free scales where ecosystem anagement will be implemented. We recommend adoption of a new hierarchical vegeta- tion classification with seven structural ...
Technical Commentary
A Structural
Classification
for Inland
Northwest Forest Vegetation Kevin L. O'Hara and Penelope A. Latham, Schoolof Forestry, University of Montana, Missoula, MT 59812; Paul Hessburg, and Bradley G. Smith, Pacific Northwest Research Station, USDA Forest Service, Wenatchee, WA 98801.
Existing approaches tovegetation classification range from thosebasedonpotentialvegetationto othersbasedon existmgvegetationcomposition, orexistingstructuralorphysiognomiccharacteristics. Examplesof theseclassificationsare numerous,and in somecases,date back hundredsof years (Mueller-Dumbois andEllenberg1974).Small-scale orstandlevel multipleresourcemanagement hasusedpotentialvegetation/siteclassificationsfor severaldecades(Daubenmire andDaubenmire1968, Layser 1974, PfisterandArno 1980, Fergusonet al.1989). At broaderscales,ecosystemmanage-
menteffortsareaidedby classifications of forestvegetation that providesimplerepresentations of vegetationcomposition andstructuralattributesthatchangeovertime andspace (Ohver 1992, Swansonand Franklin 1992, McComb et al. 1993, Oliver et al. 1994, Turneret al. 1995). However,over broad areasin the Inland Northwest,ruggedmountainous
topography,contrastinggeologicsubstrates,and a highly variable maritime
influence from the Pacific coast combine
to createwide varietyin vegetationtypesandproductivities. Ecosystemmanagersin the Inland Northwest need a classificationsystembasedon biologicallysignificantvegetativecharacteristics thatcapturethisvariation,andonethat is comprisedof variablesthat describestructuralconditions important to achieving managementobjectives.Existing potentialvegetationclassificationsystemsprovide useful representations of theeffectsof soilsandclimateon vegetanoncomposition,butdonotreflectdifferencesin disturbance h•story.In mostwesternforests,disturbances are a primary factoraffectingthe compositionandstructureof vegetation (Oltver 1981,Habeck1987,Agee 1993).In addition,potential vegetationclassescannotbe directly identifiedthrough remotesensingandthereforerequiremodelingandexpensive groundsurveysfor validation. NOTE:
The developmentof the structuralstageclassification presentedin thispaperwas supportedby the ColumbiaRiver BasinEcosystem Assessment ProjectandtheUSDA ForestServiceNorthernRegion. Helpfulreviewsof thismanuscript wereprovidedby PaulAlaback, StephenArno,RaganCalloway,RichardHutto,andRay Shearer.
In this paper, we examine some alternativevegetation classificationsystemsfollowingtheframeworkpresentedby Kimmins (1987) for vegetationclassificationexceptthat we onlydiscusscategories thatincludesystemscommonlyused in the Inland Northwest.
We advocate the use of a structural
or physiognomicvegetationclassificationbasedon the biological processof standdevelopmentwhich can be used acrossvariablespatialscales. Structureis anecologicallysignificantattributeof vegetation consideredto have threemajor components (Kershaw 1964): (a) vertical structure;(b) horizontalstructure;and (c) quantitativestructure.Quantitativestructuremay be further specifiedaccordingto life forms, floristics, or size-class distributions(Mueller-Domboisand Ellenberg 1974). The distributionof life formshasbeensuggested asa fundamental axiscontributingto thecoexistence of speciesandthe organizationof ecosystems (Cody 1986). Mueller-Domboisand Ellenberg (1974) state that structuralsimilaritiesbetween communitiesprovidea basisfor the comparisonof functions. We believea structuralvegetationclassificationbasedon standdevelopment processes reflectsfine- andcoarse-grained processeswhich operateacrossstandsand landscapes.At small scalessuchas at the standor patchlevel, a structural classificationwould be useful for creation of vegetation structures whichmeetspecificresourcemanagement objectives. At larger spatialscales,a structuralclassificationcan serveasthe basisfor predictingandplanningfor vegetation changeover time. Further, a simple classificationwith a small numberof classescanreducethe complexityof landscape-scalemodeling and planning.For example,a hypotheticalareawith 10forestvegetationtypesand10vegetation developmentclasseshas100possiblevegetationunits.Compounded by variations in rate of growth or progression throughthese10 classescausedby site,disturbance, management treatment, and spatial characteristicsof vegetation units,a sizableplanningproblemiscreated.Finally,at global scales,remotelysensedvegetationclassifications will likely be limitedto physiognomic structuralparameters(Running WJAF11(3)1996 97
et al. 1994). Theseglobal-scalestructuralclass•ficanons can be linkedto smaller-scale classifications facilitatingconsistency in hierarchicalecosystemplanningand implementation efforts.
