'-':"""""""""--

Technical Commentary -

A Structural Classific~tion for InlandNorthwest Forest Vegetation

Kevin L. O'Hara and Penelope A. Latham, School ofForestry, University of Montana, Missoula, MT5981r;Paul Hessburg, and Bradley G. Smith, Pacific NorthwestResearch Station, USDA Forest Service, Wenatchee, .WA 98801. -

A

...!I

Existing approaches to vegetation classification range from In this paper, we examine some alternative vegetationthose based on potential vegetation to odIers based on exist- classification systems following dIe framework presented bying vegetation composition, or existing structural orphysiog- Kimmins (1987) for vegetation classification except dIat wenomic characteristics. Examples of dIese classifications are only discuss categories that include systems commonly usednumerous, and in some cases, date back hundreds of years in the Inland Northwest. We advocate the use of a structural(Mueller-Duntbois and Ellenberg 1974). Small-scale or stand- or physiognomic vegetation classification based on the bio-level multiple resource management has used potential veg- logical process of stand development which can be usedetation/site classifications for several decades (Daubenmire across variable spatial scales.and Daubenmire 1968, Layser 1974, Pfister and Amo 1980, Structure is an ecologically significant attribute of vegeta-Ferguson et al.1989). At broader scales, ecosystem manage- tion considered to have three major components (Kershawment efforts are aided by classifications of forest vegetation 1964): (a) vertical structure; (b) horizontal structure; and (c)that provide simple representations of vegetation composi- quantitative structure. Quantitative structure may be furthertion and structural attributes that change over time and space specified according to life forms, floristics, or size-class(Oliver 1992, Swanson and Franklin 1992, McComb et al. distributions (Mueller-Dombois and Ellenberg 1974). The1993, Oliver et al. 1994, Turner et al. 1995). However, over distribution of life forms has been suggested as a fundamentalbroad areas in the Inland Northwest, rugged mountainous axis contributing to dIe coexistence of species and the orga-topography, contrasting geologic substrates, and a highly nization of ecosystems (Cody 1986). Mueller-Dombois andvariable maritime influence from the Pacific coast combine Ellenberg (1974) state that structural similarities between

'" to create wide variety in vegetation types and productivities. communities provide a basis for the comparison of functions.\ Ecosystem managers in the Inland Northwest need a We believe a structural vegetation classification based on

classification system based on biologically significant veg- standdeyelopmentprocessesreflectsfine-andcoarse-grainedetative characteristics that capture this variation, and one that processes which operate across stands and landscapes. Atis comprised of variables that describe structural conditions small scales such as at the stand or patch level, a structuralimportant to achieving management objectives. Existing classification would be useful for creation of vegetationpotential vegetation classification systems provide useful structures which meet specific resource management objec-representations of the effects of soils and climate on vegeta- tives. At larger spatial scales, a structural classification can

..tion composition, but do not reflect differences in disturbance serve as the basis for predicting and planning for vegetationhistory. In most western forests, disturbances are a primary change over time. Further, a simple classification widI afactor affecting the composition and structure of vegetation small number of classes can reduce dIe complexity of land-(Oliver 1981, Habeck 1987, Agee 1993). In addition, poten- scape-scale modeling and planning. For example, a hypo-tial vegetation classes cannot be directly identified through theticalarea with 10 forest vegetation types and 10 vegetationremote sensing and therefore require modeling and expen- development classes has 100 possible vegetation units. Com-sive ground surveys for validation. pounded by variations in rate of growth or progression

through dIese 10 classes caused by site, disturbance, manage-Nam: The development of the structural stage classification presented in ment treatment, and spatial characteristics of vegetation

this paper was supported by the Columbia River Basin Ecosystem units, a sizable planning problem is created. Finally, at globalAssessment Project and the USDA Forest Service Northern Region. scales remotely sensed vegetation classifications will likelyHelpful reviews of this manuscript were provided by Paul Alaback, .'. " .Stephen Arno, Ragan Calloway, Richard Hutto, and Ray Shearer. be limited to phYSIOgnOmIC structural parameters (Runnmg

WJAF 11(3) 1996 97

Reprinted from the Western Journal of Applied Forestry, Vol. 11, No.3, July 1996:

~"_._~ ~ ,.

