eral burn-window frequency pattern exhibited at Baron RAWS is consist-ent overall highest frequencies in spring and autumn lowest duringsummer Little Valley RAWS had the highest frequencies and StampedeRAWS had the lowest during winter (December to February) As withdaily frequencies (Additional file 1) Dog Valley RAWS had highest fre-quencies in March and April and second only to Baron RAWS in May andJune Markleeville RAWS had lowest frequencies April to November

Additional file 3 Annual burn-window frequency by month for four re-mote automated weather station (RAWS) compared to Baron RAWS inthe Lake Tahoe Basin and surrounding region USA Days with simultan-eous occurrence of permission to burn by the air quality regulators suffi-cient resources needed for implementation and weather within burnplan prescription criteria were designated as burn windows for our studyassessing the patterns of potential prescribed burning feasibility in theLake Tahoe Basin from 1999 to 2019 The seasonal patterns exhibited fordaily and monthly frequencies generally apply (eg low frequencies insummer and highest frequencies in spring) However a high degree ofannual variation is apparent Notable is the consistency between stationsfor relatively high burn-window frequency during summer 2019 as wellas July 2015

AcknowledgementsNot applicable

Authorsrsquo contributionsRS and MP developed the basic methodology and conducted the dataanalysis RS HS and SM contributed to the manuscript All authors read andapproved the final manuscript

FundingWork on the project was carried out as part of the authorsrsquo employment andRSrsquos graduate studies and was not funded by any specific grant or contract

Availability of data and materialsThe corresponding author will provide data and the Excel spreadsheet usedfor calculation upon request

Ethics approval and consent to participateNot applicable

Consent for publicationNot applicable

Competing interestsThe authors declare they have no competing interests

Author details1USDA Forest Service Pacific Southwest Region 1323 Club Drive VallejoCalifornia 94592 USA 2Department of Geography University of NevadaReno 1664 North Virginia Street Reno Nevada 89557 USA 3Nevada State

Climate Office and University of Nevada Extension 1664 North VirginiaStreet Reno Nevada 89557 USA 4Department of Environmental Science andPolicy University of California One Shields Avenue Davis California 95616USA 5USDA Forest Service National Forest in Florida Ocala National Forest40929 State Road 19 Umatilla Florida 32784 USA

Received 8 October 2019 Accepted 22 April 2020

ReferencesAbatzoglou JT and AP Williams 2016 Impact of anthropogenic climate

change on wildfire across western US Forests Proceedings of the NationalAcademy of Sciences 42 11770-11775 wwwpnasorgcgidoi101073pnas1607171113 httpsdoiorg101073pnas1607171113

Agee JK WH Romme JF Franklin MD Hurteau SL Stephens N Johnson TWSwetnam P Morgan J van Wagtendonk 2016 Letter to EPA USDA USDOICEQ The Fire Challenge Increasing Fire Use for Natural Resource BenefitsCarbon Stability and Protection of Public Health Letter to EPA USDA USDOICEQ httpsdoiorg101017CBO9781107415324004

Agee JK and CN Skinner 2005 Basic principles of forest fuel reductiontreatments Forest Ecology and Management 211 83-96 httpsdoiorg101016jforeco200501034

Barbour MG E Kelley P Maloney D Rizzo E Royce and J Fites-Kaufmann2002 Present and past old-growth forests of the Lake Tahoe Basin SierraNevada US Journal of Vegetation Science 13 461-472 httpsdoiorg1016581100-9233(2002)013[0461PAPOGF]20CO2 httpsdoiorg101111j1654-11032002tb02073x

Beaty RM and AH Taylor 2008 Fire history and the structure and dynamics ofa mixed conifer forest landscape in the northern Sierra Nevada Lake TahoeBasin California USA Forest Ecology and Management 255 707-719 httpsdoiorg101016jforeco200709044

Blandford TR KS Humes BJ Harshburger BC Moore VP Walden and H Ye2008 Seasonal and synoptic variations in near-surface air temperature lapserates in a mountainous basin Journal of Applied Meteorology and Climatology47 249-261 httpsdoiorg1011752007JAMC15651

