Challenges and Approaches inPlanning Fuel Treatments acrossFire-Excluded Forested Landscapes

IBrandon M. Collins, Scott L. Stephens, Jason J. Moghaddas,and John Battles

fire

F

uel reduction treatments are increas-ingly becoming the dominant forestmanagement activity throughout

western US forests, particularly in foreststhat historically burned under frequent,low- to moderate-severity fire regimes. Theprimary objectives of such activities are tomodify wildland fire behavior to minimizenegative impacts on forests (Agee and Skin-ner 2005), enhance suppression capabilities(Agee et al. 2000), improve firefighter safety

(Moghaddas and Craggs 2007), and help re-store ecological structure and function(McKelvey cc al. 1996, Weatherspoon andSkinner 1996, North et al. 2007). Stand-scale studies have established the effective-ness of various fuel treatment alternatives atchanging the behavior and reducing the im-pacts of both modeled fires (Fulé et a]. 2001,Fiedler et al. 2004, Stephens and Moghad-das 2005, Schmidt et al. 2008, Stephens ccal. 2009) and actual wildiand fires (Martin-

son and Omi 2002, Skinner et al, 2005,Ritchie et al. 2007, Strom and Fulé 2007).These studies document tradeoffs amongtreatment types and provide guidance on de-signing prescriptions for forest stands. How-ever, the extensive tracts of relatively ho-mogenous, fire-excluded forests (Hessburget at. 2005) throughout the western UnitedStates and the large wildfires that can occur

in these forests (e.g., Rodeo-Chediski, Ari-zona 2002; Hayman, Colorado 2002; Bis-cuit, Oregon/California 2002; Murphy com-

plex, Idaho 2007) show the pressing need to"scale up" insights gained at the stand level tolandscapes. However, simply implementingfuel treatments across an entire landscape is in-feasible. Lack of infrastructure (e.g., work-force, equipment, and roads), lack of funding,and management restrictions collectively limitthe amount of area available for treatment.

In response, fire scientists and managershave conceptually developed and are refin-ing methods for the strategic placement offuel treatments across landscapes (Weather-spoon and Skinner 1996, Finney 2001,2004, Stratton 2004, Finney et al. 2007).

Placing fuel reduction treatments across entire landscapes such that impacts associated with high-intensity fire are lessened is a difficult goal to achieve, largely because of the immense area needingtreatment. As such, fire scientists and managers have conceptually developed and are refiningmethodologies for strategic placement of fuel treatments that more efficiently limit the spread andseverity of fire across forested landscapes. Although these methodologies undoubtedly improvemanagers' ability to plan and evaluate various landscape fuel treatment scenarios, there is still aconsiderable gap between modeling landscape fuel treatments and actually implementing thesetreatments "on the ground." In this article we explore this gap in light of decisions managers makewith regards to the type, intensity, placement/pattern, and size of fuel treatments. Additionally, wehighlight several critical constraints acting on managers when implementing fuel treatments acrosslandscapes and offer some suggestions for dealing with these constraints.

Keywords: fire exclusion, fire suppression, fire management, fire policy

Received October 3,2008; accepted July 13, 2009.

Brandon M. Collins (bmcollins@fi.fed.us) is currently research forester, US Forest Service, Pacific Southwest Research Station, Davis, CA. Scott L. Stephens(sstephens@berkeley.edu) is associate professor, Jason J. Moghaddas (moghad@berkeley.edu) is staff researcher, and John Battles (jbattles@berkeley.edu ) is professor,Ecosystem Sciences Division, Department ofEnvironmental Science, Policy. and Management, 137 MulJbrd Hall, University of California, Berkeley, CA 94720-3114. The Sierra Nevada Adaptive Management Project, a joint effort between US Forest Service Region 5, the US Forest Service Pacific Southwest Research Station,US Fish and Wiia'l/è Service, California Department of Water Resources, California Department ofFish and Game, the California Department ofForesny and FireProtection, the University of California, and the University ofMinnesota to investigate the effects oflandscapejliel treatments on Sierran forested ecosystems, fundedthis effort. In addition, this project was the source of valuable insight and discussion for several of the topics discussed.

24 Journal ofForestry • Jan uary/February 2010

The basic idea is that an informed deploy-ment of treatment areas, a deployment thatcovers only part of the landscape, can mod-ify fire behavior for the entire landscape.However, this technique requires the syn-thesis of spatially explicit data to use process-based models of fire behavior and spread(FARSITE, Finney 1998; FlamMap, Finney2006, Stratton 2006). Recent efforts havealso integrated a stand-level vegetation

model with the fire spread and behaviormodels (ArcFuels, Ager et al. 2006), provid-ing a more accessible platform for testingvarious stand-level treatment alternatives(thinning, prescribed burning, or a combi-nation of the two) across landscapes. Al-though these models and approaches un-doubtedly improve managers' ability to planand evaluate various landscape fuel treat-ment scenarios, there is still a considerable

Stand-Level Fuel Treatment OptionsTypes of fuel reduction treatments in- goals: (I) minimize the threat to hu-

elude fire (either prescribed or managed man life and property (including en-

wildiand fire), mechanical (e.g., thinning, hancing firefighter safety), (2) create/

mastication, and chipping), or a combina- maintain more fire-resilient structure

tion of the two. In field-based experiments and ecosystem function, and (3) re-

Stephens and Moghaddas (2005), duce the cost of fire suppression. Of-

Schmidt et al. (2008), and Stephens et al. ten, the intensity of a fuel treatment is

