See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/236585548

The scientific basis of forestry

Article in Annual Review of Ecology and Systematics · November 1998

DOI: 10.1146/annurev.ecolsys.29.1.435

CITATIONS READS

143 3,766

1 author:

David A Perry
Oregon State University
113 PUBLICATIONS 7,648 CITATIONS

SEE PROFILE

All content following this page was uploaded by David A Perry on 04 June 2014.

The user has requested enhancement of the downloaded file.

P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

Annu. Rev. Ecol. Syst. 1998. 29:435–66

Copyright °c 1998 by Annual Reviews. All rights reserved

THE SCIENTIFIC BASIS

OF FORESTRY
David A. Perry
Department of Forest Science, Oregon State University, Corvallis, Oregon 97331,
and Ha o Ka ’Aina, Kapa’au, North Kohala, Hawai’i

KEY WORDS: forest management, sustainability, intensive forestry

A sufficiently wise and flexible silvicultural art can be developed on the ground only by
practitioners who understand the forest as a biological entity.

F. S. Baker (10)

...the existing level of knowledge about forests is inadequate to develop sound forest
management policies.

National Research Council (114)

ABSTRACT
Over the past two decades forestry in the United States has diverged into two
approaches with quite different objectives and scientific priorities. The manage-
ment focus of most industrial lands is on increasing productivity of wood fiber via
plantations and various cultural tools, especially genetic selection, fertilization,
and control of noncrop vegetation. Federal forest management has shifted from
a similar focus to greater emphasis on protecting diversity and water. Issues of
long-term sustainability are important regardless of ownership. Science has played
and continues to play a fundamental role in all aspects. Selection for fast-growing
genotypes has increased yields on the order of 10% to 20% depending on species.
Fertilization often increases growth significantly but responses are variable and
difficult to predict. Significant questions remain concerning the sustainability of
intensive forestry, particularly when practiced over wide areas. Soils are heavily
impacted by some harvesting practices, and the degree to which damage can be
repaired by fertilizers is an important scientific issue. Intensive forestry often
results in increased pest problems. In at least one case (fusiform rust in southern
pines), a pest has been contained by selecting resistant cultivars, a situation that
may or may not be evolutionarily stable. Species diversity is clearly reduced under

435
0066-4162/98/1120-0435$08.00
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

436 PERRY

intensive management, raising questions about the functional role of species with
no commercial value. Many of the questions facing forestry science—particularly
those dealing with the relation between complexity and function—are precisely
the ones confronting basic ecology. Over the past decade scientists have labored
to develop ecosystem-based management approaches that maintain system com-
plexity and function, and scientists have increasingly played nontraditional roles
at the interface between biology, sociology, and policy.

INTRODUCTION
Forestry has been defined as “the scientific management of forests for the con-
tinuous production of goods and services” (10), though, as with agriculture,
forestry that reflects the specifics of place as well as the generalities of science
necessarily involves art. Biologic, physical, social, management, and engineer-
ing sciences all play an important role in forestry. Over the past 20 years the
scope of biologic and environmental sciences contributing to forest manage-
ment has expanded beyond ecophysiology, genetics, and vegetation manage-
ment to encompass soil processes, ecosystem structure and dynamics, hydrol-
ogy, wildlife biology, fisheries, restoration ecology, conservation biology, and
landscape ecology. The social sciences, once relegated to a backseat (except for
economics), are now seen as critically important, at least on public lands in the
United States. Research into innovative engineering techniques and the develop-
ment of a broad array of forest products are essential parts of the contemporary
package, and management science is playing an increasingly important role in
helping to integrate science, economics, and politics. A new dialogue among
science, philosophy, and religion is exploring the esthetic/spiritual dimensions
of forests—and nature as a whole—that have been common to humans for
millenia.
In this paper I deal mostly with biology, soils, and hydrology, but with
the understanding that sociology, esthetics, ethics, spirituality, economics, and
history intertwine with virtually all aspects of the biologic and physical sciences
to produce the complex systems that foresters work with and within. Ignoring
that basic truth during much of the twentieth century has resulted in social
and political turmoil throughout the world over the way forests are managed—
eventually triggering a fundamental reevaluation on the part of foresters and
forest scientists of what forestry is all about (88). The need for integrating a
broad array of scientific disciplines to guide forestry, indeed all natural resource
management, has never been more acute.
This paper is basically a story of the changing role of science in forestry and an
assessment of the current state of knowledge and priority needs. I discuss what I
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

SCIENTIFIC BASIS OF FORESTRY 437

believe to be the most important challenges facing forestry science; however, the
breadth of the topic does not allow all aspects to be given their rightful due. My
choice of emphases reflects my own experience and biases; someone else may
have made different choices. I focus mostly on North America—though both
the practice and science of forestry are largely global affairs—because, until
recently, the basis strategy of forestry has largely been the same everywhere
during the latter half of the twentieth century.