""0 -.:i;:

et~. 1994). These global-scale structural classifications can Many of the questions regardinO' the adequacy of the su -be ~ed t.o sm~er-scale classifications facilitating consis- sion/climax paradigm revolve .:round the implied stab;~e~ft~ncy ill hierarchical ecosystem planning and implementa- climax ecosystems: succession towards climax was .

d -tton efforts.' COnsl

.ered the norm, and dIsturbances were disruptions of the...normal successional process. In many environments, and

EXIsting VegetatIon Approach particularly the Inland Northwest, disturbances are the norm,Forest cover types, or dominant tree vegetation types and ?evelopment to a ~y stable climax is rare or absent.

were among the first classifications based solely on existin ' HabItat types and potentIal ~atural vegetation then become,

vegetation in the United States. Historically, forest cove; ~ ~tated by Mueller-Dombols ~dEllenberg (1974), "ideal-types were often named for an economically important spe- ISttC s~hemes tha~ cannot~ ex~ected to fit reality." Hence,

cies which might be present at a fairly lowlevel of abundance, ~oo~e;~alcr:getatton 'classifica~ons ~e based on a trendignoring a more abundant but less valuable species At ax rather than the climax Itself (Pfister and Arnopresent, cover types are identified by the predominant.tree 1980). Recent stand development studies in the inland North-species based on basal area, and must be of "distinctive west and el~e~he~ suggest this logic may be flawed as well.

.character" and occupy a large area (Eyre 1980). Althou h These studies illdicat~ many shade-tolerant trees previously .these methods are easily understood and stands are eas~y thought to have esta~lished beneath ~e overstory and, there-classified, they are not considered to be "geared to ecological fore, to be reproducmg successfully m the understory, were

management of forest lands" (W ellne 1989) Add.. al actuall! trees of the same ~ge or cohort as the overstory that

.r .ltton were SlInply outgrown (Oli d L 1990 C bbshortcomings arise when one cover type covers large, diverse 1993 O'H 1995 ) Ins v~r~ arsdon. ,0 etal.

geographic areas or when stands with man ecies have no ,.ara. tea 0 repro ucmg success~y,clear dominant constituent. y sp these spec~e.s are outcomp~ted and relegated to subordinate

crown POSItiOns where their greater shade tolerance enablesP .al V .". " them to survive. They will not dominate unless through

otenti egetation ( Climax) Approach differential longevity (Egler 1954) they outlast their over-

The earliest vegetation classification specific to the Inland sto~ competitors. This is unlikely in many Inland NorthwestNorthwest was Daubenmire' s habitat type system envIrOnments where shade intolerant species are long-lived,(Daubenmire 1952, Daubenmire and Daubenmire 1968). and s~.e tolerant species, particularly those subjected toThis system, and the many biogeographic and plant associa- competItIve s~ess (c:astello et al. 1995), tend to be moretion studies that followed, were attempts to characterize prone to a variety of InSects and pathogens (Hessburg et al.vegetation found in diverse inland environments (see Ferguson 1?94). Thus "climax" vegetation is often more prone to1989 and Krajina 1965 for examples). These systems follow disturbance th~ pre~limax vegetation. Hence classifyingthe prevailing successional paradigm where vegetation change one set of specIes as inherently more stable than another isis assumed to be directional towards a theoretically stable tenuous.

climax plant association. A series of communities are as-sumed to ~ucceed each other in a relay fashion, converging Stand Structure or Physiognomic Approachestoward cllmax. The habitat type classification essentially

..attetnptstodiscerndominantover-andunderstoryspeciesin Vegetation classifications based on physiognomic char-the climax community through evaluations of the relative acteristics represent one of the oldest fonDs of vegetationabwidance and reproductive success of various indicator classification (Mueller-Dombois and Ellenberg 1974). Onespecies. These species are then grouped and associated with of the first efforts to characterize vertical forest structure inspecific site attributes permitting the vege~tion association the Inland Northwest was Thomas's (1979) description ofto be used as a site or land classification tool (Daubenmire structural development for forest stands in the Blue Moun-and Daubenmire 1968, Layser 1974, Pfister and Arno 1980). tains of Oregon and Washington. These stages described the

The emphasis in these climax vegetation-based classifica- sequential development of stands following clearcutting, andtions is on potential vegetation rather than current vegetation. barring disturbance, as proceeding through dense seedlingsDaubenmire and Daubenmire (1968) considered currently and saplings, saplings and poles, poles, small sawtimber,dominant trees in the overstory of minor importance com- large sawtimber, and old growth. Whereas these stand condi- r

pared to the species reproducing successfully in the under- tions are primarily representations of vertical structure, theirstory. Present vegetation was assumed to be too unstable and quantification is usually based on average diameter at breast J '

unpredictable for classification: a view still held for some height.ecosystem management uses (PfISter 1993). Other structural classifications followed. O'Hara et al.