Bradshaw Larry and Erin McCormick 2000 FireFamily Plus users guide Version 20Gen Tech Rep RMRS-GTR-67 Ogden US Department of Agriculture ForestService Rocky Mountain Research Station httpsdoiorg102737RMRS-GTR-67

California 2010 California Code of Regulations Title 17 subchapter 2 Smokemanagement guidelines for agricultural and prescribed burning httpsww3arbcagovregsregs-17htm

California Department of Forestry and Fire Protection California State ParksCalifornia Tahoe Conservancy Fallen Leaf Fire Department Lake Valley FireProtection District Meeks Bay Fire Protection District Nevada Division ofForestry Nevada Division USDA Forest Service 2014 Lake Tahoe Basin Multi-Jurisdictional Fuel Reduction and Wildfire Prevention Strategy

Cayan DR EP Maurer MD Dettinger M Tyree and K Hayhoe 2008 Climatechange scenarios for the California region Climatic Change 87 21-42 httpsdoiorg101007s10584-007-9377-6

Elliott-Fisk DL TC Cahill OK Davis L Duan CR Goldman GE Gruell R HarrisR Kattelmann R Lacey D Leisz S Lindstrom D Machida RA Rowntree PRucks DA Sharkey SL Stephens and DS Zeigler 1997 Lake Tahoe casestudy Sierra Nevada Ecosystem Project Final Report to Congress AddendumUniversity of California Davis Centers for Water and Wildland Resources

Estes BL EE Knapp CN Skinner and FCC Uzoh 2012 Seasonal variation insurface fuel moisture between unthinned and thinned mixed conifer forestnorthern California USA International Journal of Wildland Fire 21 428-435httpsdoiorg101071WF11056

Fettig CJ SR McKelvey DR Cluck SL Smith and WJ Otrosina 2010 Effects ofprescribed fire and season of burn on direct and indirect levels of treemortality in Ponderosa and Jeffrey Pine Forests in California USA ForestEcology and Management 260 207-218 httpsdoiorg101016jforeco201004019

Gilbert RO 1987 Statistical Methods for Environmental Pollution Monitoring NYWiley

Hardy CC KM Schmidt JM Menakis and NR Samson 2001 Spatial data fornational fire planning and fuel management International Journal of WildlandFire 10 353-372 httpsdoiorg101071WF01034

Harrington M 1993 Mortality from dormant season and growing-season fireinjury International Journal of Wildland Fire 3 65-72 httpsdoiorg101071WF9930065

Hayhoe K D Cayan CB Field PC Frumhoff EP Maurer NL Miller and JHVerville 2004 Emissions pathways climate change and impacts onCalifornia Proceedings of the National Academy of Sciences 101 12422-12427httpsdoiorg101073pnas0404500101

Holden ZA A Swanson CH Luce WM Jolly M Maneta JW Oyler DAWarren R Parsons and D Affleck 2018 Decreasing fire season precipitationincreased recent western US forest wildfire activity Proceedings of theNational Academy of Sciences 115 E8349-E8357 wwwpnasorgcgidoi101073pnas1802316115 httpsdoiorg101073pnas1802316115

Jolly WM MA Cochrane PH Freeborn ZA Holden TJ Brown GJ Williamsonand DMJS Bowman 2015 Climate-induced variations in global wildfiredanger from 1979 to 2013 Nature Communications 6 7537 httpsdoiorg101038ncomms8537

Kendall MG 1975 Rank Correlation Methods 4th ed London Charles GriffinKerns BK WG Thies and CG Niwa 2006 Season and severity of prescribed

burn in ponderosa pine forests implications for understory native and exoticplants Ecoscience 13 44-55 httpsdoiorg1029801195-6860(2006)13[44SASOPB]20CO2

Kilgore BM and RW Sando 1975 Crown-fire potential in a sequoia forest afterprescribed burning Forest Science 21 83-87 httpsdoiorg101093forestscience21183

Knapp EE and JE Keeley 2006 Heterogeneity in fire severity within earlyseason and late season prescribed burns in a mixed-conifer forestInternational Journal of Wildland Fire 15 37-45 httpsdoiorg101071WF04068