A SHORT HISTORY
Forestry as a science originated in Europe during the latter eighteenth and early
nineteenth centuries, largely in response to the poor condition of Europe’s re-
maining forests and an impending wood famine (126). From early on there were
competing philosophies about the proper approach. Aldo Leopold, who once
supervised a US National Forest, described two types of forestry: Type A sees
“. . . land as a commodity and trees as cellulose to be grown much like cabbages”;
Type B “. . . treats land as a community of interacting and interdependent parts,
all of which must be cared for” (quoted in 12). Type B, which I refer to as
ecosystem-based forestry, has always had a strong philosophical representation
within the forestry profession; however, Type A has predominated throughout
the world during the twentieth century. Like modern agriculture, its focus has
been on the properties of individual ideotypes rather than communities and
ecosystems (119). Type A is commonly referred to as intensive forestry (where
“intensive” refers to cultural inputs).
Intensive forestry had its beginnings in Germany during the mid-1800s.
German forest scientists, motivated by the ideas of the English economist,
Adam Smith, formulated an economic approach called soil rent theory, which
held that interest should be earned on land, timber capital, and silvicultural
expenses (in opposition, “forest rent” theory held that interest charges against
these assets were inappropriate). Plochmann (126) describes the result. “The
soil rent method furnished foresters with an ideal planning tool for calculating
the species with the highest monetary return and the financial rotation with the
highest internal rate of interest on a given site. It fit perfectly with classical
liberal economic theory, which set the maximization of profit as the general
objective of economic activities and therefore the general objective for forestry
as well.” In central Europe, native hardwood and mixed hardwood/conifer
forests were cleared and planted to monocultures that grew fast and produced
high value—mostly Norway spruce. Because interest was charged on land and
other capital assets, the rotation length (time period between regeneration of a
stand and final harvest) yielding the highest rate of return was much shorter
than the normal life span of forests.
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

438 PERRY

Despite some misgivings on the part of both foresters and the public (12),
intensive forestry took root in North America following World War II, and by
the 1950s was firmly established on both industrial and public lands. Over time
it became the dominant theme in most or all of the nation’s forestry schools. The
reasons behind the ascendancy of intensive forestry were clear. A primary re-
sponsibility of foresters was to supply the fiber needs of a rapidly expanding
economy, and this would best be done through plantation monocultures of fast-
growing tree species. The full growth potential of land could be ultimately
achieved by planting healthy, genetically superior seedlings, controlling non-
crop vegetation, enhancing site quality through fertilization (and other tech-
niques where necessary and affordable), and clear-cutting when either rate of
tree growth or economic returns calculated for a series of rotations were maxi-
mized (e.g. 40). The responsibility for meeting the demands of a rapidly growing
market without overharvesting or otherwise degrading future productive capac-
ity fell to scientists, and intensive forestry came to be described by many of its
proponents, and more recently some critics, as scientific forestry. In a narrow
sense that characterization is accurate, as the tools of science are used, albeit in
the service of a particular economic model. However, like crop-centered agri-
culture, intensive forestry was not derived in the context of testable hypotheses
about alternative approaches, nor was there much open debate about alterna-
tives (until lately); thus in a more fundamental sense it is not scientific. Not
surprisingly, forestry shared one of the central distinguishing features of in-
dustrialism: uncritical acceptance of the untested hypothesis that maximizing
economic efficiency in the short term was the path to maximizing social and
economic benefits into the future.
The last 35 years have seen much change in forestry and a significant diver-
gence in approaches. During the 1960s, Germany, for social, environmental,
and economic reasons, began converting state lands back to the native hard-
wood/conifer mixed forests and growing high-value trees on long rotations
(126). In the United States, beginning in the late 1960s and continuing to the
present, public opposition to clear-cutting and herbicides joined mounting sci-
entific evidence that, in contradiction to federal laws, native diversity was not
being maintained on federal lands; together these led the US Forest Service
(USFS) and the Bureau of Land Management (BLM), after significant prodding
by the courts to adopt ecosystem management as official policy. Although some
companies have adopted aspects of ecosystem management, intensive forestry
remains by far the most commonly used approach on industrial lands. (They do
not necessarily incorporate all aspects; fertilization and herbicides in particular
may or may not be used, depending on perceived economic benefits.) Thus,
over the last decade in the United States what was once a single approach has
diverged into two paths that, while still sharing some aspects, differ significantly
in objectives, approach, and scientific priorities.
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