In recent years, a reexamination of climax theory and the (1990) used ten structural stages similar to Thomas's stagesconcept of stability as it applies to ecosystems suggests the to describe stand development in a silvicultural expert systemclassification of vegetation based on potential climax may be in eastern Washington. Similar structural stages have beenarbitrary (Egler 1954, Drury and Nesbit 1973, Connell and used as labels in successional pathway descriptions by DavisSlatyer 1977, Christensen 1988, Pickett and McDonnell et al. (1980), Kessell and Fischer (1981), Arno et al. (1985),1989, Botkin 1990, Oliver and Larson 1990, Spruge11991, Bradley et al. (1992),' and others in the Inland Northwest.Cook 1996, as well as the initial criticism of Gleason 1926). Their principal shortcoming is that such classifications were

98 WJAP 11(3) 1996

.Ii

1'-I based priIriarily on arbitrary tree diameter classes rather thani on any biologically significant structural features such as! canopy cover or diameter size classes linked to wildlifeI. habitat. Alternatively, canopy cover alone is difficult toI measure with consistency, and little information is available: on structural requirements for most wildlife species.

A series of four process-based stand development stageswere developed by Oliver (1981) to describe even-aged(single-cohort) stand development following stand replace- A.Stand Initiationment disturbances. These stages were defined primarily by Definition: Gro.win~ space is_reoccupied following

ail bili. f d ..., .a stand replacIng dl~rbance.av a ty 0 an competItIon lor grOWIng space. The stand Description: 1 canopy stratum (may be broken orinitiation stage, for example, begins with a stand replacing continuous); 1 cohort of seedlings or saplings;disturbance and ends when growing space is fully occupied. grass, forbs, shrubs may also be present.

Growing space refers to all the resources needed by a tree to -exist on a given site (Oliver and Larson 1990). Closed stem ~exclusion is the period when intense competition from the -

existing trees precludes new regeneration. During under-story reinitiation, the even-aged s~and begins to break down,and a new cohort or age class becomes established. The fmalstage, old growth, is defmed by a uniformity of processes andan absence of n-ees originating from allogenic disturbances(Oliver 1981, Oliver and Larson 1990). These stages arebiologically based and observable in most forest types devel-oping from catastrophic disturbance; however, they have notpreviously been defined by any quantitative criteria. Quanti-fication would differ by forest cover type as stands from B. Open Stem Exclusiondifferent cover types will move through the development Definition: Underground competition limits

..establishment of new individuals.stages at different rates and possess a vanety of structural Description: One broken canopy stratum whichfeatures depending upon initial conditions. In addition, stands includes poles or smaller trees; grasses shrubs ormay not move sequentially from one stage to another. For forbs may also be present.example, a multi-aged stand might be foimed when a standalternates between stem exclusion and understory reinitiationstructures. Although Oliver's (1981) stages were originallynot intended as a classification tool, no other stand structuralclassification system offers stages defined by processes.Recently, when discussing landscape management, Oliver(1992) proposed using these structural stages to defme veg-etation types over landscape units.

Oliver's (1981) stand development stages have an advan-tage in that they describe processes leading to structures withthe potential to be quantified. McNicoll (1994) used discrimi-nant analysis to identify variables and values of variables toclassify western redcedar (Thuja plicata) stands in westernMontana to Oliver's four stand development stages. Valida-...C. Closed Stem ExclusiontIOn of these functions on an mdependent data set demon- Definition: New individuals are excluded throughstrated that canopy cover by size class could be used to light or underground competition.classify stands correctly 92% of the time. McNicoll's results Description: Continuous closed canopy, usuallyprovide a decision system for classification of stands into one cohort; poles, small or medium trees present.

