al scales applications the physical environment as a...

The Physical Environment asa Basis for Managing EcosystemsFrederick J. Swanson, Julia A. Jones, and Gordon E. Grant

Landscape Patterns and Processes 231History of Landscape Change in the Pacific Northwest 232Regional Ecological Assessments 233Future Landscape Management 233Conclusions 236Acknowledgments 237Literature Cited 237

The physical environment sets the stage on whichecological phenomena operate. Our understandingof the physical environment, especially disturbancesin forest and stream systems, has evolved throughtime. How we have used what we have learned haschanged in tandem with changes in managementfocus. Early in the 20th century, when resource ex-traction was preeminent, hydrology and geomor-phology were used to support activities such as repairof landslide-damaged roads. In the mid-20th cen-tury, environmental concerns grew, expressed in part

by a series of laws—the Clean Water Act, the Na-tional Environmental Policy Act, the National ForestManagement Act, and the Endangered Species Act—which directed public land managers to analyze anddisclose the effects of proposed actions and to protectspecies and ecosystem productivity. Information onthe physical environment was used to minimize un-desired environmental effects through actions suchas surveying proposed road locations to avoid poten-tial landslide sites.

Late in the 20th century, growing concerns about

229

.al Scales

applications

e manage-g. Ecology

Whitcomb,ects of for-deciduous

nated land-New York:

w environ-U2 tics 7:81-

ma! Ecology

pe bound-boundaries:flows, ed. A.er-Verlag.5. Boundarydying land-

unities near

i. 1990. Pat-in differentEcology and

230 Section III. Approaches to Management at Larger Spatial Scales

the effects of management actions dispersed overlarge watersheds (considered under the rubric of cu-mulative watershed effects) has led to more holisticviews of watershed function. Understanding of thephysical environment dominates watershed analysesdesigned to identify and mitigate cumulative water-shed effects (Washington Forest Practice Board 1992,FEMAT 1993).

A parallel development during the late 1980s and1990s has been concern with potential loss ofspecies, especially those associated with old-growthforests. This has led to conservation strategies fo-cused on recovery of forest habitat for wildlifespecies. Knowledge about the physical environmenthas played a significant role in conservation strate-gies for fish species, but less so for terrestrial species.

What will the 21st century bring? We expect un-derstanding of the physical environment to gain amore prominent position in land-use planning,forming the starting point for ecosystem manage-ment on public lands. Emphasis will shift from com-modity production and single species to sustainingthe dynamic ecosystems that support life. Develop-ment of the concept of managing within a range ofnatural variability is a harbinger of this shift. Thisconcept recognizes that ecosystems are naturally verydynamic and subject to periodic disturbances. Assuch, they have exhibited a range of conditions (e.g.,successional states) in the past that resulted in im-portant controls on geographic patterns of distur-bances and biotic resources.

The management implication of this perspective isthat ecosystem patterns and processes should bemanaged with adequate similarity to those of thepast so that diversity and productivity of assemblagesof native species can be maintained in conformancewith federal laws governing public forest land wheresome cutting and active fire management are permit-ted. Such an approach contrasts with past landscapemanagement, which was largely a de facto conse-quence of numerous stand-level decision's. In con-trast, an approach considering ecos ystem dynamics isbased on an understanding of landscape history andnatural disturbance regimes. It is intended to fit themanaged biotic landscape with the natural biologicaland physical character of landscapes, including dis-turbance regimes. The rationale is that native species

have adapted to the natural disturbance regime overthousands of years. The survival potential of nativespecies would be expected to be reduced if their en-vironment is pushed outside the range of conditionsthat existed over this time period. Karr and Freemark(1985:167), for example, argue that "disturbanceregimes ... must be protected to preserve associatedgenetic (Frankel and Soule 1981), population (Frank-lin 1980), and assemblage (Karr 1982a, 1982b; Kush-lan 1979) dynamics."

