EcologicalStewardshipA Common Referencefor Ecosystem Management

Volume !!

• BiologicalandEcologicalDimensions• HumansasAgentsof EcologicalChange

Editors

R.C. Szaro, N.C. Johnson, W.T. Sexton & A.J. Malk

A practical reference for scientists and resource managers

.¢

ELSEVIERSCIENCE_

ELSEVIER SCIENCE Ltd

The Boulevard, Langford LaneKidlington, Oxford OX5 1GB, UK

© 1999 Elsevier Science Ltd, All fights reserved.

This work is protected under copyright by Elsevier Science, and the following terms andconditions apply to its use:

PhotocopyingSingle photocopies of single chapters may be made for personal use as allowed by nationalcopyright laws. Permission of the Publisher and payment of a fee is required for all otherphotocopying, including mulliple or systematic copying, copying for advertising or promotionalpurposes, resale, and all forms of document delivery. Special rates are available for educationalinstitutions that wish to make photocopies for non-profit educational classroom use.

Permissions may be sought directly from Elsevier Science ILights & Permissions Department, POBox 800, Oxford ©X5 1DX, UK; phone: (+44) 1865 843830, fax: (+44) 1865 853333, e-mail: per-missions@elsevier.co.uk. You may also contact Rights & Permissions directly through Elsevier'shome page (http://www.elsevier.nl), selecting first 'Customer Support', then 'General Inform-ation', then 'Permissions Query Form'.

in the USA, users may clear permissions and make payments through the Copyright ClearanceCenter, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; phone: (978) 7508400, fax: (978)7504744, and in the UK through the Copyright Licensing Agency Rapid Clearance Service(CLARCS), 90 Tottenham Court Road, London W1P 0LP, UK; phone: (+44) 171 631 5555; fax:(+44) 171 631 5500. Other countries may have a local reprographic rights agency for payments.

Derivative Works

Tables of contents may be reproduced for internal circulation, but permission of Elsevier Science isrequired for external resale or distribution of such material. Permission of the Publisher is requiredfor all other derivative works, including compilations and translations.

Electronic Storage or UsagePermission of the Publisher is required to store or use electronically any material contained in thiswork, including any chapter or part of a chapter. Except as outlined above, no part of this workmay be reproduced, stored in a retrieval system or transmitted in any form or by any means,electronic, mechanical, photocopying, recording or otherwise, without prior written permission ofthe Publisher. Address permissions requests to: Elsevier Science Rights & Permissions Depart-ment, at the mail, fax and e-mail addresses noted above.

Notice

No responsibility is assumed by the Publisher for any injury and/or damage to persons or propertyas a matter of products liability, negligence or otherwise, or from any use or operation of anymethods, products, instructions or ideas contained in the material herein. Because of rapid advancesin the medical sciences in particular, independent verification of diagnoses and drug dosages shouldbe made.

First edition 1999

Library of Congress Cataloging in Publication Data26.catalog record from the Library of Congress has been applied for.

British Library Cataloguing in Publication DataA catalogue record from the British Library has been applied for.

ISBN: 0-08-043206-9 (Set: Volumes 1-11I)

(_) The paper used in this publication meets the requirements of ANSt/NISO Z39.48-1992(Permanence of Paper).

Printed in The Netherlands.

353

Principlesfor EcologicalClassification

Dennis H. Grossman, Patrick Bourgeron, Wolf-Dieter N. Busch,

David Cleland, William Platts, G. Carleton Ray,

C. Richard Robins, and Gary Roloff

Key questionsaddressedin this chapter

• Reviewof historicandcurrentapproachesto ecologicalclassification

• Relationshipbetweenecologicalclassificationsand managementobjectives

• Strengths and weaknessesin the applicationsof current ecologicalclassificationsystems

• Roleof dataandquantitativemethodsfor the developmentof ecologicalclassifications

• Conceptualandpracticaldifferencesin the developmentandapplicationof ecologicalclassificationsfor terrestrial,freshwaterandmarineecosystems

0 Challengesfor the future development of integrated ecological classificationapproaches

Keywords: Biogeography, ecoregions, vegetation, watersheds, climate, geology

354 D.H.Grossmanet al./Principlesfor EcologicalClassification

l INTRODUCTION integrative, multi-factorial, and synthetic. Consequent-ly, it is clear that a paradigm shift is required in how

The principal purpose of any classification is to relate classifications may become ecosystem-oriented, dyna-common properties among different entities to mic, and process-oriented. Also, they must evolvefacilitate understanding of evolutionary and adaptive according to increases in knowledge and to the evolv-

processes. In the context of this volume, it is to facilitate ing needs of ecosystem management.

ecosystem stewardship, i.e.,to help supportecosystem A fundamental question is whether a commonconservation and management objectives, classification system can be developed that would meet

This chapter has three purposes. The first is to tire needs of all ecosystem managers. The answer

provide a broad, scientific overview of the theory and seems to lie in the ecosystem concept and the clari-methods of ecological classification• The second is to fication of what is "common" to all ecologically based

review past and current efforts to shed light on the classifications. For the complex variety of conservationcharacteristics of current classifications and how they and management questions that will always exist it

have evolved. The third is to provide the scientific would appear that a variety of biophysical classifica-

foundation for applying ecological classification to I tionswill alwaysby needed.resource planning and management in the ecosystem This chapter attempts to respond to the challenge ofcontext, with a primary emphasis on new paradigms developing common ecosystem properties for ecolo-

• for classification. In all three aspects, the focus is on the gical classifications. It is structured in five sections.

United States because the intent is to help meet the goal After providing a definition of ecological classification,

of improved ecological stewardship of federal lands below, Section 2 describes the theory, conceptual ap-and waters, proaches, and methods that have been developed, and

It is now widely accepted that federal management that continue to be the basis for, ecological classifi-

activities should take an ecosystem approach. The goal cation. Section 3 is a summary of existing classifications

ef this approach to management has been defined by for the major ecological realms terrestrial, fresh-the IEMTF (1995) as follows: "to restore and sustain the water, and coastal-marine. It will become apparent that

health, productivity, and biological diversity of eco- each of these realms has been historically treated quite

systems and the overall quality of life through a natural differently, conceptually, and methodologically. Inresource management approach that is fully integrated some cases, management applications are made diffi-with social and economic goals. This is essential to cult by these contrasting approaches, as for examplemaintain the air we breathe, the water we drink the when aquatic systems are subsumed within terrestrial,food we eat, and to sustain natural resources for future floristic provinces.

populations. "e Section 4 describes how classifications may be

To this somewhat a.nthropocentric definition, one derived by means of multi-factor, integrated methods.may add the importance of maintaining entire land- This approach is required for ecosystems, which are

scapes (e.g., Franklin 1993), the conservation of evolu- complex, hierarchical, integrated systems involvingtionary and ecological processes, and the protection of both biotic and abiotic elements. Section 5 summarizes

species and ecosystems in protected areas (e.g., Grum- major existing classification systems and their applica-bine 1994). Virtually all of the approaches have explicit tions. Section 6 presents conclusions and future needs.

or implicit requirements in common for the application This chapter is written in recognition that no ecosys-

of the l_est of theoretical science, and for clearly articu- tem classification system yet exists that truly integrateslated conservation and management goals. Historically, all ecosystem attributes, nor that can suffice for allthese two requirements have not necessarily been plants and animals, nor for all lands and waters, nor

congruent; in fact, they have often been at odds. Yet that can fully describe or predict ecological change.only through the alignment of ecosystem concepts with Notwithstanding these limitations, our intent has been

realistic management goals may the necessary research to cull out the essential elements of the subject withand monitoring activities emerge to meet the challenges future needs in mind, trusting that the reaaer will referinvolved in sustaining productive ecosystems, with to the literature cited for further insight and

their full complement of biological diversity, information.Solutions to problems related to the common inter-

ests of science and management directly affect ecologi- 1. l What !$ Ecological Classification?

cal classification. In the past, problems were expressed

narrowly according to site- or resource-specific goals. The objective of any classification is to group togetherThis led to narrowly defined classifications specific to sets of observational units based on their common attri-

those problems. The ecosystem approach is broad and butes (Kent and Coker 19921. Thus, ecological classi-

Biologicaland EcologicarDimensions 355

fication refers to the development and characterization features, as sole descriptors for those units. If coastalof a set of units that are assigned by analyzing environ- systems are to be classified, they must be fully based on

mental and biological variables. The variables that can the interactions of land, sea, and atmosphere, whichbe used for classification are determined by available define the coastal zone.

data and other resources. The end product of a classi-

fication is a set of groups derived from the units of |.2 Classification Must be Driven byobservation where units within a group share more Objectiveattributes with one another than with units in other

groups. The ecological stewardship implications are Different approaches to ecological classification havethat all representatives of an ecological unit should been developed to address specific objectives. As re-

respond predictably to both natural changes and viewed in this chapter, numerous ecological classifica-resource management practices (Driscoll et al. 1984). tion systems have been applied in many different areas

Four important historical facts have set the stage for over the years. There is no one correct way to classify

present-day ecological classification. First, classifica- ecosystems; the success of an approach is measured bytion in the past has been dominated by static approach- its ability to meet management and/or scientific object-es; ecological dynamics were poorly understood and ives, preferably both when they exist. Classifications

environmental change was generally not considered, have primarily been constructed for three purposes: (1)

This is especially apparent on maps, ,,,,'here the poly- to help develop and represent the science of biogeo-

gons do not represent spatial, environmental and graphy itself; (2) to help identify "representative"temporal dynamics. Second, by far the majority of ecosystems for protection, restoration production,work on ecological classification has been land-based, management, research, monitoring, and inventory,

resulting in the omission of aquatic features. Third, most notably of living resources; and (3) conservation

aquatic systems have been classified generally by of biological diversity. With respect to managementmeans of physical features, such as temperature, itself, many possible applications exist for classifica-

salinity, topographic features that define watersheds, tions, of which Frayer et al. (1978) have identified four:the dendritic structures of river systems, estuarine (1) a basis for cataloging the status of current resources;

structure, and the like. This is in strong contrast to the (2) a means of transferring experience and knowledge

more biotic approach of land systems. These segre- of a studied area to a similar but unstudied area; (3) a

gatedtreatmentsofterrestrialandaquaticsystemsdefy framework for assessing local management opportu-attempts at integrating these ecosystem classification nities and predicting the outcomes of treatments orapproaches, as is obvious in the case of omitting rivers actions; and (4) a vocabulary for communication

from terrestrial biotic provinces or not considering between managers, between managers and re-land-sea interactions when describing coastal searchers, and between managers and the public.

classifications. (An assumption for the land has been The principal factor that controls the relationship

that vegetation is a surrogate for other biota. An between classification and objectives is scale. Forassumption for the sea is that zoogeography or example, ecological classification requirements for

physical features can be used to define oceanographic songbirds (functioning at the vegetation stand level)regions, or vice versa. These assumptions have met are different than those for amphibians (functioning atwith uncertain results.) the micro*habitat level). However, under a properly

Fourth, a host of classifications has resulted from designed system, ecological units that are appropriate

both the fragmented approach of management and the for management objectives related to amphibians can

disciplinary approach of science. As a result, some be aggregated into or nested within units meaningfulclassifications are defined at a single level, whereas for songbird management. Thus, the first step in ecolo-others contain multiple levels, which may or may not gical classification is to determine the spatial scales

be hierarchical. Most classification systems, and levels needed to address management objectives.

within classification systems, are associated with a Historically, management objectives have empha-particular geographic area and with specific spatial sized single resources as separate entities in the

scales for implementation, landscape or seascape. Contemporary land managers.It is now abundantly clear that biotic-abiotic however, are askingnew sets of questions that require

terrestrial-aquatic integration is required for an eco- integration of multiple factors. These new questions

system approach to classification. If ecological classi- include:fication systems are generally to be used to delineate

sets of biophysical units across landscapes, they must • What are the cumulative effects of a managementgo beyond climate, floristics, zoogeography or physical activity on multiple components of a system?

356 D.H. Grossmanet al./Principlesfor EcologicalClassification

• Will a management activity disrupt the inherent most important contemporary advances have utilized

ecological processes and functions of the system? the greater availability of biotic and abiotic data and

• What level of a management activity',','ill direct the improved spatial technologies to create multiplehierarchical classifications that can be applied to

system into an aberrant ecological trajectory?different objectives at multiple scales. What may be

Such questions are the essence of ecosystem steward- described as "local" vs. "regional" differs markedly

ship because they place resource management in the i among various systems. For the land, the difference iscontext of multifaceted, hierarchical, ecological rela- apparent. But for rivers, the watershed is the usual

tionships. This is a particular challenge for the inte- "regional" frame of reference, with "local" referring togration of science and management, as science is not third- or fourth-order tributaries. For coastal systems,

able to give answers with high degrees of certainty to "regional" is the biotic province, usually defined by the

such questions. An instructive example concerns ranges of endemic fauna; "local"refers to physical sub-fisheries management, which has been treated from units such as estuaries, bays, tributaries, and coastal

the point of view of simple "yield" models, out of formations. For open marine waters, including bothcontext of ecosystem properties. A growing concern is coastal oceans and ocean basins, "regional" may refer

that of sustained ecosystem functions and processes to zoogeographic provinces or water masses; "local" isover time in light of management actions (question 3). usually referred to as "sample sites" or "patchiness."

Orians (1975) defines the following measures of Ecological classifications have historically beenecosystem"stability": (1) constancy, (2) persistence, (3) based on abiotic, biotic, or integrated approaches.

inertia, (4) elasticity, (5) amplitude, (6) cyclical stability, Abiotic classifications have defined sites and mapping

and (7) trajectory stability. Which measures are units by singular or multiple components of the abiotic

operational can be very difficult to identify, but these environment. ;Fhese components include land formsterms define a way to approach ecological classification (Hammond 1964), physiographic provinces (Penne-from a functional-process viewpoint, man 1928), and climate (Trewartha 1968), along with

other variables such as geology, watersheds, and ele-vation. Biotic classifications have primarily been

2 HISTORICAL OVERVIEW OF ECOLOGICAL developed to portray vegetation classes and animalCLASSIFICATION habitats.

Examples of terrestrial biotic classifications include

Biogeographic classification has a very long history, potential natural vegetation (Kiichler 1964), vegetationwith strong links to biological diversity and physiognomy (UNESCO 1973, Grossman et al. 1994b),

biogeography. Since the time of the Greek scholars Society of American Foresters cover types (Eyre 1980),

Aristotle and Theophrastus during the third century plantcommunities(Grossmanetal. 1994b),andnaturalB.C., and of the Latin scholar Pithy the Elder during the vegetation of the Southwest (Brown et al. 1980).Hybrid

first century A.D., descriptions of species have been approaches have integrated abiofic and biotic compo-accompanied by classifications, however rudimentary, nents to classify and map actual or potential land units.

and data on site locations. Over subsequent centuries Examples of integrated approaches include habitat

to the present, many questions concerning the charac- types (Krajina 1965, Pfister et al. 1977), ecoregionsterization of biological species and communities, (Bailey 1976, Omernik 1987), forest regions (Braunincluding their relationships to environmental and 1950, Rowe 1972), ecological land units ('Haufler 1994'1,

latitudinal gradients, have emerged. As a result of and wetlands tCowardin et al. 1979).

rapid discoveries and the increasingly disciplinary In Europe, land use planning assessments promp-

nature of the inquiry, many different approaches to ted the development of ecological classification in thebiological and ecological classification have been early 1930s. Ecological, as opposed to purely biotic.developed, along with a massive literature, classifications attempt to interpret the ecological mean-

Historical approaches to ecological classifications ing of the distribution of biotic types by relating this

may be broadly characterized as being either locally distributior, to one or more features of the era'iron-specific or regionally simplified. For local classifica- merit. Tools suct7 as "ecological vegetation maps,"

fions, there has generally been a highly focused object- which not:ed the environmental features believed to beb,'e and a large quantity of field information; this has indicated by the vegetation, were used for large-scale

resulted in classifications and maps _hat are specific assessmentsofthesuitability of sites for par_cular uses

and detailed. In contrast, regional treatments have ,Zonneveld 1988). Similarly. the North American con-tended to generalize data from few points over large cept of "ecosystem" became the basis for evaluatingareas and, thus, had limited use at a local level. The suitability in that site characteristics for a given

Biological,Ind EcologicalDimensions 357

mapping unit were used to determine appropriate nentalshelvesoftheworld. Thiszoneisnowestimated

land use categories and for other resource manage- to include sea level plus and minus 200 m. It en-

ment needs. In the United States from 1900 to the late compasses 18 percent of the earth's land surface and 8

1960s, resource management was mainly concerned percent of the oceans" surface (but less than 0.5 percent

with protection from catastrophic events (e.g., fire, of the oceans' volume). However, its importance isfloods), and with the efficient extraction of commodi- disproportionate to its size: it supports 60 percent ofties. Throughout the I970s and I980s, however, an the human population, and accounts for a quarter ofincreased environmental awareness on the part of the the global primary productivity and more than 90

public and within government confirmed and percent of fisheries (LOICZ 1996). The recognition of

strengthened the need for structured, ecological data the coastal zone warrants a new ecological classifi-assessment tools. Single-factor classification systems cation that not only transgresses traditional land--sea

were ineffective at addressing these "nev,," issues and boundaries, but also functionally integrates terrestrial,

did not capture land functions and processes. As a marine, and atmospheric processes.result, multi-factor ecological classification systems The need for a global classification of coastal andthat recognized system inputs, outputs, complexity, marine systems became especially clear when IUCN

and inter-connectedness were developed to address and UNESCO recognized during the 1970s that there

multiple-use and biodiversity issues. ",','as no comprehensive, global classification for con-Terrestrial classification systems have historically servation purposes. Udvardy's (1975) "world" classifi-

been given more attention and are considerably more cation was intended to meet that need for the land, butadvanced than other classification systems. As a result, this "world" did not include freshwater and marine

aquatic systems tend to be included within terrestrial systems. Consequently, IUCN and UNESCO support-

systems. For example, Bailey (1995) described ed an effort to develop a matching, coastal-marine

"eeoregions" on the basis of land surface, climate, classification. The result was Hayden et al. (1984),vegetation, soils, and fauna. This approach neglects which reviewed the"state-of-the-art" and presented awatersheds and their dynamics and even the existence tripartite scheme for comparing coastal geomorpho-

of the coastal zone, which includes a major portion of logical provinces, coastal biotic provinces, and marinethe land surface. Additionally, Cowardin et al. (1979) realms. It was recognized that this scheme needed

described wetland and deepwater habitats in significant extensions, particularly into the third di-traditional, typological ways that did not take into mensionofdeeperoceanwatersandintosmallerscales

account the dynamic approaches employed by coastal of interaction. However, to date, this global treatment

biologists and oceanographers, has not been expanded.Coastal-marine physiographic and biogeographic In summary, most work on ecological classification

classification has its own long history, perhaps dating and the related field of biogeography has focused on

from the scientific voyages of the 19th century. Modern specific systems or elements of systems. There are

zoogeography was initiated by Ekman (1953) and manyexamples of classifications that represent specificHedgpeth (1957), who took zoogeographic and taxa, landforms, vegetation and climate, physical geo-

physical-ecological approaches to sub*dividing the graphy, the phytoplankton of oceanicwaters, estuariesnear-shore seas and oceans into now-familiar zones and habitats of continental shelves, coral reefs, and

(from littoral to abyssal). Somewhat later, Briggs (1974) other factors. It has become increasingly clear thatre-evaluated the zoogeography and proposed a system many of the challenges presented by management of

of coastal biogeographic provinces primarily based on whole ecosystems cannot be addressed by these frag-endemic species. This method allows clear identifi- merited classification approaches.

cation of province divisions, but does not create arelational basis among provinces for evaluation as all 3 ECOLOGICAL CLASSIFICATION: THEORYprovinces defined by endemism become, by definition, AND bfETHOD$unique. Nonetheless, Briggs's approach does confirmdivisions that have long been apparent among coastal From this brief history, it is clear that ecosystem man-

biotic provinces (e.g., as given for North America in agement requires a reversal of trends, in both scienceRobins and Ray 1986, Robins 1992). and management, _way from fragmentation and

A major, relatively recent development has been the reductionism. The Ecological Society of America fESAIrecognition of the "coastal zone" as _ distinct earth has led attempts to define concepts for holistic man-

realm where land, sea, and atmosphere uniquely inter- agement application, e.g., for "sustainability" (Lub-act. Ketchum (1972) essentially defined this zone to chenco et al. 1991j. More recently, the ESA has reportedinclude the entirety of the coastal plains and conti- on the scientific basis for ecosystem management,

',r I fr_11f-_-_ --_---" .............