..Suppressed trees, grasses, shrubs, and forbs maystructural stages that are biOlOgIcally based and can be used be absent in some cover types.to classify stands over broad areas for ecosystem analysis

efforts.Despite the utility of a stand structure approach based on

Oliver's (1981) stages of stand development, m~y .common Figure 1. Schematic representation of proposedInland Northwest forest structures do not fit Within any of stand structure classes with definition andOliver's (1981) four stages. We expanded Oliver's classifi- description.Standinitiation,closedstemex~lusion,cation to seven classes to include a greater variety of struc- understory reinitiati~n, and old forest multi-strata

al d.. (F. 1) " Cl " d t descn'be are adapted from Oliver (1981).tur con ltions 19ure .asses are use 0 .these structural categories because "stages" implies sequen- (contmued)

WJAF 11(3) 1996 99

~,c ",",,-_C"OC" ,',.., l'!'!IIn"I"l '! ccF1nc'i"J

E. Young Multi-StrataD. Understory Reinitiation Definition: Two or more cohorts present throughDefinition: Initiation of new cohort as older cohort establishmenta~erperiodicdisturbancesincludingoccupies leSs than full growing space. harve~ :vents. .Description: Broken overstory canopy with Description: MultI-aged (multi-cohort) stand with ..formation of understory stratum; two or more assortment of tree sizes and canopy strata presentcohorts. Overstory may be poles or larger trees; but very large trees absent. Grasses, forbs, andunderstory is seedlings, saplings, grasses, forbs, shrubs may be present.or shrubs.

.G. Old Forest Single-StratumF. Old. ~orest Multi-Strata Definition: Single stratum of medium to large, oldl?efim!/on: Two or more cohorts and strata present trees of one or more cohorts. Structure maintainedInclud!n~ large, o!d trees. .through nonlethal burning or management.Description: MultI-aged stand with asso~en! of Description: Broken or continuous canopy oftree sIzes and canopy strata present IncludIng medium to large, old trees. Single or multi-cohort.large, old trees. Grassses, forbs,and shrubs may Understory absent or consisting of some seedlings,be present. saplings, grasses, forbs, or shrubs.

Figure 1 [continued)

, tial development. The additional classes include: open stem includes a subset of the seven classes since all classes do notexclusion where crown ~over appears to be constrained by occur in all cover types. ct"below-ground competition (Figure IB); young multi-stratacreated by a series of minor disturbances (including harvest Recommendations Itreatments) which may not fit into an old-forest class becauseof the absence of large trees (Figure IE); and old forest Broad-scale ecosystem planning and analysis of land-single-stratum that consists of multi-aged trees in a single- scapes require vegetation classification tools that: (1) allowcanopy layer maintained by frequent low intensity surface elucidation of vegetation patterns through time; (2) are re-fIre, grazing, or both (Figure I G). "Old forest" is used to sponsive to biotic and abiotic processes that lead to thoseavoid alternative connotations of the term "old growth," patterns; and (3) are sufficiently flexible to reflect ecosystemwhich has become more of a social rather than an ecological change due to changing climate, invasion of exotic species,construct. These seven classes are used within cover type disturbance, migration of species, and anthropogenic influ-designations to describe structural change. Each cover type ences. A common vegetation classification system to facili-

100 WJAF 11(3) 1996

if"I""

-tate understanding and implementation of ecosystem man- ecological classifications are ultimately interpretive, it isagement objectives at different scales should be available to essential that a scheme be chosen that clarifies the ecologi-resource managers and planners. This system should allow cal attributes, processes, and dynamics appropriate to thediscrimination among existing vegetation structures such objectives and values for which landscapes are managed.that discrete structures can be quantified, change over time Ecosystem management objectives include not only tradi-can be predicted, and resource values assigned. tional socioeconomic objectives related to commodity

All three types of vegetation classification described here production, but more comprehensive objectives such asI have utility for ecosystem management and the potential to fit maintenance of biodiversity, promotion of forest health,into an integrated ecological land classification framework and maintenance of ecosystem processes (Oliver 1992,I(e.g., Bailey 1988). We believe an open-ended, O'Hara et al. 1994, Sa1wasser 1994, Turner et al. 1995, .nondeterministic, biologically based structural classification Castello et al. 1995).