Biotic landscapes include the patterns of vegeta-tion patches of various ages and compositions andthe fauna that find habitat in such vegetation pat-terns. Natural biotic patterns reflect, in part, thephysical landscape of landforms, climate, and soilpatterns. Managed biotic landscapes may be forcedinto poor fits with physical landscapes, creating bio-logical and social problems. Fire suppression inecosystems with naturally high fire frequency, for ex-ample, has led to fuel buildups that increased fireseverity when it occurred. Forest cutting has reducedthe extent of old-growth forest conditions, thus re-ducing habitat for a host of species. On the otherhand, efforts to create areas of old-growth forestwhere it would not naturally occur because of fre-quent fire, wind, or other disturbances, have eitherfailed or created unnatural habitat patterns. Mid-slope roads, a notably unnatural feature in land-scapes, can contribute to increased peak flows indownstream areas (Wemple 1994, Jones and Grant1996) and be major sources of debris slides (Sidle etal. 1985).

In this chapter, we consider some implications ofthe physical environment and ecosystem dynamics inland management. We begin with a brief descriptionof pattern-process relations in landscapes as a basisfor managing ecosystems. We follow with a discus-sion of the history of landscape management in thePacific Northwest and its relation to our understand-ing of the physical environment. To conclude, weconsider implementation of these concepts in the21st century.

These perspectives and the examples consideredare derived from the Pacific Northwest, an area ofstrong interactions between biota, physical environ-ment, and instructive pre- and post-settlement landuse histories. Rich vegetation and fire histories can be

PATTERN - GENERATIONPROCESSES

BIOTICPATTERNS

PATTERN PROCESS

RESPONSESTO PATTERN

GEOPHYSICAL

PATTERNS

15. The Physical Environment as a Basis for Managing Ecosystems 231

described, for example, from dendrochronologicaldata (records extending back to 500-800 years beforethe present), paleoecological data (records extendingback to 40,000 years before the present), and satelliteremote-sensing data (records extending back to 1972to the present).

Z:=12=IMST,.-WM

Landscape Patterns and ProcessesTo see the value of using the physical environment asa basis for ecosystem management, it is necessary tohave a conceptual framework. Such a frameworkconsiders forest landscapes and riverscapes in termsof a hierarchy of spatial and temporal scales. Ques-tions regarding evolution of landscape patterns innatural and managed systems must be posed and an-swered at the correct temporal and spatial scale (Del-court et al. 1983, O'Neill et al. 1986, Gregory et al.1991). Many terms are useful for identifying scale,depending on subject matter—site, landscape, prov-ince, and region are common geographic scales. Thespecific definition of these terms depends on objec-tives: A region, for example, may be defined by theranges of critical species of interest. Many toughenvironmental problems today are a result of man-agement decisions made at fine spatial scales, whichhad undesired, aggregated consequences at coarserscales.

Interactions between biotic and physical patternsand processes (Figure 15.1) operate at each scale.Landscape patterns are dynamic as a result of inter-actions between vegetation succession and distur-bance processes. Biotic patterns can take the formof patchworks of vegetation in various stages ofdevelopment. Vegetation patchworks partly reflectpattern-creating (disturbance) processes and geo-physical patterns, such as landform controls ondisturbance or variations in soil properties that influ-ence vegetation. Vegetation patterns in natural land-scapes include both the indelible imprint of patchescreated by resource limitations (sensu Forman andGodron 1986), such as lack of soil on rock outcrops,and ephemeral patches initiated by biotic distur-bances. Landform influences vary between distur-bance processes that follow local gravitational flowpaths (e.g., landslides, floods) and those that do not

Figure 15.1 Dominant pattern-process relations.Geophysical patterns (e.g., soils, topography) underliebiotic patterns (e.g., vegetation type, productivity) andmay influence landscape-related processes. Distur-bance processes generate patterns; other processesmay simply respond to patterns without modifying ex-isting patterns.

(e.g., wildfire, windthrow). Human activities (e.g.,cutting units) may overprint natural patterns, creat-ing intentional or unintentional patterns (e.g., thosearising from escaped slash fires). Some processes,such as generation of low streamflows, respond to bi-otic and landform patterns without creating newlandscape patterns.

Landscapes are composed of patchwork and net-work structures, each with characteristic dynamicsand functions. In forest landscape studies, patchworkstructures are typically a collection of patches of var-ied vegetation type and seral stage. Natural networkstructures include streams, riparian areas, and ridge-lines. Some disturbance processes, such as floods anddebris flows, propagate through networks, while oth-ers, such as wildfire and windthrow, move throughpatchworks.