Existing Vegetation Approach
Forest covertypes, or dominant treevegetation types, wereamongthefirst classifications basedsolelyon existing vegetationin the United States.Historically,forestcover typeswereoftennamedfor an economically importantspecieswhichmightbepresentatafairlylowlevelof abundance, ignoring a more abundantbut less valuable species.At present,covertypesare identifiedby the predominanttree speciesbasedon basal area, and must be of "distinctive character"and occupya large area (Eyre 1980). Although thesemethodsare easilyunderstoodand standsare easily classified,theyarenotconsidered to be"gearedto ecological managementof forest lands" (Wellner 1989). Additional shortcomings arisewhenonecovertypecoverslarge,diverse geographicareasor whenstandswith manyspecieshaveno clear dominant constituent.
Potential Vegetation ("Climax") Approach Theearliestvegetationclassification specificto theInland Northwest was Daubenmire's habitat type system (Daubenmire 1952, Daubenmire and Daubenmire 1968). This system,andthemanybiogeographicandplantassociation studiesthat followed, were attemptsto characterize vegetation foundin diverseinlandenvironments (seeFerguson 1989 andKrajina 1965 for examples).Thesesystemsfollow
theprevailingsuccessional paradigm wherevegetation change is assumedto be directionaltowardsa theoreticallystable climax plant association.A seriesof communitiesare assumedto succeedeachotherin a relay fashion,converging toward climax. The habitattype classificationessentially attemptsto discerndominantover-andunderstoryspeciesin the climax communitythroughevaluationsof the relative abundanceand reproductivesuccessof variousindicator species. Thesespeciesarethengroupedandassociated with specificsiteattributespermittingthevegetationassociation to be used as a site or land classificationtool (Daubenmire
andDaubenmire1968,Layser1974,PfisterandAmo 1980). The emphasisin theseclimaxvegetation-based classificationsisonpotentialvegetationratherthancurrentvegetation. Daubenmireand Daubenmire(1968) consideredcurrently dominanttreesin the overstoryof minor importancecomparedto the speciesreproducing successfully in the understory.Presentvegetationwasassumed to betoounstableand unpredictable for classification: a view still held for some ecosystem management uses(Pfister1993). In recentyears,a reexaminationof climax theoryandthe conceptof stabilityas it appliesto ecosystemssuggeststhe classification of vegetationbasedonpotentialclimaxmaybe arbitrary(Egler 1954, Drury andNesbit 1973, Connelland Slatyer 1977, Christensen1988, Pickett and McDonnell 1989, notkin 1990, Oliver andLarson 1990, Sprugel1991, Cook 1996, as well as the initial criticismof Gleason1926). 98
WJAF11(3)1996
Many of thequestions regardingtheadequacyof thesuccession/climaxparadigmrevolvearoundtheimpliedstabilityof climax ecosystems: succession towardsclimax was considered the norm, and disturbanceswere disruptionsof the normal successionalprocess.In many environments,and particularlytheInlandNorthwest,disturbances arethenorm, and developmentto a truly stableclimax is rare or absent Habitattypesandpotentialnaturalvegetationthenbecome, asstatedby Mueller-DomboisandEllenberg(1974), "idealistic schemesthatcannotbe expectedto fit reality."Hence, potential vegetation classificationsare based on a trend towardsclimax ratherthantheclimax itself (PfisterandArno
1980).Recentstanddevelopment studiesin theinlandNorthwestandelsewheresuggestthislogicmay be flawed aswell Thesestudiesindicatemany shade-tolerant treespreviously thoughtto haveestablished beneaththeoverstoryand,therefore, to be reproducingsuccessfullyin the understory,were actuallytreesof the sameageor cohortastheoverstorythat