(2009) all found that prescribed fire alone manifested as a weather threshold un-

effectively reduced surface fuels, thus re- der which the residual stand is cx-

ducing modeled rate of spread, fireline in- peered to withstand wildfire-causedtensity , and flame length under a range of change. When the treatment involves

weather conditions. In addition, these prescribed burning, managers can

studies also showed substantial reductions choose to modify prescription param-

in ladder fuels in areas treated with pre- eters (seasonality, weather conditions

scribed fire. However, as fire-killed trees during burning, fuel moisture condi-

fall and contribute to surface fuel pools, tions, and ignition pattern) to achieve

the overall effectiveness in reducing poten- desired effects. More aggressive pre-

tial fire behavior is lessened (Skinner 2005, scriptions may include the following:

Keifer et al. 2006). It is likely that in dense late-season burning, lower relative hu-

fire-excluded stands multiple burns will be midity, lower fuel moisture, and strip

needed to achieve more long-lived effects. headfire ignition pattern. Conversely,

Thinning effectiveness depends on the more conservative prescriptions in-

type of thinning performed and the suhse- volve early season burning, higher rel-

quent treatment of activity fuels (Agee and ative humidity, higher fuel moisture,

Skinner 2005). In fire-excluded forests fuel and a backing or dot ignition pattern.

reduction prescriptions often aim to both Aggressive, late-season burns will con-

reduce ladder fuels (increase canopy base sume a greater amount of surface and

height) and increase crown spacing (reduce ladder fuel than a burn implemented

crown hulk density), in combination with under more conservative, early season

removing activity and surface fuels (e.g., conditions (Knapp et al. 2005), result-

piling and burning or broadcast under- ing in stands being more resistant toburning; Agee and Skinner 2005). Whole- wildland fire under more severe fire

tree harvests [3] have also been shown to weather conditions. However, more

effectively reduce modeled fire behavior aggressive prescriptions increase the

(Schmidt et al. 2008, Stephens et al. risk of a fire escaping and possibly

2009). Data on tree mortality in thinned causing unintended damage to other

areas burned by actual wildfires, which resources. The intensity of thinning

show greater survivability in areas under- treatments is often determined based on

burned after thinning, serve as real-world the residual stand structure desired after

tests on the importance treating activity fu- treatments. This can be expressed in terms

els after thinning (see Raymond and Peter- of basal area, tree density, crown spacing,son 2005, Ritchie et al. 2007). ladder fuels, and/or canopy cover. One im-

The management decision with re- portant aspect of desired conditions is spa-

gard to the intensity of fuel treatments is tial heterogeneity in forest structure (Ste-driven by the following fire management phens and Fulé 2005).

gap between modeling landscape fuel treat-ments and actually implementing thesetreatments "on the ground." In this articlewe explore this gap in light of decisions man-agers make and constraints acting on man-agers when implementing such activities.We also synthesize some of the current ap-proaches, along with the requirements forusing these approaches and offer some sug-gestions for dealing with the constraints.

Planning/ManagementDecisions

When designing landscape fuel treat-ments managers develop prescriptions tomanipulate individual forest stands so thatnot only the treated stands withstand wild-land fire under more extreme fire weather(e.g., 60-80% of overstoty trees survive afire tinder 90th percentile fire weather), butthe fire-caused effects over the landscape arelessened as well. In developing stand-levelprescriptions there are several decisions thatmust be made with respect to the type andintensity of the treatment (see side bar). Co-ordinating multiple stand-level treatmentsacross a landscape involves making decisionson size of individual treatment units, theplacement/pattern of the treatments, andthe proportion of the landscape treated.

Size of Individual TreatmentsThe decision on how large to make in-

dividual treatments at the stand level alsorelates to the performance of the treated areawhen it encounters wildland fire. The largerthe individual fuel treatment the greater thepotential for tree survival, because of re-duced edge effect mortality (Ritchie et al.2007). Furthermore, larger individual treat-ments have a greater potential to reduce firebehavior and slow fire spread, which ulti-mately impacts adjacent untreated standsand should enhance suppression opportuni-ties and increase firefighter safety. However,there are a number of constraints that limitthe size of individual treatments that will bediscussed in the following section. In addi-tion, when planning multiple stand-leveltreatments across landscapes there aretradeoffs between the size of individualtreatments and the dispersion of treatmentsacross the landscape.

Treatment PlacementUsing computer-based modeling, re-

cent studies have explored various scenariosof treatment placement across forested land-scapes (see Ager et al. 2007a, 2007b, Finney

Journal of Foresoy January/February2010 25

et al. 2007, Schmidt et al. 2008). Finney etal. (2007) show that treatment locationsbased on optimization algorithms (Finney2004, 2007) more effectively reduce simu-lated fire growth across several landscapescompared with random location of fueltreatments, Schmidt et al. (2008) also reportthat regular arrangement of treatments out-performed random arrangement with re-spect to reducing fire spread and areaburned. Planning fewer, larger individualtreatments across the landscape appears tobe a better strategy when human communityprotection is a primary concern (Schmidt et al.2008). These larger treated stands can also beused as suppression or other fire managementactivity anchor points (Omi 1996, Agee et al.2000, Moghaddas and Craggs 2007).