SCIENTIFIC BASIS OF FORESTRY 439

STATE OF KNOWLEDGE
The biological sciences have played three distinct roles in forestry: improv-
ing growth through intensive cultural practices; researching the environmental
impacts and sustainability of intensive forestry; and, most recently, develop-
ing a science of ecosystem-based management, which includes a broad array
of basic and applied research dealing either directly or indirectly with alter-
native management approaches. Each of these will continue to be important
in forestry, though their relative importance is shifting from crop-centeredness
to a greater balance between traditional forestry research and the science of
ecosystem-based management. Questions of sustainability remain of central
importance regardless of management approach.
Improving Growth
The basic approaches to increasing stand growth are (a) rapid establishment of
a new stand following harvest (nursery and planting practices); (b) maximiz-
ing the flow of site resources to the crop (controlling competing vegetation);
(c) improving soil resources (fertilization, bedding, and drainage on some sites);
(d ) selecting and breeding fast-growing genotypes; and (e) minimizing losses
due to insects, diseases, fire, and wind (stand protection). All approaches have
been actively researched over the past 30 years, except that comparatively lit-
tle research has been done on stand protection, and that almost exclusively in
reaction to specific problems. The theoretical potential for productivity gains
through intensive culture is significant. Farnum et al (40) calculated growth
rates of unmanaged stands of Douglas-fir and loblolly pine (two of the most im-
portant commercial species in the United States) at, respectively, only 23% and
12% of what could be achieved. In this section I focus on genetics, vegetation
control, and fertilization, which are the centerpieces of intensive forestry. I deal
with vulnerability of managed stands to biotic and abiotic disturbances later.
GENETICS Most commercially important tree species contain large within-
population genetic variation, providing a rich source for selection. For exam-
ple, approximately 92% of the genetic diversity in loblolly pine, the dominant
timber species in the southern United States, occurs within stands (177). A
further boon to selection programs is that many individual genotypes of com-
mercially important species have wide ecological amplitude. Loblolly pine and
Douglas-fir families tend to maintain a constant ranking across a wide range of
environments (101, 157); when significant GXE interactions do occur in those
species it is usually due to a few families.
Much tree breeding in the United States and Canada has been through co-
operatives involving industry, universities, and in some cases public agencies
(e.g. 164). Although there is growing interest in cloning (155), most geneti-
cally improved conifers (by far the most important commercial species in North
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

440 PERRY

America) come from seed orchards. The basic strategy is to select trees in the
field with desired characteristics and propagate them in both open-pollinated
production orchards and breeding orchards where select trees are interbred
through controlled pollination (40). At the end of the first breeding cycle,
seed from the production orchard is outplanted, while seed from the breeding
orchard is used (after progeny testing) to establish second-generation produc-
tion and breeding orchards, and so on. The use of elite breeding populations
(progressive selection and interbreeding of the best performing genotypes), a
strategy adopted from agronomic and animal programs, is increasingly com-
mon in conifer breeding (177). Various technological advances have reduced
breeding cycles considerably over the past two decades—in loblolly pine from
22 years in 1974 to 7.5 years in 1994 (177). In 1992, 90% of all plantings in the
southern United States, amounting to 1.5 million acres, were with genetically
improved seedlings (164). The first genetically improved trees are now be-
ing harvested in the southern United States, with estimated volume gains of
12% for loblolly pine (32% in harvest value) and 18% for slash pine (73, 102).
Weyerhauser Company estimates a 10% gain in juvenile height growth for select
Douglas-fir in Oregon and Washington (157).
The importance of maintaining genetic diversity is much discussed among
forest geneticists (e.g. 38, 102, 110, 135). Erosion of genetic diversity through
random drift is an especially significant concern in forest trees, which carry
high levels of lethal recessive alleles and are particularly vulnerable to inbreed-
ing depression (93, 111). Generally one of two approaches is used to maintain
diversity within breeding populations (39): the hierarchical open-ended sys-
tem (HOPE) and the multiple population breeding system (MPBS). The HOPE
strategy, developed originally by agronomic breeders, periodically introduces
genetic material from populations early in the selection cycle to populations at
later stages of selection. MPBS, developed by forest geneticists (110), sets up
a number of different breeding populations and exchanges genes in controlled
crosses among these—establishing what is in essence a controlled metapopula-
tion structure. Allozyme studies in loblolly pine show that, although both strate-
gies maintain relatively high levels of allozyme diversity within elite breeding
populations, neither maintains the diversity found in natural stands (176). As
would be expected, rare alleles are the most heavily impacted. Moreover, be-
cause seedlings are culled in the various steps between controlled breeding
populations and the eventual planting stock, the diversity of resulting planta-
tions will be lower yet (177). Studies of other commercial species show similar
patterns: Breeding programs to date have had little or no influence on overall
allozyme diversity; however, rare alleles may be sharply reduced (37, 156).
Major questions remain concerning effects of selection on genetic diversity at
the landscape level, or what Friedman & Foster (48) refer to as beta genetic
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

SCIENTIFIC BASIS OF FORESTRY 441

diversity, i.e. the allozyme diversity of an elite breeding population may differ
little from that of wild stands (alpha diversity), but planted over large areas the
elite population could replace a fine-scale genetic mosaic, resulting in geneti-
cally simplified landscapes (48).
Any reduction of genetic diversity raises the presently unanswerable ques-
tion of how much is sufficient to maintain resistance to pests and the capacity
to adapt to changing environments (102, 132, 177). The adaptive value of rare
alleles is unknown at present (38). Some forest geneticists argue they contribute
little to overall fitness; others, however, suggest that loss of rare alleles could
compromise long-term adaptive flexibility (177). Inability to relate allozyme
measurements to phenotypic traits with potential adaptive value seriously lim-
its the usefulness of allozyme measures as predictors of ecological response
(135).