358 D.H. Grossmanet aL/Ptinciplesfor EcologicalClassification •

defined by: "...explicit goals, executed by policies, causal factors (Urban et al. 1987, Bourgeron and Jensenprotocols, and practices, and made adaptable by moni- 1993, Jensen et al. 1996). Once a correlation or a cause-

toring and research based on our best understanding of effect relation between pattern and process is deter-

the ecological interactions and processes necessary, top mined, predictions are made using the summa D' ofsustain ecosystem composition, st_acture, and function" data and information performed during the classifi-(Christensen et a1.1996).This amplifies the definition of cation process and/or the generation of maps, statisticalthe IETMF (1995) to include specific scientific goals, as or simulation ecological models, and/or a combination

well as social ones. Among the elements included in of the first two approaches. Such interpretations are

ecosystem management, as defined by this ESA report, reasonably well developed for terrestrial systems, butare: sustainabi]ity, ecological modeling, adaptability and are only at a primitive stage of development for coastalaccountability, spatial and temporal scale, ecosystem (Ray 1991) and marine (Steele 1989, Ogden et al. 1994)function and dynamics, ecosystem integrity, and the systems.uncertainties of our knowledge.

3.1.2 Syscem Dynamics3. ! Theoretical Basis for Ecological

Classification Most existing ecological classifications are concerned

with patterns or static structures. However, ecosystemsIn this context, it is abundantly clear that most present may exhibit several trajectories (Orians 1975), and futureclassifications are incomplete and possibly even mis- ecosystem classifications cannot ignore these dynamics.

leading in cases where the approaches are based on A new approach is that of dynamic biogeography,narrowly defined ecosystem components. To address which merges the large-scale approaches of traditional

future ecosystem-oriented needs, seven areas of biogeographywithsmallersca]e approachesofecology.ecological theory need to be considered: (1) pattern Dynamic biogeography concerns the study of biological

recognition; (2) system dynamics; (3) hierarchy theory; patterns and processes on broad geographical and time(4) systems limits; (5) the biotic component of eco- scales, and represents spatial patterns at different scalessystems; (6) the abiotic component of ecosystems; and of variation (Hengeveld 1990).(7) ecosystem characterization, Each of these areas is Ecological classifications are used for many put-

described in greater detail below. The treatment of poses, such as assessing spatial relations and compara-

these issues affects the selection of the concepts rive featuresofecologicalsystems. However, ecologicalguiding ecological classifications, the most appropriate systems exhibit temporal changes along various dev-data for such classification, the best techniques for elopmental pathways that result in different types of

classifying ecosystems and their components (e.g., organization. Many conceptual models used for plan-assemblages of species, landscapes, seascapes), and the ning purposes assume that all ecosystems reach a state

best methods for predicting ecosystem properties and of equilibrium or quasi-equilibrium (e.g., the assump-their responses to various management scenarios, finn that succession leads to climax). However, we now

know that ecosystems are rarely in equilibrium.3. |. | Pattern Recognition Although it is convenient for scientific (e.g., modeling)

and planning purposes to assume equilibrium at a

To describe ecological systems, pattern recognition tech- given scale, we must recognize that ecological systemsniques are often employed. But because ecosystems are are fully dynamic (e.g., Kay 1991,Constanza et ah 1993,.

complex systems, defined by exchanges of energy, Thus, ecological classifications at all scales need to bematerials, and information, which may be described at designed to account for change, which may be slow ormany different scales (Levin 1992), their boundaries rapid, in terrestrial and aquatt¢ systems.

may be difficult to define and may even be consideredarbitrary. Nevertheless, boundaries are determinable, 3,l.3 Hierarchical Organizationfor example, by recognizing gradients or "'ecotones."which also provide a context for understanding Complex ecosystem patterns, and the multi_de ofecological function (Hansen and di Castri 1992). processes that form them. exist within a hierarchical

Ecological classification requires that spatial and framework (Allen and Start 1982, Allen et M. 1984temporal relationships need to be defined clearly O'Neill et al. I986,,as is readily observed inlandscapes(Bourgeron and Jensen 1993, Jensen el al. 1996). or seascapes. In recent years, considerable attention

Furthermore, ecosystem management requires making has been directed to describing this hierarchical

predictions about ecological systems, that is, determin- organization. Hierarchy theory (Allen and Starr 1982,ing relationships between patterns and hypothesized O'Neill et al. 1986) views multi-scaled systems as a

Biologica[andEcologicalDimensions 359

series of constraints in v,,hich higher levels of organiza- functioi'_s at an "optimum." For example, it has been

tion providetheenvironmentinwhich thelowerleveIs assumed that, eventually, ecological succession

function. However, constraints are not necessarily continues until species stop replacing each other, at

"top*down," as Holling et al. (1993) have described, which point the processes that lead to change balanceEcosystems at any level may cycle through functional the processes that lead to ecological organization. At

states of "exploitation," "conservation," "release," and this stage, called "climax," the ecosystem presumably"reorganization," in which both "top-down" and would be operating at "optimum." Yet actual field

"bottom-up" interactions are evident, observations reveal that this point does not remainFour tenets of hierarchy theory are critical to under- constant. Environmental conditions may change (e.g.,

standing landscape patterns and their dynamics, and because of local, regional, or global climate change),

to using that knowledge for ecosystem management new species may appear, and, even in the absence of(Allen et al. 1984; O'Neill et al. 1986). major disturbances such as regional fires or coastal

1. Every component of a system ecological or other- storms, the "climax" will change to a new community.wise, is both a whole and a part at the same time. The result is an"altered ecosystem state" with different

This concept is called "whole/part duality." For ex- optimal conditions (Hayden et al. 19_?1).

ample, a forest (a whole) is made up of trees (the Obviously, the concept that optimal limits exist for

parts), each of which would consist of multiple any ecological system deserves re-examination. Thecommunities for many organisms. At a larger spa- observed state of an ecosystem may be only one oftial scales, the forest is part of a larger scale re- several possible and, hence, that the concept of an

gional landscape. "optimum operating point" at "climax" may be mis-leading. It has been assumed that if an ecological svst-

2. Patterns, processes and their interactions can bedefined at multiple spatial and temporal scales, em fo ows one path rather than another in response to

These scales need to be identified clearly, accord- disturbance or environmental change, it may operate

ing to the question being asked or the function be- far below its optimal level. An example is the mainte-nance of prairie grasslands under the natural dis-

ing examined, turbance regime of low intensity, high frequency fires.

3. There is no single scale of ecological organization When such fires are suppressed, succession leads to the

that can be used for all purposes. This is important development of a forest stage (e.g., Arno and GruelIto consider because ecological systems are often 1985). Thus, the conclusion was that the ecosystem

interpreted at a single scale or at a limited number functioned at suboptimal levels under presettlement

of scales, disturbance regimes. On the other hand, disturbance is

4. The definition of the component patterns and pro- known to increase the biological diversity of some

cesses of any particular ecological hierarchy is dic- systems, notably tropical rain forests and coral reefstared by the objectives of the study or by the (Connell 1978), and recent research has shown that an

objectives of management, increase in biological diversity may increase the

Most classifications are static and "taxonomic," where- productivity of ecosystems (Naeem et al. 1994).Ecological classifications need to be designed to

as hierarchy theory challenges them to become dynam- reflect these factors, especially if they are to be used inic and integrative. The "whole/part duality" of eco-systems (Koestler 1967,Allen and Starr 1982.Allen et a]. ecosystem management or assessments. This is esp-

ecially important because many of the ecological units1984_ clarifies ecological classifications because that require protection and management do nothierarchy theory expresses how different levels of eco-

logical organization are linked. The findings at any one represent late successional or "climax" stages and/orlevel may assist the understanding of another level, but are not in a state of equilibrium.

can never fully explain phenomena occurring at other 3. ! .S Biotic Componentlevels (Odum 1971). That is. each level of a hierarchy

has characteristics that can only be explained with Species patterns over time and space may be used to

knowledge of other levels, define the biotic component of ecosystems in two

3. 1.4 System Limits ways: by means of the continuum or community con-cepts. The continuum concept states that species ass-

As noted above, ecological systems mav follow emblages are temporary and fluctuating phenomena

different trajectories of change. Much attention has along regional gradients. In contrast, the communitybeen given to whether natural change is limited to concept maintains that repeatable assemblages ofapproaches to a "climax" state in which an ecosystem species occur in discrete habitats with characteristic

360 D.H. Grossmanet al./Princtplesfor EcologfcalClassification

properties. The conflict between these two approaches and conduct landscape assessment using such corn-has far-reaching implications for ecological classifica- munities can be successful only if the intra-community

tion, For example, if the continuum concept is correct, pattern of gradual change is correlated with gradualthe biotic component of ecosystems cannot be changesin tile enviromnent. Therefore, effectiveland-

classified because species respond individualistically to scape surveys need to take into account this depend-environmental change over space and timff. Most ence of biotic patterns on abiotic variables (Bourgeronresearchers implicitly accept the continuum concept, et aL 1993).

even avoiding the term community and referring to The extent to which regional climatic/floristic

the more neutral term "assemblage" (Austin 1991). The classifications may apply to the fauna or to aquaticsame researchers, cartographers, and other practi- systems is highly dependent upon the strengths of the

tioners, however, continue to recognize homogeneous ecological linkages or dependencies among these diff-units, hence implicitly using the community concept erent components. It would be incorrect to assume, for

for pragmatic management purposes, example, that different taxa respond in the same way toIn defining the biotic component of ecosystems, a particular set of environmental variables. Bibby et al.

finding a practical approach has faced major limita- (1992) reported that areas of high endemism for birdstions. Weak and ambiguous classifications have re- are not the same as those for other vertebrates. Fresh-

suited from unspecified or inconsistent criteria, vague water fishes are distributed according to'the confines ofdefinitions of key concepts, unspecified minimal areas watersheds, not floristics, and aquatic-coastal biotic

of reference, and undocumented sorting strategies provinces (Hayden et al. 1984) appear only roughly(Whittaker 1978, KLichler and Zonneveld 1988). In related to adjacent terrestrial provinces (Bailey 1995)_addition, existing data have not supported rigorous

testing among the various continuum alternatives _3.|.6 Abiot|c Component(Austin 1991). Nevertheless, various aspects of both thecontinuum and the community views appear to Two major categories of abiofic environmental

complement rather than exclude each other (Westhoff variables have been useful for ecological classificationand van der Maarel 1978, Whittaker 1980, Austin 1991). (Austin et al. 1984, Austin 1985, Austin and Smith 1989):

It has been shown, at least for terrestrial vegetation, • indirect factors -- known to be correlated to direct

that species can be individually distributed along factors that exert physiolog'_calinfluence on species

gradients (Austin 1987, Austin and Smith 1989) and that (e,g., elevation, bathymetry);the distribution pattern of controlling environmental

factors constrains the pattern of species combinations, ' direct factors -- known to have a direct physiologi-their distribution in the landscape, and their frequency, cal influence on species (e.g., temperature, pH, nu-

Even if some form of the community concept is trients).Both of these have been variouslybeen usedaccepted, the problem of identifying the full set of for classification of terrestrial, freshwater, and ma-environmental factors shaping the composition of the rine systems.

biotic community remains. For terrestrial classification, Numerous assumptions are made in developing ecolo-the concept of a floristlc or biotic province is defined gical classifications that involve the abiotic components

within the context of a region with a distinct pattern of of ecosystems. The first maintains that biological corn-climate and landscape characteristics (Bailey et al. munities are surrogates for the environment. This

1994). This is because such classifications are useful for approach, For example, uses tlle vegetation as adefining regional ecosystem management guidelines surrogate for the environment as a whole based on the

only if vegetation units, defined as biotic or floristic assumption that vegetation is a faithful expression ofprovinces, or any variant (e.g., Kfichler 1967, Brown et site characteristics _Troll 1941, 1943, 1955, 1956; K'achleral. 1979) are modified to correlate with climatic/land- 1988). Therefore, vegetation maps have been central to

scape regions (Burger 1976, Rowe 1980, Bailey and describing ecological patterns (Whittaker 1980). 8imi-Hogg 1986, Ki.tchler 1988). A primary purpose of such larly, for deep oceanic systems, plankton diversity has

regionally defined ecosystems is to serve as a reporting been used for decades as an indicator of environmentalstructure for information about regional resources and conditions for the delineation of ocean zonesenvironment (Bailey and Hogg 1986, Bailey et al. 1993). (McGowan 1971. Dunbar 19791.

Another purpose is to define homogeneous regions The second assumption is that ecological unitswithin which to characterize finer scale ecosystems contain recurrent patterns and characteristics that can

and/or landscape properties (Westhoff and van Der be delineated. This approach uses broad enwronment-Maarel 1978, Austin and Smith 1989). Similarly, al patterns alone to describe and delineate abiotic

attempts to characterize landscape level ecosystems elements in both terrestrial (Rowe and Sheard 1981,

Bi0IogicalandEcologicalDimensions 361

Zonneveld 1989) and coastal s),stems (Hayden and agents of pattern formation at all appropriate eco-Dolan 1976). logical and planning scales may then be defined. The

The third assumption is that environmental proper match of patterns and agents of patternpatterns may be integrated with biotic patterns to de- formation is important because any incorrect coupling

scribe and delineate habitats of both plant and animal of pattern to agents impedes the ability to make pre-communities. For terrestrial systems, classifications dictions about the future state of the ecological system.have been developed using climatic attributes either As a result, the value of conservation and management

alone or in conjunction with other attributes (e.g., actions and prescriptions is limited.Bailey 1976, Austin and Yapp 1978, Walter 1985,

Omernik 1987, Bailey et al. 1994). For aquatic systems, 3,2 Goals and Methods of Multivariatevery, different approaches are used for freshwater, Classificationestuarine, and marine systems (Section 3.2}.

A fourth assumption is based on the argument that, The introduction to this chapter defined the purpose of

to be meaningful, ecological evaluation should be ecological classification in the context of meeting the

based on species" niche-habitat relationships (e.g., goal of ecosystem management. Sections 2.1 and 2.2Hutchinson 1959, Whittaker 1972, Nix 1982, Brown described past approaches and presented some aspects1984). The aim Js to summarize environmental vari- of modern ecosystem theory as a setting for future

ability, identify the distribution of major environ- development. Clearly, classifications are nowchalleng-mental gradients, and indicate where significant shifts ed to become integrative, dynamic, and adaptive.in ecological variability might occur (Mackey et aL 1988, Accordingly, this section reviews some multi-factor

1989). To accomplish this, species' responses to a methods and their applications.limited set of dominant environmental variables that Developing ecological classification systems is an

comprise primary niche dimensions must be estimated iterative process invoMng field observation, statistics,

(Nix 1982; Mackey et al. 1988, 1989). Site-specific data and numerical modeling. At fine spatial scales, eco-are used to generate classes of sites sharing similar logical units are often identified in the field, based onranges of values of the environmental variables. A map such things as flora, soils, and physiography (Host et al.of these classes, or bioenvironments, can be used alone 1988). At coarser scales, data from remote sensing or

in the ecological assessment stage of an area for given other sources are commonly analyzed (Denton andpurposes (DeVelice et al. 1993), and/or in conjunction Barnes 1988). Classification units are refined as

with vegetation data for quantifying biotic-abiotic additional sampling and analyses are completed, and

correlations (Mackey et al. 1989). new relationships among components are identified.In all cases, the classification units arguably Additionally, sampling of ecosystem components aids

represent natural levels of ecosystem integration with in the recognition of such patterns as soil-vegetation

respect to environmental regimes and key processes - correlations. This information is then integrated intoand there is evidence that this is the case (Swanson et the classified units

al. 1988). For example, geomorphic pattern, through Once large data sets are on hand. a variety of multi-erosional/sedimentation processes, controls carbon, variate methods may be employed to identify' patterns

nitrogen, and phosphorus cycles in soils of riparian and to reduce the number of variables. Multivariateforests in southern France (Pinay et al. 1992). methods have been used to detect patterns in different

types of data, including overstory basal areas, ordinal3.1.'7 Ecosystem Characterization ground flora coverages, and nominal soil texture

classes CCleland er al. 1993}. Also, statistical methods

In the field, various ecological classifications are often can be employed to simplify data sets by extracting

used together to delineate the boundaries of ecological their "principal components," as has been done for

systems. For example, a classification of the vegetation estuaries by Ray er al. (1997) through the analysis ofmay be used with classifications of land forms and NOAA's (1985) large data set. Several methods havedisturbances. The process of using ecological classifica- been widely applied in recent years to the analysis of

tions for matching patterns and processes is called I flora, fauna, plankton, abiotic features, and whole

ecosystem characterization (Levin 1992). This processis J environments (e.g., estuariest, especially where it isused for mapping (e.g., Avers et al. 1994, Zonneveld necessary to condense and summarize extremely large1989) and landscape evaluation from the site to the data matrices.continent level In defining ecological land types (e.g., Multivariate ecological classifications for terrestrial

Wertz and Arnold 1972), the idea is to draw boundaries I and aquatic environments have changed dramaticallyaround ecological systems. Relevant patterns and I over the past few decades. Historically, analysis was

362 D.H.Grossmanet aL/Principlesfor EcologicalClassification

carried out through the tabular analysis of plot data. analysis techniques may be applied to the same data

Braun-Blanquet's procedure of 1921 used tabular me- set. The communication of results is promoted bythods in successive approximations to identify groups employing a moderate number of commonly used,of species occurring in similar samples, and to identify relatively standardized methods (Pielou 197_.samples with similar species composition. These early

tabular classification techniques were informal and in- 3.2,2 Clusteringherently subjective (Whittaker 1962, Mueller-Dombois

and EUenberg 1974). As a result, recognizing different The objective of clustering is to identify naturallyspecies groups as well as groups of similar samples occurring groups based on all variables in a data set.

depended heavily on the individual investigator's Both the process and the choice among techniques are

understanding of species-species and species--environ- more complex and more subjective than those of ordi-mental relationships withir_ a study area. The results nation. The most commonly used clustering methodshave been variously expressed in tables, dendrograms, are: (1) non-hierarchical, (2) polythetic hierarchical

and maps. agglomerative, and (3) polythetic hierarchical divisive.The past few decades have seen the proliferation on Non-hierarchical clustering assigns data to clusters,

more objective procedures and refined multivariate placing similar samples or species together. This is ananalyses, in some cases leading to models expressive of excellent way to handle redundancy and outliers, but is

ecosystem function (Mueller-Dombois and Ellenberg limited in its ability to analyze relationships. Hier-1974, Gauch 1982, Spies and Barnes i985, Denton and archical clustering also groups similar entities togetherBarnes I988, Hix I988, Host and Pregitzer I991, Bulger into classes, but additionally arranges these within a

et al. 1993). These procedures include explorative data hierarchy such that a single analysis may be viewed onanalyses involving descriptive statistics and graphical several levels, with relationships expressed among the

displays that are used to: (1) detect intercorrelations entities classified. The utility of a given technique isamong variables, thus reducing in the number of vari- judged in relation to others, and often several

ables that need to be considered; (2)check assumptions classification techniques are applied to the same dataabout the data structure underlying particular anal- sets with results compared afterwards.yses; (3) suggest appropriate transformations; and (4) The term "polythetic" means that information on all

identify sample outliers, variables is used to assign observations to a cluster, asopposed to earlier monothetic methods that used

J,2.1 Ordination single variables in a non-multivariate analyses. Poly-

thetic agglomerative clustering has two steps: (1) the

Ordination is a powerful technique used to identify samples-by-species data matrix is used to compute aimportant variables, to summarize data, and to reduce samples-by-species dissimilarity matrix using any ofthe complexity of the data set. The results may be several distance measures such as Euclidean distance

displayed diagrammatically or mapped. In ecological or percent dissimilarity; (2) an agglomeration pro-studies, ordination is also used to discover the latent cedure is applied successively to build up a hierarchy

structure of data by analyzing species' responses to ofincreasinglylargeclusters, startingwithdusterscon-underlying environmental gradients (Prentice 1977, sisting of a single member, and agglomerating these

Bulger et al. 1993). hierarchically until a single cluster contains all thePrincipal component analysis (Gauch 1982,Morrison samples or species. The polythetic hierarchical divisive

1976), correspondence analysis (Hill 1974, Greenacre method similarly computes the dissimilarity matrix in

1984), and det:rended correspondence analysis are the first step, and then successively applies a divisiveamong the most commonly used ordination techniques procedure to a single large cluster to create a hierarchy

in modern ecological studies. They employ different of individual members.methods to account for the variance in the data. The simplest polythetic-divisive classification is

Ordination is often used in an exploratory sense to ordination space partitioning (e.g., Noy-Meir 1973,detect trends as well as outliers, to screen variables, to Peer 1980). Two-way indicator species analysis (TWIN-

reduce dimensionality, and to summarize community SPAN) (Hill 19791is another polythetic divisive cluster-and environmental patterns. Ordination is often acc- ing technique. TWINSPAN begins with all species or

ompanied by clustering procedures to clarify natural samples (depending on the objectives) in a singlegroupings of sampled data. Results of ordination and cluster and divides these into smaller clusters by first

clustering may be compared, and subsets of data may ordinating data by reciprocal averaging. The reciprocalbe interrogated to elucidate relationships further and averaging procedure is repeated until each cluster has

to develop hypotheses. Several complementary nomore thanachosenmm_mumnumberofmembers.