Treatment Proportions/RatesAlthough only a few studies have ex-

plicitly modeled landscape fuel treatment ef-fectiveness at different proportions oftreated area, there are some common find-ings: (1) although noticeable reductions inmodeled fire size, flame length, and spreadrate across the landscape relative to un-treated scenarios occurred with 10% of thelandscape treated, the 20% treatment levelappeared to have the most consistent reduc-tions in modeled fire size and behavior acrossmultiple landscapes and scenarios (Ager etal. 2007a, Finney et al. 2007, Schmidt et al.2008); (2) increasing the proportion of areatreated generally resulted in further reduc-tions in fire size and behavior, however, therate of reduction diminishes more rapidlybeyond 20% of the landscape treated (Ageretal. 2007a, Finney et al. 2007); (3) randomplacement of treatments requires substan-tially greater proportions of the landscapetreated compared with optimized or regulartreatment placement (Finney et al. 2007,Schmidt et al. 2008), however, Finney et al.(2007) note that the relative improvementof optimized treatment placement breaksdown when larger proportions of the land-scape (-40-50%) are excluded from treat-ment because of land-management con-straints.

To our knowledge, Finney et al. (2007)is the only published study that compareseffectiveness of different treatment rates overseveral decades. Their findings indicate thattreatment rates beyond 2% of the landscapeper year, based on optimized treatment place-ment, yielded little added benefit. This in-cludes both the maintenance of previouslytreated units and the installation of new treat-

ments throughout the 50-year simulation pe-riod. However, Finney et al. (2007) do notethat "higher rates might be desirable in the firstdecade followed by later decreases."

Forest DynamicsFinally, the dynamic nature of forest

ecosystems imposes an important temporalconsideration on landscape fuel planning. Asuite of fuel treatments deployed strategi-cally across the landscape will have a charac-teristic lifecycle. As the time since treatmentlengthens, tree growth responses rebuild fuelload and fuel continuity (Agee and Skinner2005, Collins et al. 2009). Thus, as stand-level treatments mature, the performance atthe landscape level will decline (Vaillarit2008). Therefore, the design of landscape-level fuel treatments involves a tradeoff be-tween maximizing the fraction of the land-scape area treated at least once or treating alimited area repeatedly to maintain treat-ment effectiveness (Finney et al. 2007).

Constraints

Habitat PreservationIn choosing among the options for

type, intensity, size, and placement/patternof fuel treatments across a landscape, thereare often conflicts between reducing poten-tial fire behavior and protecting/conservingother resources. One of the conflicts oftenon the forefront is habitat for wildlife speciesof concern. In some cases these species prefermultistoried stands and/or closed canopies(e.g., northern spotted owl, Strix occic/en-ta/is; Pacific fisher, Martes pennanti; Solisand Gutiérrez 1990, Spencer et al. 2008).Although it has been argued that fire sup-pression and past harvesting practices havecreated much of the habitat that is beingcalled "desirable" for species such as thespotted owl and fisher (see Spies et al. 2006),the species-specific approach toward manag-ing forests continues to prevail (Stephensand Ruth 2005). This approach not onlylimits the intensity of fuel treatments, butthe size and location of treatments as well. Asa consequence, managers' ability to modifypotential fire behavior, especially crown firebehavior, in forests with prolonged fire ex-clusion is restricted. Furthermore, regula-tions on forest management within andaround nesting centers or natal dens (pro-tected activity centers [PAC]) and riparianbuffer zones affect the placement and/orpattern of fuel treatments (Figure 1). Assuch, the optimal placement of fuel treat-

ments to maximize the reduction in firespread and intensity across the landscape,such as the regular pattern described byFinney (2001), is limited in its applicability.Additionally, these protected areas are oftenhighly productive and contain large amountof live and dead fuel; likely resulting in exac-erbated fire behavior creating effects notonly within these protected areas (Spies et al.2006), but also in adjacent stands.

Human CommunitiesWildland—urban interface (WUI) com-

munities located within a given landscape orplanning area can change the fuel treatmentplanning considerably. In most settings whereWUI communities are adjacent to large tractsof publicly owned forests, placing fuel treat-ments around the WUI is a high priority. Thisnot only affects the placement/pattern of fueltreatments across a landscape, but the harvestintensity and size of individual treatments. Forexample, because of exceptional potentiallosses from an escaped fire, prescribed firetreatments in the "IWUI would likely be smallerand involve more conservative prescriptionscompared with more remote locations. Airquality concerns influence whether or when aprescribed burn is conducted. Smoke produc-tion from prescribed burns is monitored andregulated by regional air quality districts, andthese districts need to approve burn plans andidentify days for a prescribed burn to takeplace. This can further restrict an already lim-ited "burn window," for which weather andfuel conditions are at a level that will allow fordesired fire spread, intensity, and ecological ef-fects while minimizing fire escapes.

Regulations and AppealsA necessary step to successfully imple-

menting any landscape-level fuel reductionproject on federal public lands is for thatproject to satisfy the requirements of the Na-tional Environmental Policy Act of 1969(NEPA) planning process. With respect tolandscape fuel treatment planning theNEIA process requires comprehensive eval-uation of the effects and impacts of varioustreatment scenarios, or alternatives, includ-ing a no treatment alternative. This processcan be completed in a reasonable time framefor small projects but has been shown to takeseveral years for landscape-level fuel treat-ment projects. Forest managers are oftenlimited in time, and some cases expertise,when conducting these comprehensive eval-uations. All projects contain a degree of riskor uncertainty that can not be reduced or

26 Journal of Forestry • January/February 2010

explained with additional analysis. This isespecially true when attempting to assess cu-mulative effects of multiple projects across aplanning area. Those opposed to a particularfuel treatment project can appeal the plan onthe basis of insufficient evaluation of eitherthe alternatives or the associated impacts ofthe alternatives (Germain et al. 2001).which in many cases can never be fullyknown until the project is implemented.When appeals are denied, litigation can en-sue, significantly delaying or even stoppingfuel treatment projects entirely. The puhliccan comment on the alternatives proposed.However, most projects generate few com-ments from the general public; the majorityof comments received come from engagedinterest groups who either support or opposethe project. The Healthy Forest RestorationAct (Healthy Forest Restoration Act[HFRA] 2003) was intended to streamlineand expedite the planning and review pro

cess for fuel treatment projects (Stephensand Ruth 2005). However, it is unclear towhat extent the intended effects of HFR \have actually been realized.