Most landscape ecology studies have examinedpatchwork structures; few have considered networkcomponents of landscapes; and still fewer have dealtwith network–patchwork interactions. Understand-ing interactions between landscape structures of thesame or different type is critical to forest and water-shed management. Road segments, for example, may

LAND OWNERSHIP1972 1988


follow ridges and large streams (fourth-order andlarger). But roads in steep terrain are likely to inter-sect groundwater flow paths and first- through third-order streams at right angles. This may accentuatethe effects of roads in increasing the density ofstream channels and thereby the potential of water-sheds to generate peak flows (Wemple 1994, Jonesand Grant 1996). An extensive network of roadditches and stream flow paths may help to transportincreased runoff from clearcut patches (Han 1981,1986) to downstream areas where peak flows may beincreased (Jones and Grant 1996). Such interactionsbetween network and patchwork structures are sig-nificant but often overlooked because of disciplinaryinstincts; stream ecologists tend to see landscapes asnetworks, while forest and wildlife ecologists seethem as patchworks.

SIZOWESISZUMNISI

History of Landscape Changein the Pacific NorthwestPostglacial landscape patterns in the Pacific North-west have developed sequentially through three

stages: (1) a wild landscape dominated by naturaldisturbances and actions of native people, spanningmost of the Holocene until the early 1800s; (2) thelandscape managed by European settlers throughmuch of the 19th and 20th centuries; and (3) the pre-sent period of regional ecological assessments andassociated management plans focused on speciesconservation on public lands. The physical environ-ment dominated ecosystem change in the first stage;human forces attempted to override some influencesof the physical environment in the second stage; andthe physical environment will be an increasing con-sideration in management through the present andfuture.

The transition from the wild landscape of the firststage to the present, managed landscape createdduring the second stage has occurred over a centuryof piecemeal land-use decisions. A simple, visual ex-pression of this is revealed in maps of closed-canopyforest and of disturbance patterns (mainly harvestcutting) from 1972 to 1988 in an area of mixed own-ership and varied land-use designations in the cen-tral Cascade Mountains of Oregon (Figure 15.2)(Spies et al. 1994). Here we see patterns created by

CENTRALCASCADE

MOUNTAINS

LEGEND■ Conifer Type

Other Types

III Water

SCALE=■=1

0 10 km

r-sir J

oREOON

Figure 15.2 (a) Maps of closed-canopy conifer (>40 years inage) forest type in 1972 and 1988 for a 259,000-ha study area inthe Oregon Cascade Range (from Spies et al. 1994). (b) Landownership and land-use designations (modified from Spies et al.1994).

(b)

Public Land

■ Private Land

III Wilderness andResearch Natural Area

(a)

1


large, aggregated cutting units on private industrialands cut on approximately 50-year rotations,smaller, dispersed cutting units on public landswhere the cutting cycle was 80 to 100 years, and veryow levels of disturbance in Wilderness, ResearchN.atural Areas, and other reserves where logging doeslot occur and fire is suppressed. Note how the re-serves and other areas of limited forest cuttingemerge as distinctive landscape features over this 16-vear period.

Management of landscape patterns on publiclands in the Pacific Northwest has changed dramati-cally in the past few years. After 40 years of dis-Persed-patch clearcutting, brief consideration was3iven to more aggregation of cutting units, assuming:hat the rate of cutting was unlikely to change(Franklin and Forman 1987, Swanson and Franklin1992). Yet, from the early 1990s, cutting rates havedramatically decreased, and concepts about how wemanage landscape patterns have changed. For exam-ple, there is growing interest in allowing more wild-:ire in the wilderness areas, and state regulations are.imiting the size of cutting units on private lands. Themost profound change will probably be in areas for-merly designated as general forest lands of the U.S.Forest Service and Bureau of Land Management.Management in those areas will be altered by theNorthwest Forest Plan developed by the ForestEcosystem Management Assessment Team (FEMAT1993).