BiologicalandEcobgicalDimensions 363

3.2,3 Summary classifica fions have the goal of determining the relative

degree of simiIarity and difference among units. How-Presently', ecological classification is often approached ever, it is apparent that none comprehensively

in an combined program of ordination, clustering, di- addresses the holistic character of ecosystems. This is

rect gradient analysis, and tabular synthesis of results, not to say that they are not useful, but tl_at integrated,A five-step procedure has become routine, especially ecological classification still lies ahead. It is also app-for terrestrial systems where procedures are most E arent that a classification is most useful to managementadvanced. The analysis progresses by successive stages when it can be mapped and related to biogeographyof refinement (Cleland 1996). First, exploratory data (e.g., Pielou 1979, Brown and Gibson 1983).

analyses are conducted to ensure that the assumptionsunderlying particular methods are met, and to reduce 4. I Terrestrial Classificationsthe number of variables. Second, ordination is used to

summarize community and environmental patterns. Most terrestrial classifications are biotic in nature,

Third, clustering is used to identify groupings of samp- although many incorporate references to physiognom-Ies and variables, and to corroborate patterns detected ic regions and climatic zones. Terrestrial classifications

through ordination. Fourth, community patterns are are often simpler than aquatic classifications because of

compared with environmental information to eluci- the comparably static nature of landforms anddate congruent changes and to produce an integrated vegetation. Also, more research has been completed ininterpretation of the ordination and clustering results, terrestrial systems than in aquatic systems. As a result,

Finally, hypothesis testing methods may be used a post- terrestrial classifications are more advanced in theoryeriori to assess the relationship between classification and in practice than are aquatic classifications.

and mapping as well as the ecosystem-level differencesin structure and function. Such differences might in- 4, 1. | Vegetation Classification

dude biomass production and productivity (Host et al.

1988), successional pathways (Host et al. 1987, Johnson Most terrestrial, biotic classifications have been based1992),and nitrogen dynamics(ZaketaL 1986, 1989). on vegetation. Beginning with the viewpoint of

Thus, multivariate methods are extremely useful in Gleason (1917, 1926) and extended by others (e.g.,

ecological classification because of the large number of Whittaker 1956, 1962; Curtis 1959), it was generallyvariables and observations commonly analyzed, the believed that vegetation units could not be defined.

difficulty in detecting patterns due to the complex inter- The approaches taken often became polarized betweenrelationships involved, and the need to verify, the taxo- the "continuum" and the "community unit" concepts,

nomic and spatial hypotheses represented by classifica- described above. But despite differing viewpoints,tion and mapping. Multivariate analyses alone, several features became widely recognized (Mueller-however, are insufficient to develop classification Dombois and Eflenberg 1974): (1) similar species

systems. Even though the analysis methods themselves combinations recur; (2) no two sampling units are

may be relatively objective, the selection of particular exactly alike; and (3) species assemblages change moremultivariate mefl_ods, decisions about how to stand- or lass continuously if a geographically widespread

ardize data, the selection of important variables or the community is sampled throughout its range. Thus,

removal of superfluous, redundant, or rare variables are recurring species combinations may be correlated withsubjective choices that can influence the resulting elassi- their environments, and these combinations shift geo-ficafion. In addition, the experience gained through graphically, meaning that a limited degree of orderingreconnaissance, observation, and field sampling, and zspossible.

through the thought processes required to develop anintegrated ecosystem classification is valuable and Potential Versus Existir_gVegetationshould ",eused to augment strictly mechanical methods.The role of multivariate analysis, therefore, is that of a Two differing approaches to the classification of

tool to be used in conjunction with knowledge of vegetation are the portrayalofexisfingversuspotentialecological relationships gained in the field, natural vegetation ,PNV). Classifications emphasizing

existing vegeration determine vegetation units fromthe current characteristics of the vegetation. Classifi-

4 ECOLOGICAL CLASSIFICATION SYSTEMS cations emphasizing potential natural vegetation usecharacteristics that represent the most mature

Various types of classifications are summarized in this conditions of vegetation development _,T_xen 1956,section, to illustrate their objectives and priorities, All K'3chler 1964_ It is, however, difficult to model the

l_r rI 7F711Tirlln1 _ r--l' _... ...... ,,_ '_, _, ......

364 D.H. Grossmanet al.IPrinciplesfor EcologicalClassification

relationships between existing vegetation and poten- approach, which in practice, defines formationstim natural vegetation. This modeling is limited by the defined by varied, conventionally accepted combina-current knowledge of vegetafion-site relationships, lions of growth-form dominance and characteristics ofand the ability of the observer to infer these relation- the environment.

ships (Cook 1995). These models also emphasize FlorBtic methods characterize the species them-hypothesized climax vegetation, a concept that is selves. Early 20th century ecologists who favored afraught with theoretical and practical difficulties. For strict floristic system included members of what has

example, most existing schemes assume that exotic been termed the Zurich-Montpellier Tradition inspecies will be replaced by native species when, in fact, central Europe (Shimwell 1971). The most well known

the behavior of exotics is notoriously complicated to arnong them is that of Braun-Blanquet (1928), whopredict, called plant associations with common diagnostic

species "alliances." Nowadays, the most commonlyBasic Classification Approaches defined floristic unit is k_own as the "association,"

defined by Flahault and Schroter (1910) as "a plantThree vegetation classification systems have gained community ... presenting a uniform composition and

widespread acceptance: physiog_omic classifications, physiognomy, and growing in uniform habitatfloristie classifications, and combined physiognomic- conditions." This definition implies that associations

floristic classifications (Howard and Mitchell 1988). All sharing a certain physiognomy would be grouped to-three approaches provide a systematic ordering of gether into the same formation. Some floristic methods

vegetation or ecosystem pattern and relate these focus on species that occur constantly throughouta set

patterns to ecological processes. Beginning in the 19th of stands; others emphasize indicator or diagnosticcentury with the work of plant geographers such as species, which are dominant in or restricted to these

Humboldt, Warming, and Grisebach, vegetation stands. Floristic methods require intensive field samp-

classification focused on the physiognomy of the ling, detailed knowledge of the flora, and carefulvegetation. Broadly speaking, physiognomy refers to tabular analysis of stand data to determine diagnostic

the structure (height and spacing) of the vegetation species groups. These methods reflect local andand to the life forms of the dominant species (i.e., gross regional patterns of vegetation and are more detailedmorphology and growth aspect of the plants). In than physiognomic methods. They also provide

addition, physiognomy refers to characters of detailed descriptions of biotic communities regardlessseasonality, leaf shape, phenology, duration. These of their successional stage or origin. As such, they are

features are relatively easy to recognize in the field typified by an agglomerative, "bottom-up" approach.with limited knowledge of the flora. Physiognomy Thus, florisfic units have been used fre,:tuent_y as indi-provides a fast, efficient way to categorize vegetation cators of ecosystem processes and are a useful compo-and can often be linked to remote sensing signatures, nent of ecosystem classifications (Mueller-Dombois

These characters are also useful for initial reeonnais- and EUenberg 1974, Rowe 1984, Strong et aL 1990).sance of areas to be surveyed. In addition, they permit Combinations of these two approaches have also

generalizations about the vegetation at a coarse, often been developed on the basis that vegetation is mostworldwide, scale, thoroughly described by both structure and floristic

The principles underlying physiognomic classifiea- composition. As stated previously, physiognomic

lion are that each specificlife form represents astrategy, systems are easily recognized in the field, can be(Stearns 1976), and that the composition of life forms in applied with limited knowledge of the flora, permit

a vegetation type is governed bv these strategies generalizations of vegetation patterns over large areas,(Raunkier 1937, Monsi 1960, Walter 1973. Whittaker and can be linked to remotely sensed data to facilitate

1975). The predominance of certain physiognomic vegetation mapping. In contrast, floristicinformationis

types in a region tends to correspond to major climatic almost always used for detailed site analyses, whetherzones.Thus, physiognomiccategoriesareoftenexpres- forstudying environmental gradients, ecological sttesions of macroclimate, soils, and vegetation (Holdridge factors or describing and forming classification urals.1947, Walter 1985, Howard and Mitchell 1985_. Furthermore, studies have shown that a very good fit

The basic unit of physiognomic classifications is the exists between floristie and physiognomlc classifica-

"formation," i.e., a "community type defined by dotal- tmns because both types of attributes are borne bynance of a given growth form in the uppermost individual species (e.g., Rfibel 1930,WesthoffandHeldstratum ... of the community, or by a combination of 1969, Webb et al 1970 Wergner and Sprangers 1982,

dominant growth forms" (Whittaker 1962). These Borhidi 1991). The Nature Conservancy (Grossman etclassifications emphasize a "top-down/" divisive aL 1994b; Grossman et aI 1998) have refined the

BiologicalandEcologicalDimensions 365

CLASSIFICATION HIERARCHY animals, and (3) the mutual relationships of geographic

ranges with humankind (Anee and Schmidt 1937,SYSTEM -- Terrestrial Muller i974).

PHYSIOGNOMIC CLASS-- Woodland Classifying animals with reference to theirgeographic locations dates back to the first century

PHYSIOGNOMIC SUBCLASS-- Mainly evergreen A.D. when faunal lists were created for specific areaswoodland (cf. Allee and Schmidt 1937). Faunal classifications have

focused on numerous factors, including morphological

PHYSIOGNOMIC GROUP -- Temperate evergreen traits, geologic influer_ces (e.g., continental drift), andneedle-leavedwoodland ecological relationships (e.g., animal distributions seen

FORMATION -- Needle-leaved evergreen seasonally as dependent upon localized environments, as op-flooded/saturated woodland with rounded crowns posed to geographic distribution and range). Regard-

less of emphasis, all zoogeographic approaches reeog-ALLIANCE -- Pseudotsugamenziesiiwoodland nize the link between habitat and animal distributions.

As is the case for plants, animal life is unequally' ASSOCIATION -- Pseudotsugamenziesii/Festucaidahoensis

woodland distributed within any zoogeographic area. Primaryfactors influencing the distribution of animals include

(1) the means of dispersal; (2) inherent or introduced

Fig. I. An example of the TNC terrestrial physiognomic- landscape barriers; (3) the degree of adaptability

floristic hierarchy (Grossman et al. 1994b). exhibited by the species (Allee and Schm_dt 1937); and(4) competition with other species. Animal distri-

physiognomic classifications of UNESCO (1973) and butions are further complicated by the fact that manyDriscoll et al. (1984) and combined this with two species(e.g., most insects, crustaceans and fishesJ havefloristic levels to create an hierarchical, physiognomic- several life history stages, each with its own discrete

floristic approach (Fig. 1) whichis now widely used for habitat. Another major complicating factor is thatconservation and resource planning across the United different taxonomic animal groups have strikinglyStates. different geographical affinities. It is possible to

The common question for terrestrial classification construct classifications for birds mammals, and

approaches is whether the obiective is to portray exist- invertebrates that have only a superficial resemblance

ing vegetation, potential natural vegetation, or both. in geographic pattern of distribution to one another.Classifications emphasizing existing vegetation Even when "hot spots" for endemic species with limit-

determine vegetation units based ou existing structure, ed ranges are identified for conservation purposes, the

composition, and successional status (Haufler 1994). resulting maps for differenl taxonomic groups are notClassifications emphasizing potential natural vegeta- well aligned (see the case of birds vs. amphibians and

tion use vegetation characteristics that represent the mammals in Bibby et al. 1992/. Thus, the concept of amost mature conditions of vegetation development "formation" or "assooation" does not apply to fauna,

(Kiichler 1964_. In the words of Ttixen (1956, in except for species that are highly restricted to specificMueller-Dombois and Ellenberg 1974), potential veg- plant units at small scales. In these cases, the term

etation becomes "the vegetation structure that would "community" may be used. Additionally, the habitatbecome established if all successional sequences were requirements of most species are multi-scale in accord

completed without interference by man under the with differing life-history requirements, meaning thatpresent climatic and edaphic conditions." the distributions of most animals are best viewed in an

hierarchical framework in which distributions at

4.1.2 Zoogeographic Classification different life-history stages are depenoeru on sets ofhabitat requirements defined at those scales. For many

Zoogeography is the scienti.fic study of the distribution species, the scales may vary from inter-continentalof animals on the earth and the mutual influence ot during migration to highly local during reproduction.

environment and animals on each other (Allee and Lawton et aL (1993) explored numerous hypothesesSchmidt 1937}. It is more difficult to summarize and to describe the relationships between range sizes or

represent that plant geography, simply because distribution limits and body size, latitudinal patterns,animals are much less easily observed and are highly and metapopulation theories Multiple approaches to

mobile, Approaches to zoogeography include: (1) the coarse-scale zoogeographic classification have also

faunalcompositions of landscapes and regions, (2) the taken place, for example, zoogeographic realms

evolutionary dynamics of the geographical ranges of (Muller 1974_ and life zones (Merriam 1898). However,

366 D,H.Grossmanet a]./Principlesfor EcologicalClassification

these coarse-scale classifications are not very useful for watershed modeling remain a major challenge foroperational land managers. Scott et aL (1993) hydrologists. These include the need to deal with sca-

developed a Gap Analysis approach more conducive to ling and the linkages among hydrology, geochemistry,operational planning; it depicts species distributions environmental biology, meteorology, and climatologybased on vegetation alliances using small-scale satellite (Hornberger and Boyer 1998).

imagery. Even in this case, however, significant scale- The development of freshwater classification has

related problems emerge, occurred in phases. Early efforts to classify riversystems identified whole river system types (Davis

4.2 Aquatic Classifications 1890, Shelford 1911). The recognition of biological

zones led to a second phase of classification (Carpenter

As compared with terrestrial systems, aquatic systems 1928, Picker 1934). The utility of this approach >,'asare highly dynamic. Classification is facilitated more by limited, however, because the classes were onlyabiotic than by biotic features. The water's surface also applicable in basins where zoogeography, geology,

presents a screen to direct viewing by visible light. As a and climate were held constant; biological zones could

result, aquatic classification is much less advanced than not be compared across regions (Naiman et al. 1992).terrestrial classification. It is apparent that aquatic The third phase, which continues to the present day,classification systems should not be assumed to nest attempts to describe systems in terms of general

easily, if at all, into terrestrial classification systems, ecosystem processes (lilies 1961). Concurrently, geo-As aquatic systems in general are highly dynamic morphologists worked to describe the physics of chan-

and subject to intense natural perturbations, the con- nel formation (Horton 1945, Leopold and Wolmancept of a stable "climax" has not been useful. Coastal- 1957, Strahler 1957); their findings led to a recognitionmarine systems shed some light on the question of the need to consider geologic and climatic processeswhether ciassifications of existing biology or biotic (Naiman et al. 1992).

potential apply to other than terrestrial systems. Valiela The most recent efforts classify river systems within(1984) devoted an entire chapter to colonization and surrounding landscapes to account for the dynamics of

succession in marine communities; it is significant that ecosystem processes. Building on earlier work (Warren

the concept of"climax" does not appear. In fact, Peters 1979, Lotspeich 1980, Lotspeich and Platts 1982),(1976) pointed out that many properties associated Frissell et al. (1986) applied principles of hierarchywith succession are tautological, creating doubt on the theory to describe stream ecosystems in a watershed

validity of the concept of "potential" ecosystem states, context. This model has two important elements. First,Indeed, as any recent marine biology or oceano- ecosystem processes are regionally scaled (i.e., whole

graphy text willillustrate, marine ecology is dominated stream systems are formed by climatic and geologicby process concepts that emphasize change. Succes- events), whereas variation in channel form results fromsion is seen as the composite result of a complex of more local events, such as storms and landslides.

factors. Perhaps it is for this reason, that few "potential" Second, large-scale processes constrain the develop-

or"climax" maps equivalent to those for terrestrial veg- ment of small-scale stream features. In other words, theetation exist. An exception is the work of Hayden et al. genesis of the stream system limits the range of micro-(1984), which brought together mapped classifications habitat systems.of coastaI-biogeographic, coastal-physical, and Much testing of hierarchical models has focused on

oceanographic environments. Although there was the constraint of the upper level ecological processessome concurrence in the boundaries, enough differ- on stream systems, particularly their biologicalences were present to illustrate that these three composition. For example, Larsen et al (1986 found

approaches express different attributes, that ecoregions corresponded to the distribution of fishassemblages in Ohio. Similar studies ,,','ere conductedin Kansas IHawkes et al. 1986), Arkansas Rhom et al.

4.2. I Freshwater Classifications -- Rivers 1987},Oregon (Hughes et al. 1987 Whittier et aI. 19881_and Wisconsin (Lyons 1989, At a coarse scale these

Approaches to classifying freshwater ecosystems range studies showed that ecoregional classification providesfrom those focusing on single physical, chemical, or a useful means to stratify variability in fish distribu-

biological variables to efforts using complex combi- tions (Hughes et al. 1994, However, prediction of fishnations of all three types of variables (Naiman et al species occurrence at a given site will require a classi-1992, Hudson et al. 1992', However, despite manv fication system that accounts for local-scale habital

recent advances, issues related to parameterizing the features that are not apparent at the regional level

many variables and physical processes involved ir (Lyons 1989).

_ltl ................................ ,11mnlm I

BiologicalandEcologicalDimensions 367

4.2,2 Freshwater Classifications- Lakes caused by winds are common, These dynamicsencourage more distinctive zonation and prevent the

The development of lake classification has followed a aging process usualh" found in inland wetlands

path similar to that for rivers. Typically, they have been (Herdendorf et al. 19S6). Inland wetlands respond tobased on single or multiple variables (Leach and rapid water-level increases from runoff, and ground-Herron 1992) that generally fall into four groups: tro- water contribution can be significant. During hot and

phic or productivity levels (e.g., Hooper 1969; Shannon dr), periods, the survival of some inland wetland plantand Brezonik 1972), chemical and/or biological charac- communities may depend on groundwater inflow'.

teristlcs(eg.,Moyle1956, Bright1968),basinoriginand Groundwater is of lesaer significance for coastal

physiographic characteristics (e.g., Hutchinson 1957, wetlands.Winter 1977), and hydrologic setting (Winter 1977). Wetlands are often classified according to their

Current efforts to classify lakes relate biological plant communities. For example, coastal wetlands may

composition to large-scale ecosystem processes (Tonn be referred to by a major plant species, e.g., mangrove,and Magnuson 1982; Tonn 1990; Schupp 1992). Tonn Spartina, or ]uncus. The approach for this classification(1990) described the environmental and historical is floristic, as described above for terrestrialvegetation.factors that determine patterns of fish biogeography

and community structure in lakes as a series of filters 4,2.4 Estuarine Classificationsthat operate from continental to local scales. This inc-

orporates appropriate hierarchical levels for aquatic Estuaries represent important mixing zones of freshecosystem processes and underscores the importance and marine waters. The'," have been classified in severalof considering the interaction of zoogeography and ways, largely by combinations of geomorphology,

geoclimatic factors at the higher levels of the hierarchy, hydrodynamics, and geographic (climate). PritchardSly and Busch (1992) suggested an approach for classi- (1967) recognizes five geomorphologic categories: (1)

fying large aquatic systems based on multi- drowned river mouths, e.g., Chesapeake Bay, (2) tect-dimensional, geophysical parameters. This approach onically produced, e.g., San Francisco Bay, (3) bar-built,allows the consideration of productivity or functional e.g., coastal-lagoon systems of the southeastern and

performance criteria within geophysical boundaries. It Gulf coasts, and (4) fjord-like, e.g., glacially over-also allows flexibility in identifying three-dimensional deepened valleys along the Alaskan coast. Birdsfootboundaries, river delta or everted river mouths, e.g., the Mississippi

River Delta, may be a fifth category.