VEGETATION MANAGEMENT Grounded in the common view of plant commu-

nity ecologists that competition is the most important interaction among plants,
the focus of vegetation management in forestry has been on increasing crop
yields through the use of chemicals and tillage. Considerable research (though
not all) shows that controlling noncrop vegetation increases growth of crop
trees during the early years of stand development, and in some cases establish-
ing plantations can be difficult or even impossible without vegetation control
(54, 171). A major thrust has been to develop competition indices relating lev-
els of noncrop vegetation to potential growth response. However, the very short
lifetime of most experiments limits the reliability of indices, which have been
found to vary over time with shifting community structure (170).
The issues surrounding vegetation management are among the more complex
in forestry, involving ecological effects of herbicides, long-term and indirect
effects of altering community structure, and the functional roles of different
plant species. As pointed out by some prominent researchers in the field, little
attention has been paid to long-term and ecosystem-level effects of vegetation
management (130, 170). The basic scientific issue is one of the relationship
between community structure and ecosystem function, something ecologists in
general are just beginning to grapple with.
In North American forestry, where conifers are the major commercial species,
noncrop vegetation is virtually always broadleaved trees, forbs, shrubs, and
grasses. A variety of studies have either conclusively demonstrated or strongly
suggested that these plants perform numerous important ecological functions,
including providing unique food (e.g. nuts, nectar), enhancing nutrient availabil-
ity (9, 49), replenishing nitrogen capital through biological fixation (14, 67), sta-
bilizing soil nutrients and biology following disturbances (5, 16), and increasing
resistance of conifers to herbivorous insects, pathogens, and fire (e.g. 52, 123,
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

442 PERRY

136, 146, 161, 181). The complexity of interactions among plant species is
only beginning to be understood but goes far beyond simple competition. For
example, it is now well established that different plant species within at least
some communities—including broadleaved trees and conifers—participate in
a network of shared resources mediated by mycorrhizal fungi (131, 147).
The major challenge for the science of vegetation management is to find
balance between competition and other important ecological functions, and
this will require a much improved understanding of interactions among plant
species and between plants, animals, and microbes and how these interactions
are affected by relative density, environmental factors, and time. The few stud-
ies that have manipulated relative density of crop and noncrop species show
complex, shifting competitive relationships; as Shainsky & Rose (142) put it,
“competition is not linear and unidirectional.” Moreover, protective and stabi-
lizing functions may come into play only during critical periods, such as during
wildfire or recovery from disturbance, which means the importance of these
functions could easily be missed in short-term studies (125).
Scientists working with vegetation management in forestry increasingly take
an ecosystem view (the theme of the 1998 International Conference on For-
est Vegetation Management is “Forest Vegetation Management and Ecosys-
tem Sustainability”). Experiments that manipulate relative density or spatial
patterns of noncrop plants or that separate shrub from herb competition are
becoming more common (e.g. 103, 141). There is movement toward an “inte-
grated vegetation management,” which, like integrated pest management, seeks
ways to reduce pesticide use or replace it altogether (170).

FERTILIZATION Hundreds and perhaps thousands of fertilizer trials have been

installed over the past 30 years in forests of the United States and Canada;
these vary widely in experimental rigor, with later installations generally better
designed and more intepretable than earlier ones. A substantial number in the
United States are university-industry-USFS cooperatives. Fertilization exper-
iments are frequently coupled with other cultural treatments such as thinning
and vegetation control. Most research has focused on nitrogen and, in the south,
phosphorus (15), though experiments have been installed using NPK (112) and,
in at least one case, slow-release multinutrient tabs (168).
N fertilization frequently produces significant, occasionally spectacular, but
often highly variable growth responses. Volume gains from N average 16% to
26% in coastal Oregon, Washington, and British Columbia (104). In a regional
study employing N and P in a factorial design, loblolly pine volume growth
in response to N and P averaged 25% greater than controls, with NP yield-
ing 2–3 times greater response than did the addition of N and P alone (112).
A recent research trend has been to employ frequent, small additions of N, a
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

SCIENTIFIC BASIS OF FORESTRY 443

technique shown by Swedish researchers to better mimic natural processes and

produce significantly greater growth response than infrequent, large additions
(80). Using that approach in a North Carolina study resulted in exceptional
growth responses in young loblolly pine, doubling leaf area over a four-year
period and increasing stem volume 180% compared with controls (3). Part of
the growth response was due to reallocation of CHO from fine roots to above-
ground tissues, a dynamic seen in other studies and probably quite general.
Despite a droughty site, irrigation had a relatively small effect in the North
Carolina study, suggesting that other cases in which water is thought to be
limiting may actually be N limited.
Predicting fertilizer response has been a problem area and the subject of
considerable research. Although N has enhanced tree growth in 70% of trials in
Oregon, Washington, and British Columbia, the magnitude of response varies
widely (104). Similarly, researchers have had limited success in predicting
growth response to N and P in the southeastern coastal plain (66). Standard
measures of soil nutrient availability are seldom useful (66, 144), while foliar
levels may or may not be. Evidence from both the south and the west indicates
that limited response to N and P is often due to deficiencies of other nutrients,
with potassium, sulfur, boron, magnesium, and iron identified in one or more
studies (e.g. 112, 144, 168). Recently, DRIS norms have been employed with
some success to classify the Douglas-fir stands with regard to their response
to N fertilization (144). DRISS, originally developed by agronomists, uses
nutrient ratios in foliage or soil to predict fertilizer response.