Regional Ecological AssessmentsRecently for the first time, broad-scale patterns ofsorest vegetation in the Pacific Northwest were sub-ect to conscious design. Regional ecological assess-ments, such as FEMAT (1993), the Columbia RiverBasin/East-Side Ecosystem Project, and the SierraNevada Ecosystem Project, are defining broad tem-plates for ecosystem management on major pieces ofpublic lands in the United States. Some of these as-sessments derive from legal and other imperatives toConserve threatened and endangered species and3ld-growth forest ecosystems.

These regional management plans use a variety ofand designations to define future patterns of forestdevelopment. Existing reserve lands, such as con-

gressionally designated wilderness areas, are com-bined with new designations to create an interactingnetwork of reserves. Guidance for management oflands between reserves and for designation of ripar-ian corridors is intended to provide dispersal habitatbetween reserves. In these and other ways, reservesestablished in the past as independent entities are in-tended to fit within a larger design to meet ecologicalobjectives. Physical processes and landscape featureshave minor roles in the design.

The focus on species conservation and the cumu-lative effects of past ecosystem alterations severelylimit opportunities to reintroduce features of naturaldisturbance regimes in some areas. This is particu-larly acute in areas characterized by infrequent,stand-replacement wildfire and where managementis focused on conserving old-growth forest habitat.Reserve placement and design in the Northwest For-est Plan, for example, are structured to provide old-growth habitat for the northern spotted owl and toaccommodate aquatic resources, but with little refer-ence to landscape patterns created by historic, nat-ural disturbance regimes. Under this regional plan,only a small fraction of the forest landscape of publicland is subject to cutting—but the forest patterns cre-ated on those lands are highly fragmented and unlikenatural patterns. Perhaps once certain levels ofecosystem recovery have been achieved, it will bepossible to more nearly match management patternswith natural patterns.

In other regions, ecological assessments and plansmay provide for a somewhat better fit of managed bi-otic patterns to patterns controlled by natural distur-bance regimes and landforms. Where wildfire wasfrequent and has been suppressed for several naturalfire recurrence intervals (with resulting problems ofinsect outbreaks and buildup of fuels), there isgreater attention to matching management practiceswith natural disturbance regimes. This may be thecase in the Columbia River Basin assessment andplan.

Future Landscape ManagementWe expect th y.' management of forest and riverineecosystems in the 21st century, at least on publiclands, will depart substantially from earlier and cur-

Early-Seral =•3=

Late-Seralpark-like '

standsStream bank

stability 0

^ Natural conditionsCurrent conditions

• Current median

A

40 60 80 10020

Sizedistribution


rent approaches. These traditional approaches havebeen based either on site-specific best managementpractices, which lack long-term, large-scale perspec-tives, or on the habitat-restoration emphasis of con-servation strategies for threatened and endangeredspecies. Emerging new approaches will define arange of desired ecosystem states based on naturalstates and disturbance regimes interpreted fromanalysis of past events and controls of climate, land-forms, soils, etc., on process location and rates.Hence, management interventions are treated asmodifications of the natural disturbance regime. Theycan be applied at the site or landscape level—the firstis most suitable to sustaining productivity and thelatter to sustaining elements of the larger ecosystem.

Characterization of landscape dynamics is essen-tial to planning ecosystem management. This can beapproached in several complementary ways. Onemethod is to define the range of past ecosystem con-ditions, such as extent of old-growth forests, earlysuccessional habitat, or unstable stream banks (Fig-ure 15.3)(Caraher et al. 1992). A second approach fo-cuses on disturbance processes in terms of the rangeof frequency severity, and geographic pattern (suchas size distribution) of patches created by theseprocesses before and during periods of intensiveforestry management (Figure 15.4)(Swanson et al.1993). Mapping forest conditions or the extent of in-dividual disturbance events through time can helpdefine either the range of vegetation conditions orthe disturbance regime.

An even more useful, spatially explicit characteri-zation can be developed by mapping disturbanceregime units across landscapes (Figure 15.5). This is

Range of Conditions (% area)

Figure 15.3 Example of range of interpreted naturaland present ecosystem conditions for the Silvies Riverarea (from Caraher et al. 1992, p. 22).

accomplished by interpreting event histories for indi-vidual sites and then compiling site histories across alandscape. Disturbance regime patterns are inter-preted by considering effects of topography, vegeta-tion type, soil, climate, or other relevant factors thatexercise long-term controls on the frequency andseverity of disturbance processes.