weresimplyoutgrown(OliverandLarson1990,Cobbet al 1993, O'Hara 1995). Insteadof reproducingsuccessfully, thesespeciesare outcompeted andrelegatedto subordinate crownpositionswheretheir greatershadetoleranceenables them to survive.They will not dominateunlessthrough differentiallongevity(Egler 1954) they outlasttheir overstorycompetitors. Thisis unlikelyin manyInlandNorthwest environmentswhereshadeintolerantspeciesare long-lived, and shadetolerant species,particularlythosesubjectedto competitivestress(Castello et al. 1995), tend to be more proneto a varietyof insectsandpathogens(Hessburget al 1994). Thus "climax" vegetationis often more prone to disturbancethan preclimaxvegetation.Hence classifying one set of speciesas inherentlymore stablethan anotheris tenuous.
Stand Structure or PhysiognomicApproaches Vegetationclassifications basedon physiognomic characteristicsrepresentone of the oldestforms of vegetation classification(Mueller-Domboisand Ellenberg1974). One of the first efforts to characterize
vertical forest structure •n
the Inland Northwestwas Thomas's (1979) descriptionof structuraldevelopmentfor foreststandsin the Blue Mountainsof OregonandWashington.Thesestagesdescribedthe sequential development of standsfollowingclearcutting, and barringdisturbance,as proceedingthroughdenseseedlings
and saplings,saplingsand poles,poles,small sawtimber, largesawtimber,andold growth.Whereasthesestandconditionsareprimarilyrepresentations of verticalstructure,their quantificationis usuallybasedon averagediameterat breast height. Other structural
classifications
followed.
O'Hara
et al
(1990) usedten structuralstagessimilarto Thomas'sstages to describestanddevelopmentin a silviculturalexpertsystem in easternWashington.Similar structuralstageshave been usedaslabelsin successional pathwaydescriptions by Daws et al. (1980), KessellandFischer(1981), Amo et al. (1985), Bradley et al. (1992), and othersin the Inland Northwest Their principalshortcoming is thatsuchclassifications were
basedprimarily on arbitrarytree diameterclassesratherthan on any biologically significantstructuralfeaturessuch as canopy cover or diameter size classes linked to wildlife
habitat. Alternatively,canopycover alone is difficult to measurewith consistency,and little informationis available on structuralrequirementsfor mostwildlife species.
A seriesof four process-based standdevelopment stages were developedby Oliver (1981) to describeeven-aged (single-cohort)standdevelopmentfollowing standreplacementdisturbances. Thesestageswere definedprimarilyby availabilityof andcompetitionfor growingspace.The stand initiationstage,for example,beginswith a standreplacing disturbance andendswhengrowingspaceis fully occupied. Growingspacerefersto all theresources neededby a treeto existon a given site(0liver and Larson1990).Closedstem exclusionis the periodwhen intensecompetitionfrom the existingtreesprecludesnew regeneration.During m•det'stocvreinitiation.theeven-agedstandbeginsto breakdown. anda newcohortor ageclassbecomesestablished.The final stage.oldgrowth.isdefinedby a uniformityof processes and an ab.•enceof treesoriginatingfrom allogenicdisturbances (0liver 1981, Oliver and Larson 1990). These stagesare biologicallybasedandobservable in mostforesttypesdevelopingfromcatastrophic disturbance; however,theyhavenot previouslybeendefinedby anyquantitativecriteria.Quanti-
ficationwould differ by forestcovertype as standsfrom differentcover typeswill move throughthe development stagesat differentratesand possess a varietyof structural featuresdependinguponinitialconditions.In addition,stands may not move sequentiallyfrom one stageto another.For example,a multi-agedstandmightbe formedwhena stand alternatesbetweenstemexclusionandunderstoryreinitiation structures. AlthoughOliver's (1981) stageswereoriginally
A. Stand
Initiation
Definition: Growing space is reoccupied foilowl ng a stand replacing disturbance. Description: 1 canopy stratum (may be broken or continuous); I cohort of seedlings or saplings; grass, forbs, shrubs may also be present.