4.2.3 Wetland Classifications The most simplistic estuarine classification is theso-called "Venice System," which classifies estuaries

Wetland classification became an issue in the early 20th solely on the basis of salinity regimes (Anonymous

century through efforts related to the identification of 1959). This has long been considered an over-

peatlands (Mitsch and Gosselink 1986). More recently, simplification. In contrast, estuaries may be classifiedscientists and environmental regulators have develop- hydrodynamically as: (1) stratified, (2) partiallyed various wetland classifications to identify and stratified or salt-wedge, and (3) completely mixed. In

quantify areas for management and protection stratified estuaries the freshwater floats over saltwater;purposes. The goal, as stated in Cowardin et al. (1979), river influence is greater than tidal influence. Littleis to create "boundaries on natural ecosystems for the mixing occurs, and the wedge may move far inland. In

purpose of inventory, evaluation and management." partially stratified estuaries, freshwater flow equalsNonetheless, this classification system is a generalized tidal flow as an agent of turbulence; the resultant layer-

"taxonomy" that is not georeferenced and that lacks ing may be complex. Completely mixed estuaries arethree-dimensional boundaries- a much-needed ex- vertically homogeneous, with tidal currents dotal-tension and refinement, nating the freshwater inflow. All these classificat|ons

Water levelsdictatethetypeofwetlandcommunity, are entirely physical. Bulger et al. (1993) used a

Wetlands usually have their water table near or above principal components analysis to demonstrate thatthe terrestrial surface, so that the area is seasonally or biologically relevant salinity reglmes can be derived.

permanently covered by shallow water, in which Geographic (climatic. classifications representhydrophytes dominate. The water source and duration attempts to categorize seasonal effects of climate onoften differs for coastal versus inland wetlands. Coastal ran-off, temperature, and daylight, allof which trigger

wetlands are usually located on gentle, sloping topog- biological responses, such as migration and spaw'ning

raphy, and are exposed to regular, seasonalwater-level Runofi is especially subject to seasonal shifts. Forfluctuations. However, abrupt, short-term changes example, in the tropics too little rainfall occurs in the "

[lff11_mrln ........

368 D,H. Grossmanet al./Principlesfor EcologicalClassification

dry season for any runoff to reach coastal areas. 4.2.6 Hierarchical ApproachesLagoons may become sealed by bars formed by long-shore currents and quickly become hypersaline. With The Nature Conservancy is developing an aquaticthe onset of the rainy season, freshets break through classification system to provide a framework for the

the bars, and flush the lagoon. As a result, the lagoon inventory, identification, and characterization of the

may become entirely fresh. Such dynamics are freshwater biodiversity of North America (Higgins et

important considerations for estuarine classification, al. 1996). This classification is hierarchical and allowsRoy (1984) recognized three main estuarine types for for the characterization of aquatic communities onNew South Wales, Australia: (1) drowned river valley, both abiotic and biotic levels at multiple scales (Fig. 2).

(2) barrier, and (3) saline coastal lake, all of which This approach captures information on the ecologicalevolve by processes of infilling. He documented context of community types, including physical pro-

unique biotic assemblages in each estuarine type. cesses. It also characterizes the physical environment,which provides an indirect way to identify community

4.2.5 Coas_l-Marine Classifications types where biological information is not sufficient(Angermeier and Schlosser 1995).

It has been recognized for some time that a useful Another recent development is that of Maxwell et al.

ecological integrator for aquatic systems is the water- (1995), which calls for an hierarchical framework of

shed (e.g., Sheldon 1972, Lotspeich 1980, Seaber et al. aquatic units (Fig. 3). This system is based on the1984). This approach may be extended into the sea as a aggregation of hydrologic units and the geographic"seashed." The advantage of using this approach is that distribution of aquatic biota (particularly fish species).watersheds reflect ecological properties across terres- The system is therefore not merely typological or de-

trial-aquatic boundaries, and thus act as controls on scriptive, but ecological. It provides insights intothe distribution of biological diversity. This approach is ' system dynamics and how these units may be used for

driven by multi-scale ecosystem factors, as exemplified management and other purposes.

by Lotspeich (1980); i.e., large-scale controlling forcesare climate and geology, reacting forces are soils and 4,3 Integrated Classification Approaches

vegetation, and the streams themselves respond to allfactors of the system. Although watershed and seashed The introduction to this chapter presented the overallmodels are relatively well-developed, they are also purposes of ecological classification. From previous

limited by the lack of available information on ecologi- sections, as well as Tables 4 and 5, it is apparent thatcal interactions of the coastal zone. Significant research many different classifications exist. As has beenis needed to establish ecological relationships that are emphasized previously, terrestrial classifications have

useful for coastal-marine ecological classification, dominated the field. The classifications have many

Ray and Hayden (1992) adopted concepts from propertiesmcommon, namely the organization of dataSeaber et al. (1984) to develop a watershed- or into categories, which facilitates retrieval and analysisseashed-based classification that recognized five and allows comparisons among systems. Most of these

coastal-zone subdivisions: uplands, coastal plain, and systems do not share a common approach totidelands on the terrestrial side and the shoreface en-trainment volume and offshore entrainment volume ECOREGIONAL

over the continental shelf. Hydrological interactions "_ _[Lorg_scaleare key to this system, which is specifically intended as Lcmdscope

1a comparative-hierarchical classification that is adapt- Cor,t_,×r_l 1 ECOREGIONAtable to both scientific and conservation purposes. This S.;CT_ON

hierarchical approach is necessary to address the many 1-

scales that are involved. For example, the ability of any lt_CROHAa'_jJ--q aZLIANCE Ilocation to provide suitable habitat foi a diversity of _ tENTICp.OTI I_

species depends on an array of physical and biotic Small-scalePhvs_¢alT j CommunayEiornems

factors that vary at different temporal and spadal scales Contex, _ H Jfrom whole ocean basins to local estuarine and _lJ TAT

j UNITS ASSOClAnONfreshwater tributaries. (Regional distributions of fishes

have been described by Robins [1971], Robins and Ray ABIOTIC.

[1986 and Robins [1992]; the multi-scale dynamics of

these relationships have been examined by Ray et al, Fig. 2. TNC Aquatic Classification Hierarchy (Higgins et al,

[19971.) 19961.

: Biologicaland EcologicalDimensions 369

=#

[ NEARCTIC ZONE (NORTH AMERICA) 1d I i

SURFACE-WATER OEOCLIMATIC _GROUNlY.WATER --SYSTEMS SETTING SYSTEM

{so zo..sI iOOMA,.Sj{.EO,O.] IO,v,S,o.s]

I WATERSH_OS ]I

I

I I t

V ...... I f !

I R'VERtNE I [ LACUST"'NE ] ' [ I

GROUND-WATER

SYSTEM SYSTEM I SYSTEM

Sites I

Fig. 3. US ForestServiceAquaticEcologicalUnit Hierarchy(laxweB et at. 1995).

classification, and this is of utmost importance for and the soils or landform factors are established during

ecosystem management. Nor is it clear, at this point, if and after the classification process, but these factors arethe different classifications may be reconciled into or not used to define the vegetation units (Komarkovaaround a common classification approach. 1983). The units described are natural ones, but

emphasis is placed on determining vegetation units4.3.t $i¢e Classifications that represent "ecologically equivalent landscapes"

(Kotar et aL 1988). Insofar as they describe the floristicSingle Factor composition of part of the natural vegetation, namely

climax stands, the units of the habitat type are fairly

Single-factor site classifications generally assume that equivalent to the plant association concept used in thethe chosen characteristic best reflects spatial patterns of western United States (Komarkova 1983). The intent isthe environmental resources of interest (Omernik and to use these descriptions to classify sites that are not atGallant 1989). For terrestrial sites, these classifications climax and, b). examining their understory composi-

are primarily intended to reflect plant and animal tion, to infer their ecological potential.potentials. Several site classification systems have used Somewhat different from the habitat type approach

only vegetation to determine site potentials, usually J is the ecological species group approach, which mare-with reference to successional trends or productivity, i rains that species show similar "ecological behavior."In this sense, these systems focus on potential natural i Generally, these species belong to the same layer of

vegetation (e.g., Fenneman 1946; K'3chler 1967; vegetation (e.g., the herb layer, nonvascular layer, orUdvardy 1975; Brown et aL 1979, 1980). A widespread shrub layer). The method presumes that communities

approach to site classification using vegetation is the are combinations of plant species whose compositionhabitat type classification system (Daubenmire I952, depends on the local environment (Mueller-DomboisPfister and Arno 1980, Kotar et aL 1988). This system and Ellenberg 1974). The community unit identified

focuses on natural climax or near climax vegetation, can, at times, be very similar to the plant association.

and recognizes all understory species as a reflection of whereby the ecological species groups are thesite characteristics. Relationships between vegetation diagnostic species for the association. However, it is

370 D.H. Grossmanet al./Principlesfor Ecological Classification

also possible that the same association could contain similarly by Sukachev (1945) as "biogeocoenosis."

several ecological species groups (Mueller-Dombois Central to the application of the approach is that alland Ellenberg 1974). The ecological species group parts of the ecosystem are included. In someinformation can either be used by itself to indicate site ecosystems, each part -- vegetation, soils, climate and

characteristics, in which case the system partially landform -- is first studied independently and then

resembles the habitat type system, or it can be inte- combined (Jones et al. 1983, Driscoll et al. 1984, Sims et

grated with other measured site factors as part of an aI. 1989). In others, the parts are combined at the outsetecosystem classification (Pregitzer and Barnes 1982, because it is their joint interactions on the landscapeCleland et al. 1994). that define the units. Because it is difficult to integrate

multiple factors and understand their interactions

Multiple factor beyond the local level, the units are usually consideredhypotheses in need of further testing (Albert et al.

Mulfi-factor classifications are based on the idea that 1986). Mapping is a key product from this process

ecosystems and their components display regional (Rowe 1984, Zonneveld 1989). Bailey's ecoregionalmap

patterns that are reflected in a combination of causal of the United States (1976, 1995) is more similar to theand integrating factors (Bailey 1983; Austin and independentapproachbecauseBaileyreliesheavilyon

Margules 1986; Omernik and Gallant 1987). Existing separate climatic, physiographic, and vegetation mapsmulti-characteristic frameworks include, but are not and then reconciles their boundaries. The works of

limited to, the British Columbia biogeoclimatic system Albert et al. (1986) and Cleland et al. (1994) represent

(Pojar et al. 1987), Cajender's (1926) forest types, the combined approach.Daubenmire's (1968) habitat types, the ecoregions of The biogeoclimatic zone system of Krajina (1965, inthe conterminous United States (Omernik 1987, Mueller-Dombois and Ellenberg 1974) is another syst-

Gallant et al. 1989, Omernik and Gallant 1989), Bailey's em in which vegetation is emphasized when defining

(1976, 1980) ecoregions of the United States, and the landscape or ecosystem units. These zones are defined

major land resource areas (USDA Soil Conservation as geographic areas that are predominantly controlledService 1981). Many similarities are found among these by the same macroclimate and contain similar soils andclassi fica tion systems. For example, site classifications (climatic climax) vegetation. The definition of the zonesinclude floristic information that is collected in the at lower scales uses vegetation units that are defined by

same way as that which would be used for vegetation the plant association concept (Pojar et al. 1987). Atclassification (Mueller-Dombois and Ellenberg 1974, higher levels, climatic zones and topographic position

Pregitzer and Barnes 1982). Similarly, habitat type are used to group vegetation units into biogeoclimaticclassifications define plant associations in a similar zones.manner to that of the floristic system of Braun-

Blanquet. Furthermore, site classifications that bring 4.:3.2 Ecoregional Classifications

together independent vegetation, soil, and landformclassifications rely on the independent classification of Ecoregions are defined as geographical zones that

these variables as their starting point (Jones et al. 1983, represent groups or associations of ecosystems thatSims et al. 1989, Host 1993). Site classification systems function in a similar way. Ecoregions are assumed to be

that use multiple factors subdivide land into major and highly uniform with respect to their abiotic and bioticminor land types or landscape ecosystems. They have factors, the processes controlling their ecosystems, andbeendevelopedprimarilyforlandmanagerswhoneed their limits and prospects for human exploitatmn.

to integrate resource management, biological conser- Consequently, classifying the world into an orgamzedvafion, and restoration planning. These systems are geographic system and subdividing the earth's surface

most appropriate for classifying ecosystems, which are into consistent regions is difficult.defined by the dynamic interactions of the biotic and In general, two distinct concepts have been used.

physical components. As with vegetation classifica- Th.e first is the "controlling factor(s)" concept (Baileytions, emphasis is placed on units that are more or less 1983, 1996L m which it is assumed that one or more

homogeneous both as to form and structure, but in this enwronmental factors act as primary controls or limita-case with respect to all factors of the land and the tions for a particular ecoregmn as defined at a parti-

vegetation (Rowe 1961). cular scale. The second is the "synthesis approach"An ecosystem approach to classification, namely ,Omernik 1987), whict_ assumes that a holistic integra-

that the plant community is considered together with tion of all predominant and stable components must beits environment, wasimplicitin Clements work (1916), considered in the regionalization. This approach

but was defined explicitly by Tansley (1935) and defines ecoregions by perceived patterns in a corn-

: BiologicalandEcologicalDimensions 37t

bination of causal and integrative factors, including climatic criteria. Divisions are subdivided intoland use, land surface form, potential natural veg- provinces on the basis of the climax plant formation

etation, soils, and so oil. The distinction between these that geographically dominates tile upland area of tile

approaches is obscured when they are mapped province. Provinces are further subdivided intobecause mapping always requires a reduced set of sections by differences in the composition of the climaxdifferential and associated characteristics, vegetation type. The sections correspond generally to

Through over 100 years of development of biogeo- the potential natural vegetation types of Kfichler

graphy, many ecoregion systems have been produced (1964). As an extension of this system, an Ecologicalat various scales from global to local. Many approaches Classification and Mapping Task Team (ECOMAP),

and applications havebeen developed, includingsome was formed in the U.S. Forest Service to develop awhich integrate soils, land forms, and vegetation (see consistent approach to ecosystem classification andHerbertson 1905, Austin 1972, Udvardy 1975, Bailey mapping at multiple geographic scales. A national hier-

1976, Omernik 1987, Bailey 1995). Below is a brief archical framework of ecological units was developed

description of four ecore_onal classification systems by the ECOMAP team with four new levels underthat are presently used in North America two for Bailey's section: subsection, landtype association, land-Canada and two for the United States. Also included is type, and landtype phase (McNab and Avers 1994).

a brief description of a new ecoregional classification ECOMAP is described in greater detail in the

system that was developed by the World Wildlife Fund "Synthesis Approaches" section.to evaluate conservation priorities at an ecoregionallevel in Latin America and the Caribbean. SynthesisApproaches

Controlling Factor Approaches Terrestrial ecozones oj' Canada: Using the synthesisapproach to integrate a different set of assumptions in

Ecoclimatic regions of Canada: Environment Canada mapping and characterizing regional land units,has developed the "Ecoclimatic Regions of Canada, Wiken (1986) developed "Terrestrial Ecozones of Can-First Approximation" (1989). This is a traditional ada." Although its nomenclatureissimilar to that of the

"control factor(s)" approach to creating a hierarchy of Ecoclimatic Regions of Canada, this system is based on

regionalizations. In order of decreasing size and gen- predominant and stable biophysical characteristics,erality, the map includes ecozone, ecoprovince, rather than a singular controlling mechanism. Eco-

ecoregion, and ecodistrict levels. Climate guides the zones (15 units) are areas of the earth's surface repre-development of the ecoprovinces, although physio- senting large and very generalized ecological units,

graphy, such as large mountain rar_ges, has major characterized by various abiotic and biotic factors. Eco-modifying influences. Ecoregions were subdivided provinces (45 units) are parts of ecozones characterizedbased on the effects of macroclimate on the vegetation, by major assemblages of structural or surface forms,

Ecoclimatic regions were defined as broad areas on the faunal realms, and vegetation, soil, hydrological and

earth's surface characterized by distinctive ecological climatic zones. Ecoregions (177 units) are parts of

responses to climate as expressed by vegetation and ecoprovinces characterized by distinctive regionalreflected in soils, wildlife, and water. Within eco- ecological responses to climate, as expressed by the

climatic regions, similar trends in vegetation succession development of vegetation, soil, water, and fauna.•,,.'ill be found on similar soils occurring on similar Ecoregions of the contcrnlinous United States: Omernik

parent materials and positions on the landscape. Ten has produced ama F entitled "Ecoregaons of the Con-ecoclimatic provinces and 77 ecoclimatic regions were terminous United States" (Omernik 1986, 1987) thatidentified in Canada. shows 76 eco:egions at a scale of 1:7.500.000. Omernik's

Ecoregio_s of the Lhffted States: Bailey (1976, 1980, method grew ov,t of an effort to create a stream classi-1983) developed a national regionalization of eco- fication system for the U.S. Environmental Protection

system units for the U.S. Forest Se_'ice. It is a hier- Agency to plan water resource management. Thearchical system, and similar in concept to the premise behind Omernik's approach is that ecologicalEnvironment Canada system, Four broad ecological regions can be identified by analyzing the patterns andlevels (domain, division, province, section) are composition of biotic and abiotic pnenomena that

distinguished, based on climate and biogeography, reflect differences m ecosystem quality and integrity

Domains are identified by broad climatic similarity, CWiken 1986: Omernik 1987, 1995). These phenomenasuch as lands having dry climates at the subcontinental include geology, physiography, vegetation, climate,scale. The domains are quite heterogeneous and are soils, land use, wildlife, and hydrology. The relative

further subdivided into divisions through additional importance of each characteristic varies from one

372 D.H. Grossrnan et al./Principles for Ecological Classifica',ion

ecological region to another, regardless of the States. ECOMAPusesasingle factor approach at somehierarchicallevel, levels, and a synthesis approach at others. Because

ecological variables and processes operate at different

ECOMAP spatial spaces, each level of the hierarchy is associated

with particular design criteria for its scale (see Table I).

Until the early i990s, no single system had been dev- At the ecoregion scale, the domains, divisions, and

eloped with the structure and flexibility needed to dev- provinces are adapted from Bailey (1980, 1995), and are

elop ecological units from continental to local scales, recognized by differences in global, continental, and

The National Hierarchical Framework of Ecological regional climatic regimes and gross physiography.

Units (Avers et al. 1994) was developed by the U.S. These upper levels are based on regional climatic types

Forest Service to provide a consistent, hierarchical of KOppen (1931), as modified by Trewartha (1968). In

framework for application throughout the United addition to climate and physiography, the units at the

Table I. ECOMAP: Principal map units and design criteria (Avers et. al. 1994).

Scale Ecolog/caI Principal Map Unit Design Criteria* Map Scale Range General PolygonUnit Size

Eooregional Domain • Broad climatic zones or groups (e.g. dry, humid 1:30,000,000 or 1,000,000's oftropical), smaller square miles

Division • Regional climatic types (Koppen 1931, Trewatha 1:30,000,000 to 100,000's of square1968). 1:7,500,000 miles

• Vegetational affinities (e.g. prak,'ie or .forest).• Soil order.

Province • Dominant potential natural vegetation (Kuchler 1964). 1:15,000,000 to I0,000's of square• Highlands or mountains with complex vertical 1:5,000,000 miles

climate-vegetation-soil zonation.

Subregion Section • Geomorphicprovince, geologic age, stratigraphy, 1:7.500,000 to 1.000's of squarelithology. 1:3.500,000 miles

• Regional clima tic data.• Phases of soil orders, suborders or great groups• Potential natural vegetation.• Potential natural communities (PNC_ (FSH 2090).

Subsection • Geomorphic process, surficial geology, lithology. 1:3.500.000 to 10's to low 1.000's• Phases of soil orders, suborders, or great groups. 1:250,000 of square miles• Subreglonalclimaticdata.• PNC--formation or senes.

Landscape Landtype • Geomorphic process, geologic iormatior., surficial l:_-_t 000 to ] O0's to I D00'sofAssodation geology., and elevation 1:60,000 acres

• Phases of soil suborders, famines, or series.• Local climate.

• PNC--series, subseries, plant associations.

Land. unit Landtype • Landform and topography (elevation, aspect, slope 1:60,000 to 1:24,000 10's to 100's ofgradient, and position}, acres

• Phases of soil subgroups, families, or series.- Rock type, geomorphic process.• PNC-plant associations.

Landt)_e • Phases of soil families or series. ] :24.000 or }arger < 1DOacresPhase • Landform and slope posl,tion.

• PNC--plant associations or phases.

•Note: Criteria listed are broad categories of environmental and landscape components. The actual classes of components chosen

for designing map units depend on the objectives for the map.

Biologicaland EcologicalDimensions 373

_-_ ,,_ _-_ I >._ _,_ _ o

_ _'__r.q i I va

._o _ ' _ i _'

._ _ _-,_ . .-u, ..._ ..J _, . ,_v ._..v

_ _'_ _ :* _ '_'_ "_" 82,01= Q,-II N nl

i e_ i _10_._ _1 nl I_ _o _1 (_ O _,_ I _ nl i i _.<N '4

..oo .'_ o,: .=.-,.-,oo._i ,11 _ f_ *,-: r_ ,-< ¢:

I

!