Knapp EE JE Keeley EA Ballenger and TJ Brennan 2005 Fuel reduction andcoarse woody debris dynamics with early season and late season prescribedfire in a Sierra Nevada mixed conifer forest Forest Ecology and Management208 383-397 httpsdoiorg101016jforeco200501016

Knapp EE DW Schwilk JM Kane and JE Keeley 2007 Role of burning seasonon initial understory vegetation response to prescribed fire in a mixedconifer forest Canadian Journal of Forest Research 37 11-22 httpsdoiorg101139x06-200

Knowles N M Dettinger and D Cayan 2006 Trends in snowfall versus rainfall inthe western United States Journal of Climate 19 4545-4559 httpsdoiorg101175JCLI38501

Kolden CA 2019 Were not doing enough prescribed fire in the western UnitedStates Fire 2 30 httpsdoiorg103390fire2020030

LTBMU 2015 Lake Tahoe Basin Management Unit Visitor Use Monitoring ReportMallek CR HD Safford JH Viers and J Miller 2013 Modern departures in fire

severity and area vary by forest type Sierra Nevada and southern CascadesCalifornia USA Ecosphere 4 153 httpsdoiorg101890ES13-002171

Manley PN JA Fites-Kaufman MG Barbour MD Schlesinger and DM Rizzo2000 Biological Integrity In Lake Tahoe watershed assessment Volume IPSW-GTR-175 ed DD Murphy and CM Knopp 403-600 Albany US Dept ofAgriculture Forest Service Pacific SW Research station

Mann HB 1945 Non-parametric tests against trend Econometrica 13 163-171httpsdoiorg1023071907187

Maxwell R A Taylor C Skinner H Safford R Isaacs C Airey and A Young 2014Landscape-scale modeling of reference period forest conditions and firebehavior on heavily logged lands Ecosphere 5 32 httpsdoiorg101890ES13-002941

McIver JD SL Stephens JK Agee J Barbour REJ Boerner CB EdminsterKL Erickson KL Farris CJ Fettig CE Fiedler S Haase SC Hart JEKeeley EE Knapp JF Lehmkuhl JJ Moghaddas W Otrosina KWOutcalt DW Schwilk CN Skinner TA Waldrop CP Weatherspoon DAYaussy A Youngblood and S Zack 2013 Ecological effects ofalternative fuel-reduction treatments Highlights of the National Fire andFire Surrogate study (FFS) International Journal of Wildland Fire 22 63-82httpsdoiorg101071WF11130

McLeod AI 2011 Kendall Kendall rank correlation and Mann-Kendall trend testR package version 2 2 httpsCRANR-projectorgpackage=Kendall

Miller JD HD Safford M Crimmins and AE Thode 2009 Quantitative Evidencefor Increasing Forest Fire Severity in the Sierra Nevada and Southern CascadeMountains California and Nevada USA Ecosystems 12 16-32 httpsdoiorg101007s10021-008-9201-9

Nagel TA and AH Taylor 2005 Fire and persistence of montane chaparral inmixed conifer forest landscapes in the northern Sierra Nevada Lake TahoeBasin California USA The Journal of the Torrey Botanical Society 132 442-457httpsdoiorg1031591095-5674(2005)132[442FAPOMC]20CO2

National Wildfire Coordinating Group 2017 Interagency Prescribed Fire Planningand Implementation Procedures Guide PMS 484-1 httpswwwnwcggovpublications484

Nelson RMJ 2000 Prediction of diurnal change in 10-h fuel stick moisturecontent Canadian Journal of Forest Research 30 1071-1087 httpsdoiorg101139x00-032

North M A Brough J Long B Collins P Bowden D Yasuda J Miller and NSugihara 2015a Constraints on mechanized treatment significantly limitmechanical fuels reduction extent in the Sierra Nevada Journal of Forestry113 40-48 httpsdoiorg105849jof14-058

North M BM Collins and S Stephens 2012 Using fire to increase the scalebenefits and future maintenance of fuels Journal of Forestry 110 392-401httpsdoiorg105849jof12-021

North M S Stephens B Collins J Agee G Aplet J Franklin and PZ Fuleacute2015b Reform forest fire management Agency incentives undermine policyeffectiveness Science 349 1280-1281 httpsdoiorg101126scienceaab2356

