The Consequences of Delay

The Biscuit Fire began July 13,2002. During the next 54days, it burned approximately

400,000 acres on the Siskiyou Na-tional Forest (fig. 1). The Biscuit Firewas the nation’s most expensive firesuppression effort of 2002, reportedlycosting $154 million in federal and

state funds. Burned were congression-ally reserved lands (152,900 acresmostly within the Kalmiopsis Wilder-ness Area), administratively with-drawn lands (64,100 acres), and underthe Northwest Forest Plan, late-succes-sional reserves (133,700 acres) andmatrix lands (33,000 acres).

The Biscuit Fire poses a policy ques-tion regarding areas designated to func-tion as mature and old-growth forests.After intense fires, are managers to letvegetation develop naturally, or arethey to invest in regeneration of foreststo achieve the intended late succes-sional status quickly? The NorthwestForest Plan and the spotted owl finaldraft recovery plan indicate that man-agement interventions are encouragedif natural vegetative recovery will notproduce desired habitat conditions(USDI 1992; USDA/USDI 1994;Thomas 2003).

To understand the opportunities to

38 Journal of Forestry • April/May 2004

John Sessions, Pete Bettinger, Robert Buckman, Mike Newton, and Jeff Hamann

Over 54 days in 2002, the Biscuit Fire, the largest fire in recorded Oregon history, burned morethan 400,000 acres. Much of the burned land was being managed under the federal NorthwestForest Plan to provide habitat for species that live in complex, older conifer-dominated forestsas well as for recreation purposes. Only a narrow window of opportunity exists to hastenconifer restoration to complex forest conditions in a cost-effective manner, to reduce risks ofinsect epidemics and future fires, and to capture some economic value that could offsetrestoration costs. Delays in decisionmaking and implementation will likely destine much of themost intensely burned area to cycles of shrubs, hardwoods, and recurring fires for manydecades. This is the opposite of what current management plans call for—maintenance of mature forests.

Keywords: biodiversity; forest health; Northwest Forest Plan; old-growth; policy; restoration

AB

ST

RA

CT

The Consequences of Delay

Above: In July 2002, the Biscuit Fire burnedthrough the 1987 Silver Burn. Note the snags igniting ahead of the fire front.Photo by Tom Link, Siskiyou National Forest

Hastening the Return of

Complex Forests Following Fire

Hastening the Return of

Complex Forests Following Fire

hasten regrowth of structurally com-plex conifer-dominated forests, we ex-amine prefire forest conditions, whatexists now in the aftermath of the fire,and the likely consequences and trade-offs of natural ecosystem recovery ver-sus management interventions. Thisexamination of forest policy wasprompted by Douglas County com-missioners who asked Oregon StateUniversity to examine costs of manage-ment delay following the large fires insouthwest Oregon during 2002.

Landscape ConditionThe area of the Biscuit Fire is a geo-

logic patchwork, characterized byrough terrain, deeply incised valleys,and geologic strata that vary in resis-tance to erosion. Soils in the burnedarea can be loosely grouped into twocategories, those derived from serpen-tine base rock and those that are not.The serpentine-derived soils are themost erosion prone. They are high inmagnesium and low in calcium, andtheir water-holding capacity is low.Vegetation on these soils is sparse, andthey are low productivity for coniferforests. About 25 percent of the soils inthe Biscuit Fire perimeter are serpen-tine-derived. The remainder are de-rived primarily from sandstones andschists; their productivity for coniferforests is low to medium.

Within the Biscuit Fire, mature for-ests (>100 years) comprised Douglas-fir (Pseudotsuga menziesii), white fir(Abies concolor), tanoak (Lithocarpusdensiflorus), Jeffrey pine (Pinus jeffreyi),and Port-Orford-cedar (Chamaecyparislawsoniana), with scattered westernwhite pine (Pinus monticola), sugarpine (Pinus lambertiana), ponderosapine (Pinus ponderosa), madrone (Ar-butus menziesii), and other species. Fewstands reached great ages (>300 years)because of frequent fires, and most ofthe older stands were on north slopes,which do not burn readily. The prefiretimber volume within the fire perime-ter was estimated to have been about10 billion board feet, Scribner scale.

One-third of the burned area mayreturn to the desired forest conditionswith minimal human intervention be-cause it retains seed sources and ele-ments of structural complexity. But on

about 345,000 acres, according to theUSDA Forest Service (2003), therenow are more dead than live conifertrees (fig. 2, p. 40). At least 25 percentof the canopy was killed on more thantwo-thirds of the fire area.