Mapping units for characterizing wildfire distur-bance regimes may describe the frequency and sever-ity of disturbances—such as infrequent, stand-re-placement wildfires or 60-year rotation clearcuttingon private land holdings of the forest products in-dustry. In some areas, landscapes can be stratified byvegetation types (Barrett and Arno 1991, Agee 1993)to provide a basis for mapping fire regimes. Wherevegetation types are not highly differentiated withrespect to fire regimes, topography may be more use-ful in interpreting fire regimes over landscapes. In ei-ther case, information on landform effects on firepatterns assists in interpreting long-term landscape

Frequency

Figure 15.4 Hypothetical representation of a nat-ural disturbance regime as the large, irregular "cloud"showing a probability distribution of wildfire events.Box 1 represents a management system of dispersedclearcuts with broadcast burning, assuming no land-scape disturbance by wind or fire. Box 2 represents thcdisturbance regime resulting from the interaction of

the management system (box 1) with disturbanceprocesses that could not be suppressed (e.g., wind-throw at stand edges and wildfire). The black area represents a range of conditions resulting from a management system within the range of natural variability, but

not a mimic of the full natural range (Swanson et al

1993).

Fr

1,

1

historic fires from ages ofregenerated trees, scars

•II.125 150

age of tree, scar

10-

A6

T 2 4C

,2 2-

H 0

low seventy, high frequency fireC3 moderate seventy, moderate

frequency fireE high severity,

low frequencytire

1.375 400

N

E

T

0

R

K

IN low debris flow potential

high debns flow potentialdebris slide and flow from roadbased on dendrochronology,geomorphic interpretation


dynamics beyond the time scale reflected in currentvegetation (Figure 15.5).

Disturbance regimes imposed by geomorphicprocesses are another important aspect of fittingmanaged biological landscapes with the physical en-vironment. Examples of mapable disturbance re-gimes include landslide hazards and the potential forlateral change in river channels (Grant and Swanson1995). Some elements of watershed analysis are de-signed to provide such information (Washington For-est Practice Board 1992, FEMAT 1993, Montgomeryet al. 1995). Information on geomorphic disturbanceregimes is used not only to minimize effects of man-

agement on accelerated erosion but also to identifyparts of the landscape where natural disturbancesshould not be impeded—such as stream processesthat create and maintain secondary channels on widevalley floors.

There are, however, some important challenges todefining natural ranges of variability and disturbanceregimes: First, interpretation of past disturbanceregimes has limitations. Second, native people influ-enced fire regimes to varying extents for much or allof the period that the present dominant forest specieshave been significant components of the flora of theregion. And finally, exotic species, such as undesired

1. Site level interpretation 2. Landscape history 3. Disturbance regime

event typeevent date

c. event severity

compile site data on mapgeneralize points to areas, lines

c. create mapping of events

generalize from events to regimesconsider controlling factors suchas topography, vegetation, soils

Figure 15.5 Schematic example of steps for going from site-level interpretation of distur-bance history to landscape history of disturbances to disturbance regime mapping. Fire pat-terns are an example of patch-creating disturbances, and debris flows down stream channelsare an example of disturbances in networks.


weeds and planted fish stocks, and engineered struc-tures, such as roads and dams, are unnatural parts ofthe ecosystem that are likely to stay in most cases;hence, they must be accommodated in managingecosystems at large scales (Swanson et al. 1993).There are concerted efforts to improve understandingof each of these factors. Although some uncertaintysurrounds each of these considerations, we believethat the range of natural variability concept for man-agement remains important and viable.

Adopting a management approach based on dis-turbance regimes has rather different implications interrestrial habitats of uplands areas than it does instream and riparian networks. Stream and ripariannetworks are potentially affected not only by directlocal disturbances, but also by disturbances in uplandareas that may have effects transported downslopeby streamflow and other processes. Furthermore, dis-turbances that may be limited in uplands areas mayhave substantial effects on the disturbance regime ofstream and riparian networks. Roads, for example,may affect less than 5 percent of uplands, but theymay significantly alter the size distribution of peakflow events (Jones and Grant 1996).