I combined with existing vegetation (cover type) has the great- Integration of various classification schemes has beenest potential to meet these objectives. Such a shift has the a part of vegetation classification for decades. It can bepotential benefit of placing the emphasis of classification on seen in use of structural stages and potential vegetation by

~ existing conditions and processes and specifically on vegeta- Arno et al. (1985), vegetation layers and potential vegeta- .tion structure which increasingly is the basis for describing tion by Steele (1984), and cover types and structural stagesresource management objectives (O'Hara et al. 1994). by the Columbia River Basin Assessment Project. Such

No single variable will adequately describe vegetation integrations lead to some new uses of terminology: forcharacteristics. An effective vegetation classification system example, changes in structure or physiognomic conditionwill incorporate a set of variables to characterize vegetation have been described as successional changes. We suggestinto groups or classes which can be arranged hierarchically that changes in structure be described as structural or standcorresponding to their utility in a landscape planning format: development changes to avoid confusion with changes inthe variables chosen must be responsive to changes in scale. community composition that have been traditionally de-However, what is important at one scale may not be important scribed as successional.at another, and variables important for one resource value The most commonly used classification systems in themay not be important for another (Bailey et al. 1978, Wiens Inland Northwest are based on potential natural vegeta-and Milne 1989, Turner et al.1995). For example, the primary tion. The basic premise of these systems is a deterministicvariables in the classification system should be those most succession to a climax community represented by speciesimportant for distinguishing vegetation at intermediate or present in the understory (Cook 1996). Christensen (1988),meso scales (1:24,000 to 1:100,000) such as dominant cover in discussing the "demise" of the classical succession!type and structural characteristics. At fmer scales, variables climax paradigm stated "we may now conclude that ashould reflect vegetation characteristics important to imple- grand unified theory of community succession such as thatmentation of resource objectives but should be linked to envisioned by Clements and Odum is not possible or evenbroader-scale vegetation characterization. At broader scales, desirable. The bottom-line message to those who muststructural classes can be collapsed into larger physiognomic manage natural ecosystems is that the world is consider-groups. This vegetation classification stroctln'e would facilitate ably less tidy than we thought. Furthermore, this untidynesseffective ecosystem management by building direct linkages may be an integral part of maintenance of many ecosys-between landscape classifications for ecosystem planning and tems." Although many researchers have tried to modifyfine scales where ecosystem management will be implemented. the succession paradigm and to rederme its terminology,

We recommend adoption of a new hierarchical vegeta- we believe it is time to move towards a new paradigmtion classification with seven structural classes (Figure 1) centered on a vertical structure and linked to floristicthat reflect vegetation development. Lower levels in the composition (cover type).hierarchy can expand these seven classes to derIne specific Potential natural vegetation classification systems mayliner grained stand structure attributes. For example, a have value if the presence of certain species is interpreted aswildlife habitat classification at a fine scale might reline an environmental indicator not linked to the climax concept.the structural classes to include more detail on crown For example, the well-recognized progression of different -cover to describe specific habitat requirements of certain "climax" species with elevation in Inland Northwest environ-wildlife. A timber classification might reline all classes by ments is indicative of an environmental gradient. These, oradding levels of relative density. Such a classification of other, plant species can serve as useful indicators of the .forest stand structure may serve as a unifying tool to meet presence of environmental or site conditions which the!timber, wildlife, water, or recreation objectives, and for require, but not for the future development of any hypotheU-describing stand or ecosystem conditions for long-term cal community. In this sense, the interpretation may beassessment and monitoring efforts. correct. but the underlying concepts are flawed. .

A feature common to all vegetation classification sys- A structural classification system b.ased on ~e re~ltytems is their inability to satisfy all potential users. All of vegetation development charactenzed by bIologicalthree types of Inland Northwest vegetation classifications potentials and ecological constraints is recommended ratherdescribed have shortcomings that make one classification than one based on an idealized community type that sel-system more appropriate for some uses than another. As dom occurs on Inland Northwest landscapes.

WJAF 11(3) 1996 101

..;.~,;'J

Literature Cited LA ~SER, E.F. 1974. Vegetative classification: Its application to forestry ~I..In the northern Rocky Mountains. 1. For. 72:354-357. ~

AGEE, I.K. 1993. Fire ecology of Pacific Northwest forests. Island Press, McCoMB WC T A SPIES W H E 1993 D .:

W h DC' .., .., AND ..MMINGHAM. .ouglas-fIras., ..forests: Managing for timber and mature-forest habitat 1 ForARNo, S.F., D.~. SIMM~AN, AND R.E. KEANE. 1985. Forest succession 91(12):31-42. ...