TREATMENT COMBINATIONS Cultural practices are usually applied as sets, i.e.

a single site may be drained, fertilized, herbicided, and planted with genetically
improved trees. Some research has shown non-additive responses to treatment
combinations. For example, trees may not respond to fertilization unless stands
are thinned or competing vegetation is controlled (153). Even though GXE
is minimal in loblolly pine, McKeand et al (101) argue that planting certain
responsive families will significantly increase gains from cultural treatments.
The degree to which intensive cultural factors interacts with ecologic factors
such as long-term soil fertility and resistance to pests and pathogens is the
subject of the next section.

Sustainability
Concern with sustainability in forestry dates back to the late seventeenth cen-
tury; in fact, Wiersum (174) argues forestry was the first science in the western
world to explicitly acknowledge the need to safeguard finite natural resources for
future generations. However, achieving sustainability within an economic sys-
tem that devalues the future is a particular challenge, especially for a long-term
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

444 PERRY

endeavor like forestry (13). Negative impacts of practices on future yields

of wood can, in theory, be dealt with by incorporating them within economic
calculations, though the further in the future such costs are realized the less in-
fluence they have on present net value. Impacts on nonmarketable goods (e.g.
habitat, water) and risks with high uncertainty must be dealt with in other ways.

What To Sustain?
For much of the history of forestry, sustainability meant maintaining a steady
supply of wood (174). In the United States considerable discussion occurred
prior to and in the years following World War II about the role of forests in
wildlife habitat, watershed protection, and even regional climate, but this had
relatively little effect on practices (e.g. 92). The Multiple Use Sustained Yield
Act of 1960 specified five things to be sustained on public lands: timber, fish
and wildlife, outdoor recreation, range and fodder, and watersheds (174). More
recently, sustaining biodiversity has been center stage, though that term is itself
subject to various interpretations. Angermeier & Karr (9) argued that policies
for achieving sustainability should focus on biological integrity, which they
define as “a system’s wholeness, including presence of all appropriate ele-
ments, and occurrence of all processes at appropriate rates.” Franklin (44)
defines sustainability as “. . . maintenance of the potential for our land and wa-
ter ecosystems to produce the same quantity and quality of goods and services
in perpetuity.”
The following discussion focuses on four critical components of ecological
sustainability: soils, water, species diversity (habitat), and resistance to distur-
bances (118). A fifth, evolutionary potential, was discussed in an earlier section.

Soils
German foresters knew as early as the mid-1800s that export of nutrients as-
sociated with removing forest litter reduced tree growth. Outside of Europe,
however, little attention was given to the effects of forestry practices on soils
and nutrient capital until the mid-1960s, when data from intensively managed
Pinus radiata in Australia and New Zealand showed yields had declined signif-
icantly between the first and second rotations, a phenomenon eventually traced
to loss of soil organic matter (SOM) and nutrients during site preparation for
planting. As the same practices—hot slash burns or use of heavy equipment
to push logging residues (and often topsoils) into piles—were widely used in
North America, considerable research was stimulated that continues today (e.g.
11, 53, 68, 120). A group of leading forest soil scientists expressed the current
view of many soil scientists and ecologists: “. . . intensive site preparation, in-
creased utilization, and shortened rotations that accompany domestication carry
high potential risks to the site’s capability to sustain growth” (129).
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

SCIENTIFIC BASIS OF FORESTRY 445

Soil fertility (broadly, the capacity of soils to support plant growth) is an

emergent property arising from dynamic interactions among nutrients, organic
matter, soil structure (a function of both physical and biological processes),
soil organisms, and plants (121). Most forestry-related research to date has
focused on nutrient balances, comparing losses associated with harvest and
site preparation with inputs to replace those losses. Nutrient losses associ-
ated with clear-cutting have been well documented and include three primary
pathways: (a) direct removal in harvested biomass; (b) losses associated with
site preparation techniques such as burning, which impacts volatile elements
(especially carbon and nitrogen), or windrowing (also called raking or pile-and-
burn), which displaces residual organic matter from the majority of a site onto
small portions; and (c) export in erosion and leaching. In most cases, losses in
harvested biomass and site preparation far exceed leaching and erosion (32, 97);
an important exception is where early successional vegetation is prevented from
recovering, especially in forest types characterized by low C/N organic matter
(19).
Not surprisingly, the more biomass removed over a given period of time and
the greater its nutrient content, the greater the probability that rates of nutrient
export will exceed rates of input through natural processes. Short rotations
combined with practices that remove a high proportion of site biomass result
in exceedingly large nutrient drains. By far the largest losses of nutrients and
either present or future soil carbon are associated with harvesting tree crowns
and site preparation (81). Various studies show that removing crowns, which
have relatively high nutrient concentration, increases nutrient loss by 50% to
400% over harvesting boles only (21, 87, 97, 149).
Throughout North America, sites are often prepared for planting by either
windrowing or burning residues in place (“broadcast burning”). Windrowing
as traditionally practiced compacts soils and removes large amounts of SOM
and nutrients; Morris et al (108) estimated nutrient losses during windrowing as
equal to six bolewood harvests. Hot broadcast burns, once the norm in western
North America, volatilize large amounts of carbon and nitrogen (81).
Reductions in SOM and soil pore space (the latter due to compaction) are a
significant concern (129). Soil organic matter (SOM)—including humus, for-
est floor, and both fine and coarse woody debris—stores nutrients (especially
nitrogen), provides cation exchange and water-holding capacity, and serves as
substrate for the belowground food chain; hence SOM is a keystone resource
for numerous soil functions. SOM falls into two broad groups that differ both
functionally and with respect to potential management impacts. Fine litter (fo-
liage, small branches, epiphytes) in various stages of decay forms the majority
of surface organic layers and (eventually) humus within the soil profile, and
it functions in nutrient cycling, cation exchange capacity, and water retention.
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