These concepts actually can be applied to reallandscapes. One example is the Augusta LandscapeProject on the Willamette National Forest in Oregon,a test area for ecosystem management planning (Cis-sel et al. 1994, in preparation, Wallin et al. in press).Planning for this 7,600-hectare area began withmapping disturbance event histories for wildfire,based on dendrochronology (A.D. 1500-1900) andlandslides and the use of field and aerial photographinterpretation (1950-1994). To develop maps of dis-turbance regimes, event histories have been inter-preted in light of landform influences on disturbancepatterns. This information, combined with informa-tion on present landscape conditions and land-usedesignations, has been used to develop a landscapemanagement plan that sustains aquatic and terres-trial habitat more within the range of historic varia-tion. Moreover, the plan has been modeled for sev-eral centuries into the future.

Key features of this disturbance-based Augustalandscape design include variable rotation lengths(100-300 years), varying levels of tree removal at eachcutting, and size distribution of cutting units. Selec-

tion of silvicultural prescriptions is based on wildfirefrequency, severity, and geographic patterns of thepast. The range of cutting rotations, levels of live treesretained, and unit sizes are much more variableunder this management system than they wereunder earlier patch clearcut systems or the North-west Forest Plan.

Approaches to riparian zone management also dif-fer significantly. The Willamette National Forest Planand the Northwest Forest Plan prescribe streamsidebuffer strips throughout the stream and riparian net-work. The Augusta landscape design uses a smallnumber of large reserves to protect aquatic and ripar-ian species. These are connected by wide buffersalong only the major channels. Some cutting is per-mitted within other riparian areas.

The result of the Augusta landscape design is a fu-ture landscape with forest habitat patterns more sim-ilar to the historic patterns than would be the caseunder the previous patch clearcut system or theNorthwest Forest Plan (Wallin et al. in press, Cissel etal. in preparation). This conclusion has been exam-ined in a variety of respects, including the extent ofinterior forest habitat in mature and old-growthclasses and the density of edge between forest andopen areas. Edge density is a useful measure of land-scape pattern because it may serve as an index ofblowdown potential—that is, freshly exposed forestedges are susceptible to wind damage. Also, somespecies that favor interior forest habitat, such asnorthern spotted owls, may be more susceptible topredation in landscapes with higher edge densitiesThe Augusta landscape design, therefore, has ,1higher likelihood of sustaining terrestrial and aquati,species.

En••••••••1.

ConclusionsThe use of understanding of the physical environment, including natural disturbance regimes, 11,1"evolved through the stages of helping to repair damage to commodity production systems, to minimizineffects of further management, to using the physk.c1environment and ecosystem dynamics as a basis te'management design. In the 21st century, we expecland management to better fit the managed biolo;z'

'I


cal landscape with the physical landscape and its his-toric disturbance processes and patterns. The majorimpetus for doing so is that past poor fits haveproven socially untenable; consequences of poor fitsinclude the actual or potential listing of threatened orendangered species as a result of habitat alterationwell outside the historic range of conditions.

Prototypes of landscape management designs arebeing developed that illustrate use of natural distur-bance regimes and other aspects of the physical en-vironment to plan ecosystem and watershed man-agement. In some regions, however, the need toprovide for habitat recovery following a prolongedperiod of intensive land use may slow implementa-tion of an ecosystem-based approach to manage-ment. This seems to be the case in the range of thenorthern spotted owl. In areas where landscapeshave been strongly influenced by suppression ofwildfire, a focus on restoring the disturbance regimethrough management may be more immediately ac-ceptable.

Viewed from the mid-1990s, it is interesting toconsider how difficult changing to a more ecosys-

tern-based approach to management may be. The re-cent growth of understanding about how ecosystemsand watersheds function and the major revamping ofnatural resource management approaches wouldseem to set the stage for improving the fit of man-aged landscapes to natural landscape patterns andfunctions. However, it may be well into the 21st cen-tury before much progress is made. Social changesare critical—such as increased public understandingof the role of disturbances in ecosystems. Changes inpolicy also are needed to direct planning to regionaland watershed scales, rather than restricting it to thescales of projects (e.g., timber sales) and agencymanagement units (e.g., an individual national for-est) where cumulative ecological and watershed ef-fects may be obscured.