B. Open Stem Exclusion Definition: Underground establishment
of new
competition
limits
individuals.
Description: One broken canopy stratum which includes poles or smaller trees; grasses shrubs or forbs may also be present.
not intended as a classification tool, no other standstructural
classificationsystemoffers stagesdefined by processes. Recently, when discussinglandscapemanagement,Oliver (1992) proposed usingthesestructuralstagesto definevegetationtypesover landscapeunits. Oliver's ( 1981)standdevelopment stageshaveanadvantagein thattheydescribeprocesses leadingto structures with thepotentialto bequantified.McNicoll (1994) useddiscriminant analysisto identify variablesand valuesof variablesto
classifywesternredcedar(Thujaplicata) standsin western Montanato Oliver's four standdevelopmentstages.Validation of thesefunctionson an independentdata set demonstratedthat canopy cover by size class coukl be used to classifystandscorrectly92% of thetime. McNicoll's results provide a decision systemfor classificationof standsinto structuralstagesthat are biologicallybasedandcan be used to classifystandsover broad areasfor ecosystemanalysis
C. Closed
Stem
Exclusion
Definition: New individuals are excluded through light or underground competition. Description: Continuous closed canopy, usually one cohort; poles, small or medium trees present. Suppressed trees, grasses, shrubs, and forbs may be absent in some cover types.
efforts.
Despitethe utility of a standstructureapproachbasedon Oliver's ( 1981) stagesof standdevelopment,manycommon Inland Northwestforest structuresdo not fit within any of Oliver's (1981) four stages.We expandedOliver's classificationto sevenclassesto includea greatervariety of structural conditions(Figure 1). "Classes"are usedto describe thesestructuralcategoriesbecause"stages"impliessequen-
Figure 1. Schematic representation of proposed stand
structure
classes
with
definition
and
description. Stand initiation, closed stem exclusion,
understory reinitiation, and old forest multi-strata are adapted from Oliver (1981).
(continued) WJAF 11(3)1996 99
D. Understory Reinitiation Definition:
Initiation
of new cohort
as older cohort
occupies less than full growing space. Description: Broken overstory canopy with formation of understory stratum; two or more cohorts. Overstory may be poles or larger trees; understory is seedlings, saplings, grasses, forbs,
E. Young Multi-Strata Definition: Two or more cohorts present through establishment after periodicdisturbancesincluding harvest
events.
Description: Multi-aged (multi-cohort) stand with assortment of tree sizes and canopy strata present but very large trees absent. Grasses, forbs, and shrubs may be present.
or shrubs.
F. Old Forest
Multi-Strata
Definition: Two or more cohorts and strata present including large, old trees. Description: Multi-aged stand with assortment of tree sizes and canopy strata present including large, old trees. Grassses, forbs,and shrubs may be present.
G. Old Forest Single-Stratum Definition: Single stratum of medium to large, old trees of one or more cohorts.
Structure
maintained
through nonlethal burning or management. Description: Broken or continuous canopy of medium to large, old trees. Single or multi-cohort. Understory absent or consisting of some seedlings, saplings, grasses, forbs, or shrubs.