The affected area includes about 20percent of the entire natural range ofPort-Orford-cedar and is also an im-portant area for sugar pine and westernwhite pine (Don Goheen, pers. com-mun., July 2, 2003). If naturallyseeded, these five-needle pines have lit-tle resistance to an exotic blister rust,and naturally seeded Port-Orford-cedars have almost no resistance toPort-Orford-cedar root disease, an-other exotic. These species can be suc-cessfully established on sites within the

Biscuit Fire area only if disease-resis-tant seedlings from improved stock areplanted.

Of the 300 wildlife species that meetat least part of their yearly needs on theRogue River–Siskiyou National Forests,three are listed as threatened or endan-gered: the bald eagle, the marbled mur-relet, and the northern spotted owl.Habitat for the owl is the most signifi-cant consideration, as about 25 percentof the 202 known spotted owl activitycenters on the Siskiyou are within thefire area. The fire transformed 23 of the40 “functional home ranges” to “non-functional” habitat and made 75,000 to80,000 acres of nesting habitat unsuit-able (USDA Forest Service 2003). Inaddition, some 460 miles of streams

39April/May 2004 • Journal of Forestry

Figure 1. The 2002 Biscuit Fire, encompassing 500,000 acres, affected critical habitat for the north-ern spotted owl, as well as other species . Source: USDA Forest Service (2003).

Siskiyou National ForestWildernessBiscuit Fire boundaryBLM landsOther ownership in Siskiyou NF

and rivers within the Biscuit Fire areacontain both resident and anadromoussalmonids, the most sensitive being thefederally listed coho salmon.

The choices facing forest managersand society can be summarized:

• Supplement natural ecologicalprocesses with investments to regener-ate conifer species, at stocking ratesand with supplemental managementactions that hasten return to complexforest structure, or let nature take itscourse.

• Salvage some fire-killed timber(outside the wilderness area) to helpachieve desired future conditions inspecies composition, diversity, and re-silience or resistance to fires, storms,insects, or weeds, or leave all dead anddying trees.

• Pay for necessary investments bysalvaging some commercially valuablewood from the area, and/or seek appro-priations from Congress.

• Use herbicides to regenerate thelargest number of acres with the great-est probability of achieving complexforest structures, or attempt mechani-cal control of vegetation competingwith conifers, or do not attempt tocontrol competing vegetation.

MethodsWe developed databases to help us

evaluate the economic, ecological, andsocial effects of regeneration, standmaintenance, insect infestation risk,heavy fuels, and salvage operations re-lated to potential management actions.

For prefire stand attributes, timber

volumes, and accessibility, we obtainedtree and plot data from the USDA For-est Service current vegetation survey.Within the study area, 210 randomlyselected plots were used to describe veg-etation types. A circular training area of2 hectares was created around each plotlocation. Landsat 5 Thematic Mapperimages from 1995 were classified byusing the sequential maximum a poste-riori estimation (Schowengerdt 1997)with the GRASS software package (USArmy 1993). The classified image,showing areas of similar vegetationcharacteristics, was then imported intoArc/Info for further processing.

Polygons were created from the clas-sified grid and intersected with otherpolygon databases for the Siskiyou Na-tional Forest, including land-use allo-cation and soils GIS layers. Areaswithin two miles of a road were delin-eated for potential timber salvage andregeneration cost analysis.

The postfire vegetation was derivedby overlaying a GIS layer of canopymortality developed by the Siskiyouforest using aerial photography inter-pretation (USDA Forest Service 2003).

Four site index values were assignedfor nonserpentine-derived soils basedon aspect. The wetter, west side wasprojected at higher site index values.High and low ranges were used to sim-ulate competing shrub conditionsbased on height and diameter growthresults from regeneration experiments(Newton and Cole 2004). King’s siteindex values were used for projectingthe basic 210 vegetation types.

We identified trees that were mostlikely killed by the fire and decayedthem over time, using relationshipsfrom Lowell et al. (1992) to determinethe amount of sound wood availablefor salvage within five years.

Additional mortality from stressedtrees believed to have high insect infes-tation risk was also estimated using theresults of a postfire survey (USDA For-est Service 2003). Risk to stressed treeswas assumed to be concentrated in the low-moderate to high-moderateburned areas. These trees were not re-moved from the tree lists for growthprojections.