446 PERRY

Coarse debris (logs and large branches) is abundant in natural forests and func-
tions as sites for N fixation, water reservoirs (decayed logs are sponge-like in
their ability to retain water), and habitat for numerous fungi, invertebrates, and
vertebrates such as salamanders, which are at the top of the soil food chain
(62, 71). Immediate impacts of forest management on soil C are mainly asso-
ciated with extreme site preparation techniques such as hot fires, windrowing,
and tilling (70, 81), whereas practices that remove sources of future soil C (trees
and litter layers) have longer-term effects that remain to be determined. Pre-
sumably losses in those components of soil C that derive from fine litter will be
replenished so long as plant communities regrow vigorously, a condition tied
in a self-referential loop to the status of soil nutrients, soil C, and soil biology
after harvest. Components of soil C derived from large debris—especially tree
boles—are more problematic. Until recently, forestry never considered leaving
tree boles on site; in fact—quite the opposite—doing so was seen as waste. As
I discuss later, retaining trees on site as future sources of large dead wood is
a major component of new silvicultural approaches. However, the question of
how many trees to leave turns on gaining a better understanding of the functions
of large dead wood within the ecosystem.
Though soil biology is undoubtedly the key to understanding soil function
(124, 138), the enormous diversity belowground has made research into func-
tional aspects slow and expensive. Consequently, the science of the below-
ground is in its infancy. Existing studies related to forestry have focused either
on total microbial biomass, particularly as nutrient sinks following harvest
(27, 169), or on specific functional guilds such as mycorrhizal fungi, nitrogen-
fixing and other rhizosphere bacteria, invertebrates, or some combination of
these (e.g. 5, 7, 69, 71, 84, 105, 121, 122). Soil bacteria act as important nutri-
ent sinks following clear-cutting, a phenomenon tied to the availability of labile
C and therefore likely to be impacted by intensive site preparation (169). Man-
agement effects on specific guilds range from minor to severe, depending on the
management practice employed and the guild of organisms studied. Significant
change in soil biology as a result of clear-cutting and site preparation is always
found, including increased bacterial relative to fungal biomass, greater rates
of nitrification, reduced concentrations of microbially produced iron chelators,
shifts in mycorrhizal types, and sharp declines in invertebrate numbers. How-
ever, interpretation is complicated by poor understanding of functional and
dynamic aspects of the belowground ecosystem.
Translating impacts on soils to longer-term soil fertility, tree growth, and
other ecosystem processes is not always straightforward. Nutrient inputs in
precipitation are reasonably well known, but inputs from weathering—the major
natural source of all essential elements except N, C, and O—are difficult to
measure and poorly understood (31). Moreover, recent research shows conifers
(and probably other non-nodulated trees) have some capacity to renew soil
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