Figure 1 (continued)
tial development.The additionalclassesinclude:openstem
includes a subset of the seven classes since all classes do not
exclusionwherecrowncoverappearsto be constrainedby below-groundcompetition(Figure lB); youngmulti-strata createdby a seriesof minor disturbances (includingharvest treatments)whichmaynotfit intoanold-tbrestclassbecause
occur in all cover types.
of the absenceof large trees (Figure I E); and old Jbrest single-stratumthat consistsof multi-agedtreesin a single-
Broad-scaleecosystemplanning and analysis of landscapesrequirevegetationclassificationtoolsthat: (1) allow elucidationof vegetationpatternsthroughtime; (2) are responsiveto biotic and abiotic processesthat lead to those patterns;and(3) are sufficientlyflexible to reflectecosystem changedueto changingclimate,invasionof exotic species, disturbance,migrationof species,and anthropogenicinfluences.A commonvegetationclassificationsystemto facili-
canopylayer maintainedby frequentlow intensitysurface fire, grazing,or both (Figure IG). "Old forest" is usedto avoid alternative connotationsof the term "old growth." which hasbecomemore of a socialratherthanan ecological construct.These seven classesare used within cover type designations to describestructuralchange.Each covertype 100
WJAF 11(3)1996
Recommendations
can be predicted,andresourcevaluesassigned. All threetypesof vegetationclassiftcation described here haveutilityfor ecosystem management andthepotentialtofit into an integratedecologicalland classificationframework
ecologicalclassificationsare ultimately interpretive,•t is essentialthata schemebe chosenthatclarifiestheecological attributes,processes,anddynamicsappropriateto the objectivesand valuesfor which landscapesare managed. Ecosystemmanagementobjectivesincludenot only traditional socioeconomicobjectives related to commodity production,but more comprehensiveobjectivessuch as maintenanceof biodiversity,promotionof forest health, and maintenanceof ecosystemprocesses(Oliver 1992,
(e g., Bailey 1988). We believe an open-ended, nondeterministic, biologicallybasedstructural classiftcation
O'Hara et al. 1994, Salwasser 1994, Turner et al. 1995, Castello et al. 1995).
combined withexistingvegetation (covertype)hasthegreat-
Integrationof variousclassificationschemeshasbeen a part of vegetationclassificationfor decades.It can be seenin useof structuralstagesandpotentialvegetationby Arno et al. (1985), vegetationlayers andpotentialvegetation by Steele(1984), andcovertypesandstructuralstages by the ColumbiaRiver Basin Assessment Project. Such integrationslead to somenew uses of terminology: for example,changesin structureor physiognomiccondition havebeendescribedas successional changes.We suggest thatchangesin structurebe describedasstructuralor stand developmentchangesto avoid confusionwith changesin communitycompositionthat have been traditionally de-
tate understandingand •mplementanonof ecosystemmanagementobjectivesat differentscalesshouldbe availableto resourcemanagersandplanners.This systemshouldallow d•scriminationamongexistingvegetationstructuressuch
thatdiscretestructures canbe quantified,changeovertime
est potentialto meet theseobjectives.Such a shift has the potentialbenefitof placingtheemphasisof classification on existingconditionsandprocesses andspecificallyonvegeta-
tion s•tructure whichincreasingly is the basisfor describing resourcemanagementobjectives(O'Hara et al. 1994). No singlevariable will adequatelydescribevegetation characteristics. An effectivevegetation classification system will incorporatea setof variablesto characterize vegetation into groupsor classeswhichcanbe arrangedhierarchically corresponding to theirutilityin a landscape planningformat: thevariableschosenmustberesponsive to changes in scale. However,whatisimportantatonescalemaynotbeimportant at another,andvariablesimportantfor oneresourcevalue may not be importantfor another(Baileyet al. 1978,Wiens andMilne 1989,Turneretal.1995).Forexample,theprimary variablesin the classificationsystemshouldbe thosemost importantfor distinguishingvegetationat intermediateor mesoscales(1:24,000 to 1:100,000) suchasdominantcover type and structuralcharacteristics. At finer scales,variables