The basic 210 vegetation types werethen expanded to account for the five

40 Journal of Forestry • April/May 2004

Figure 2. Canopy mortality within the Biscuit Fire perimeter. Source: USDA Forest Service (2003).

Biscuit Fire boundaryNon-Forest-Service land

< 10% (unburned)10–25% (light)25–50% (low and moderate)50–75% (high moderate)> 75% (heavy)Not applicable

41April/May 2004 • Journal of Forestry

levels of canopy mortalitythat determined the amountof direct fire-induced mor-tality and expanded again bythe four site index values, re-sulting in 4,200 stand types.The vegetation types wereprojected for 100 years withthe ORGANON DynamicLinked Library growth-and-yield model (Hann et al.1997).

To estimate the decay ofstanding fire-killed trees andtheir contribution to downwood over time, we used asnag-and-down-wood decaymodel (Mellen and Ager1998).

Regeneration. Historicallyin southwestern Oregon,large burned areas regeneratenaturally from seed. Whereseed is readily available and site condi-tions are conducive, natural stands ofDouglas-fir begin with more than1,000 seedlings per acre (Hermann andLavender 1990). However, Douglas-firseed crops occur at irregular intervals.On drier sites, restocking can takedecades, sometimes a century or more.Wetter areas can develop more rapidly,but shrub and hardwood competitioncan be severe.

To estimate regeneration costs, weassume that the goal is to establish 200young conifer trees “free to grow” oneach acre of future complex coniferforest. This is considerably belowstocking levels on private industrialforests (300 to 400+ seedlings peracre). Species composition would re-flect aspect and elevation. We assume

natural seeding will occur on nonser-pentine-derived soils that have not losta large part of their canopy. We also assume natural seeding as the regener-ation method on the very low produc-tivity serpentine soils.

The biggest problem facing coniferregeneration in the region is competingvegetation. With limited soil moisture,competition from woody and herba-ceous vegetation greatly reduces thesurvival and growth of conifers.Hughes et al. (1990), for example, in-dicate that shrub stands reach highcompetitive vigor earlier than coniferstands and that shrub species canrapidly occupy a site. Climate changemay also be a factor (e.g., Cromack etal. 2000). Current and likely future cli-mates may be more favorable to root-

sprouting shrubs than whenthe burned conifer forestsoriginated (Tom Atzet, pers.commun.).

Aerial seeding canquickly achieve reforestationof large roadless areas but isnot currently an option forthe Biscuit Fire area becauseof lack of approved bird androdent repellents. Seedingand planting are most effec-tive if done immediately(Zavitkovski et al. 1969,Tappeiner et al. 1992). Eventhe best nursery-grownstock is affected by competi-tion during the first fewyears after planting; initiallyhigh levels of herbaceousand shrub cover can increaseseedling mortality and re-duce growth (e.g., Harring-

ton and Tappeiner 1997). Tools to reduce shrub competition

on tens of thousands of acres are lim-ited. Prescribed fire cannot be usedwhen shrubs are too small to provideadequate fuel. Personnel and budgetsare generally inadequate to provide ef-fective manual or mechanical control,and success of these techniques is atbest limited. For each 10,000 acres,manual treatment would require 300to 500 person-years, is difficult work,and is hazardous (e.g., Stavins et al.1981).

Using results from the Forestry In-tensified Research (FIR) Program (acollaboration between Oregon StateUniversity and the Forest Service PacificNorthwest Research Station), Newtonand Lavender (unpublished) have esti-

Grasses sprouting from hay bales dropped by helicopter to prevent soilerosion will compete with regenerating conifers in the late-successionalreserve outside Onion Camp.

John

Ses

sion

s

Table 1. Estimated regeneration cost (dollars per acre) to successfully establish 200 conifer trees per acre considering initial cost, probability of success, and cost of restocking failures.

North slope South slope

Regeneration method 2004 2005 2006 2007 2004 2005 2006 2007

Plant plugs $ 250 $ 286 $ 667 $ 1,000 $ 333 $ 400 $ 1,000 $ 2,000Plant plugs + 1 259 293 367 733 314 367 550 2,200Plant plugs + chemical prep 335 335 357 383 383 383 447 536Plant plugs + 1 + chemical prep 320 339 360 360 360 384 443 443Plant plugs + 1 + chemical prep + chemical release 372 394 394 394 419 479 479 479

NOTES: Values displayed in bold show the most cost-effective method for year of establishment. Cost-effectiveness considers only tree survival, not thecosts of later controlling shrub competition.