SCIENTIFIC BASIS OF FORESTRY 447

fertility. Bacteria associated with conifer rhizospheres and mycorrhizal fungi

can fix appreciable amounts of N (17, 94). Both ectomycorrhizal fungi and
rhizosphere bacteria significantly accelerate rock weathering and are likely to
speed recovery of fertility in soils with unweathered rock in the rooting zone
(18, 83). [Soils in the northern United States and Canada contain abundant un-
weathered rock, but that is not the case with some old soils in the southern
United States (107).]
Better understanding the capacity of trees and their symbionts to renew
soil fertility, and how that varies with site, species, and environmental condi-
tions, promises to add a significant new dimension to our picture of ecosystem
nutrient budgets. A key question that has received little research is whether
site impacts can disrupt these biologically mediated renewal processes. The
few studies that do exist suggest the renewal process can be influenced by
management. N fixation associated with Douglas-fir and its mycorrhizae is
stimulated by proximity to certain hardwood species (8), suggesting that in
systems where that relationship occurs, excessive vegetation control will re-
duce associative N fixation. N fixation in ponderosa pine rhizospheres is de-
pressed in both clear-cuts and patch-cuts (<0.2 ha) relative to undisturbed forest
(79).
Predicting long-term impacts on soil fertility has been further complicated
by the fact that pines, in particular, often grow better during their early years
on sites that have been intensively prepared than on sites that have not (23).
Although the ability of trees to renew fertility may be a partial explanation,
most researchers agree better growth on intensively prepared sites is a transi-
tory phenomenon due to various factors, especially reduced competition from
herbs and grasses (23, 106, 129). As plantations have reached crown closure,
relatively poor growth on intensively prepared sites has become widely evident
(20, 33, 43, 128). In one of the few studies in which trees were approaching
rotation age, stem volume per hectare of 31-year-old loblolly pine was 23%
lower on plots that had been windrowed prior to planting than on plots that had
been broadcast burned (43).
The multirotation predictions required for assessments of sustainability, cou-
pled with the complexity and nonlinearity of forest resilience, require experi-
ments and models based in ecosystem-level processes (89, 152). Several models
exist that link nutrients (mainly nitrogen) to tree growth (1, 86, 112, 116). Al-
though the importance of such models is clear, they are also potential traps
because their long-term predictions cannot be validated in the short term, and
there is a history of managers using models as substitutes for observation and
critical thought. Yarie (178) compared two models and found they predicted dif-
ferent outcomes for the same management practices. He concluded “. . . neither
(should) be used as a . . . decision tool without the help of sufficient expertise
in ecosystem ecology to correctly interpret the results.”
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

448 PERRY

Several long-term experiments dealing with management impacts on soils

and tree growth have been installed within the past few years, each nearly unique
in approach and objectives. Now 14 years old, a study installed by the North
Carolina State University Nutrition Co-operative manipulated levels of harvest,
site preparation, and vegetation control factorially (112). A widely replicated
experiment sponsored by the US Forest Service includes different levels of com-
paction, biomass removal, and vegetation control in a nested design within clear-
cuts (129). Several studies in the United States and Canada have experimentally
manipulated forest structure (via levels of harvest and size of harvest units) and
are considering an array of soil chemical and biological variables (6, 46, 76).
Information on potential negative impacts of intensive site preparation has
reached field foresters through symposia and other sources. Hot broadcast burns
have become less common in the west, and managers in some areas have either
abandoned windrowing or adopted windrowing techniques with lower impact.
Intensive site preparation is still common, however, at least in part because of
faith in the remedial effects of fertilizers. A central challenge for research is to
determine the degree to which that faith is justified.

Water
Forests are keystone modulators of the water cycle; hence forestry and water are
inseparable. Reflecting that fact, watershed protection was a primary rationale
for the creation of the National Forest System, and hydrologic research by the
USFS dates back at least 50 years. By the mid-1960s approximately 150 gaged,
forested watersheds existed in the United States, many of which involved man-
aged and unmanaged pairs (78). Other studies have either followed single basins
as they are harvested over time, or, particularly in the case of large basins, com-
pared data among basins operationally harvested to different levels (e.g. 82).
Hydrology exemplifies many other issues in forest science—large scale and
high background variability (temporally and spatially) make generalizations
that are both broad and accurate extremely difficult, expensive, and in some
cases impossible. Only dramatic impacts on stream hydrology are statistically
detectable in short-term studies. Moreover, effects of clear-cutting on stream
hydrology tend to be highly variable for various reasons, including bedrock geo-
logy, topography, climate (e.g. rain or snow dominated, importance of fog-drip
from canopies), and harvest practices. Clearest effects are from experimentally
paired, small watersheds, and the ability to detect effects diminishes as one
scales up to larger basins (133). Decades-long records are often necessary to
discern trends, especially in larger basins.
Small watershed studies generally show clear-cutting increases water yield,
though in areas with significant fog-drip (water raked from clouds or fog by tree
crowns), yields may decline following harvest (133). Peak flows, which are of
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