0

.,. ..... o,,, =o ._o .._

I I I i

I I @ l @, o ,! O, -_ I 0,o I

t

8 _°,o_ ,:°°. ' _ ..-,_ -- -u

J ,',<

374 D.H. GrossmanecaklPrinciplesfor EcologicalClassi_cation

Table 3. ECOMAP;Nationalhierarchyof ecologicalunitsassociatedwith purpose,

Planning and Analysis EcologicalUnits Purpose, Objectives, andScale General Use

Ecoregions

Global Domain Broad applicability for modeling and sampling. RPA assessment.

Continental Division Internatonal planning.

Regional Province

Subreglons Sections RPA planning, Multi-forest, statewide, and multi-agency analysis and

Subsections assessment.

Landscape Landtype Association Forest or area-wide planning, andwatershed analysis.

Land Unit Landtype Project and management area planning

Landtype Phase and analysis.

province level are further characterized and classified Taxonomic and geographic scales: implications for

by soil orders and potential natural communities (from classt[ications and their integration: The different scalesK(_chler 1964). inherently associated with the levels of the National

At the subregion scale, sections and subsections are Hierarchical Framework of Ecological Units have

characterized by combinations of climate, geomorphic implications for the specific types of planning

process, topography, and stratigraphy. These factors objectives for which each can be used. Table 3 showsinfluence moisture availability and exposure to radiant the general utility of each ECOMAP level.solar energy, which in turn directly control hydro-

logical function, soil-forming processes, and potential Biodiversity PatternApproaches

plant community distributions (Avers et al. 1994). The

classification and descriptions have now been corn- Terrestrial ecoregions of Latin America and thepleted throughout the United States at the sectionlevel Caribbean: With the objective of setting conservation(McNab and Avers 1994). Numerous regions (the east- priorities at an ecoregional level, the World Wildlife

ern United States, California, and the Columbia Basin) Fund developed a hierarchical, terrestrial ecoregion

have completed classification, characterization, and system for Latin America and the Caribbeanmapping at the subsection level (see Keys et al., in (Dinerstein et al. 1995). This system is based on an

press). An example of the level of detail for the map estimation of the "original/" pre-Colombian distribu-units is shown in Table 2. In the upper midwest tion of habitats in this area, and differs significantly

(Minnesota, Wisconsin, and Michigan)and other areas from the approaches used in the United States andthis work has been completed to the subsubsection Canada. Three hierarchical levels were identified:

level (see Albert 1994). maior ecosystem types, major habitat types, and eco-

At the landscape scale, the landtype association is regions. Major ecosystem types (5 units) are defineddefined by general topography, geomorphic process, primarily by dynamic properties and spatial patterns of

surficial geology, soil, potential natural community patt- biodiversity, not wholly by vegetation structure. Majorerns, and local climate (Formann and Godron i986; ecosystem types are subdivided into major habitat

Avers et al. 1994). At this level, terrestrial features and types (11 units/ based on general habitat structure.processes may also have a strong influence on ecological climatic regimes, and major ecological processes. The

characteristics of aquatic habitats (Platts 1979, Ebert et al. level of species turnover with distance fDeta diversity)1991). At the land unit scale, landtypes and landtype is also considered, with flora and fauna showing slml-

phases are designed and mapped in the field based on lar guild structures and life histories. Nested within theproperties of local topography, rock types, soils, and major habita{ types, ecoregions are identified to

vegetation. These factors influence the structure and represent geographically distinct assemblages ofcomposition of plant communities, hydrologic function, natural communities that share a large majority of their

and basic land-use capability (Avers et al. 1994). species, ecological dynamics, and similar envzron-

BiologicalandEcologicalDimensions 375

mental conditions. The ecological interactions critical S. | Data Issues

for the long-term persistence of these COmmunities are

also considered. Four questions relating to data need to be addressed

when developing an ecological classification system:4.3.3 Land Cover ClasslncatJons (1) what kind of data are needed, (2) how many data

are needed, (3) can existing data be used, and (4) whatLand cover classifications are primarily intended for are the costs associated with preparing data for theland management or resource planning. They emph- classification _-ystem? All four questions must be

asize conspicuous features of the land surface, and can answered in relation to the specific management

be combined with land-use maps to convey an overall objective(s) to be addressed.perspective of what is visually perceived on the land. To respond to the first two questions, it must beMany classification systems derive land cover units recognized that the different applications of ecological

from general structure and composition of the vegeta- classification may not necessarily be achievabletion (Anderson et al. 1976). For example, cover types, through the use of one data set or analysis, Char-

named by the dominant tree (Eyre 1980), are used as a acterization data are used grouping similar patterns

descriptive land cover classification of forest lands, and delineating areas of interest. Pattern recovery dataRecent land cover systems have increasingly relied are used to extrapolate or interpolate the classification

on factors that can be characterized through remote units to areas that have not been sampled. The amount

sensing imagery (Witmer 1978). Examples include the of data needed to satisfy either of these applications is

Gap Analysis (GAP) and the Multi-Resolution Land related to the required accuracy and precision of theCharacterization (MRLC) Programs. GAP is a national classification system. Additional data may be needed to

program that uses remote sensing to derive landcover characterize the properties of each classification unit.types, which are then used to model the distribution Therefore, the total amount of data needed depends on

and protection status of animal species. Similarly, the the objective of the classification; the required accuracyMRLC Program is creating a consistent, remote and precision for pattern characterization andsensing-based land cover characterization approach recovery; and the type of description required for eachacross the nation that can be used for both one-time classification unit.

and repeatable resource inventory and environmental To answer the third question, it is important toassessment objectives, determine what types of data are already available and

whether they will be useful and cost effective beforeundertaking ori_nal research. Existing data sources

S USE OF ECOLOGICAL CLASSIFICATION may include remote sensing imagery, herbarium and

SYSTEMS museum species records, and plot/transect data thathave been collected during past ecological surveys.

Many ecological classification systems use existing Data, maps, and charts for vegetation, soils, lakes,

biological and environmental information to predict geology, fresh waters, estuaries, salinity, temperature,patterns, processes, and occurrences that will support coastal geomorphology, and climate also exist in many

conservation and management objectives. Classifica- geographical areas. Because using existing data oftentions have been able to integrate more information in involves compiling disparate and interdisciplinary

recent years, and the resulting products can increasing- data, problems related to differences in sampling dates]y be applied to multiple purposes (e.g., strategic and design, data quality and storage, spatial repre-planning, environmental monitoring, protected-area sentation and resolution, and standardization must be

identification). Caution must be exercised, however, solved (see Davis et al 1991, Davis 1995_.Nevertheless

when using classifications and maps to make resource existing data that are well organized and appropriateconservation and management prescriptions when to the objective can, in certain instances, be an effectivethose products were developed through combining way to help develop ecological classifications and

preexisting data sets. The products derived from such maps _e.g, Reid et al. 1995).integrated data sets are, as a rule, much loss precise in The fourth :tuestion regarding costs, can on|v be

terms of class and spatial accuracy than their compo- answered following the first three. Cost is determinednent parts, by the amount of eqmpment, techniques. _"kills. and

Care must be taken to develop classifications t_me necessary to acquire the h, pe and amount of data

tailored to their intended applications and to choose needed. It should be noted that. although it might bethe most appropriate classifications for addressing more expensive, it is often easier to build ecologicalspecific management needs, classifications and prepare maps using new data.

-'-N

376 D.H. Grossmanet aL/Principlesfor Eco{ogicalClassification

rather than attempting to derive them from existing 5.3 Ecological Boundaries: What Do the

products and data. An example of this concerns tile Lines on Maps Mean?high costs and questionable utility of digitizing old

maps of undocumented spatial accuracy into Ecological variables are generally semi-continuous, as

geographic information systems (GIS). opposed to being categorical, and are expressed asgradients. The concept of the "ecotone" reflected this

5.2 Ecological Classification and Mapping (di Castri et al. 1988, di Castri and Hansen 1992): "..,the

ecotone is a 'zone of transition' between adjacent

Therelationshipbetweenclassificationandmappmgls ecological systems, having a set of characteristicsoften confusing for scientists and managers alike. In its uniquely defined by space and time scales, and by thepurest form, ecological classification is independent of strength of the interactions between adjacent ecologic

mapping. It is a scientific process of methodically systems." Ecotones may be expressed at any scale. As aarranging units of quantitatb,,e information into classes result, distinct lines between ecological units are theor groups that possess common properties. Mapping is exception rather than the norm.

a representation of these units which is constrained by The problem of where to draw a line on a map isthe scale of the map and by the spatial information that complex and requires an explicit set of rules and

must be related to the ecological units. Additional disclaimers. Two untested assumptions are typicallycomplexity is introduced with the interpretation of made: (1) the mappable attribute is a clearly defined

"spectral signatures" or an aggregation of available and a predictable property of the ecological unit, andgeoreferenced"data layers." (2) the mappable attribute does not vary over the

Maps will never exactly represent ecological units known distribution range of the ecological unitunless those units were developed through the (Bourgeron et aL 1993). This is difficult enough foraggregation of spatial thematic layers. A map of earth simple variables, but wiren multiple variables arecover that is created through the analysis of remotely brought into the classification framework, the "fuzzi-

sensed data or interpretation of aerial photography hess" of the lines increases dramatically.

will not portray an ecological classification system. If Consequently, several important factors must becover types are determined to be an important factor of considered when developing a map of ecological units.an ecological classification system, additional research First, interpreting ecological units depends in part onmust be conducted to define the relationship between knowledge of the distributions of environmental

the cover classes and the ecologicalclassification units, attributes along gradients. Second, the ecological rela-As reviewed earlier in this chapter, ecological tionsamong bioticand abiotic components need to be

classifications may be developed by using a variety of stated and tested explicitly. A clear link between scales,variables. Using spatial data (maps) that include patterns, and processes must be established; andvariables needed for a classification system can greatly temporal variability must be considered (Bourgeron et

simplify the process of creating an ecological classi- al. 1993). The question becomes how much fuzziness infication map. Spatial data sets have variable thematic the transition zone is acceptable to meet the elassi-

and locational accuracy, so that when a map is com- fication and mapping objectives. Strategic planningpiled from multiple spatial products, its overall may tolerate more fuzziness, whereas operational

accuracy is unpredictable and interpretation must be planning typically tolerates less. Regardless of thetentative. Furthermore, the boundaries of separate management needs, any ecological map must producemap themes rarely coincide, largely because of real boundaries that are ecologically significant. One way

differences among attributes. This is as true for land to become explicit is to map the gradients themselves.units (Bailey et a1.1994), as it is for coastal (e.g., physical The concept of ecoregions is intuitively attrac tire forvs. biogeographic units), and marine units (e.g., marine all systems -- terrestrial, freshwater, and marine. The

phytoplankton vs. physical oceanographic regimes reconciliation of boundaries among different environ-(McGowan 1972, Hayden et al. 1984). mental attributes, as has been done for terrestrial

In addition to these technical challenges, care'must systems (e.g., Bailey 1995), remains problematic,be taken not to substitute a composite map of ecolo- Ecological classifications, especially when presented

gical components for a map of ecological units. Overlay on maps, give the appearance of ecosystem stasis andmaps do not represent the ecological relationships offers few opportunities for further analysis.between the individual layers. Anecologicalunitinthe Furthermore, when classification approaches are

natural environment is not merely a compilation of dominated by few factors, they tend to address someindependent components; but is functionally systems better than others. Each environmental factor

integrated, and/or system exhibits its own behavior, which in

BiologicalandEcologicalDimensions 377

mapped terms may be expressed as gradients across 5.5 Management Applicationsthe land- or seascape. Combining these may be poss-

ible only through statistical modeling approaches of Regulatory and management agencies at all levels of"dynamic biogeography" (e.g., Hengeveld 1990). The government, have struggled to apply ecological

science of ecosystem classification and geography still classifications to resource planning and managementhas a long way to go, especially in the "critical areas of challenges. The diversity of ecological classificationstreatment of heterogeneity and scaling" (Hornberger has been reviewed in earlier sections. Each system has

and Boyer t995), been developed for a specific set of purposes and manyhave proven extremely valuable for management

applications (see Management Chapter). However,

5,4 Key A¢¢ribuces of Representative having value for one set of applications does not makeClassification Syscems a classification system useful for other applications. For

example, classification systems that focus on site

Forty-four classification systems are presented in Table potential (e.g., ECONLa.P Land Type Associations) may

5, and key attributes in six topic areas are listed for each be appropriate for gross, strategic analyses of resourcesystem. The table can be used to evaluate what each management. For operational planning or statisticallyeffort classifies (e.g., terrestrial vs. freshwater systems), defensible estimates, analyses that are conducted at thehow the classification was developed (e.g., use of stand level and that integrate existing and potentialabiotic vs. biotic factors), and at what scale the system is vegetation are recommended (Roloff 1994).

classified (e.g., global vs. local). One classification Table6providesanevaluationoftheapplicabilityof

system often addresses more than one attribute in a the major ecological classification systems for differenttopic area; for example, that of Grossman et ah (1994b), management applications. The management applica-Grossman et al. (1998) considers both terrestrial and tions are divided into four general areas: land manage-

wetland systems. Each classification effort is described ment and use, biodiversity conservation, resource

by: inventory and management, and research and assess-ments. Additional discussion concerning the applica-

• The ecologicalsystem that theeffortclassifies ter- tions of ecological classification systems to meet

restrial, freshwater, coastal/marine, estuarine, and/ management objective can'be found in Carpenter et al.or wetland; (this volume).

"Land management and use" activities focus on the• The geographical coverage that the effort encom- land unit and its development by humans for a variety

passes --global, continental, national, regional, and/ of purposes, including

or local; • Infrastructure siting for roads and other "conduits"

• The overall objectives of the classification effort -- and buildings;research, management, and/or conserva6on;

• Resource planning, site prescription, and manage-

. The environmental factors used to develop the clas- ment for forestry, wildlife, fisheries, recreation, agri-sification effort climate, soils, elevation, geology, culture, erosion control, minerals and water, and

landform/land position, and/or hydrology/ multiple use designations;

hydrography; • Desired future conditions analysis; and

• The biological factors used to develop the classifica- • Sustainable development.

tion effort -- existing vegetation structure, existing

vegetation composition, potential natural vegeta- B'odivers ty conservat'on" refers to the identifica-tion, zoological guilds, and/or zoological composl- don and protection of natural biota as it occurs ivtion; and, ecological systems, including:

• Biological diversity inventory;• Other factors that are importam in Jescribing the • Identification of conservation sites:

classification effort, such as: are the classification • Sustainable design for conservation; and

units geographically referenced? Is the classification • Restoration planning.hierarchical (i.e., do multiple levels exist']? Were a

varie D" of factors used to develop the classification "Resource inventory and management" includeseffort? Were extensive data needed to develop the _ctivities intended to determine the existing number,classification? abundance, and condition of natural resources to

_,_ [ .rl r'lrlrllll r'II _l_r[ [TIT

378 D.H.Grossmanet al./Principlesfor EcologicalClassification

_lt_CI _A!suolx_I

-,,g

I• . I

I

v_'_ oJV _l!u_ uo!lnaU!SS_l D _ _" _ _ _ _ _ _" _r

uo! ]sodmo_ p:_l_OlOO Z _ "_I

spl!n 9 p:o!_olooz _ _

I

uo]l!sodtuoD

I

"_ _anl_rulg

• I

I

0 _ uo!l[SOd/tuaoJpur?I _- _ _v",3 J

g_ ,(_o[o_o

_ st!os_ '_ I

_ _ _tuq Do I

u= Po " I¢.,

--G e., I

©

._

_ l_lU_U]lUO_)

m (=_ I_qOlD

>,

= I

I-_ _ Ir_ seoDl_uu_l_! _ _-._' I

I

BiologicalandEcologicalDimensions 379

380 D.H. Grossman et al.IPrinciples for Ecological Classification

-----__-- .......... "_q_ I n"H : qllTTl_lllllll " ' -'--- ""

Biological and Ecological Dimensions 381

"-I

......... __ [ll_/i/i-[tlrlr_1

382 D.H, Grossman et al./Principles for EcologicalClassification

inform management decisions. These activities Espec a "..for terrestna s)stems, class f catlonsha_.einclude: been developed to identify boundary conditions which

can be employed to 0npose restrictions on land use. For• Resource sampling; wetlands connected to larger aquatic systems, the• Resource inventory; boundary between vegetated and non-vegetated areas• Resource management forestry, wildlife, fisher- is of considerable manageme1_.t interest. The federal

ies, agriculture, recreation, soil, and minerals and government has the mandate to monitor wetland con-water; dition and abundance in relation to defined levels for

• Habitat suitabifityanatysis; and healthy ecosystems and abundant fish and wildlife• Resource reporting and mapping, resources. Wetland classification is the first step in

"Research and assessments" include the evaluation identifying these areas for inventory and management;

and prediction of the effects of natural and human subsequent steps include protection, regulation,

changes on natural resources and systems, including: evaluation, restoration, and rehabilitation.Ecological classification has assisted aquatic eco-

• Environmental impact assessments; system management in several areas, including water• Ecosystem and landscape monitoring; quality assessment (Meador et al. 1993, Paulsen and• Study of pattern and process in relation to scale; Linthurst 1994), fish productMty modeling (Fausch et• Representativeness assessment; and al. 1988), fish habitat requirement modeling (Nelson et• Predictive modeling, al. 1992, Hill and Platts 1995), and adjacent land

Perhaps the most common application of terrestrial management (Platts i980, Maxwell et al. 1995). Because

vegetation and site classifications is the determination of the complexity of lakes and their typically large size,of appropriate uses for agricultural and forest lands• as ,,,,'ellas the historical focus on fisheries management,

Integrated units derived from climatic, soil, and classification for productivity has received consider-vegetation provide a useful tool for the determination able attention.of site potential. In addition, the combination of One example of aquatic ecosystem management is

existing and potential vegetation classifications have the National Water Quality Assessment Programbeen used to describe current conditions, predict (NAWQA). In 1990, the U.S. GeologicalSurvey (USGS)

successional processes, and characterize disturbance initiated this Program as a comprehensive survey of

regimes (Haufler 1994, Roloff 1994). The limited degree the status and trends of ground and surface waterof success using exclusively potential vegetation in quality in the United States. Physical, chemical, andaccomplishing these goals is based on limited know- biologicaldataarebeingcollected fromstudyareasthat

ledge concerning vegetation-site relationships, and correspond to hydrologic units based on the drainagesthe abiH ty of the dependence on the observers to infer of major rivers and aquifers (USGS 1982), which will besite characteristics from the vegetation, further stratified according to the classification frame-

A standard application for ecological classification work of Frissell et al. (1986) (Meador et al. 1993).systems, terrestrial and aquatic alike, is conservation NAWQA researchers are currently assessing the use ofplanning. Ecological classifications have been used ecoregions as a stratification tool for national samplingextensively to inventory and protect terrestrial and (J. Higgins, pers. comm.).

aquatic areas. A direct biolo_cal classification will For wetlands, management initially focused onalways provide the greatest confidence of capturing waterfowl and furbearer production, which had pro-the conse_'ation target, but coarser biological and site found implications for wetland classification. In the

classifications can be used to develop a high degree of past two decades, however, other services provided bypredictive accuracy. Vegetation classifications have wetlands, such as biological functions, habitat,

been the primary approach for the identification and sanctuaries, hydrological functions, and cultural val-delineation of terrestrial conservation sites by resource ues, have been identified _Greeson et al. 197% Reppertmanagement agencies (Scott et al. 1993, Grossman et al. and Sigleo 1979, Tiner 1984 _.In addition, society often1994a). Aquatic researchers and managers have assigns shoreline marshes a higher ecological valueattempted to classify standardized descriptors for habi- than other types of wetlands because of their direct

tat sub-units for conservation planning (Busch and Sly contribution to fish and wildlife populations, rather1992). Boundary delineations alone are often used in than their role in larger ecosystems (Maltby et al. 1983jterrestrial and aquatic systems to protect delineated Estuarine wetlands and wetlands located _t nver

habitat units that are in short supply, unique, or mouths are also highly valued by society becausesupport known species of special interest, wildlife (bird) and fish usage is usually high, rare plants

Biologicaland EcologicalDimensions 383

or unusual wildlife colonies may occur, and these areas izations and classifications are abundant throughout

are often near urban centers (Crispin 1990). the scientific literature. However, the existing systems

Descriptive attributes have been used to develop a do not fully meet the current expectations or needs o f

"taxonomy" of coastal wetlands, iocludir_g vegetation, resource managers.hydrology, geography, climate, soils, stratigraphy, and Most ecological classifications have been descriptivelandscape position (topography, aspect, slop@ Funo and have focussed on either terrestrial, freshwater,

tional attributes are also used as insights into system coastal, or marine systems. Up to now, terrestrialdynamics and management, including species life systems have dominated ecological classification

histories, multi-species interactions, landscape inter- efforts; these generally do not incorporate adequateactions, hydrological processes (flood storage and information on aquatic systems (e.g.,Avers et al. 1994,storm-flow modification), nutrient retention and trans- Bailey 1995). This is partially an artifact of traditional

formation, sediment and toxicant trapping, and sedi- training in ecology and resource management, but alsomeat stabilization. Both sets are then modeled to reflects the variable status of data and knowledge

provide decisions on wildlife habitat and public use associated with the different systems.(James E. Perry, Virginia Institute of Marine Sciences, Aquatic classifications have primarily been based on

pers. comm.), biophysical factors, whereas most terrestrial elassifi-Estuaries represent high levels of complexity and cations emphasize vegetation (potential and existing),

conservation urgency. They house major fish climate, and physiography. For example, Cowardin etpopulations, and provide vital habitat for early al. (1979)focus on wetland topography, while Bailey

life-history stages of many marine fishes and in- (1995) emphasizesphytogeographyandelimate. Coast-vertebrates. All anadromous and catadromous fishes al classifications are often approached with a more

must cross estuaries. Estuaries are the sites of most of comprehensive "ecosystem" perspective as they includethe world's large cities and ports, and also provide components of both terrestrial and aquatic systems.important industrial and commercial sites. As a result, The increased need for sound management of

they are centers for coastal pollution and are, in natural resources and prioritizaflon of conservationgeneral, heavily disturbed, which has implications for action has resulted in an increased dependence upon

coastal living resource and fishery sustainability, existing ecological classification systems. The existingClassification of estuaries requires information on ecological classification systems can address some, but

where various types of habitats occur, which are the not all, management and conservation concerns. Eachbest examples, and whatactivities most threaten them. system was developed to address a specific set of

objectives. It is important to understand the intendedand appropriate uses for each classification system and

6 PRESENT SITUATION AND FUTURE its associated products. No individual system will ever

VISION meet the full spectrum of potential applications.