North M P Stine KO Hara W Zielinski and SL Stephens 2009 An EcosystemManagement Strategy for Sierran Mixed-Conifer Forests General TechnicalReport PSW-GTR-220 49 Albany US Department of Agriculture Forest ServicePacific Southwest Research Station httpsdoiorg102737PSW-GTR-220

Quinn-Davidson LN and JM Varner 2012 Impediments to prescribed fireacross agency landscape and manager An example from northernCalifornia International Journal of Wildland Fire 21 210-218 httpsdoiorg101071WF11017

R Core Team 2016 R A language and environment for statistical computingVienna R Foundation for Statistical Computing httpswwwR-projectorg

Raybould S and T Roberts 2006 A matrix approach to fire prescription writingFire Management Today 66 79-82

Ritchie MW CN Skinner and TA Hamilton 2007 Probability of tree survivalafter wildfire in an interior pine forest of northern California Effects ofthinning and prescribed fire Forest Ecology and Management 247 200-208httpsdoiorg101016jforeco200704044

Ryan KC EE Knapp and JM Varner 2013 Prescribed fire in North Americanforests and woodlands history current practice and challenges Frontiers inEcology and the Environment 11s1 e15-e24 httpsdoiorg101890120329

Safford HD MP North and MD Meyer 2012a Climate change and therelevance of historical forest conditions In Managing Sierra Nevada forestsGeneral Technical Report PSW-GTR-237 ed MP North 23-46 Albany USDAForest Service Pacific Southwest Research Station

Safford HD and JT Stevens 2017 Natural Range of Variation (NRV) for yellowpine and mixed conifer forests in the Sierra Nevada southern Cascades andModoc and Inyo National Forests California USA General Technical ReportPSW-GTR-256 229 Albany US Department of Agriculture Forest ServicePacific Southwest Research Station

Safford HD JT Stevens K Merriam MD Meyer and AM Latimer 2012b Fueltreatment effectiveness in California yellow pine and mixed conifer forestsForest Ecology and Management 274 17-28 httpsdoiorg101016jforeco201202013

Schmidt KM JP Menakis CC Hardy WJ Hann and DL Bunnell 2002Development of Coarse-Scale Spatial Data for Wildland Fire and FuelManagement General Technical Report RMRS-87 Fort Collins USDepartment of Agriculture Forest Service Rocky Mountain Research Station41 p + CD httpsdoiorg102737RMRS-GTR-87

Schwilk DW EE Knapp SM Ferrenberg JE Keeley and AC Caprio2006 Tree mortality from fire and bark beetles following early and lateseason prescribed fires in a Sierra Nevada mixed-conifer forest ForestEcology and Management 232 36-45 httpsdoiorg101016jforeco200605036

Skinner CN and C-R Chang 1996 Fire Regimes Past and Present SierraNevada Ecosystem Project Final report to Congress vol II Assessments andscientific basis for management options Davis University of CaliforniaCenters for Water and Wildland Resources

State of California 2020 California Climate Investments Prescribed FireProgram httpwwwcaclimateinvestmentscagovprescribed-fireAccessed 8 Mar 2020

Stephens SL JDM McIver REJ Boerner CJ Fettig JB Fontaine BRHartsough PL Kennedy and DW Schwilk 2012 The effects of forest fuel-reduction treatments in the United States Bioscience 62 549-560 httpsdoiorg101525bio20126266

Stephens SL and JJ Moghaddas 2005 Fuel treatment effects on snags andcoarse woody debris in a Sierra Nevada mixed conifer forest Forest Ecologyand Management 214 53-64 httpsdoiorg101016jforeco200503055

Stephens SL JJ Moghaddas C Edminster CE Fiedler S Haase M HarringtonJE Keeley EE Knapp JD McIver K Metlen CN Skinner E Fiedler and MHall 2009 Fire treatment effects on vegetation structure fuels and potentialfire severity in western US forests Ecological Applications 19 305-320 httpsdoiorg10189007-17551