Despite the sound conceptual undcu-pinning of strategic fuel treatments (Finney2001), there is uncertainty regarding effec-tive implementation. Specifically, it is un-clear given current planning and operationalconstraints whether managers could succes.-fully treat the amount of area recommendedby fire modeling studies to appreciably mod-ify fire behavior and effects over large land-scapes (Finney 2001, Finney et al. 2007).The constraints on planning and operationinclude NEPA planning regulations, costs,lack of infrastructure (locally available con-tractors, smaliwood processing facilities, andbiomass-based power plants), and land des-ignation restrictions. It is also unclear howlandscape-level fuel treatments affect wild-life and water resources, both positively andnegatively. Although there is general con-sensus that reducing fire severity is desired,there is no quantitative landscape data show-ing the positive and negative effects of thesetreatments on wildlife and water resources.

FundingFunding is still another limitation on

landscape fuel treatment planning and im-plementation. The expected revenue or bud-get for a given project influences decisionson the type of fuel treatment that is con-ducted, as well as the size and placement oftreatment units. For prescribed burning,costs can vary tremendously depending on

,1? pF

I,.

F

1r

JJ

factors such as treatment size, accessibility oftreatment units, and complexity of terrain.In general, the costs per unit area decreasewith increasing unit size because of the fixedcosts of planning, staffing, and acquiringnecessary resources for a prescribed burn(Hartsough et al. 2008). Given a limitedbudget and relatively high NEPA planningcosts for small burn units (Hartsough et al.2008), there may be tendencies to planfewer, larger prescribed burns and imple-ment them over a long period. This may re-sult in fewer disparate large prescribed burnsbeing performed than might otherwise be

r

desired for optimal reduction of fire spreadand behavior at a landscape level.

The costs/gains of thinnings dependnot only on the size, accessibility, and terrainof treatment units, but also on the amountof merchantable timber harvested and itscurrent market value, the amount and mar-ket value of biomass removed, and the treat-ment of activity fuels (Hartsough et al.2008). In the fire-excluded forests through-out the western United States many of thetrees to he removed in fuel reduction opera-tions arc below 15-in. dbh. In some cases therevenue from harvesting merchantable tim-

iLand allocations

Protected habitat

Limited activity habitat area NStream buffer zone

0 5 ,000Ofibase/deferred '

All other lands Meters

Figure 1. An example of area constraints on fuel treatment planning from the MeadowValley area, Plumas National Forest, California. These are the actual land allocation layersprovided by Plumas NF personnel. The protected habitat is for the California spotted owl(Strix occidenta!is occidentalis).

Journal ofForestry January/ February 2010 27

':-AWI

P

r If

, lvI

I I Mastication

0 ctc[, I Thin

Figure 2. Conditional bum probability estimates for the Last Chance Project, Forest Hill RangerDistrict, Tahoe National Forest, California. These estimates were generated for pretreatment (upper)and modeled posttreatment (lower) landscapes using FlamMap and are based on 1,000 randomignitions. The treatments are planned as part of the Sierra Nevada Adaptive Management Project.This project is a collaborative effort between land managers, researchers, and interested groups andis designed to explicitly explore the effects of coordinated landscape fuel treatments.

Planned treatmentsConditionalburn probability

0.83A

0 3,000A If

I I

___ I I Underburn

her can pay for the service contracts nccddto remove biomass and treat activity fuk.However, recently, funding such fuel redu-tion activities in Sierra Nevada forests withrevenues generated from removing and sc]I -ing merchantable timber has been callcdinto question by the US Court of App:for the Ninth Circuit (Noonan 2008). l I1H

decision on whether fuel treatments shoiilJbe subsidized or can be allowed to pay hrthemselves will impact the extent and ratc ol

implementation of fuels projects.

ApproachesGiven the decisions that must be made

and the constraints on decision space whenplanning landscape fuel treatments, manag-ers attempting to reduce fire spread and ef-fects across a landscape are often chargedwith a difficult task. The following subsec-tions provide both modeling and real-worldapproaches that can help in dealing with thiscomplexity. The ideas presented are targetedtoward management of large, contiguoustracts of forest where mitigation of unchar-acteristic wildland fire effects is a dominantmanagement priority. As such, these ideasmay be more applicable to federal land man-agement; however, some principles can beapplied to state and private forests as well.

ModelingArcFuels (Ager et al. 2006) allows man-

agers to explore the effects of numerous fueltreatment alternatives and scenarios. BecauseArcFuels incorporates the Forest VegetationSimulator (FVS; Dixon 2002) and the Fireand Fuels Extension (FFE; Reinhardt andCrookston 2003), it can simulate a wide arrayof prescribed burning, thinning, and combina-tion treatments for any number of standsacross a planning landscape. This tool allowsusers to tailor simulated treatment type, inten-sity, size, and placement/pattern relatively pre-cisely to address the constraints imposed For agiven landscape. Additionally, incorporatingFVS allows for modeling the recovery/devel-opment of forests after treatments. [1]