8665T2TIMMIIBIECEMI

AcknowledgmentsWe thank John Cissel and David Wallin for man y thought-ful discussions of these topics and leadership in develop-ment of the Augusta Landscape Project.

Literature CitedAgee, J. K. 1993. Fire ecology of Pacific Northwest forests.

Washington, DC: Island Press.Barrett, S. W., and S. F. Arno. 1991. Classifying fire regimes

and defining their topographic controls in the Selway-Bitterroot Wilderness. In Proceedings of the 11th Confer-ence on Fire and Forest Meteorology, April 16-19, 1991,Missoula, Montana, ed. P. L. Andrews, and D. F. Potts.Boise, ID: USDA Forest Service, Boise National Forest.

Caraher, D. L., J. Henshaw, F. Hall, et al. 1992. Restoringecosystems in the Blue Mountains: A report to the regionalforester and the forest supervisors of the Blue MountainForests. Portland, Oregon: USDA Forest Service, PacificNorthwest Research Station.

Cissel, J. H., F. J. Swanson, W. A. McKee, and A. L. Burditt.1994. Using the past to plan the future in the PacificNorthwest. Journal of Forestry 92(8):30-31, 46.

Cissel, J., F. Swanson, D. Olson, G. Grant, S. Gregory, S.Garman, L. Ashkenas, M. Hunter, J. Kertis, J. Mayo, M.McSwain, D. Wallin. In preparation. A disturbance-basedlandscape design in a managed forest ecosystem: The Au-gusta Creek study.

Delcourt, H. R., P A. Delcourt, and T. Webb III. 1983. Dy-namic plant ecology: The spectrum of vegetationalchange in space and time. Quarterly Science Review1:153-175.

FEMAT (Forest Ecosystem Management Assessment Team).1993. Forest ecosystem management: An ecological, eco-nomic, and social assessment. Report of the Forest Ecosys-tem Management Team (FEMAT), 1993-793-071. Wash-ington, DC: GPO.

Forman, R. T. T., and M. Godron. 1986. Landscape ecology.NewYork: John Wiley & Sons.

Frankel, 0. H., and M. E. Soule. 1981. Conservation and evo-lution. New York: Cambridge University Press.

Franklin, I. R. 1980. Evolutionary change in small popula-tions. In Conservation biology: An evolutionary-ecologicalperspective, ed. M. E. Soule and B. A. Wilcox. Sunderland,MA: Sinauer.

Franklin, J. F., and R. T. T. Forman. 1987. Creating landscapepatterns by cutting: Ecological consequences and prin-ciples. Ecology 1(1): 5-18.


Grant, G. E., and F. J. Swanson. 1995. Morphology andprocesses of valley floors in mountain streams, westernCascades, Oregon. In Natural and anthropogenic influ-ences in fluvial geomorphology: The Wolman volume, ed. J.E. Costa, A. J. Miller, K. W. Potter, and P. R. Wilcock.Geophysical monograph no. 89. Washington, DC:American Geophysical Union.

Gregory; S.V., F. J. Swanson, W. A. McKee, and K.W. Cum-mins. 1991. An ecosystem perspective of riparian zones.BioScience 41 (8) :540-551.

Han, R. D. 1981. Some characteristics and consequences ofsnowmelt during rainfall in western Oregon. Journal ofHydrology 53:277-304.

1986. Effects of clearcutting on rain-on-snowrunoff in western Oregon: A new look at old studies.Water Resources Research 22(7):1095-1100.

Jones, J. A., and G. E. Grant. 1996. Peak flow responses toclearcutting and roads in small and large basins, west-ern Cascades, Oregon. Water Resources Research 32:959-974.

Karr, J. R. 1982a. Avian extinctions on Bano Colorado Is-land, Panama: A reassessment. American Naturalist119:220-239.

. 1982b. Population variability and extinction in theavifauna of a tropical land bridge island. Ecology63:1975-1978.

Karr, J. R., and K. E. Freemark. 1985. Disturbance and ver-tebrates: An integrative perspective. In The ecology ofnatural disturbance and patch dynamics, ed. S. T. A. Pick-ett and P. S. White. NewYork: Academic Press, Inc.

Kushlan, J. S. 1979. Design and management of continen-tal wildlife reserves: Lessons from the Everglades. Bio-logical Conservation 15:281-290.