Journal of Forestry • April/May 200442

mated the initial cost of re-generation options, the de-clining probability of successrelated to time, and the differ-ences of success on north-versus south-facing slopes.Federal costs are likely to behigher than those on privateforests, but the relationshipsbetween methods and delayof regeneration should remainvalid. Rogue River–SiskiyouNational Forest managers(USDA Forest Service 2003)estimate reforestation costs at$1,000 per acre, without useof herbicides. We estimatethat aggressive use of the mostpromising techniques, includ-ing planting only 200 treesper acre, could reduce thesecosts by up to two-thirds. Table 1 con-siders cost per acre to reestablish coniferforests by year of planting, includingprobability of failure and cost of re-stocking to achieve success by slope ori-entation. Cost-effectiveness depends onyear of planting—that is, on delay inestablishment.

Three things stand out from an ex-amination of regeneration costs: Themost cost-efficient method of estab-lishing conifers is immediate regenera-tion; planting delays beyond 2005 cansubstantially increase costs if weed con-trol is not adequate; and when delaysare unavoidable, herbicides for sitepreparation and release will dramati-cally reduce costs.

Stand maintenance. Once a site is re-forested, seedlings benefit tremen-dously from several years of vegetationmanagement, as illustrated by the re-forestation experiments by Newtonand Cole (2004). Even when plantingwas completed immediately after a fire,failure to control shrubs in the first twoyears afterward reduced tree growth by75 percent, substantially delaying at-tainment of tree sizes desired for late-successional wildlife habitat. These dif-ferences increase even more if hard-woods have one or more years to de-velop before planting.

Atzet et al. (1992) describes a short-term growth reduction of up to 45 per-cent from failure to control competi-tion. Longer-term experiments from

the FIR program showed that whenshrubs were completely removed, thegrowth rate of conifers saw a four-foldincrease over 23 years. If wildlife asso-ciated with late-successional forestsneed big trees for habitat, such findingshave important implications.

Insect infestation. Trees weakened byfire lose their ability to repel and sur-vive insect attacks. Bark beetles oftenkill many trees that might otherwisesurvive, and the resulting snags andfine fuels create high rates of firespread. Insect buildups can threatenlive trees in adjacent unburned forests,leading to even higher fuel loadings.The largest numbers of fire-stressedtrees are likely to be infested in the yearafter a fire.

To estimate the additional mortalityfrom fire-stressed trees, we used the re-sults of a postfire survey (USDA ForestService 2003) that provided probabilityof infestation by species and tree diam-eter. Risk to stressed trees was assumedto be concentrated in the low-moderateto high-moderate burned areas. Trees in unburned and lightly burned areaswere assumed not to be fire-stressed.

Fuel loads. To estimate the numberof snags per acre and down wood fromdead trees, we combined the ForestService photo-interpreted canopymortality estimates with our vegeta-tion overlay. The number of trees bydiameter class and species that werekilled by the Biscuit Fire was then

entered into the Mellenand Ager (1998) snag decaymodel and their conditionsprojected over time. Weadded to the fire-killedtrees those still-living treesthat would be expected todie naturally over the next100 years.

Salvage opportunities.Ground-based methods oftree harvest (rubber-tiredand tracked skidders) arenormally used where slopesare accessible and less than30 percent, and cable sys-tems (skylines) are used forsteeper slopes. Ground-based systems are not lim-ited in the distance they canoperate from roads, but

costs increase with distance. Cable sys-tems are limited to about one-half mileor less. At longer distances on steep ter-rain, helicopters must be used or addi-tional roads must be constructed. Theeconomically feasible transport distanceis limited by the value of the timber. Iftrees of mixed sizes and values are ex-tracted, the maximum economic dis-tance for helicopters is about two miles.If only the most valuable logs are re-moved, the maximum economic dis-tance increases. Helicopter logging, al-though expensive, permits immediatesalvage without additional roads andwith little soil disturbance. Helicopterlogging capacity in Oregon is sufficientto deliver more than 2 million boardfeet per day.

Improperly done, however, timbersalvage can contribute to increased sur-face runoff and soil erosion (see McIverand Starr 2001). Poorly located andimproperly constructed roads, im-proper choice of harvesting systems,and inadequate road maintenance cancontribute to erosion. Yet mineral soilis the ideal seedbed for establishingmost conifers. If adequately plannedand controlled, logging disturbancecan produce excellent results and maybe the best method of site preparationin shelterwood, selection, and partial-cut silvicultural systems where use offire or chemicals is limited (Cleary etal. 1978).