on four habitat types In western Montana. USDA For. Serv GenTech. Rep. INT -177. ..McNICOLL, C. 1994. ~truc:ural development ~d classification of west-

B R G 1988 E ...em redcedar (ThuJa phcataDonn.) stands In northwestern MontanaAn.EY, ...cogeographlc analysIs: A guide to the ecological M t ' th . U . f M Mi I ....as er s eSIS mv.o onto ssou a.dIvIsIon of land for resource management. USDA For. Servo Misc. M D '

D H .Publ. 1465. UELLER- .OMBOIS, ., AND: .ELLENBERG. 1974. Aims and methods ofBAn.EY, R.G., R.D. PFISTER, AND I.A. HENDERSON. 1978. Nature of land , vegetation ecology. WIle.y,New York.

and resource classification-a review. 1. For. 76:650-655. 0 HARA, K.L. 19?5. Early height development and species stratificationBOTKIN, D.B. 1990. Discordant harmonies: A new ecology for the across five ~llmax series in..the east_ern Washington Cascade Range.

twenty-first century. Oxford Univ. Press, New York. New For. 9.53-60.BRADLEY, A., W.C. FISCHER, AND N.V. NOSTE. 1992. Fire ecology of the O'HARA, K.L., C.~: OLIVER, S.G. PICKFORD, AND 1.1. TOWNSLEY. 1990. A

forest habitat types of eastern Idaho and western Wyoming. USDA P!ot.otype decISI.o~ ~le base for planning and anticipating effects ofFor. Servo Gen. Tech. Rep. 1NT-290. sIlvlculturai aCtIvIties over broad areas. AI App1. in Natur. Resour.

CASTl!LLO, I.D., D.I. LEOPOLD, AND P.I. SMALLIDGE. 1995. Pathogens, , 4:25-34. -patterns, and processes in forest ecosystems. Biosci. 45(1)16-24. 0 HA;u,' K.L., R.S. SEYMPUR,. S.D. TESCH, AND I.M. GULDIN; 1994.

CHRISTENSEN, N.L. 1988. Succession and natural disturbance: Para- SI1vlc.ulture and our changing profession: Leadership for shiftingdigms, problems, and preservation of natural ecosystems. P. 62-86 paradigms. 1. For. 92(1):8-13.in Agee, I.K., and D.R. Iohnson (eds.) Ecosystem management for OUVER, C.D. 1981. Forest development in North America followingparks and wilderness. Univ. of Wash. Press, Seattle. major disturbances. For. Ecol. Manage. 3:153-168.

COBB, D.F., K.L. O'HARA, AND C.D. OLIVER. 1993. Effects of variations OUVER, C.D. 1992. A landscape approach: Achieving and maintainingin stand structure on d.evelopment of mixed-species stands in eastern biodiversity and economic productivity. 1. For. 90(9):20-25.Washington. Can. 1. For. Res. 23:545-552. OUVER C.D., AND B.C. LARSON. 1990. Forest stand dynamics. McGraw- 'i#"-" .

CoDy,M.L. 1986. Structural niches in plant communities. P. 381-405 in Hill, New York. 467 p. .Diamond, I., and T.I. Case (eds.) Community ecology. Harper & OLIVER, C.D., W.H. KNAPP, ANDR. EVER1m'. 1994. A system forimp1e-Row, New York. menting ecosystem management. P. 355-366 in Vol. II. Ecosystem

CONNELL, I.H., AND R.O. SLATYER. 1977. Mechanisms of succession in management: Principles and applications. USDA For. Servo PNW-natural communities and their role in community stability and GTR-318.organization. Am..Na~r. 111:1119-1144. ..PFISTER, R.D. 1993. The need and potential for ecosystem m8;nagement

Cook, ~ .E. 1996. ~mpllcauons. of modem successional theory for habitat in forests of the Inland West. P. 217-239 in Def'lning sustainabletyping: A review. For. SCI. 42:67-75. forestry, Aplet, G., et al. (eds.) The Wild. Society/Island Press,

Daubenmire, R. 1952. Forest vegetation of northern Idaho and adjacent Wash., DC.Vfashington, and its bearing on concepts of vegetation classifica- PFISTER, R.D., AND S.F. ARNo. 1980. Classifying forest habitat typestion. ~co1. Monogr. 22:301-33~. .based on potential climax vegetation. For. Sci. 26:52-70.