shouldreflectvegetationcharacteristics importantto implementationof resourceobjectivesbut shouldbe linked to broader-scale vegetationcharacterization. At broaderscales, structuralclassescan be collapsedinto largerphysiognomic groups. Thisvegetation classification structure wouldfacilitate effectiveecosystem management by buildingdirectlinkages betweenlandscape classifications for ecosystem planningand freescales whereecosystem management willbeimplemented. We recommendadoptionof a new hierarchicalvegetation classificationwith sevenstructuralclasses(Figure 1) that reflect vegetationdevelopment.Lower levels in the hierarchycanexpandthesesevenclassesto definespecific finer grained stand structureattributes. For example, a wildlife habitatclassificationat a fine scalemight refine the structural
classes to include
more detail
on crown
cover to describespecifichabitatrequirementsof certain wildlife. A timberclassificationmightrefine all classesby addinglevels of relative density. Such a classificationof foreststandstructuremay serveasa unifyingtool to meet timber, wildlife, water, or recreationobjectives,and for describingstandor ecosystemconditionsfor long-term assessmentand monitoring efforts. A featurecommonto all vegetationclassificationsystems is their inability to satisfy all potential users. All three typesof Inland Northwestvegetationclassifications describedhave shortcomings that make oneclassification systemmore appropriatefor someusesthan another.As
scribed as successional.
The most commonlyusedclassificationsystemsin the Inland Northwest are basedon potential natural vegetation. The basicpremiseof thesesystemsis a deterministic succession to a climax communityrepresentedby species presentin theunderstory(Cook 1996). Christensen(1988), in discussingthe "demise" of the classical succession/ climax paradigm stated "we may now conclude that a grandunifiedtheoryof communitysuccession suchasthat envisionedby Clementsand Odum is not possibleor even desirable. The bottom-line messageto those who must managenatural ecosystemsis that the world is considerablylesstidy thanwe thought.Furthermore,thisuntidyness may be an integral part of maintenanceof many ecosystems." Although many researchershave tried to modify the successionparadigmand to redefine its terminology, we believe it is time to move towards a new paradigm centered
on a vertical
structure
and linked
to floristic
composition(cover type). Potentialnatural vegetationclassificationsystemsmay
havevalueif thepresence of certainspecies is interpreted as an environmental indicatornotlinkedto theclimaxconcept. For example,the well-recognizedprogressionof different "climax"specieswith elevationin InlandNorthwestenvironmentsis indicativeof an environmentalgradient.These,or other, plant speciescan serveas useful indicatorsof the presenceof environmentalor site conditionswhich they require,butnotfor thefuturedevelopment of anyhypothetical community.In this sense,the interpretationmay be correct,but the underlyingconceptsareflawed. A structuralclassificationsystembasedon the reality of vegetation developmentcharacterizedby biological potentialsandecologicalconstraintsis recommendedrather
than onebasedon an idealizedcommunitytype that seldom occurson Inland Northwest landscapes. WJAF11(3)1996 101
Literature
Cited
AGEE,J.K. 1993. Fire ecologyof Pacific Northwestforests.IslandPress, Wash., D.C. ARNO, S.F., D.G. SIMMERMAN,ANDR.E. KEANE. 1985. Forest succession
on four habitat types in western Montana. USDA For. Serv. Gen. Tech. Rep. INT-177. BAILEY,R.G. 1988. Ecogeographicanalysis:A guideto the ecological division of land for resourcemanagement.USDA For. Serv. Misc. Publ. 1465.
BAILEY,R.G., R.D. PEISTER, ANDJ.A. HENDERSON. 1978. Nature of land and resource classification--a
review.