On slopes less than 30 percent, skid

A young plantation on private land (background, left) received vegetationmanagement using herbicides; the federal land near Gold Hill, Oregon, established following the 1994 Hull Mountain Fire has had mechanical release.

John

Ses

sion

s

trails could be limited toabout 10 percent of the areaand treated after use to con-trol runoff and erosion. Pre-liminary results from aneastern Oregon study ofcarefully planned salvage-logging operations withground-based machines areencouraging (USDA ForestService 2002). Althoughsome soil disturbance wasobserved after the opera-tions, little or no sedimentleft the harvest units, andthe reburn hazard was esti-mated to be either reducedor unchanged. Any increasedsediment yield would betemporary and dwarfed byincreases due to the recentfire and by natural geologicerosion in the Biscuit area.

Markets for salvaged tim-ber depend on quality, quan-tity, and price. Processing ca-pability in southern Oregonis approximately 2.75 billionboard feet per year (PaulEhinger, pers. commun.,2003). Additional process-ing centers exist in northernCalifornia. The actual effectof Biscuit salvage harvests onregional employment woulddepend on whether fire-killed timber is additive orsubstituted for green timber from otherforestlands. In the current forest prod-ucts market, substitution would proba-bly be more likely.

ResultsWith human assistance, we estimate

that large conifer trees (>18 inches di-ameter)—those that provide much ofthe character of a complex mature for-est and most of the habitat for old-growth-dependent wildlife—will take50 years or more to develop and sup-plement the surviving larger trees andup to 100 years to approach prefireconditions. Without planting and sub-sequent shrub control, however, itcould take more than 100 years to evenestablish conifer forests. This is wellbeyond the guidelines in the draft re-covery plan for the northern spotted

owl (USDI 1992) to provide futurelarge green conifers and future largesnags.

Wilderness areas and areas with ser-pentine soils aside, if the remainingareas with 25 percent or more canopymortality were considered for reforesta-tion, more than 137,000 acres wouldbe candidates for reforestation. Promptand successful regeneration on theseareas could produce nesting habitat forspotted owls within 80 years.

On average, the fire killed morethan 160 trees per acre. These trees willfall over time, and while providinghabitat for many species and slowly re-turning organic matter to soils, the de-bris could also fuel the intensity of fu-ture fires . Significant portions of deadand dying trees in a largely shrub andhardwood plant community will leave

the landscape susceptible tolarge, intense wildfires for atleast 60 years into the future,further jeopardizing the po-tential of remaining conifersand newly planted conifersto reach late-successionalconditions.

We estimate that the Bis-cuit Fire destroyed approxi-mately 4.2 billion board feet(conifer and hardwood), or40 percent of the prefire treevolume within the fireperimeter, and that the aver-age standing live tree vol-ume in the area (excludinglow-productivity serpentinesoils) has declined from26,000 to 14,000 board feetper acre. We believe thatfire-stressed conifers con-taining an additional 0.8billion board feet are at highrisk of insect attack in thenear future. Ongoing ForestService studies will refinethese estimates.

The recovery value offire-killed timber will de-crease as trees deterioratefrom checking, fungal decay,and woodborer activity.Based on data in Lowell etal. (1992), we estimate thatapproximately 22 percent ofthe fire-killed volume that

existed immediately after the fire waslost during the first year, and by thefifth year, only volume in the lowerlogs of the larger trees will have eco-nomic value. The economic loss due totimber deterioration is already in thetens of millions of dollars.

Access to fire-killed trees across theburn varies. In total, approximately 75percent of the fire-killed timber in thenonwilderness fire area is within twomiles of existing roads. Almost all thefire-killed timber volume within thematrix lands is accessible, as are about60 percent of the volume in adminis-tratively withdrawn areas and 80 per-cent in the late-successional reserves.

Depending on the scope and timingof a salvage program, we estimate that,at a stumpage value (mill value minuslogging costs) of $100 per thousand

43April/May 2004 • Journal of Forestry

Much of the burned area looks as this did in April 2003.

Road access exists in much of the matrix and some of the administrativelywithdrawn and late-successional reserves, making possible managementaction to regenerate the damaged areas.