ArcFuels can use LANDFIRE data(Landflre 2009) and forest inventory data(e.g., from Forest Inventory and Analysisplots), along with a map of individual foreststands, as inputs to then develop the specificfiles needed to run FARSITE and FlamMap.One of the main strengths of this tool is tocompare modeled fire behavior and effectsestimates for various alternatives of the"treated" landscape (i.e., after running userassigned treatments to stands) to those for

r

28 Journal of Forestry January/ February 2010

the untreated landscape, as well as compareamong treatment alternatives (Stratton2006,Ageretal. 2007a, 2007b, Finney etal.2007). This would allow for simulating sce-narios such as placing fuel treatments adja-cent to protected habitat (e.g., Ager et al.2007a) or placing fuel treatments whollywithin protected habitat, and then compar-ing such scenarios to one with no fuel treat-ments to examine potential fire behavior andeffects across a given landscape. Similarly,one could evaluate the impact of includingor not including various land ownerships ina landscape fuel treatment planning. Thisinformation on various fuel treatment alter-natives is suited for the planning process thatis mandated for management of federal andsome state agencies (e.g., NEPA process).Establishing a substantial reduction in po-tential fire behavior and effects across a land-scape for a chosen alternative can serve asdefensible justification for performing fuelreduction treatments that may be neededwhen plans are questioned.

In addition to providing fire behaviorestimates, the Minimum Travel Time fea-ture within FlamMap can yield conditionalburn probability estimates, as well as opti-mized locations for fuel treatments. Condi-tional burn probability estimates involvesimulating a number of randomly ignitedfires across a given landscape and can helpidentify areas most likely to burn if large firesoccur (Finney 2006; Figure 2). The Treat-ment Optimization Model (TOM) withinFlamMap identifies areas to treat, such thatfire spread across a landscape (given a user-defined ignition and wind direction) is min-imized (Finney 2004, 2006, 2007). Usingthe TOM requires a defined set of posttreat-ment fuel conditions for all areas where fueltreatments are possible, which can helpmanagers address some of the fuel treatmentplacement constraints discussed earlier (e.g.,not within PACs or WUI communities) andyield more realistic and useable outputs.Both of these tools within FlamMap can alsohelp managers choose stands or areas andjustify their selection for landscape fueltreatment planning (Figure 2).

The major limitation of the FlamMapand ArcFuels approaches for aiding manag-ers in designing landscape fuel treatments,aside from the inherent model assumptionsand limitations explained by Stratton(2006), is the expertise not only to gatherand manipulate the necessary spatial data forthese tools, but to execute the models andinterpret the results. The Stewardship and

Fireshed Assessment process currently un-derway in US Forest Service Region 5 is ad-dressing this by bringing this expertise to in-dividual forests, and even districts, in theform of interactive workshops that use localdata to generate usable results (Bahro et al.2007). This approach begins with the iden-tification of a "problem" fire. The problemfire is one that has the greatest potential im-pact on human and natural resources basedon historical weather patterns and terrain.Often, this fire is an actual wildland fire thatoccurred in the past under weather condi-tions that rendered suppression actions inef-fective. Based on the problem fire the"fireshed" is delineated (25,000-100,000ac) such that it includes areas with similarfire regimes, fire history, and wildland firerisk issues (Ager et al. 2006). The firesheddefines the scale at which fires and fuel treat-ments are considered and can be viewed asconceptually analogous to a watershed. Us-ing the weather conditions, location, size,and effects of the problem fire, a treatmentpattern is developed such that modeled firebehavior and anticipated effects are miti-gated.

Assuming the expertise exists, this mod-eling framework used to meet immediatemanagement needs can also be used to in-form longer-term planning. For example,ArcFuels could be used to model forestgrowth in the future for both treated anduntreated stands across a landscape. Subse-quent analysis with FlamMap would pro-vide insight into the lifecyle of landscape-level fuel treatments [2]. Although muchwork remains, these spatially and temporallyexplicit fire behavior models represent themost promising way to assess the longevityand effectiveness of fuel treatments. Such in-formation is essential to the development ofa comprehensive, long-term forest manage-ment strategy.

Real WorldOutside of the modeling environment

there are management decisions that can ad-dress some of the previously listed con-straints. Planning to conduct prescribed firetreatments across multiple watersheds con-currently may be an approach for mitigatinglocal air quality impacts. This would allowmanagers to incrementally implement treat-ments over a period of several years by rotat-ing burning between various planning areas.This approach would effectively spread smokeimpacts on affected communities out overtime, and limit the potential for inundating a

particular community with heavy smoke. Asfor funding constraints, particularly regardingfuel treatments in more remote and inaccessi-ble areas, prescribed burning and managedwildland fire may be the most appropriate al-ternatives (Mills 2006, Collins et al. 2007,Collins and Stephens 2007). When prescribedburning in these areas, larger burns may benecessary to use landscape features as naturalfuel breaks (rocky ridges, rivers, talus slopes,and more), thus keeping costs of fireline con-struction to a minimum.

Adaptive resource management is stillanother approach for managing landscapeswhere uncertainty exists in the response of aparticular resource to management actions.It is a transparent process that involves test-ing, monitoring, and evaluating appliedstrategies and incorporating new knowledgeinto management approaches that are basedon scientific findings and the needs of soci-ety. This process requires long-term com-mitments from scientists, stakeholders, andmanagers, which ultimately leads to moresound and informed decisionmaking. Werecognize that outputs from adaptive man-agement projects can be subject to changingeconomic and political priorities, and com-prehensive results can take a decade orlonger to generate. However, faced with theuncertainty of natural systems (Millar et al.2007) and the current scale of fire hazardthroughout the western United States, it isone of the few ways to move forward.