J. For. 76:650-655.
BOThtIN,D.B. 1990. Discordant harmonies: A new ecology for the twenty-first century. Oxford Univ. Press,New York. BRAI)LEY, A., W.C. FISC.ER,ANDN.V. NOSTE.1992. Fire ecologyof the foresthabitattypesof easternIdaho andwesternWyoming. USDA For. Serv. Gen. Tech. Rep. INT-290. CASTELLO, J.D., D.J. LEOPOLD, ANDP.J. SMALLIDGE. 1995. Pathogens,
patterns,andprocesses in forestecosystems. Biosci.45(1)16-24. CHRISTENSEN, N.L. 1988. Succession and natural disturbance: Para-
digms,problems,and preservationof natural ecosystems.P. 62-86 in Agee, J.K., and D.R. Johnson(eds.) Ecosystemmanagementfor parks and wilderness.Univ. of Wash. Press,Seattle. COBB,D.F., K.L. O'HARA, ANDC.D. OLIVER.1993. Effects of variations
in standstructureon developmentof mixed-speciesstandsin eastern Washington.Can. J. For. Res. 23:545-552. CODY,M.L. 1986. Structuralnichesin plant communities.P. 381-405 in Diamond, J., and T.J. Case (eds.) Community ecology. Harper & Row, New York.
CONNELL, J.H., ANDR.O. SLATYER.1977. Mechanisms of successionin natural communities and their role in community stability and organization.Am. Natur. 111:1119-1144.
Cook,J.E. 1996.Implicationsof modernsuccessional theoryfor habitat typing: A review. For. Sci. 42:67-75. Daubenmire,R. 1952. Forestvegetationof northernIdaho and adjacent Washington,and its bearingon conceptsof vegetationclassification. Ecol. Monogr. 22:301-330. DaubenmireR., and J.B. Daubenmire. 1968. Forestvegetationof eastern Washingtonand northernIdaho. Wash. Agric. Exp. Stn. Tech. Bull. 60.
DAVIS,K.M., B.D. CLAYTON, ANDW.C. FISCHER. 1980. Fire ecology of Lolo National Forest habitat types. USDA For. Serv. Gen. Tech. Rep. INT-79. DRURY,W.H., ANDI.C.T. NESBITT.1973. Succession.J. Arnold Arbor. 54:331-368.
EGLER, E.F. 1954. Vegetationscienceconcepts.I. Initial floristic composition--a factor in old-field vegetationdevelopment.Vegetatio 35:95-105.
EYRE,F.H. (ED.). 1980. Forest cover types of the United Statesand Canada. Soc. Am. For., Bethesda, MD. FERGUSON, D.E., P. MORGAN,ANDF.D. JOHNSON, (EDS.) 1989. Proceed-
ings--land classificationsbased on vegetation: Applications for resourcemanagement.USDA For. Serv. Gen. Tech. Rep. INT-257. GLEASON, H.A. 1926. The individualistic conceptof the plant associations. Bull. Torr. Bot. Club 53:7-26.
HABECK, J.R. 1987. Present-dayvegetationin the northernRocky Mountains. Ann. Missouri
Bot. Gard. 74:804-840.
HESSBURG, P.F., R.G. MITCHELL,ANDG.M. FILIP. 1994. Historical and current roles of insectsand pathogensin easternOregon and Wash-
ington. USDA For. Serv. Gen. Tech. Rep. PNW-327. KERSHAW, K.A. 1964. Quantitative and dynamic ecology. American Elsevier Publishing Co., Inc., New York. KESSELL, S.R., ANDW. C. FISCHER. 1981. Predictingpostfireplantsuccession for fire managementplanning.USDA For. Serv. Gen. Tech. Rep. INT-94. KIMMINS,J.P. 1987. Forest ecology. Macmillan, New York. KRAJINA, V.J. 1965. Biogeoclimaticzones in British Columbia. Ecol. West. N. Am. (Univ. of British Columbia, Vancouver)l: 1-17.
102
WJAF11(3)1996
LAYSER, E F 1974 Vegetative classification Its applicationto forestry in the northernRocky Mountains. J. For. 72:354-357. MCCOMB,W.C., T.A. SPIES,ANDW.H. EMMING.^•4.1993. Douglas-fir forests: Managing for timber and mature-forest habitat. J. For 91(12):31-42.