The understanding of ecological dynamics at the com-

munity, landscape, and ecosystem levelsis undergoing 6,2 The l)evelopmen¢ of New Ecologicalrapid growth. Spatial and temporal organization and Classificationsdynamics are better understood through the applica-tion of hierarchy theory and remotely-sensed inform- Existing classifications provide the framework to go

ation. Biogeographic and environmental information beyond mere description of the distribution ofare being developed at an unprecedented rate and biological species and communities to focus on theirwith a high level of quality control. Technological relationships to one other and to environmental gra-

advances in the management and analysis of spatial dients. There are many challenges associated w_.th thedata are increasingly available to managers and development of the nexl generation of classificationscientists. This remarkable progress provides a suite of [ systems.

new opportunities for the science and application of The dynamics of environmental change are central

ecological classification, to the concept of the ecological unit However. it has

been difficult to apply ecosystem concepts and6.1 Existing Ecological Classifications practices ro most ecological classification systems, as

few of them emphasize ecological process and theMany ecological classification systems have already dynamics of change. In addition, ecological classifi-

been developed. Accounts of the development and cations must reflect our increasing knowledge aboutapplication of biogeographic and ecological character- the biological and ecological processes that function-

384 D.H, Grossmanet al./Principlesfor EcologicalClassification

ally integrate terrestrial, freshwater, and coastal/ be represented and interpreted spatially. Simulta-

marine systems. Marine classification must integrate neously, there have been remarkable advances in thethe coastal systems which, in turn, must integrate technological capabilities for managing, aggregating,

terrestrial systems. Until more classifications integrate analyzing, and portraying these data. These advances

aquatic and terrestrial features and processes, they i have stimulated rapid testing, refinement and imple-cannot be seen as truly "ecosystemic" and they will be mentation_,of numerous classification and assessment

restricted in applications and future value, approaches that are used to address various conser-vation and management objectives, as well as basic

6.2. l A Set of National Ecological Units research @estions.This increased capacity to integrate information for

Many managerswould derive great benefit from access targeted application will greatly augment the devel-to one common ecological classification system with opment of a common classification approach. Specific

explicit standards and application guidelines. In many classifications can be developed to address specificcases, access to one standard framework of resource management and conservation objectives.

classification units would represent the most efficient Users can determine the appropriate scales for analysesand cost-effective solution to many shared resource and identify the specific biophysical variables and

management challenges, ecological processes that apply to their specificThe successful development of a common set of questions and objectives. They can draw from multiple

ecological units will require sufficient consensus on data sources to compile and analyze the appropriatecommon objectives, appropriate scales of analysis, and spatial data and to develop very specific solutions to

the critical ecological processes and biophysical vail- their questions.ables that function at those scales. This system ideally Where multiple approaches to ecological classi-

wouldintegrateacross multiple scales, multiple factors, fications are appropriate, the ability to share stand-and be relevant for terrestrial, freshwater and coastal/ ardized data layers becomes increasingly important.

marine systems. The development of this common eco- For this to occur, standards for ecological inventory,

logical framework would not restrict the development data management, and analytical approaches must beor use of other systems that better address specific developed and documented. Ecological classifications

needs and applications, and associated products (e.g., keys, assessments, maps,

This concept of a common set of ecological units has etc.) should be accompanied with appropriate meta-

already gathered momentum. In December 1995, a data that fully disclose the methods, data sets, scales,Memorandum of Understanding (MOU) was signed variables, analyses, classes, and other information

by the U.S. Department of Agriculture (Natural required to fully interpret the utility of the product. AllResources Conservation Service, Forest Service, and data and data products must meet minimal standards

Agricultural Research Service), the U.S. Department of so that they can be broadly interpreted and applied tothe Interior (Bureau of Land Management, Geological multiple objectives. Data standards and inclusion of

Survey, Fish and Wildlife Service, National Biological metadata files will allow partners to confidently assess

Service, and National Park Service), and the U.S. the appropnateness of the data products forEnvironmental Protection Agency. This MOU was addressing their individual objeefives.

designed to develop a spatial framework of ecological It is critical that resource managers, conservation-units for the United States. A National Interagency lsts, and researchers will be able to use the information.

Steering Committee and a National lnteragency Potential users willneed to have theability to access all

Technical Team have been established to implement data so they may identify the appropriate informatmnthis MOU. As part of the initial and ongoing effort, and ecological classification systems that would help

maps of common ecological units will be developed them meet their specific objectives. This ,,,,'ill alsoand published at standard scales. Digital data sets in require a high level of terminological consistency;

formats meeting available Federal Geographic Data working definitions must be widely agreed upon andCommittee standards will also be published, implemented to guard against inappropriate applica-

tions and faulty interpretatmn of ecological data,6.2.2 User-Defined Classification Systems classifications and associated products.

The quanti W and quali_ of ecological and biological 6.3 Human Factorsdata has dramatically increased over the past few

decades. More and more of these data are Many appropriate applications of ecological classifi-geographically referenced, allowing the information to cation systems have been identified in this report, but

Biological and Ecological Dimensions 385

perhaps the most importartt objective for these ecolo- REFERENCES

gical classification systems has not been specifically

mentioned: the ability to use these systems for the pur- Albert, D.A. 1994. Regional landscape ecosystems of Michi-• gan Minnesota, and Wisconsin: a working map and classi-

poses of education and communication. Ecological fication. Fourth revision. USDA Forest Service General

classification systems help managers of ecological sys- Technical Report NC-178, North Central Forest Expefi-terns understand the functions and processes of these ment Station, St. Paul, MN.

ecosystems. This is an important step in breaking the Albert, D.A., S.R. Denton, and KV. Barnes. 1986. Regio_zal

cycle of managing for one component of an ecosystem Landscape Ecosystems qf Michigan. University of Michigan,

at a time. Ecological classificatior_ systems and associ- School of Natural Resources, Ann Arbor, MI.Allee, W.C., and K.P. Schmidt. 1937. EcologicalAnimal Geogra-

ated products can also provide invaluable commu-phy. Wiley, New York, NY.

nications tools conveying critical land management Allen, T.F.H., T.W. Hoekstra, and R.V. O'Neilh 1984.

and conservation information to the general public. Interlevel relations in ecological research and manage-Finally, ecological classification efforts have histo- ment: some working principles from hierarchy theory.

' rically avoided the integration of the human USDA Forest Service General Technical Report RM-110,

dimension into the system, even as reported system Rocky Mountain Research Station, Fort Collins, CO.

attributes. Management and conservation is a human Allen, T.F.H., and T.B. Start. 1982. Hierarchy: Perspectives forEcological Complexity. University of Chicago Press, Chi-

endeavor and people have an increasingly profound cago, IL.impact on the structure and function of ecological Anderson, J.R., E.E. Hardy and J.T. Roach. 1976. Land use

systems. Therefore, it is critical that we learn how to and land cover classification system for use with remoteintegrate key social aspects of politics, economics, sensing data. Geological Survey Professional Paper 964 (a

anthropology, and sociology into future ecological revision of the land use classification system as presentedclassification efforts, in U.S. Geological Circular 671/, U.S. Government Printing

Office. Washingtor D.C.Angermeir, P.L, and I.J. Schlosser. 1995. Conserving aquatic

biodiversity: beyond species and populations. AmericanFisheries Society Symposium 17402-17414.

ACKNOWLEDGMENTS Anonymous. 1959. Symposium on the classification of brack-ish waters. Venice, Italy, April 8-14, 1958. Archivio di

This chapter represents a synthesis of materials that Oceanografia e Limnologia, Volume 11. Supplemento

was developed, reviewed and edited by an extensive Amo, S.F., and G.E. Gruell. 1985. Douglas-fir encroachmentinto mountain grasslands in southwestern Montana. ,rour-

team of contributors. Some of these contributors pro- hal of Range Management 39: 272-276.vided scientific input for the text. others developed Austin. M.E. 1972. Land resource regions and major land re-material for figures and diagrams, while others source areas of the United States. Agricultural Handbookprovided invaluable review and editing. 296. Revised edition. U,DA'q Soil Conservation Service,

The primary contributors were members of the Washington, D.C.

au thor team, who provided materials and review as Austin, M.P. 1985. Continuum concept, ordination methods

requested. Special recognition is due to Carleton Ray, and niche theory. Annual Review of Ecology and Systematics

who, through great effort succeeded in balancing the 16: 3941.Austin, M.P. 1987. Models for the analysis of species response

presentation of terrestrial, freshwater and manne to environmental gradients. Vegetat*o 69: 35-_5.materials throughout the chapter. The authors on the Austin. M.P 1991 Vegetation theory m relation toscience chapter teambenefited from thediseoursewith cost-efficient surveys, pp 17-22. In: C.R. Margules, andand constructive review from the authors in the M.P. Austin (eds.), Nature Co tse_vation: Cost Effective Bio-

management chapter team, particularly their team logical Surveys and Data Analysis. CSIRO Publishing,

leader Constance Carpenter. Collingwood. Vic. AustraliaAustin. M.P.. R.B. Cunningham, and P.M Fleming. 1984. NewCritical scientific contribution and review were

approaches to direct gradient analysis using envtronmen-received from Don Faber-Langendoen for the history tal scalars and statisticatcurve fitting procedures. Vegetanoof vegetation classification, and on freshwater classifi- 55: 11-27.

cation ss'stems from Jonathan Higgins and Mary Austm, M.P., and C.R. Margules. 1986. Assessing representa-

Lammert. Mark Bryer and Xiaoiun Li provided con- fiveness, pp. 45-67. In: M.B. Usher _ed.), Wildlife Conserva-structive scientific revtew ot various sections and tion Evaluation. Chapman and Hail, London, UK.

helped to develop some of the fign,lres. Cynthia Austin, M.P., and T.M. Smith. 1989. A new model for the con-

Swinehart provided early editorial assistance. Kat tinuumconcept. Vegetatio83: 35-47.Austin, M.P., and G.A. Yapp. 1978. Definition of rainfall re-

Maybury provided editorial review and coordinated gions of southeastern Australia by numerical classificationthe complex process of getting the final manuscript methods. Archly fur meterologie, Geophysik und"out the door." Bioklimatologie, Series B, _: 121-142.

386 D.H. Grossman et al./Principles for Ecological Classification :_

Avers, P.E., D.T. Cleland, W.H. McNab, M.E. Jensen, R.G. tern managemertt: principles and applications. USDA For-Bailey, T. King, C.B. Goudney, and W.E. Russell. 1994. Na- est Service, Pacific Northwest Research Station, Portland,tional hierarchical framework of ecological units. USDA OR.Forest Service, Washington, D.C. Braun, E.L. I950. Deciduous Forests of Eastern North Anlerica.

Bailey, R.G. 1976. Ecoregions of the United States (nT,ap, scale Blakiston Press, Philadelphia, PA.1: 7,500.000). USDA Forest Service, intermoun*;ain Region, ! Braun-Blanquet, J. 1928. Pflanzenso:iologie. Grundzuge der vege-

Ogden, UT. i tafiolzskunde. Springer, Berlin, Germany. [English transla-Bailey, R.G. 1980. Descriptions of the Ecoregio_ls qf the United I tion published in 1932 by McGraw-Hill, New York, NY.]

States. USDA Forest Service Miscellaneous Publication i Briggs, J.C. 1974. Marine Zoogeography. McGraw*Hill, New1391, Washington, D.C. i York, NY.

Bailey, R.G. 1983. Delineation of ecosystem regions. ]ol_rlzalof Bright, R.C. 1968. Surface-water chemistry of some MinnesotaEnvironmelltal Managcmet2t 7: 365-373. lakes, witb preliminary notes on diatoms. Interim Report

Bailey, R.G. I989a. Ecoregions of the continents (map, scale 1: 3, Limnological Research Center and Bell Museum of Nat-30,000,000). USDA Forest Service, WashingtQn, D.C. ural History, University of Minnesota, Minneapolis, MN.

Bailey, R.G. 1989b. Explanatory supplement to the ecoregions Brov,'n, D.E., C.H. Lowe, and C.P. Pase. 1979. A digitized sys-map of the continents. EnviromnentaI Conservation 15: 307- tematic classification system for the biotic communities of309. North America, with community (series) and association

Bailey, R.G. 1995. Description of the ecoregions of the United examples for the Southwest. Journal of the Arizona-NevadaStates (with separate map at a scale of 1: 7,500,000). USDA Academy of Science 14 (suppl. 1): 1-16.Forest Service Miscellaneous Publication 1391 (revised), Brown, D.E., C.H. Lowe, and C.P. Pase. 1980. A digitized sys-

Washington, D.C. tematic classification for ecosystems with an illustratedBailey, R.G. 1996. Ecosystem Geography. Springer-Verlag, New summary of the natural vegetation of North America.

York, NY. USDA Forest Service General Technical Report RM-73,

Bailey, R.G, P.E. Avers, T. King, and W.H. McNab (eds.). 1994. Rocky Mountain Forest and Range Experiment Station,Ecoreginns and subregions of the United States (map, Fort Collins, CO.scale 1: 7,500,000). USDA Forest Service, Washington, D.C. Brown, J.H. 1984. On the rela lion between the abundance and

Bailey, R.G., and H.C. Hogg. 1986. A world ecoregions map distribution of species. American Naturalist 124: 255-279.for resource reporting. Enviromnental CorIservation 13: 195- Brown, J.H., and A.C. Gibson. 1983. Biogeography. C.V. Mosby,202. St. Louis, MO.

Bailey, R.G., M.E. Jensen, D.T. Cleland, and P.S. Bourgeron. Bulger, A3., B.P. Hayden, M.E. Monaco, D.M. Nelson, and1993. Design and use of ecological mapping units, pp. 105- M.G. McCormick-Ray. 1993. Biologically-based estuarine116. In: M.E. Jensen, and P.S. Bourgeron, editors. Eastside salinity zones derived from a multivariate analysis. Estu-forest ecosystem health assessment. Volume 1I. Ecosystem aries 16: 311-322.management: principles and applications. USDA Forest Burger, D. 1976. The concept of ecosystem regions in forestService GeneralTechnical Report PNW-318, Pacific North- site classification, pp. 213-218. In: Proceedings XW IUFROwest Research Station, Portland, OR. World Congress, Divisioll l, June 20-July 2, 1976, Odo, Nor-

Barnes, B.V. 1984. Forest ecosystem classification and map- way.ping in Baden-Wurttemberg, West Germany. pp. 49-65. Busch, W.-D.N., and P.G. Sly, editors. 1992. "DieDevelopment ofIn: Forest land classification: Experience, problems, per- an Aquatic Habitat Classification System for Lakes. CRC Press,spectives. Proceedings of the symposium: March 18-20, 1984, Boca Raton, FL.Madison, WI, USA. Cajender, A.K. 1926. The theory of forest types, Acta Forestalia

Barnes, B.V., K.S. Pregitzer, T.A. Spies, and V.H. Spooner. Fennica 29: 1-108.1982. Ecological forest site classification. Journal of Forestry Carpenter, K.E. 1928. Life in Inland Waters, Macmillan, New80: 493_98. York, NY.

Bibby, C.J., N.J. Collar, M.P. Crosby, M.F. Heath, Ch. Christensen, N.L.,A.M.Bartuska, J.H. Brown, S. Carpenter, C.Imboden, T.H. Johnson, A.J. Long, A.J. Stattersfield, and D'Antonio, R. Francis, J.F. Franklin, J.A. MacMahon, R.F.SJ. Thirgood. 1992. Putting Biodiversity o_ tile Map: Priority Noss, D.J. Parsons, C, H. Peterson, M.G. Turner, and R.G.Areas GlobalConservation. International Council for Bird Woodmansee. 1996.The report of the Ecologacal Society ofPreservation (ICBP), Cambridge, UK. America Committee on the Scientific Basis for Ecosystem

Borhidi. A. 1991. Phytogeography and Vegetation Ecology of Cuba. Management. EcologicalApplications 6: 665-691.

Akademiai Kiado. Budapest. 858 pp. Cleland_ D.T. 1996. Multifactor classification of ecological landBourgeron, P.S., H.C Humphries, R.L. DeVellce and M.E units in northeastern Lower Michigan. Unpublished

Jensen. 1993. Ecological theory in relation to landscape Pb.D. dissertation. Michigan State Universi .ty, East Lan-evaluation and ecosystem characterization. PF 65-79. In: stag. MI.M.E. Jensen, and P.S. Bourgeron (eds.), Eastside Forest Cleland, D.T., J.B. Hart G.E. Host KS. Pregitzer. and C.W.Ecosystem Health Assessment. Volume II. Ecosystem Ramm 1993. Ecological class(fication system for the Hu-management: principles and applications. USDA Forest ron-Manistee Natio_al Forest: field guide. Huron-ManisteeService, Pacific Northwest Research Station, Portland. OR. National Forest, Cadillac, MI

Bourgeron, P.S. and M.E Jensen. 1993. An overview of the Cleland, D.T., J.B. Hart G.E. Host. K.S. Pregitzer. and C.W.ecological principles for ecosystem management, pp. 51- Ramm. 1994. Ecologicalclassification and hwentory system o]63 In: M.E Jensen, and P.S. Bourgeron (eds.), Eastside the Huron-Manistee National Forest. USDA Forest ServiceForest Ecosystem Health Assessment. Volume lI. Ecosvs- Region 9. Milwaukee, WI.

Biological and Ecological Dimer_sions 387

Clements, F.E. 1916. Plant succession: an analysis qflhe develop- Driscoll, R.S., D.L. Merkel, D.L. Radloff, D.E. Snyder, and ].S.

ment of vegetation. Carnegie Institute of Washington Publi- Hagihara. 1984. All ecol_,gicalland classification framework forcation, Washington, D.C. the United States. USD..\ Forest Service, Miscellaneous Pub-

Conneil, J.H. 1978. Diversity kn tropical rain forests and coral lieation 1439, Washington, D.C.reefs. Science 199: 1302-1310. Dunbar, M.J. 1979. The relation between oceans, pp. 112-125.

Constanza, R., L. Weinger, C. Faulk and K. Maler. 1993. In: S.Van der Spoel, and A.C. Pierrot-Bults (eds.), Zoogcog-

Modelling complex ecological economic systems. Bfosci- raphy and Diversity in Plankton. Halstead (Wiley), Newence 43: 545-555. York, NY.

Cook, J.E. 1995. Implications of modern successional theory Ebert, D.L, T.A. Nelson, and J.L. Kershner. I991. A soil-basedfor habitat typing: a review. Forest Science 42: 67-75. assessment of stream fish habitats in Coastal Plain streams.