ConclusionThe impacts of extensive and severe

fires in drier mid- to low-elevation forests arelargely detrimental to the local ecosystem(Schoennagel et al. 2004, Keane et at. 2008).As the occurrence of larger, higher-severityfires is increasing in some areas of the west-ern United States (Westerling et al. 2006,Miller eral. 2009) the need to mitigate thesepotential impacts is growing. We point out anumber of challenges facing managers whenattempting to plan and implement coordi-nated fuel treatments across forested land-scapes. These challenges do not make thetask of managing landscapes to preventlarge-scale change from fire impossible; theysimply mean we have to accept some uncer-tainty and resulting imperfection in imple-menting treatments, especially in initial ef-forts. The modeling approaches wesummarize provide meaningful compari-sons among treatment options/scenariosand can considerably improve the process ofplanning coordinated landscape fuel treat-

Journal ofForestry January/February 2010 29

ments. However, output from modeling is

inherently subject to a certain number of as-

sumptions and ideas put forth by the mod-

eler and is no substitute for learning from

actual treatments. Implementing landscape

fuel treatments, even based on imperfect

knowledge, and improving subsequent ap-

plications, will likely be a better alternative

than the "no-action" alternative that contin-

ues to leave vast areas of forest susceptible to

high-severity fire (Agee 2002).

EndnotesUi It is important that users of FVS/FFE evalu-

ate and adjust, where necessary, fuel modelselection based on knowledge local to thestudy site. If possible, users should calibratefuel model selection with observed behaviorand effects of an actual wildfire within ornear the study area. Without such evalua-tion/adjustment results output from FVSIFFE can be suspect.

[21 A recent study (Vaillant 2008) showed theefficacy of such an approach.

[31 If whole trees are removed, the potential ex-ists for some cost recovery by chipping smalltrees and transporting the chips to bioenergyplants. Depending on location of cogenera-tion infrastructure, as well as the status of aparticular plants biomass supply, the cost ofchipping and subsequent transport can ex-ceed that of generated revenues. It is thisavailability of infrastructure that often deter-mines the feasibility of biomass projects.However, when considering the alternativeof treating activity fuels either by pile burn-ing or by broadcast burning, which can befairly costly (Hartsough et al. 2008), the po-tential losses associated with bioenergv useonly need to be less than the costs of treatingactivity fuels to make such efforts worth-while.

Literature CitedAGEE, J.K. 2002. The fallacy of passive manage-

ment: Managing for firesafe forest reserves.Conserv. Biol. Pract, 3(l):18-25.

ACER, J.K., B. BAHRO, M.A. FINNEY, P.N. Oi,D.B. SA psis, C.N. SKINNER, J.W. VAN WAG-TENDONK, AND C.P. WEATHERSI'OON. 2000.The use of shaded fuelbreaks in landscape firemanagement. For. Ecol. Manag. l27(1-3):55-66.

ACRE, J.K., AND C.N. SKINNER. 2005. Basic prin-ciples of forest fuel reduction treatments. For.Ecol. Manag. 211(1-2):83-96.

ACER, A.A., B. BAHRO, AND K. BARBER. 2006.Automating the Fireshed Assessment Processwith ArcGIS. P. 163-168 in l'roc, of couf onFuels management-How to measure success,Andrews, P.L., and B.W. Butler (eds.). USFor. Serv. RMRS-P-4 1.

ACER, A.A., M.A. FINNEY, B.K. KERNS, AND H.MAFFEI. 2007a. Modeling wildfire risk tonorthern spotted owl (Strix occidentalis cau-

rina) habitat in Central Oregon, USA. For.Ecol. Manag. 246(1):45-56.

AGEIt, A.A., A.J. MCMAHAN, J .J . BARRETT, ANDC.W. McHu;si. 2007b. A simulation study ofthinning and fuel treatments on a wildland-urban interface in eastern Oregon, USA.Land/c. Urban Plan. 80(3):292-300.

BAHRO, B., K.H. BARBER, J.W. SHERLOCK, ANDD.A. YASUDA. 2007. Stewardship and fireshedassessment: A process for designing a landscapefuel treatment strategy. P. 41-54 in Proc. ofconf on Restoring fire-adapted ecosystems. 2005National Silviculture Workshop, Powers, R. F.(ed.). US For. Serv. Gen. Tech. Rep. PSW-GTR-203.

COLLINS, B.M., M. KELLY, J.W. VAN WAGTEN-DONK, AND S.L. STEPHENS. 2007. Spatial pat-terns of large natural fires in Sierra Nevada wil-derness area. Landsc. Ecol. 22:545-557.

COLLINS, B.M., AND S.L. STEPHENS. 2007. Man-aging natural wildfires in Sierra Nevada wil-derness areas. Front. Ecol. Environ. 5(10):523-527.

COLLINS, B.M., J.D. MILLER, A.E. THODE, M.KElLY, J.W. VAN WAGTENDONK, AND S.L. 5cR-PHENS. 2009. Interactions among wildlandfires in a long-established Sierra Nevada natu-ral fire area. Ecosystems 12(l):114-128.

DIXON, G.E. 2002. Essen tialFVS: A user c guide tothe Forest Vegetation Simulator. US For. Serv.,Rocky Mtn. Res. Sm. (revised April 2007).209 p.

FIEDLER, C.E., C.E. KEEGAN, C.W. WooDAu.,AND T.A. MORGAN. 2004. A strategic assessmentofcrown fire haza rd in Montana: Potential effec-tiveness and costs ofhazard reduction treatments.US For. Serv., Pacific Northwest Res. Stn.,PNW-GTR-622. 48 p.

FINNEY, M.A. 1998. FARSITE: Fire area simula-tor-Model development and evaluation. USFor. Sen'., Rocky Mtn. Res. Stn. RMRS-RP-4.47 p.

FINNEY, M.A. 2001. Design of regular landscapefuel treatment patterns for modifying firegrowth and behavior. For. Sci. 47(2):219-228.