Montgomery, D. R., G. E. Grant, and K. Sullivan. 1995. Wa-

tershed analysis as a framework for implementingecosystem management. Water Resources Bulletin 31(3):369-386.

O'Neill, R.V., D. L. DeAngelis, J. B. Waide, and T. F. H. Allen.1986. A hierarchical concept of ecosystems. Princeton:Princeton University Press.

Sidle, R. C., A. J. Pearce, and C. L. O'Loughlin. 1985. Hills-lope stability and land use. Water resources monograph11. Washington, DC: American Geophysical Union.

Spies, T. A., W. J. Ripple, and G. A. Bradshaw. 1994. D y-namics and pattern of a managed coniferous forest

landscape in Oregon. Ecological Applications 4(3):555-568.

Swanson, F. J., and J. F. Franklin. 1992. New forestry princi-ples from ecosystem analysis of Pacific Northwest for-ests. Ecological Applications 2(3):262-274.

Swanson, F. J., J. A. Jones, D. 0. Wallin, and J. H. Cissel.1993. Natural variability: Implications for ecosystemmanagement. In Eastside forest ecosystem health assess-ment. Vol. II, Ecosystem management: Principles and appli-cations, ed. M. E. Jensen and P. S. Bourgeron. Portland,OR: USDA Forest Service, Pacific Northwest ResearchStation.

Wallin, D., F. Swanson, and B. Marks. In press. Comparisonof managed and pre-settlement landscape dynamics in.forests of the Pacific Northwest, U.S.A. Forest Ecolog,r,and Management.

Washington Forest Practice Board. 1992. Washington ForestPractice Act Board manual: Standard methodology for con-ducting watershed analysis. Ver 1.10. Olympia, WA: De-partment of Natural Resources, Forest Practices Divi-sion.

Wemple, B. C. 1994. Hydrologic integration of forest roadswith stream networks in two basins, western Cascades, Ore-gon. Corvallis, OR: Oregon State University.

Purchased by USDA ForestService for official use

Freshwaterheart of ti-(Naiman 11995b). Becsocioeconothe comple:manageme:Changes inbon, culturelogical applthat compicstability as

1/4...-reaLing a ruresiryfor the 21st Century

Edited by Kathryn A. Kohm and Jerry F. FranklinForeword by Jack Ward Thomas

ISLAND PRESS q7c5— p.Washington, D.C. • Covelo, California

Copyright © 1997 by Island Press

All rights reserved under International and Pan-American Copyright Conventions. No part of this bookmay be reproduced in any form or by any means without permission in writing from the publisher:Island Press, 1718 Connecticut Avenue, N.W., Suite 300, Washington, DC 20009.

ISLAND PRESS is a trademark of The Center for Resource Economics.

No copyright claim is made in chapters 2, 3, 6-11, 13-19, 21, and 29, work produced in whole or in partby employees of the U.S. government.

Grateful acknowledgment is expressed for permission to publish the following copyrighted material:Figure 3.1, on page 112, from Peet, R.K. 1992. "Community Structure and Ecos ystem Function." In PlantSuccession: Theory and Prediction, edited by D.C. Glenn-Lewin, R.K. Peet, and T.T. Veblen. New York:Chapman and Hall.

Library of Congress Cataloging-in-Publication Data

Creating a forestry for the 21st century: the science of ecosystemmanagement / edited by Kathryn A. Kohm Jerry F. Franklin.

p. cm.Includes bibliographical references and index.ISBN 1-55963-398-0 (cloth). — ISBN 1-55963-399-9 (pbk.)1. Forest management—Northwest, Pacific. 2. Forest ecology—

Northwest, Pacific. 3. Forests and forestry—Northwest, Pacific.4. Ecosystem management—Northwest, Pacific. 5. Forest management.6. Forest ecology. 7. Forests and forestry. 8. Ecosystemmanagement. I. Kohm, Kathryn A. II. Franklin, Jerry F.SD144.A13C74 1997

634.9--dc20

96-32771CIP

Printed on recvcled, acid-free paper

Manufactured in the United States of America

10 9 8 7 6 5 4 3 2

al scales applications the physical environment as a...

Documents