Cowardin, L.M., V. Carter, F.C. Golet, and E.T. LaRoe. 1979. Proceedings of Warlnwater Fisheries Symposium, June 4-8,

Classification: of the wetlands and deepzeater habitats qf the 1991, Phoenix, Arizona, USA.United States. U.S. Fish and Wildlife Service, Washington, Environment Canada. 1989. Ecoclilnatic regions of Canada, firstD.C. approximation (with map at a scale of 1: 7,500,000). Ecological

Crispin, S. 1990. A biological database for monitoring the Land Classification Series No. 923,Environment Canada,Great Lakes ecosystem, pp. 43-48. In: J. Kusler, and R. Ottawa, Canada.Smardon (eds.), Proceedings of a_ International Symposli_ln Ekman, S. 1953. Zoogeography of the sea. Sidgwick and Jack-on Wetlands of the Great Lakes: Protection and Restoration Pol- son Ltd., London, UK.icies. May 16-18, 1990, Niagra Fails, New York. Association Eyre, F.H. (ed.). 1980. Forest Cover Types of the United States andof State Wetland Managers, Berne, NY. Canada. Society of American Foresters, Washington,

Curtis, J.T. 1959. Tlie vegetation of Wisconsin: an ordination of D.C.plant communities. University of Wisconsin Press, Madison, Fausch, K.D., C.L. Hawkes, and M.G. Parsons. 1988. ModelsWI. that predict standing crop of stream fish from habitat vari-

Daubenmire, R.F. 1952. Forest vegetation of northern Idaho ables: 1950-1985. USDA Forest Service General Technical

and adjacent Washington, and its bearing on concepts of Report PNW-GTR-213, Pacific Northwest Research Sta-vegetation classificatSon. Ecological Monographs 22: 301-330. lion, Corvallis, OR.

Daubenmire, R.F. 1968. Plant Communities: A Textbook of Fenneman, N.M. 1928. Physiographic divisions of the United

Synecology. Harper Rowe, New York, NY. States. Am_als Association American Geographers 18: 261-353.Davis, F.W. 1995. Information systems for conservation re- Fenneman, N.M. 1946. Physical diz_isions of the United States

search, policy and planning: better access to data for scien- (map, scale 1: 7,000,000). U.S. Geological Survey, Reston,tists, policy makers, and managers. Bioscience (suppl.) $36- VA.$42. Flahault, C., and C. Schroter. 1910. Rapport sur la nomencla-

Davis, F.W., D.A. Quattrochi, and M.K. Ridd. 1991. Environ- ture phytogeograpbique. Proceedhlgs ofthe3rd International

mental analysis using integrated GIS and remote sensed Botanical Congress, Brussels 1: 131-164.data: some research needs and priorities. Photogrammetric Forman, R.T.T., and M. Godron. 1986. Landscape Ecology.Engineering alld Remote Sensing 57: 689-697. Wiley and Sons, New York, NY.

Davis, W.M. 1890. The rivers of northern New Jersey, with Franklin, J.F. 1993. Preserving biodiversity: species, ecosys-note on the classification of rivers in general. National Geo- terns, or landscapes? Ecological Applications 3: 202-205.

graphic Magazine 2: 82-110. Frayer, W.E., L.S. Davis, and P.G. Risser. 1978. Uses of landDenton, D.R. and B.V. Barnes. 1988. An ecological climatic classification. Journal of Forestry 76: 647-649.

classification of Michigan: a quantitative approach. Forest FrisseU, C.A., W.J. Liss, C.E. Warren, and M.D. Hurley. 1986. AScience 34: !19-138. hierarchical framework for stream classification: viewing

DeVelice, R.L., G.J. Daurniller, P.S. Bourgeron, and J.O. Jarvie. streams in a watershed context. Environmental Management1993. Bioenvironmental representativeness of nature pre- 10: 199-214.serves: assessment using a combination of a G1S and a Gallant, A.L., T.R. Whittier, D.P. Larsen, J.M. Omernik, andrule-based model, pp. 51-58. In: Proceedings, First Biennial R.M. Hughes. 1989. Regionalization as a tool for managingScientific Conference on the Greater Yellowstone Ecosystem, environmental resources. EPA/600/3-89/06(3. U.S. Environ-September 1992, National Park Service, Mammoth, WY. mental Protection Agency, Corvallis, OR.

di Castri, F., and A.J. Hansen. 1992. The environment and de- Gauch, H.G., Jr. 1982. Multivariate Analysis in Community Ecol.velopment crises as determinants of landscape dynamics, ogy. Cambridge University Press, New York. NY.

pp. 3-18. In: A.J. Hansen, and F. di Castri(eds.), Landscape Gleason, H.A. 1917 The structure and development of theboundaries: consequences for biotic diversity and ecologi- plant association. Bulletin of the Torrey Botanical Club 44:cal flows. Springer-Verlag, New York, NY. 463.-481.

di Castri, F., AJ. Hansen, and M.N. Holland, editors. 1988. A Gleason, H.A. 1926. The individualistic concept of the plant

new look a t ecotones: emerging international projects on association. Bulletin o/the Torrey Botanical Club 53: 7-26.landscape boundaries. Biology lnter!_ational Special Issue Greenacre. M.J. 1984. Theory and Appficat ions of CorrespondenceI7, IUBS, Paris, France. Analysis. Academic Press London.

DinersteIn, E., D.M. Olson, D.J. Graham, A.L. Webster, S.A. Greeson. P.E., J.R. Clark, and J.E. Clark, editors. 1979.WetlandPimm, M.P. Bookbinder, and G. Ledec. 1995. A consemation functions and values: the state of understanding. Proceed-

assessment of tile terrestrial ecoregions of Latin America and tile ings National Wetlands Symposium. American Water Re-Caribbean. The World Wildlife Fund and The World Bank, source Association, Technical Publication No. TPS79-2,Washington, D.C. Minneapolis. MN.

388 D.H. Grossman et al./Principles for Ecological Classification

Grossman, D.H., D. Faber-Langendoen, A.S. Weakley, M. An- Hills, G.A. I952. The classification and evaluation of sites for

derson, P. Bourgeron, R. Crawford, K. Goodin, S. Landaal, forestry. Ontario Department of Lands and Forests, Re-K. Metzler, K.D. Patterson, M. Pyne, M. Reid, and L. source Division Report 24, Toronto, Ontario, Canada.Sneddon. 1998. International classification of ecological Hix, D.M. 1988. Multifactor classification of upland hardwoodcommunities: Terrestrial vegetation of the United States. forest ecosystems of the Kickapoo River watershed, south-Volume h The national vegetation classification standard, western Wisconsin. Canadian Journal Forest Researdz 18:The Nature Conservancy, Arlington, VA. 1405-1415.

Grossman, D.H., K.L. Goodin, X. Li, D. Faber-Langendoen, Holdridge, L.R. 1947. Determination of world plant forma-

and M. Anderson. 1994a. Standardized national vegeta- fions from simple climatic data. Science 105: 367-368.tion classification system. Prepared for the USDI National Holdridge, L.R. 1967. Life Zone Ecology. Tropical Science Cen-Biological Survey/National Park Service Vegetation Map- ter, San Jose, CA.

ping Program, Denver, CO. Holling, C.S., L. Gunderson, and G. Peterson. 1993. Com-Grossman,, D.H., K.L. Goodin, and C.L. Reuss (eds.). 1994b. paring ecological and social systems. Beijer Discussion Pa-

Rare plmlt communities of the conterminous United States: an per, Series #36. Beijer International Institute of Ecologicalinitial survey. The Nature Conservancy, Arlington, VA. Economics, Royal Swedish Academy of Science, Stock-

Grumbine, R.E. 1994. What is ecosystem management? Con- holm, Sweden.

servatio_l Biolo_! 8: 27-28. Hooper, F.F. 1969. Eutrophication indices and their relation toHammond, E.H. 1964. Classes of land-surface form in the other indices of ecosystem change. In: Eutrophication:

forty-eight states, U.S.A. (map, scale 1: 5,000,000). Annals Causes, Consequences, Correctiw,s: Proceedings of a Sympo-Association American Geographers 54: Map Supp. No. 4. slum. National Academy of Sciences, Washington, D.C.

Hansen, A.J., and F. di Castri (eds.), 1992. Landscape Bound- Hornberger, G.M., and E.W. Boyer. 1995. Recent advances in

aries: Consequences for Biotic Diversity and Ecologic Flows. watershed modelling. Reviews Geophysics (Suppl. July):Springer-Verlag, New York, NY. 949-957.

Haufler, J.B. I994. An ecological framework for planning for Horton, R.E. 1945. Erosional development of streams andforest health. Journal of Sustainable Forestry 2: 307-316. their drainage basins: hydrophysical approach to quan fi-

Hawkes, C.L., D.L. Miller, and W.G. Layher. 1986. Fish tative morphology. Geological Society of America Bulletin 56:

ecoregions of Kansas: stream fish assemblage patterns and 275-370.associated environmental correlates. Environmental Biolo,_nd Host, G.E. 1993. Field sampling and data analysis methods for

of Fishes 17: 267-279. development of ecological land classifications: an applica-Hayden, B.P., and R. Dolan. 1976. Coastal marine fauna and tion on the Manistee National Forest. USDA Forest Service

marine climates of the Americas. Journal of Biogeography 3: General Technical Report NC-162, St. Paul, MN.71-81. Host, G.E., and K.S. Pregitzer. 1991. Ecological species groups

Hayden, B.P., R.D. Dueser, J.T. Callahan, and H.H. Shugart. for upland forest ecosystems of northern Lower Michigan.1991. Long-term research at the Virginia Coast Reserve. Forest Ecology and Management 43: 87-102.BioScienee41: 310-318. Host, G.E., K.S. Pregitzer, C.W. Ramm, J.B. Hart, and

Hayden, B.P., G.C. Ray, and R. Dolan. 1984. Classification of D.T. Cleland. 1987. Landform mediated differences incoastal and marine environments. EnvironmentalConserva- successional pathways among upland forest ecosystemstion 11: 199-207. in northwestern Lower Michigan. Forest Science 33: 445-

Hedgpeth, J.W. 1957. Marine biogeography. Geological Society 447.America Memoir 67(1): 359-382. Host, G.E., K.S. Pregitzer, C.W. Ramm, D.T. Lusch, and D.T.

Hengeveld, R. 1990. Dynamic Biogeography. Cambridge Uni- Clelar_d. 1988. Variations in overstory biomass among gla-versity Press, Cambridge, UK. cial landforms and ecological land units in northwestern

Herbertson, A.J. 1905. The major natural regions: an essay in Lower Michigan. Canadian Journal of Forest Research 18:systematic geography. Geography Journal 25: 300-312. 659-668.

Herdendorf, C.E., C.N. Raphael, and E. Jaworski. 1986. The Howard, J.A., and C.W. Mitchell. I985. Phytogeomorphotogy.ecology, of the Lake St. Clair wellm-,ds: a community profile. John Wiley and Sons, New York, NY.U.S. Fish and Wildlife Service Biological Report 85(7.7). Hudson. P.L., R.W. Gtiffiths, and T.J. Wheaton. 1992. Review

Higgins, J., M. Lammert D. Grossman, and M. Bryer. 1996. A of habitat classification schemes appropriate to streams.classification framework for freshwater communities: pro- rivers, and connective channels in the Great Lakes dram-

ceedings of The Nature Conservancy's Aquatic Commu- age basin, pp. 73-107. In: W..D.N. Busch. and P.L. Slynity Classification Workshop, April 9-11 1996, New (eds.), Tile Develoumem of an Aquatic Habitat ClassificationHaven, Missouri Available from The Nature Conser- System for Lakes. CRC Press Boca Raton. FL.

vancv Arlington. VA. Hughes R.M, S A. Heiskary W.J. Matthews, and C.O. Yoder.Hill, M., and W.S. Platts. 1995 Lower Owens Riverwatershedeco- 1994 Use of ecoregions in biological monitoring, pp- 125-

syste_ 1 a age _el _plan. Ecosystem Sciences. Boise, ID. 151. In: S.L. Loeb, and A. Spacie (eds.), BiologicalMonitoringHill, M.O. 1974. Correspondence analysis: a neglected of Aquatic Systems. Lewis Publishers, Boca Raton. FL.

multivariate method Journal Royal Statistical Society, Series Hughes. R.M., E. Rexstad. and C.E. Bond. 1987. The relation-o _ - ship of aquatic ecoregions river basins, and physiographicG ~3: 340-3M.

Hill, M.O. 1979. Twinspan. A FORTRAN program for arrang- provinces to the icthylographic regions of Oregon. Copeia

ing multivariate data in an ordered two-way table by clas- 2: 423_32.sification of the individuals and the attributes. CorneB Hutchinson. G.E. 1957.A Treatise on Limnology Vol. 1. Wiley,

University. Ithaca, NY. York. NY.

Biological and Ecological Dimensions 389

4

Hutchinson, G.E. 1959. Homage to Santa Rosalia, or why are Larsen, D.P., JM. Omernik, R.M. Hughes, C.M Robin, T.R.theresomanykindsofarOanals.AmericanNaturalist93:45-159. Whittier, A.J. Kinney, A.L. Gallant, and D.R. Dudley. 1986.

IEMTF [Interagency Ecosystem Management Task Force 1. Correspondence between spatial patterns in fish assem-1995. The ecosystem approach: healthy ecosystet;is and sustain- blages in Ohio streams and aquatic ecoregions. Environ-able economies. Volumes 1-111. U.S. Dept. Commerce, Na- mental Management 10: 815-828.

tional Technical Information Service, Springfield, VA. Lawton, J.H., S. Nee, A.J. Letcher, and P.H. Harvey. 1993. Ani-lllies, J. 1961. Versuch ei.ner allgemein biozonotischen real distributions: patterns and processes, pp. 41-58. In:

Gleiderung der Fliessgewasser. Verhandlungen der P.J. Edwards, R.M. May, and N.R. Webb (eds.), Large-scaleInternationalen Verinigung fur theroetische und Ecolo_/and Conse_,atio11 Biology. Black,,vell Scientific, Ox-angewandte. Lhnnologie 13: 834-844. ford, UK.

Jensen, M.E., P.S. Bourgeron, R. Everett, and I. Goodman. Leach, J.H., and R.C. Herron. 1992. A review of lake habitat1996. Ecosystem management: a landscape ecology per- classification, pp. 27-57. In: W.-D.N Busch, and P.G. Sly

spective. Water Resources Bulletin 32: 1-14. (eds.), The Developnu'nt of an Aquatic Habitat ClassificationJohnson, P.S. 1992. Oak overstory/reproduction relations in System for Lakes. CRC Press, Boca Raton, FL.

two xeric ecosystems in Michigan. Forest Ecology and Man- Leopold, L.B., and MG. Wolman. 1957. River channel pat-

agement 48: 233--240. terns: braided, meandering and straight. U.S. GeologicalJones, R.K., G. Pierpoint, G.M. Wickware, J.K. Jeglum, R.W. Survey Professional Paper 422H, Washington, D.C.

Arnup, J.M. Bowles. 1983. Field guide toforest ecosystem clas- Levin, S.A. 1992. The problem of pattern and scale in ecology.sifioation for the clay belt, site region 3e. Ministry of Natural Ecology 73: 1942-1968.Resources, Toronto, Ontario, Canada. LOICZ. 1996. Coastal facts. Land Ocean Interactions Coastal

Kay, J.J. 1991. A nonequilibrium thermodynamics framework Zoae Newsletter 1: 2.for discussing ecosystem integrity. Environmental Manage- Lotspeich, F.B. 1980. Watershed as the basic ecosystem: thismerit 15: 483-495. conceptual framework provides a basis for a natural classi-

Kent, M., and P. Coker. 1992. Vegetation Description and Analy- fication system. N'atural Resources Bulletin 16: 581-586.sis. Belhaven Press, London, UK. Lotspeich, F.B., and W.S. Platts. 1982. An integrated

Ketchum, B.K. 1972. The Waters" Edge: Critical Problems of the land-aquatic classification system. North American JournalCoastal Zone. Massachusetts Iin.stitute of Technology Press, of Fisheries Management 2: 138-149.

Cambridge, MA. Lubchenco, J., A.M. Olson, L.B. Brubaker, S.R. Carpenter,

Keys, J. Jr., C. Carpenter, S. Hooks, F. Koening, W.H. MeNab, M.M. Holland, 8.P. Hubbell, S.A. Levin, J.A. MacMahon,W. Russell, and M.L. Smith. 1996. Ecological units of the P.A. Matson, J.M. Melillo, H.A. Mooney, C.H. Peterson,Easten'L United States -- first approximation. Part I: ARC H.R. Pulliam, L, S.A. Real, P.J. Regal, and P.G. Risser. 1991.INFO; Part II: Supporting Imagery; Part III: Text. Thesustainablebiosphereinitiative:anecologicalresearchCD-ROM. USDA Forest Service, Atlanta, GA. agenda. A Report from the Ecological Society of America.

Koestler, A. 1967. The Ghost in the Machi_l¢ Macmillan, New" Ecology 72: 371-412.York, NrY. Lyons, J. 1989. Correspondence between the distribution of

Komarkova, V.1983. Comparisonofhabitattypeclassification fish assemblages in Wisconsin s_'eams and Omernik'sto some other classification methods, pp. 21-31. In: W.H. ecoregions. American Midland Naturalist 122: 163-182.Molt, and L. Hendzel (eds.), Proceedings of the Workshop on MacKey, B.G., H.A. Nix, M.F. Hu tchin son, and J.P.Southwestern Habitat Types, April 8-10, 1983, Albuquerque, MacMahon. 1988. Assessing representativeness of placesNew Mexico. USDA Forest Service, Southwestern Region, for conservation reservation and heritage listing. Environ-

Albuquerque, NM. mental Management 12: 501-514.KOppen, W. 1931. Grundriss der Klimakunde. Walter de MacKey, B.G., H.A. Nix, J.A. Stein and S.E. Cork. 1989. As-

Gruyter, Berlin, Germany. sessing the representativeness of the wet tropics ofKotar, J., J.A. Kovach, and C.T. Locey. 1988. Fieldguide to forest Queensland World Heritage property. Biological Conserva-

habitat types of northern Wisconsin. University of Wisconsin, tion 50: 279-303.Department of Forestry and Wisconsin Department of Maltby, L.S., G.B. McCullough, and E.Z. Bottomley. 1983.Natural Resources, Madison, WI. Baseline studies of 35 sclected southern Ontario wetlands. Ca-

Krajina, V.J. 1965. Biogeoclimatic zones and classification of nadian Wildlife Service, Ottawa, Canada.British Columbia. pp. l-17. In:Krajina, V.J.(ed.),Ecologyof Maxwell, J.R., CJ. Edwards, ME. Jensen, S.J. Paustian, H.Western North Amerzca. University of British Columbia, Parrott _nd D.M. Hill. 1995. A hierarchical framework ofVancouver, British Columbia. Canada. aquatic ecological units in North America. USDA Forest

K0.chler, A.W. 1964. Potential natural vegetation of the contermi. Service Genera] Technical Report NC-176, North Centralnoua United States (with ser2aratemat;at ascale of/: 3,168,000). Forest Expenrnent Station, St. Paul, MN.

American Geographical Society Special Publication 36, McGowan, j.A.1971.Oceanographicbiogeographyofthe Pa-New York, NY. cific, pp. 3-74. In: B.M. Funnell, and W.R. Riedel (eds.), The

KO.chler, A.W. 1967. Vegetation Mapping Ronald Press. New Micropaleontology qf Oceans Cambridge University Press,York, NY. Cambridge, UK

Kfichler, A.W. 1988. Ecological vegetaoon maps and their inter- McGowan, J.A. 1972. The nature of oceanic ecosystems, pp. 9-pretation, pp. 469_80. In: A.W. Kfichler and I.S. Zonneveld 28. In: C.B. Miller red.), The Biology of the Oceanic Paez_fic.Or-(eds.), Vegetation Mapping. Kluwer. Boston, MA. egon State University Press, Corvallis, OR.

K_chler, A.W. ap,d I.S. Zonneveld, editors. 1988. Vegetation McNab, W.H.. and ?.E Avers, editors. 1994. Ecological subre-

Mapping. Kluwer. Boston, MA.. gions of the United States: section descriptions. USDA Forest

390 D,H. Grossmanet al./Principles for Ecological Classification

L

Service Administrative Publication WO-WSA-5, Washing- mon (eds.), BiologicalAssessme_ta_d Criteria: Toolsfor Waterton, D.C. Resource Planning and Decision Making. Lewis Publishers,

Meador, M.R., C.R. Hupp, T.F. Cufiney, and M.E. Gurtz. 1993. Boca Raton, FL,

Methods for characterizing stream habitat as part of the Omernik, J.M., and A.L. Gallant. 1987. Ecoregions of theNational Water-Quality Assessment Program. U.S. Geo- southwest states (map, scale I: 2,500,000). EPA/600/D-87/logical Survey Open File Report 93--408, Reston, VA. 316, U.S. Environmental Protection Agency, Environmen-

Merriam, C.H. 1898. Life zones and crop zones of the United tal Research Laboratory, Corvallis, OR.