John

Ses

sion

s Jo

hn S

essi

ons

board feet of salvage, 1 billion boardfeet of salvage by helicopter would re-cover $100 million of stumpage value.Salvage by ground-based and cable sys-tems would return higher stumpagevalues, but helicopters would minimizedisturbance to soils. One billion boardfeet of salvage would remove approxi-mately 50 percent of the total volumeof fire-killed trees from accessible areas,or about 40 percent of the combinedvolume of the fire-killed plus fire-stressed trees that are expected to die.

DiscussionChoices for management options

following the Biscuit Fire revolvearound societal goals for future forestsand the relative risks, benefits, andcosts of both action and inaction.Where society and managers choose to let nature deliver future landscapesand ecosystems, human-aided forestrestoration and timber salvage are notonly unnecessary but counterproduc-tive. If the goal is to let natural proc-esses dominate, there may be “no eco-logical need for immediate interven-tion on the postfire landscape”(Beschta et al. 1995). More recently,Everett (1995) and Ice and Beschta(1999) have provided differing per-spectives. Our understanding of forestmanagement, including the use of“light-on-the-land” harvesting meth-ods on ecologically sensitive sites, hasgreatly improved. Researchers havefound enormous variability in both the

effects of natural processes and theconsequences of human interventionon watersheds. For example, reviewingstudies with unlogged controls, McIverand Starr (2001) find that human in-tervention can reduce adverse water-shed impacts or be largely neutral aswell as aggravate impacts. Followingmonitoring on the 1987 Silver Firewithin the Biscuit Fire area, Kormeierand Park (1995) concluded that “thelack of adverse impacts from salvagelogging is attributed to protection of riparian areas, improved road construc-tion practices, and minimizing distur-bance through the use of helicopterlogging.”

If the goal is to hasten restoration ofcomplex mature conifer-dominatedforests on the Biscuit Fire landscape,careful timber salvage can be useful.Ecologically, it would allow full sun-light to reach young seedlings, reducefuture fuel loads, and reduce potentialadditional tree death from insect at-tack, all of which will hasten the re-growth and recovery of a complex for-est. Economically, it would generatesource of funds for forest restoration,reduce the costs of future fire suppres-sion, and make future helicopter stand-maintenance operations feasible. Tim-ber salvage could also provide a tempo-rary source of extra revenue for schoolsor other public services that have suf-fered under the current state budgetcrisis. Socially, timber salvage and sub-sequent forest regeneration would pro-

vide short-term local employment andenhance long-term recreational oppor-tunities.

The fixed and variable costs of har-vest are particularly important becauseof time constraints. As timber deterio-rates, there is a smaller economic baseover which to spread the fixed costs ofharvest. For low-impact, high-cost sys-tems such as helicopter logging, thewindow of opportunity for cost-effec-tive salvage closes quickly.

A Science-Based StrategyGiven the immense number of fire-

killed trees within the Biscuit Fire area,a location-specific strategy for timbersalvage would necessitate considerationof erosion control, including contourfelling; sensitive areas, including steepslopes and exposed soils; stream pro-tection; protection and reforestation ofcritical wildlife sites, especially riparianareas; log value, yarding distance, andmethod; loss in salvage value over time;regeneration activities and future ac-cess; dead wood for wildlife andstreams; probable fire-stressed treedeath from insect attack; potential firerisk from dead trees and down wood;future stand maintenance activities;and public involvement.

Alternative timber sale preparationprocedures could also be considered.Typical federal timber sale proceduresnow take up to two years. For live treetimber sales, this time investment re-flects the costs and benefits of the pro-posed actions. In timber salvage, how-ever, the costs of delay are extreme:Fire-killed trees will lose more than 40percent of their value in two years, anddelays in forest regeneration will in-crease costs (fig. 3). Alternatives such as“end-result contracting,” tested by theBureau of Land Management, offersignificant time savings. Marginal costtimber pricing to encourage salvage atlonger distances from roads could alsobe considered.

The Northwest Forest Plan at-tempts to protect and perpetuate ma-ture forests and their associated bio-logical and ecological diversity. Theoriginal intent of management on late-successional reserves was to reducestand density and clean up accumulat-ing fuels to decrease the risk of stand-

44 Journal of Forestry • April/May 2004

Figure 3. Average salvage value of fire-killed trees as a function of distance from road and year,using helicopter logging, and cost of reforestation.