Swain DL DE Horton D Singh and NS Diffenbaugh 2016 Trends inatmospheric patterns conducive to seasonal precipitation and temperatureextremes in California Science Advances 2 e1501344 httpsdoiorg101126sciadv1501344

TahoeClim Desert Research Institute 2017 httpwwwtahoeclimdrieduAccessed 5 May 2017

Taylor AH 2004 Identifying forest reference conditions on early cut-over landsLake Tahoe Basin USA Ecological Applications 14 1903-1920 httpsdoiorg10189002-5257

Thies WG DJ Westlind and M Loewen 2005 Season of prescribed burn inponderosa pine forests in eastern Oregon Impact on pine mortalityInternational Journal of Wildland Fire 14 223-231 httpsdoiorg101071WF04051

USDA Forest Service 2012 increasing the Pace of Restoration and Job Creationon Our National Forests Published Report Washington DC US Departmentof Agriculture Forest Service

USDA Forest Service California Dept of Forestry amp Fire Protection Bureau ofLand Management National Park Service Bureau of Indian Affairs U S Fishand Wildlife Service Governors Office of Emergency Services 2016 CaliforniaMobilization Guide

USDA Forest Service Pacific Southwest Region 2011 Region 5 EcologicalRestoration Leadership Intent R5-MR-048 4 Pacific Southwest Region USDepartment of Agriculture Forest Service

USDA Forest Service Pacific Southwest Region 2019 Spotlight One Region OneProgram of Work httpswwwfsusdagovlistr5homelistposition=SubFeature Accessed 29 Apr 2019

Vaillant NM and ED Reinhardt 2017 An Evaluation of the Forest ServiceHazardous Fuels Treatment Program-Are We Treating Enough to PromoteResiliency or Reduce Hazard Journal of Forestry 115 (4) 300-308 httpsdoiorg105849jof16-067

Vaillant NM and SL Stephens 2009 Fire history of a lower elevation JeffreyPine-mixed conifer forest in the eastern Sierra Nevada California USA FireEcology 5 4-19 httpsdoiorg104996fireecology0503004

van Wagtendonk J N G Sugihara SL Stephens AE Thode KE Shaffer and JFites-Kaufman editors 2018 Fire in California Ecosystems 2nd editionUniversity of California Press Oakland

Wang S-Y LE Hipps O Chung RR Gillies and R Martin 2015 Long-termwinter inversion properties in a mountain valley of the western United Statesand implications on air quality Journal of Applied Meteorology andClimatology 54 2339-2352 httpsdoiorg101175JAMC-D-15-01721

Westerling AL HG Hidalgo DR Cayan and TW Swetnam 2006 Warming andearlier spring increase western US forest wildfire activity Science 313 940-943 httpsdoiorg101126science1128834

Western Regional Climate Center 2017 SOUTH LAKE TAHOE AP CALIFORNIANCDC 1981-2010 Monthly Normals Retrieved from httpswrccdrieducgi-bincliMAINplca8762

Yurkonis KA J Dillon DA McGranahan D Toledo and BJ Goodwin 2019Seasonality of prescribed fire weather windows and predicted fire behaviorin the northern Great Plains USA Fire Ecology 15 7 httpsdoiorg101186s42408-019-0027-y

Zar JH 1999 Biostatistical analysis 4th ed 663 Upper Saddle River Prentice Hall

Publisherrsquos NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations

Abstract
- Background
- Results
- Conclusions
- - Resumen
  - - Antecedentes
    - Resultados
    - Conclusiones
    - - Abbreviations
      - Background
      - Methods
      - Study area
        
        Burn window analysis
        
        CARB burn day
        
        Days within prescription
        
        Availability of firefighting resources
        
        Burn-window occurance
        
        Results
        
        Discussion
        
        Conclusions
        
        Supplementary information
        
        Acknowledgements
        
        Authorsrsquo contributions
        
        Funding
        
        Availability of data and materials
        
        Ethics approval and consent to participate
        
        Consent for publication
        
        Competing interests
        
        Author details
        
        References
        
        Publisherrsquos Note

Page 2: Retrospective analysis of burn windows for fire and fuels ...very rare. Interannual variation was considerable. Finally, an abrupt increase in burn windows was detected in 2008. ...