FINNEY, M.A. 2004. Landscape fire simulationand fuel treatment optimization. P. 117-131in Methods for integrated modeling of landscapechange, Hayes, J.L., A.A. Ager, and J.R. Bar-bour (eds.). US For. Scrv., Pacific NorthwestRes. Stn., Portland, OR.

FINNEY, M.A. 2006. An overview of FlamMapmodeling capabilities. P. 213-220 in Proc. ofconf on Fuels management-How to measuresuccess, Andrews, P.L., and B.W. Butler (eds.).US For. Serv. RMRS-P41.

FINNEY, M.A. 2007. A computational method foroptimising fuel treatment locations. mt. J.Wild/and Fire 16(6):702-711.

FINNEY, MA., R.C. SELIA, C.W. MCHUGH, A.A.A(-,ER, B. B,.\HRO, AND J.K. ACER. 2007. Simu-lation of long-term landscape-level fuel treat-ment effects on large wildfires. intl. Wild/andFire 16(6):712-727.

FULE, P.Z., C.W. McHUc;K, T.A. HEINLEIN, ANDW.W. COVINGTON. 2001. Potential fire be-havior is reduced following forest restorationtreatments. P. 28-35 in Proc. of conf on Pon-

derosa pine ecosystems restoration and conserva-tion: Steps towards stewardship, Vance, R.K.,C.B. Edrninster, W.W. Covington, and J.A.Blake (eds.). US For. Serv. RMRS-22.

GERMAIN, RH., D.W. FLOYD, AND S.V. STEH-MAN. 2001. Public perceptions of the USDAForest Service public participation process.For. Policy Econ. 3(3-4): 113-124.

HARTSOUGFI, BR., S. At.&ias, R.J. BARBOUR,E.S. DRLws, J.D. MCIVER, J .J . MOGHADDAS,D.W. SCHWILK, AND S.L. STEPHENS. 2008.The economics of alternative fuel reductiontreatments in western United States dry for-ests: Financial and policy implications fromthe National Fire and Fire Surrogate Study.For. Policy Econ. 10:344-354.

HEALTHY FOItuSc RESIORATION Acr (HFRA).2003. Healthy Forest Restoration Act [HR.1904]. Public Law 108-148.

HESSBURG, P.F., J.K. ACRE, AND J.F. 2005.Franklin. Dry forests and wildland fires of theinland Northwest USA: Contrasting the land-scape ecology of the pre-settlement and mo-dem eras. For. Lcd. Manag. 211(1-2):117-139.

KEANF., RE., J.K. AGEE, P.Z. FULE, J.E. KEELEY,C.H. KEY, S.G. KITCHEN, R. MILLER, AND L.A.SCHULTE. 2008. Ecological effects of large fireson US landscapes: Benefit or catastrophe. list.I Wild/and Fire 17(6):696-712.

KEIFER, M., J.W. VAN WAGTENDONK, AND M.BUHLER. 2006. Long-term surface fuel accu-mulation in burned and unburned mixed-co-nifer forests of the central and southern SierraNevada, CA (USA). Fire Ecol. 2:53-72.

KNAPP, E.E., J.E. KEELEY, E.A. BALI.ENGER, ANDT.J. BRENNAN. 2005. Fuel reduction andcoarse woody debris dynamics with early sea-son and late season prescribed fire in a SierraNevada mixed conifer forest. For. Ecol. Manag.2080-3083-397.

LANDFIRE. 2009. landscape fire and resourcemanagement planning tools project. Availableonline at www.landfire.gov; last accessed Oct.15, 2009.

M.ARTINSON, E.J .,AND P.N. OMI. 2002. Perfor-mance of fuel treatments subjected to wild-fires. P. 7-13 in Proc. ofconf on Fire, fuel treat-ments, and ecological restoration, Omi, P.N.,and L.A. Joyce (eds.). US For. Serv. RMRS-P-29.

MCKEIVF.Y, KS., C.N. SKINNER, C. CHANG,D.C. EIu1AN, S.J. HUSARI, D.J. PARSONS,J.W.VAN WAGTENDONK, AND C.P. WEATHER-SPOON. 1996. An overview of fire in the SierraNevada. P. 1033-1040 in Sierra Nevada Eco-system Project: Final Report to Congress, Vol. II:Assessments and scientific basis for managementoptions. Wildland Resource Center Rep. 37,Univ. of California, Davis, CA.

MILLAR, CI., N.L. SIEm'FIENsoN, AND S.L. STE-PHENS. 2007. Climate change and forests of thefuture: Managing in the face of uncertainty.Ecol. Applic. 17:2145-2151.

MILLER, J.[)., H. D. SAFFORD, M. CRIMMINS, ANDA. E. TI lODE. 2009. Quantitative evidence forincreasing forest fire severity in the Sierra Ne-vada and southern Cascade Mountains, Cali-

30 Journal ofForesny • January/February 2010

fornia and Nevada, USA. Ecosystems12(l):16-32.

MILLS, D.P. 2006. Wildland fire use success sto-ries. Fire Manag. Today 66(4):16-19.

MOGI-IADDAS, J .J . , AND L. Cit.s;s. 2007. A fueltreatment reduces fire severity and increasessuppression efficiency in a mixed conifer for-est. Intl WildlandFire 16:673-678.

NOONAN,J.T. 2008. U.S. court ofAppeals for theNinth Circuit, Opinion on Sierra Forest Legacyv. Rey. No. 07-16892.

NORTH, M., J . INNES, AND H. ZAIn. 2007. Com-parison of thinning and prescribed fire restora-tion treatments to Sierran mixed-conifer his-toric conditions Can.]. For. Res. 37(2):331-342.