Tree value ($/tree)

Distance from road

1 mile

Reforestation cost

2004 2005 2006 2007

1.5 miles

2 miles

Reforestation cost ($/acre)

replacing fires. Fires that threatenedthese reserves were to be given thehighest priority. Continued public re-sistance to planned management,however, has resulted in very littlemanagement, and the statutory andadministrative guidelines are largelysilent about regeneration after majordisturbances. Although the potentialbenefits of timber salvage in criticalowl habitat are identified and criteriafor salvage established, less attention isdirected toward criteria for hasteningforest regrowth through regenerationand control of competing vegetation.The plan anticipated a dynamic—incontrast to a static (or no action)—strategy on severely damaged areas,such as those affected by the BiscuitFire (Thomas 2003), but inaction hasbeen the norm.

ConclusionHow can agencies speed up consid-

eration of reforestation, salvage, andfire and insect hazard reduction withinthe Biscuit Fire area? Will the land—and the people affected by it—be bet-ter served by letting nature take itscourse or by making strategic invest-ments to influence the course of futureecosystems? Will society or forest man-agers accept the consequences of inac-tion versus action for future forests?

Time is not neutral. If society orland managers choose not to expeditepostfire decisionmaking for theroughly 200,000 acres outside the des-ignated wilderness so that restorationaction can begin in 2004 and end by2006 or 2007, then nature alone willdetermine the future conditions in asmuch as 400,000 acres of the entireBiscuit area. Regardless of congres-sional or administrative intent, theseforests will likely be dominated by cy-cles of shrubs, hardwoods, and fires fora long time.

The Biscuit Fire is not unique. Soci-ety faces similar choices after virtuallyevery large burn in dry, fire-prone for-ests with high accumulations of fuels.The recent 91,000-acre Booth andBear Fire on the Deschutes NationalForest proves that point again in Ore-gon. The consequences of delay in has-tening forest regrowth are large, impor-tant, and real.

Literature CitedATZET, T.D., D.L. WHEELER, B. SMITH, J. FRANKLIN,

and D. THORNBURGH. 1992. Vegetation. Chapter 5in Reforestation practices in southwestern Oregon andnorthern California, eds. S.D. Hobbs, S.D. Tesch,P.W. Owston, R.E. Stewart, J.C. Tappeiner II, andG.E. Wells, 92–113. Corvallis: Oregon State Univer-sity, Forest Research Laboratory.

BESCHTA, R., C. FRISSELL, R. GRESSWELL, R. HAUER, J.KARR, G. MINSHALL, D. PERRY, and J. RHODES.1995. Wildfire and salvage logging: Recommendationsfor ecologically sound post-fire salvage management andother post-fire treatments on federal lands in the West.Report to Pacific Rivers Council. Available online atwww.pacrivers.org, accessed by authors March 2003.

CLEARY, B., R. GREAVES, and R. HERMANN. 1978. Re-generating Oregon’s forests: A guide for the regenerationforester. Corvallis: Oregon State University.

CROMACK, K., JR., J. LANDSBERG, R. EVERETT, R. ZE-LENY, C. GIARDINA, E. STRAND, T. ANDERSON, R.AVERILL, AND R. SMYRSKI. 2000. Assessing the im-pacts of severe fire on forest ecosystem recovery. Jour-nal of Sustainable Forestry 11(1/2):177–228.

EVERETT, R. 1995. Review of Beschta et al. (1995) doc-ument. Letter dated August 17, 1995 to John Lowe,Pacific Northwest Regional Forester, USDA ForestService, Portland, Oregon.

HANN, D., A. HESTER, and C. OLSEN. 1997. ORGANON:Southwest Oregon growth and yield model user manual,Edition 6.0. Corvallis: Oregon State University, De-partment of Forest Resources.

HARRINGTON, T., and J. TAPPEINER II. 1997. Growth re-sponses of young Douglas-fir and tanoak 11 yearsafter various levels of hardwood removal and under-story suppression in southwestern Oregon, USA. For-est Ecology and Management 96:1–11.

HERMANN, R., and D. LAVENDER. 1990. Pseudotsugamenziesii (Mirb.) Franco Douglas-fir. In Silvics ofNorth America, Volume 1, tech. coords. R. Burns andB. Honkala, 527–540. Agriculture Handbook 654.Washington, DC: USDA Forest Service.

HUGHES, T., J. TAPPEINER II, and M. NEWTON. 1990.Relationship of Pacific Madrone sprout growth toproductivity of Douglas-fir seedlings and understoryvegetation. Western Journal of Applied Forestry5:20–24.