States. U.S. Department of Agriculture Division Biological Omernik, J.M., and A.L. Gallant_ 1989. Defining regions forSurvey, Bulletin 10, Washington, D.C. evaluating environmental resources. Unpublished manu-

Mitsch, W.J., and J.G. Gosselink. 1986. Wethlnds. Van script prepared for Global Natural Resource MonitoringNostrand Reinhold, New York, NY. and Assessment, Preparing for the 21st Century, Septern-

Monsi, M. 1960. Dry matter reproduction h'_ plants. 1. Sche- her 24-30, 1989, Venice, Italy. Available from U.S. Environ-mata of dry matter reproduction. Botanical Maga._ine 73: mental Protection Agency, Environmental Research81-90. Laboratory, Corvallis, OR.

Morrison, D.F. 1976. Multivariate Statistical Methods. O'Neill, R.V., D.L. De,aa_gelis, J.B. Waide, and I'.F.H, Allen.

McGraw-Hill, New York, NY. 1986. A Hierarcliical Concept q( Ecosystems. Princeton Uni-Moyle, J.B. 1956. Relationships between chemistry of Mifine- versity Press, Princeton, N J.

sota surface waters and wildlife management. Journal of Orians, G.H. 1975.Diversity, stability and maturi_, in naturalWildlife Management 20: 303-320. systems, pp. 139-150. In: W.H. van Dobben and R.H.

Mueller-Dombois, D. and H. Ellenberg. 1974. Aims and Lowe-McConneU (eds.), Unifying Concepts in Ecology. W.Methods of Vegetation Ecology. Wiley, New York, NY. Junk, The Hague, Netherlands.

Muller, P. 1974. Aspects of Zoogeography. W. Junk, The Hague, Paulsen, S.G. and R.A. Linthurst. 1994. Biological monitoringNetherlands. in the Environmental Monitoring and Assessment Pro-

Naeem, S., L.J. Thompson, S.P. Lawler, J.H. Lawton, and R.M. grain. In: A.L. Loeb, and A. Spacie (eds.), Biological Moifi-Wood fin. 1994. Declining biodiversity can alter the perfor- torhtgofAquaticSystcms. Lewis Publishers, Boca Raton, FL.mance of ecosystems. Nature 368: 734-737. Peel, R.K. 1980. Ordination as a tool for analyzing complex

Naiman, R.J., D.G. Lonzarich, T.J. Beechie, and S.C, Ralph. data sets. Vegetatio42: 171-174,1992. General principles of classification and the assess- Peters, R.H. 1976. Tautology in evolution and ecology. Ameri-ment of conservation potential in rivers. In: PJ. Boon, P. can Naturalist 110: 1-12.Calow, and G.E. Petts (eds.), River Conservation and Mall- Pfister, R.D., and S.F. Arno. 1980. Classifying forest habitatagement. Wiley, Chichester, UK. types based on potential climax vegetation. Forest Science

Nelson, R.L, W.S. Platts, D.P, Larsen, and S.E. Iensen. 1992. 26: 572.-70.

Trout distribution and habitat in relation to geology and Pfister, R,D., B.L. Kovalchik, $.F. Arno, and R.C. Presby. 1977.geomorphology in the North Fork Humboldt River drain- Forest habitat types of Montana. USDA Forest ServiceagG northeastern Nevada. Transqctions of the American General Technical Report lNT-34, Intermountain ForestFisheries Society 121: 405-47,6. and Range Experiment Station, Ogden, UT.

Nix, H.A. 1982. Environmental determinants and evolution in Pielou, E.C. 1977.Mathematical Ecology.Wiley, New York, NY.Terra Australia. pp. 47-66. In: W.R. Barker, and PJ.M. Pielou, E.C. 1979. Biogeography. Wiley-Interscience, NewGreenslade (eds.), editors. Evolu lion of the Floraand Fauna of York, NY.Arid Australia. Peacock Publications, Frewville, Australia. Pinay, G., A. Fabre, P.H. Vervier and F. GazeBe. 1992, Control

NOAA [National Oceanographic and Atmospheric Adminis- of C, N, P distribution in soils of riparian forest. LandscapetrationJ. 1985. Natiottalestuarinefnventory: data allas. Volume Ecology6: 121-132.l:physicalandhydrologiccharacteristies. NOAA.-NOS Strate- Platts, W.S. 1979. Including the fishery system in land plan-gic Assessment Branch, Washington, D.C. ning. USDA Forest Service, General Technical Report

Noy-Meir, I. 1973. Data transformations in ecological ordifia- INT-60, Intermountain Forest and Range Experiment Sta-tion. I. Some advantages of non-centering. Ecology61: 329- lion, Ogden, UT34l. Platts. W.S. 1980. A plea for fishery habitat classification. Fish-

Odum, F.P. 1971. Fundamentals of Ecology. Third edition. W.B. eries 5(1): 2-6.Saunders, Philadelphia, PA Pojar, J., K. Klink, and J. Meidinger. 1987. Biogeoclima_c eco-

Ogden, J.C., J.W. Porter, N.P. Smith. A,M. Szmant, W.C. Jaap, system classification in British Columbia. Forest Ecologyandand D. Forcucci. 1994. A long-term interdisciplinry study Management 22: 119-154.of the Florida Keys seascape. Bulletin o/Marh_e Science 54: Pregitzer, K.S., and B.V. Barnes, 1982.The use of ground flora1059-1071. to indicate edaphic factors in upland ecosystems of the

Omernik, J.M. 1986. Ecoreglons of the conterminous United McCormick Experimental Forest, Upper Michigan. Cana-States (map, scale 1: 7,500,000J. U.S. Environmental Protec- dian Journal of Forest Research 12: 661-672.lion Agency, Environmental Research Laboratory, Prentice, I.C. 1977 Non-metric ordination methods in ecol-Corvallis, OR. ogy. Ecolo,'_.W65: 85-94.

Omernik, J.M. 1987. Ecoregions of the conterminous United Pritcharc2, D.W. 1967. What is an estuary? pp. 3.-5. In: G.HStates. Atmals oj the Association of American Gco,,_'aphers77: Lauff (ed.), Estuaries. AAA$ Press, Washington, D.C.118-125. Raunkier, C. 1937. Plant LFe Forms. Clarendon Press. Oxford,

Omernik, J.M. 1995. Ecoregions: a framework for environ- UKmental management, pp. 49-62. In: W. Davis. and T. Si-

Biological and Ecological Dimensions 391

Ray, G.C. !991 Coastal-zone biodi:ersily patterns. BioScicnce partment of Natural Resources lnvesfiga_ion Report 417,41: 490--,'98. Minneapolis, Nil.

Ray, G.C., and BP Hayden. I992 Coastal zone ecotol'_es, pp. Scott, ].M., f_. Davis, B Csuti, R. Noss, B. Buttterfield, C,403-t20 In: F di Castri, a;'_d AJ Hansen (eds), Landscape Groves, Ill Anderson,S. Caicco, F, D'Erchia, T.C, Edwards,

Bou_adarics: Consequences for Biedfi'crs!!y and Ecol_\<ic:_ jr., J. UBman, and R G WrighL 1993. Gap analysis: protect-Flou,s. Springer-Verlag, New York, NY. ing biodb,'ersity using geographic informafion systems.

Ray, GC., B.P Elayden, M.G. McCormick*Ray, and T.M. iVihB{feMono_,'raplis, 123.Sn'dth. 1997. Land-seascape diversity of Ihe U.B.East Coast Seaber, P.R., F.P. Kapinos, and G.L. Knapp. 1984. Statecoastal zone with particular reference to estuaries, pp. hydrologic unit maps. U.S. Geological Survey Open-file357-371. In: R.F.G. Ormond, I. Gage, and M.V :M'_gel Report 8-I-708, U.S. Geological Sur:,ey, Res{on, VA.(eds.), Marine Biodiversity: Causes amt Conscquc_._'cs,Cam- Shannon, E.E., and P.L. grezonik 1972, Eutrophication ana]y-

bridge Uni:ersity Press, Cambridge, UK. sis: a multivariate approach. Journal qf tlie Samtary En.gq-Reid, M.S., P.S, Bourgeron, H C. Flu mphries, and M E.jensen, neering Dit'ision qf the American Society of Civil Engineers 98

1995. Documentation of the modeling of potential vegela- (BA1): 37-57.tion at three spatial scales using biophysical settings in the Si_eldon, A.L. I972. A quantitative approach to the ctassifica-Columbia River Basin Assessrnent Area. Unpublished re- tion of inlar_d waters, pp. 203-261. In: J.V. Krutilla fed.),

port prepared for the USDA Forest Service, contract # .\'atural environmc_Its: studies in theoretical applied analysis.53_04H1-6890, by the Western Heritage Task Force, The johns Hopkins Press, Baltimore, MD.Nature Conservancy, Boulder, CO. SheIford, V.E, 1911. Ecological succession. 1.Stream fishes and

Reppert, R.T., and W. Sigleo. 1979, 1_.lithmdz,alues:concepts and method of physiographic analysis. Biological Bulletin121: 9-methods for wedlalld evaluation. U.B. Army Institute for Wa- 35.tee Resources, Fort Belvoir, VA. Shimwell, D.W. 1971. The Description slid Classification of Vege-

Rhom, C.M, J.W. Giese, and C,C. Bennett, 1987. Evaluation of ration. University of Washington Press. Seattle, WA.

an aquatic ecoregion classification of streams in Arkansas. Sims, R.A, W.D. Tov,'ill, K.A. Baldwin, and G,M. Wickware.Journal Fresha'ater Ecolo_nt4: 127-140. t989. Field guide to theforest ecosystem classification for north-

Ricker, W.E. 1934. An ecological classification of certain On- western On tario. Forestry Canada, On ratio Ministry of Na ttario streams. Publication oftlieAcademyqfNatura! Scie_icesof ural Resources, Thur_der Bay, Ontario, Canada.

Philadelphia 101: 277-341. Sly, P.G., and W.-D.N. Busch. 1992. A system for aquatic habi-Robins, C.R. 1971. Distributional patterns of fishes from tatclassificafionoflakes.pp. 15-26.h'cW.-D.N. Busch, and

coastal and shelf waters of the tropical western Atlantic. P,G. Sly (eds.), Tlie Dt,¢,elopmentofan Aquatic Habitat Classiifi-

pp. 249-255. In: Symposium on Investigations and Resources qf cation System for laTkes.CRC Press, Boca Raton, FL.the Caribbean Sea andAdjaceut Regiosls.FAO Papers on Pish_ Spies, T.A., and B.V. Barnes, 1985. A multifactor ecologicaltries Resources, FAO, Rome, ItaIy. classification of the northern hardwood and conifer eco-

Robins, C.R. 1992. Saltu'ater Fish. Smithmark, New York, NY. systems of Sylvania Recreation Area, Upper Peninsula,Robins, C.R., and G,C, Ray. I986. A FieldGuhfe to Atlantic Coast Michigan. Canadian1]ourlml of Forest Research 15: 949-960,

Fishes of.X'oriliAmerica. Houghton Mifflin, Boston, MA, Stearns, S.C. 1976. Life history tactics: a review of the ideas.Rolof, G,J. 1994. Using an ecological classification system and Quarterly Re-view of Biology 51: 347.

wildlife habitat models in forest planning. Ph,D, disserta- Steele, J.H. 1989. The ocean "landscape." Landscape Ecof%'y 3:

tion, Michigan State University, East Lansing, Mh 186-192.Rowe, J.S. 1961. The level-of-in tegrafion concept and ecology, Strahler, A.N. 1957, Quan titative analysis of watershed geo-

EcoloBnj 42: 4_427, morphology. American Geophysical Union Transactions 38:Rowe, J.S. 1972. Forest regions of Canada (with separate map at a 913-920,

scale of 1: 6,336,000). Publication No. 1300, Canadian For- Strong, W,L., E.T. Oswald, and D.J. Downing. 1990, The Ca-estry Service, Ottawa, Ontario, Canada. nadian vegetation classification system, first approxima-

Rowe, J.S. 1980. The common denominator in land classifica- tion. National Vegetation Working Grou p, Canadiantion in Canada: An ecological approach to mapping, For. Committee on Ecological Land Classification. Ecologicalestry Chronicle 6: 19-30. Land Classification Series, No. 25, Sustainable Develop-

Rowe, J.S. 1984. Forestland classification: limitations of the ment, Corporate Policy Group, Enviror_ment Canada Ot-

use of vegetation, pp. 132-147. In: J.G, Bockheim fed,), For- taws, Ontario, Canada.est Land Classification: Experiences, Problems, Perspectives. Sukachev, V. 1945. Biogeocoenology and phytocoenolog2.'.

Proceedings of the symposium, March 18-20, 1984. Uni- CR. Academy of Science, USSR, 47:429431versity of Wisconsin, Madison, WI. Swanson, F.J,, T.K. Kratz, N. Caine, and R,G. Wooamansee,

Ro;ve, J.S., and J.W, Sheard. 1981. Ecological land classification: 1988. Landform effects on ecosystem patterns and pro-a survey approach. Dlvironme,tai 3danagcment 5: 451-:i64. cesses. Bioscience 38: 92-98.

Roy, P,S. t984. New South Wales estuaries: their origin and Tans!ey, A,G. 1935. The use and abuse of vegetational con-evolution, pp. 99-12t. In: B.G. Thorn fed.), Coastal geo- cepts and terms. Ecolo_/16: 284-307.morphology in AustraBa. Academic Press, Sydney, Austra- Tiner, R.W., Jr. 1984. Wetlands of the United Stares: current statusha. and recent tren&. U.S. Fish and Wildlife Service, U.S. Gov-

Rtibel, E. 1930. Die pflanzengesellsdl*(Oen der Erde. Huber eminent Printing Office. Washington, D.C.Verlag, Bern, Switzerland. Tonn, W.M. 1990. Climate change and fish communities: a

Schupp, D.H. 1992. An ecological classificatioi-, of Minnesota conceptual framework. Transactions of the American Fish-lakes with associated fish communifies. Minnesota De- tries Society 119: 337-352.

392 D,H. Grossman et al./Principles for Ecological Classification i_

Tonn, W.M., and J.J. Magnuson. 1982. Patterns in species com-

position and richness in fish assemblages in northern Wis- THE AUTHORSconsm lakes. E_oIow 6.x 1149-I166.

Trewartha, G.T. 1968. Introduction to Climate. 4th edition. Dennis H. GrossmanMcGraw-Hill, New York, NY. Conservation Science Division,

Troll, C. 1941. Studien zur vergleichenden Geographie der The Nature Conservancy,

Hochgebirge. Bonn: Friedrich Wilhelm Universitat, 4245 N. Fairfax Drive, Suite 100,

Bericht 23: 94-96. Arlington, VA 22203, USATroll, C. 1943. Die Frostwechselhaufigkeit in den Luft - und

Bodenklimaten der Erde. Meteororologische Zeitschrift 60: Patrick Bourgeron161-171.

Troll, C. 1955. Der jahreszeitliche Ablauf des Natur- Institute for Arctic andAlpine Reserach,

geschehens in den verschiedenen Klimagurteln der Erde. University of Colorado,Studium Generale 8: 713-733. Boulder, CO 80309

Troll, C. 1956. Der Klfma- und Vegetationsau)bau der Erde fin "

Lichte neuer Forachungen. Akademie der Wissencschaften Wolf-Dieter H. Buschund Literatur, Mainz, pp. 216-229. U.S. Fish and Wildlife Service,

T0.xen, R. 1956. Die heutige potentielle natO,rllche vegetation 405 N. French Road, Suite 120A,

als Gegenstand der Vegetationskartierung. Angew. Buffalo, NY 14228, USAPflanzensoziol. fStolzenau, Weser) 13: 5-42.

Udvardy, M.D.F. 1975. A classification of the biogeographical David Clelandprovinces of the world. International Union for Conserva-tion of Nature and Natural Resources, Occasional Paper Forestry Services Lab,

No. 18. Morges, Switzerland USDA Forest Service,UNESCO [United Nations Educational, Scientific and Cul- -P.O. Box 898

tural Organization]. 1973. International Classification and Rhinelander, WI 54501, USAMapping of Vegetation, Series 6, Ecology and Conserva-tion. UNESCO, Paris, France. William Platts

Urban, D.L., R.V. O'Neill, and H.H. Shugart, Jr. 1987. Land- Platts Consulting,scape ecology: a hierarchical perspective can help scien- 1603 Sunrise Rim,fists understand spatial patterns. Bioscience 37: 119-127. Boise, ID 83705, USA

USDA Soil Conservation Service. 1981. Land resource regions

and majorlandresourceareasofthe UnitedStates. Agriculture G. Carleton RayHandbook No. 296, USDA Soil Conservation Service, Department of Environmental Sciences,Washington, D.C.

USGS [United States Geological Survey]. 1982. Hydrologic unit University of Virginia,map of the United States. U.S. Government Printing Office, Clark Hall,Washington, D.C. Charlottesville, VA 22903, USA

Valiela, I. 1984. Marine Ecological Processes. Springer-Verlag,New York, NY. C. Richard Robins

Walter, H. 1973. Die vegetation der Erde I. Tropische and Kansas Museum of Natural History and

subtropische Zonen. Third edition. Fisher, Jena, Stuttgart, Biodiversity Research Center,

Germany. University of Kansas,Walter, H. 1985. Vegetation oftheearth, and ecolog@alsystems of the 448 N. 1500 Road,

geob ospl ere.Th rd ed bon. Sprmger-Verlag, Ne_,,, York, NY. Lawrence, KS 66049-9090, USAWalter, H., and E. Box. 1976. Global classification of natural

terrestrial ecosystems. Vegetatio 32: 75-81.Warren, C.E. 1979. Toward classification and rationale for wa- Gary Roloff

tershed management and stream protection. Report No. Boise Cascade Corporation,EPA_500/3-79-059, United States Environmental Protec- 123 Wolf Court,

tion Agency, Corvalis, OR. East Lansing, M148223, USAWebb, L.J., J.G. Tracey, W.T. Williams, and G.N. Lance. 1970.

Studies in the numerical analysis of complex rainforestcommunities. V. A comparison of the properties of floristicand physiognomic-structural data. Journal of Ecology 58:203-232. esthoff V., and A.J. den Held. 1969. Plantengemenschappen

Wergner, M.J.A. and J.T.C. Sprangers. 1982. Comparison of in Nederland. Thieme, Zutphen.floristic and structural classification of vegetation. Westhoff. V.. and E. van der Maarel. I978 The Braun-Vegetatio 50: 175-183. Blanquet approach. Chapter 20. In: RH. Whittaker (ed.),

Wertz, W.A., and J.A. Arnold. 1972. Land Systems hwentory. Classification qf Plant Communities. W. Junk, The Hague,USDA Forest Service, Intermountain Region, Ogden, UT. Netherlands.

Biological and Ecological Dimensions 393

Whittaker, R.H. 1956. Vegetation of the Great Smoky Moun- Winter, T.C. 1977. Classification of the hydrologic settings oftains. EcologicalMonographs 26: 1-80. lakes in the north central United States. Water Resources Re-

Whittaker, R.H. 1962. Classification of natural communities, search 13:753-767Botanical Review 28: 1-2,39. Witmer, R.E. 1978. U.S. Geological Survey land-use and land-

Whittaker, R.H. 1972. Evolution and measurement of species cover classification system. Journal of Forestry 76: 661-668.

diversity. Taxon 21: 213-351. Zak, D.R., G. E. Host, and K.S. Pregitzer. 1989. Regional vari-Whittaker, R.H. 1975.CommunitiesandEcosystems. Secondedi- ability in nitrogen mineralization, nitrification, and

tion. MacMillan, New York NY. overstory biomass in northern Lower Michigan. Canadian

Whittaker, R.H. (ed.). 1978. Classification of Plant Comniunities. Journal Forest Research 19: 1521-1526.W. Junk The Hague, Netherlands. ZaK D.R., K.S. Pregitzer, and G.E. Host. 1986. Landscape vari-

Whittaker, R.H. (ed.), 1980. Classification of Plant Communities. ation in nitrogen mineralization and nitrification. Cana-Volume Ih W. Junk The Hague, Netherlands. dian Journal Forest Research 16: 1258-1263.

Whittier, T.R., R.M. Hughes, and D.P. Larsen. 1988. Corre- Zonneveld, I.S. I988. Landscape (ecosystem) and vegetation

spondence between ecoregions and spatial patterns in maps, their relation and purpose, pp. 481M86. In: A.W.stream ecosystems in Oregon. Canadian Journal of Fisheries K0.chler, and I.S. Zonneveld (eds.). Vegetation Mapping.and Aquatic Sciences 45: 1264-1278. Kluwer, Boston, Massachusetts.

Wiken, E.B., compiler. 1986. Terrestrial ecozones of Canada. Zonneveld, I.S. 1989. The land unit -- a fundamental conceptEcological Land Classification Series No. 19, Environraeii t in landscape ecology, and its application. Landscape EcologyCanada, Lands Directorate, Ottawa, Ontario, Canada. 3: 67-86.