McNICOLL,C. 1994. Structuraldevelopmentandclassificationof western redcedar (Thuja plicata Donn.) standsin northwestern Montana Master's thesis, Univ. of Mont. Missoula. MUELLER-DoMBOIS,D., AND H. ELLENBERG.1974. Aims and methods of
vegetationecology.Wiley, New York. O'HARA,K.L. 1995. Early heightdevelopmentandspeciesstratification acrossfive climax seriesin the easternWashingtonCascadeRange New For. 9:53-60.
O'HARA, K.L., C.D. OLIVER,S.G. PICKFORD, ANDJ,J. TOWNSLEY. 1990 A
prototypedecisionrule basefor planningandanticipatingeffectsof silviculturalactivitiesover broadareas.AI Appl. in Natur. Resour 4:25-34.
O'HARA, K.L., R.S. SEYMOUR,S.D. TESCH,ANDJ.M. GULDIN. 1994 Silviculture and our changingprofession:Leadershipfor shifting paradigms.J. For. 92(1):8-13.
OLIVER,C.D. 1981. Forest developmentin North America following major disturbances.For. Ecol. Manage. 3:153-168. OLIVER,C.D. 1992. A landscapeapproach:Achievingandmaintaining biodiversity and economicproductivity. J. For. 90(9):20-25. OLIVERC.D., ANDB.C. LARSON. 1990. Foreststanddynamics.McGrawHill, New York. 467 p. OLIVER,C.D., W.H. KNAPP,ANDR. EVERETT. 1994. A systemfor implementingecosystemmanagement.P. 355-366 in Vol. II. Ecosystem management:Principlesand applications.USDA For. Serv. PNWGTR-318.
EElSTEl,R.D. 1993. The need and potentialfor ecosystemmanagement in forests of the Inland West. P. 217-239 in Defining sustainable forestry, Aplet, G., et al. (eds.) The Wild. Society/IslandPress, Wash., DC.
EElSTEl,R.D., ANDS.F. ARNO.1980. Classifyingforest habitat types basedon potential climax vegetation.For. Sci. 26:52-70. PICKETT, S.T.A., ANDM.J. MCDONNELL. 1989. Changingperspectivesin communitydynamics:A theory of successionalforce. Trends in Ecol. Evol. 4(8):241-245.
RUNNING, S.W., T.R. LOVELAND, ANDL.L. PIERCE. 1994. A vegetation classificationlogic basedon remotesensingfor usein globalblogeochemicalmodels.AMBIO 23(1):77-81. SALWASSER, H. 1994. Ecosystemmanagement:Can it sustaindiversity and productivity?J. For. 92(8):6-10. SPRUGEL, D.G. 1991. Disturbance,equilibrium,andenvironmentalvariability: What is 'natural' vegetation in a changing environment9 Biol. Cons. 58:1-18.
STEELE, R. 1984. An approachto classifyingseral vegetationwithin habitat types. Northwest Sci. 58:29-39. SWANSON, FJ., ANDJ.F. FRANKLIN. 1992. New forestryprinciplesfrom ecosystem analysis of Pacific Northwest forests. Ecol. Appl 2(3):262-274. T.OMAS,J.W. (TEC.. ED.). 1979. Wildlife habitatsin managedforests The Blue Mountains of Oregon and Washington.USDA For. Serv Agric. Hand. No. 553. TURNER,M.G., R.H. GARDNER,ANDR.V. O'NEILL. 1995. Ecological dynamicsat broadscales.P. 29-35 in BioSci. Supplement,Science and biodiversitypolicy. WELLNER,c.g. 1989. Classification of habitat types in the western United States.P. 7-21 in Proc.Land classificationsbasedonvegetation: Applicationsfor resourcemanagement,Ferguson,D.E, P Morgan, andF.D. Johnson(eds.). USDA For. Serv. Gen. Tech. Rep INT-257.
WIENS,J.A., ANDB.T. MILNE.1989. Scalingof "landscapes"in landscapeecology,or, landscapeecologyfrom a beetle's perspective LandscapeEcol. 3:87-96.