ICE, G., and R. BESCHTA. 1999. Should salvage loggingbe prohibited following wildfires? In Proceedings of the1999 NCASI West Coast Regional Meeting, Volume II,452–60. Research Triangle Park, NC: National Coun-cil for Air and Stream Improvement.

KORMEIER, E., and C. PARK. 1995. Summary of fish habi-tat/water quality since the Silver Fire of 1987. GrantsPass, OR: Galice Ranger District, Siskiyou NationalForest.

LOWELL, E., S. WILLITS, and R. KRAHMER. 1992. Deteri-oration of fire-killed and fire damaged timber in thewestern United States. General Technical ReportPNW-GTR-292. Portland, OR: USDA Forest Ser-vice, Pacific Northwest Research Station.

MCIVER, J., and L. STARR. 2001. A literature review onthe environmental effects of postfire logging. WesternJournal of Applied Forestry 16:150–68.

MELLEN, T.K., and A. AGER. 1998. Coarse Wood Dynam-ics Model (CWDM). Portland, OR: USDA Forest Ser-vice, Pacific Northwest Region.

NEWTON, M., and E.C. COLE. 2004. 23-year responses ofponderosa pine and Douglas-fir to managed shrub den-sity in treated stands of madrone and whiteleaf man-zanita. Final report. Redding, CA: Sierra Cascade In-tensive Forest Management Research Cooperative.

SCHOWENGERDT, R. 1997. Remote sensing models andmethods for image processing. 2nd ed. San Diego, CA:Academic Press.

STAVINS, R.N, D.L. GALT, and K.L. ECKHOUSE. 1981.Economic analysis of alternative methods of vegetationmanagement and control in commercial forests of the Pa-cific Northwest Region. Washington, DC: US Envi-ronmental Protection Agency, Office of Pesticide Pro-grams.

TAPPEINER, J. II, M. NEWTON, P. MCDONALD, and T.HARRINGTON. 1992. Ecology of hardwoods, shrubs,and herbaceous vegetation: Effects on conifer regen-eration. Chapter 7 in Reforestation practices in south-western Oregon and northern California, eds. S.Hobbs, S.D. Tesch, P.W. Owston, R.E. Stewart, J.C.Tappeiner II, and G.E. Wells. Corvallis: Oregon StateUniversity, Forest Research Laboratory.

THOMAS, J.W. 2003. Sustainability of the NorthwestForest Plan—Dynamic versus static management.Draft manuscript.

US ARMY. 1993. GRASS4.1 reference manual. Cham-paign, IL: US Army Corps of Engineers, Construc-tion Engineering Research Laboratories.

USDA FOREST SERVICE (USDA-FS). 2002. Science Find-ings, Issue 47, October. Portland, OR: USDA ForestService, Pacific Northwest Research Station.

———. 2003. Biscuit post-fire assessment. Rogue Riverand Siskiyou National Forests. Available online atwww.biscuitfire.com; accessed by authors March2003.

US DEPARTMENT OF AGRICULTURE (USDA) and US DE-PARTMENT OF THE INTERIOR (USDI). 1994. Standardsand guidelines for management of habitat for late-suc-cessional and old-growth related species with the range ofthe northern spotted owl. Attachment A to the Recordof Decision. Portland, OR.

US DEPARTMENT OF THE INTERIOR. 1992. Managementguidelines for federal lands. Chapter 3 in Recoveryplan for the northern spotted owl—final draft, Volume1, 63–90. Portland, OR.

ZAVITKOVSKI, J., M. NEWTON, and B. EL-HASSAN. 1969.Effects of snowbrush on growth of some conifers.Journal of Forestry 67(4):242–46.

John Sessions (john.sessions@oregonstate.edu) is University Distinguished Professorand Stewart Professor of Forest Engineer-ing, College of Forestry, Oregon StateUniversity, 215 Peavy Hall, Corvallis,OR 97331-5706; Pete Bettinger is asso-ciate professor, Daniel B. Warnell Schoolof Forest Resources, University of Georgia,Athens; Robert Buckman is formerdeputy chief for research, USDA ForestService, Washington, DC, and retiredprofessor of forestry, Oregon State Univer-sity; Mike Newton is professor emeritus,College of Forestry, Oregon State Univer-sity, Corvallis; Jeff Hamann is principal,HamannDonald Associates, Corvallis. Adetailed version of this article, includingadditional background information, ta-bles, and complete citations, is availablefrom the corresponding author.

45April/May 2004 • Journal of Forestry