OND, P.N. 1996. The role of fuelbreaks. P.89-96 in Proc. ofconf on 17th Forest vegetationmanagement conference, Cooper, S.L. (comp.).University of California, Shasta County Coop-erative Extension, Redding, CA.

RAYMOND, CL., AND D.L. PETERSON. 2005. Fueltreatments alter the effects of wildfire in amixed-evergreen forest, Oregon, USA. Can. I.For. Res. 35(12):2981-2995.

REINITARDT, E., AND N.L. CROOKSTON. 2003.The fire and fuels extension to the Forest Vegeta-tion Simulator. US For. Serv., Rocky Mtn. Res.Stn. RMRS-GTR-116. 209 p.

RITCHIE, MW., C.N. SKINNER, AND T.A. HAM-ILTON. 2007. Probability of tree survival afterwildfire in an interior pine forest of northernCalifornia: Effects of thinning and prescribedfire. For. Ecol. Manag. 247(1-3):200-208.

SCHMIDT, D.A., A.H. TAYLOR, AND C.N. SKIN-NER. The influence of fuels treatment andlandscape arrangement on simulated fire be-havior, Southern Cascade range, California.For. Ecol. Manag. 255(8-9):3170-3184,2008.

SCHOENNAGEL, T., T.T. VEBLEN, AND W.H.ROMME. 2004. The interaction of fire, fuels,and climate across Rocky Mountain forests.Bioscience 54(7):661-676.

SKINNER, C.N. 2005. Reintroducing fire into theBlacks Mountain Research Natural Area: Ef-fects on fire hazard. 1'. 245-257 in Proc. ofconfon symp. on Ponderosa pine: Issues, trends, andmanagement, Ritchie, M.W., D.A. Maguire,and A. Youngblood (eds.). US For. Sm. Gen.Tech. Rep. PSW-GTR-198.

SKINNER, C.N., M.W. RITCHIE, I. HAMILTON,AND J . SYMONS. 2004. Effects of prescribed fireand thinning on wildfire severity: The ConeFire, Blacks Mountain Experimental Forest. P.80-91 in Proc. of conf on 25th VegetationManagement Conj., Cooper, S. (ed.). Cooper,S.L. (comp.). University of California, ShastaCounty Cooperative Extension, Redding, CA.

SOLIS, D.M., AND R.J. GUTIERREZ. 1990. Sum-mer habitat ecology of the Northern SpottedOwls in northwestern California. Condor92(3):739-748.

SPENCER, W.D., H.L. RUSTIGAN, R.M.SCHELLER, A. SYI'HARD, J . STRITTHOLD, AND B.WARD. 2008. Baseline evaluation offisher hab-itat and population status, and effects offires andfuels management on fishers in the southern Si-erra Nevada. Unpubl. Rep. prepared by Con-servation Biology Institute for US For. Serv.,Pacific Southwest Region. 133 p.

SPIES, T.A., M.A. FIEMSTROM, A. YOUNGBLOOD,AND S. HuaKIEI.. 2006. Conserving old-growth forest diversity in disturbance-pronelandscapes. conserv. Bid. 20(2) :351-362.

SFEI'HENs, S.L., AND P.Z. FULE. 2005. Westernpine forests with continuing frequent fire re-gimes: Possible reference sites for manage-ment.J For. 103(7):357-362.

STEI'HENS, S.L., AND J .J . MOGHADDAS. Experi-mental fuel treatment impacts on forest struc-

ture, potential fire behavior, and predicted treemortality in a California mixed conifer forest.For. Ecol. Manag. 215(1-3):21-36.

STEPHENS, S.L., AND L.W. RUTH. 2005. Federalforest-fire policy in the United States. Ecol, Ap-plic. 15(2):532-542.

STEPHENS, S.L., J j. MOGHADDAS, C. EDIMIN-STER, C.E. FIEDLER, S. I-IASSE, M. HAR-RINCTON, J.E. KEELEY, E.E. KNAPP, J.D.MCIVER, K. METLEN, C.N. SKINNER, AND A.YOUNGBLOOD. 2009. Fire treatment effects onvegetation structure, fuels, and potential fireseverity in western U.S. forests. Ecol. Applic.19(2):305-320.

SmAnoN, R.D. 2004. Assessing the effectivnessof landscape fuel treatments on fire growth andbehavior. I. For. 102:32-40.

STRATTON, R.D. 2006. Guidance on spatial wild-land fire analysis: Models, tools, and tech-niques. US For. Serv., Rocky Mtn. Res. Stn.,RMRS-GTR-183. 15 p.

STROM, BA., AND P.Z. FULE. 2007. Pre-wildfirefuel treatments affect long-term ponderosapine forest dynamics. mt. i Wildland Fire16(l):128-138.

VAILLANT, N.M. 2008. Sage/sen ExperimentalFo rest past, present, andfuture: An evaluation ofthe fireshed assessment process. Dissertation,Univ. of California, Berkeley, CA. 172 p.

WEATHERSPOON, CI'., AND C.N. SKINNER. 1996.Landscape-level strategies for forest fuel man-agement. P. 1471-1492 in Sierra Nevada Eco-system Project: Final Report to congress, Vol. Ii:Assessments and scientific basis for managementoptions. Wildland Resource Center Rep. 37,Univ. of California, Davis, CA.

WESTERLING, A. L., H. HIDALGO, D. CAYAN,AND T.W. SWETNAM. 2006. Warming and ear-lier spring increase western US forest wildfireactivity. Science 313(5789):940 -943.

II.

Journal ofForestsy • Januaiy/Februarv 2010 31