Daubenmlre R., and I.B. DaubenmIre. 1968. Forest vegetation of east- PI~~ S T A ANDM1 McDoNNELL1989 Ch . U..W h' d . I w. ~~'" ..., angmgperspec ves mem as mgton an northern daho. ash. Agnc. Exp. Stn. Tech. community dynamics: A theory of successional force. Trends in

Bull. 60.. Eco1. Evo1. 4(8):241-245.DAvIS, K.M., B.D. CLAYTON, AND W.C. FISCHER. 1980. Fire ecology of R G S W T R L L L D 1994 A .L 1 N . 1F h b. USDA F S G T h UNNIN, .., ..OVELAND, AND "",CE. .vegetation0 0 ationa orest a ltat types. or. erv. en. ec. 1 .fi U. 1 . b d t . f . I bal b. R INT -79 c aSSI ca on ogIc ase on remo e sensmg or use m g 0 10-

ep. ..geochemical models. AMBIO 23(1):77-81.DRURY, W.H., AND I.C.T. NESBni. 1973. Succession. 1. Arnold Arbor. S H 1994 E . Can.. di .54:331-368. ALWASSER, ...? cosystem management: It sustam versltyEGLER, E.F. 1954. Vegetation science concepts. I. Initial floristic com- and producuvlty. 1: For. 92(8):6-.1? ...

position-a factor in old-field vegetation development. Vegetatio SPRUGEL,D.G. 1991. Disturbance, equillbnum, and envIronmental van-35:95-105. ability: What is 'natural' vegetation in a changing environment?

EYRE, F.H. (ED.). 1980. Forest cover types of the United States and BioI. Cons. 58:1-18.Canada. Soc. Am. For., Bethesda, MD. STEELE, R. 1984. An approach to classifying seral vegetation within

FERGUSON, D.E., P. MORGAN, AND F.D. IOHNSON, (EDs.) 1989. Proceed- habitat types. Northwest Sci. 58:29-39.ings-1and classifications based on vegetation: Applications for SWANSON, Fl., AND I.F. FRANKUN. 1992. New forestry principles fromresource management. USDA For. Servo Gen. Tech. Rep.INT-257. ecosystem analysis of Pacific Northwest forests. Eco1. Appl.

GLEASON, H.A. 1926. The individualistic concept{)f the plant associa- 2(3):262-274.tions. Bull. Torr. Bot. Club 53:7-26. THOMAS, I. W. (TECH. ED.). 1979. Wildlife habitats in managed forests:

HABECK, I.R. 1.987. Present-day vegetation in the northern Rocky Moun- Th~ Blue Mountains of Oregon and Washington. USDA For. Serv.tains. Ann. Missouri Bot. Gard. 74:804-840. Agric. Hand. No. 553.

HESSBURG, P.F., R.G. MrrCHELL, AND G.M. FIUP. 1994. Historical and TURNER, M.G., R.H. GARDNER, AND R.V. O'NEILL. 1995. Ecologicalcurrent roles of insects and pathogens in eastern Oregon and Wash- dynamics at broad scales. P. 29-35 in BioSci. Supplement, Scienceington. USDA For. Servo Gen. Tech. Rep. PNW-327. and biodiversity policy.

KERsHAW, K.A. 1964. Quantitative and dynamic ecology. American WELLNER, C.A. 1989. Classification of habitat types in the westernElsevier Publishing Co., Inc., New York. United States. P. 7-21 inProc. Land classifications based on vegeta- r .

KEssELL, S.R., AND W.C. FISCHER. 1981.Predictingpostfireplantsucces- tion: Applications for resource management, Ferguson, D.E., P.sion for fIre management planning. USDA For. Servo Gen. Tech. Morgan, andF.D. Iohnson (eds.). USDA For. Servo Gen. Tech. Rep.Rep. INT-94. INT-257 " .I

KIMMINs, I.P. 1987. Forest ecology. Macmillan, New York. WIENS, I.A., AND B. T. MILNE. 1989. Scalmg of .landscapes m I~d-KRAJINA, V.I. 1965. Biogeoc1imatic zones in British Columbia. Ecol. scape ecology, or, landscape ecology from a beetle's perspectIve.

West. N. Am. {Univ. of British Columbia, Vancouver)1:1-17. Landscape Ecol. 3:87-96.

-

102 WIAF 11(3) 1996