SCIENTIFIC BASIS OF FORESTRY 449

more concern environmentally and economically, are often increased by clear-

cutting, though that depends on various factors, particularly the extent and rate
of logging within a basin, how logging is done, and the road network. Intensive
site preparation, vegetation control, extensive roading (especially when poorly
designed), and disruption of riparian forests have the greatest likelihood of
increasing the magnitude and duration of peak flows (133).
Studies in western Oregon show that clear-cutting and roads act synergis-
tically to alter hydrology. Clear-cutting reduces evapotranspiration (ET) and
increases snow accumulation and melt, resulting in increased deep soil water
storage that persists until leaf area of deep-rooted trees and shrubs has fully
recovered (on the order of decades) (64). In poorly drained areas, water ta-
bles rise in clear-cuts (23), and in some cases bog formation can be triggered
(DA Perry, personal observation). Roads, on the other hand, alter hillslope
flows by converting subsurface to surface flow; hence they provide pathways
for deep soil water to reach the surface and flow rapidly to streams (63). On
the HJ Andrews Experimental Forest (HJA) (Oregon Cascades), peak flows
did not differ between a watershed that was 100% clear-cut without roads and
one that was 25% clear-cut with roads, though seasonality of altered flows did
differ (82). In both, average peak discharge increased by >50% (compared
with pretreatment) for the first 5 years after treatment; 25 years after treatment,
discharge remained 25% to 40% higher than pretreatment peak flows.
Large basins are a particular challenge because experimental controls do
not exist and also because even dramatic events in small watersheds may be
statistically undetectable at larger scales without many years of record (133). In
the west, records are just now becoming sufficiently long to allow trends to be
separated from noise in larger basins. Jones & Grant (82) examined 50-to-55-
year records from three pairs of 60- to 600-km2 basins in the western Cascades
of Oregon (paired based on proximity and different rates of logging). Responses
in all cases were statistically significant (though with low r 2): a 5% difference in
cumulative area clear-cut translated to a difference in yearly peak flows ranging
from 10% to 55%, depending on basin-pair. None of the basins were heavily
clear-cut (maximum 25% of total basin area), yet Jones & Grant estimate that
peak discharges in these basins have increased by 50% to 250% compared with
prelogging. The extensive road network required where clear-cuts are in a
dispersed patchwork, as is common on federal forests, means total cutover area
significantly underestimates potential impacts on hydrology and on spatially
mediated processes in general.
Effects of forest management on sedimentation have been easier to demon-
strate than effects on water flows because background variability is much less;
very little soil is eroded from undisturbed forests. As with water, the best docu-
mented studies have been in experimental watersheds, although clear evidence
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

450 PERRY

exists from other sources, such as comparisons of logged and unlogged wa-
tersheds and historical observations. Sediments associated with forestry come
from four primary sources (159): surface erosion from roads, surface erosion
from clear-cuts, mass transport during slash burns, and landslides associated
with either roads or clear-cuts. Studies on the HJ Andrews Experimental Forest
(HJA) found that landslides, especially from poorly designed roads during ma-
jor storms, pulsed large amounts of sediment in brief episodes, while surface
erosion from roads and clear-cuts was more chronic. Eleven years after treat-
ment, suspended sediment from a roaded watershed, 25% clear-cut and burned,
averaged 57 times higher (per year) than the unroaded, unlogged control water-
shed, whereas nine years after treatment, sediment from an unroaded watershed,
100% clear-cut and broadcast-burned, averaged 23 times higher than its con-
trol (159). Total increases in surface erosion following clear-cutting are most
severe on steep slopes (generally >60%); however, proportional increases may
be more severe on shallow slopes (159). Broadcast burning greatly exacerbates
losses in both cases. Absolute losses vary widely depending on soil and rock
type.
As with many issues in forestry, short-term studies may not be adequate
to determine effects on sedimentation. Zeimer et al (179, 180) modeled the
long-term cumulative effects of harvesting on sedimentation for coastal north-
ern California to central Oregon (an area with high landslide potential). Their
simulations predicted that sediment produced during the first century following
harvest would be initially stored in small tributaries, to be washed into larger
streams during the second century. That prediction was supported by data from
Casper Creek in northern California, which is still adjusting to logging-related
sedimentation from the late nineteenth and early twentieth centuries.

Diversity and Habitat

Intensive forestry significantly alters the spatial and temporal structure of forests
and, applied widely, of forested landscapes. Aside from the fact that living
trees are killed, it has little ecological similarity to natural disturbances, which
leave a plethora of structural legacies and usually have a frequency distribution
characterized by many small and few large events, hence leave a more vari-
able patch mosaic than does intensive forestry (109, 143, 154). Stand recon-
structions show that forest types once believed to be extensively even-aged due
to infrequent, large natural disturbances are actually mosaics of age classes
resulting from relatively frequent, small disturbances that established cohorts
of young trees within a matrix of older trees (42, 47, 160). Because of their bio-
logical legacies and spatial patchiness, natural disturbances are likely to initiate
different early successional patterns than does intensive forestry, which aims
for rapid site capture by even-aged crop trees.
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

SCIENTIFIC BASIS OF FORESTRY 451

Fire exclusion, which along with roading is traditionally the first step in
bringing a forested landscape under management, has significantly altered forest
types with a history of frequent, low-intensity fires, especially ponderosa pine
in the west and longleaf pine in the south, leading in many cases to increased
problems with insect pests and pathogens, and paradoxically to a greater risk of
catastrophic fires. Roads may affect diversity in a number of ways, principally as
barriers to movement and by providing access for predators (including humans),
noxious weeds, and pathogens (123, 148). Various studies have shown that
population sizes of bears, wolves, moose, and mountain lions are inversely
proportional to road density (e.g. 22).
In short, forestry as commonly practiced places a radically new selective
filter on the landscape, structurally much simplified compared to the natural,
but also with new elements. Some species benefit, others are endangered, some
may adapt. But none has evolved with that filter. The most vulnerable habitats
in an intensive forestry regime are associated with forests older than harvest
age (20 to 100 years depending on forest type and product), hardwoods (as
most forestry focuses on conifers), riparian zones, wetlands, and streams, all of
which are unique and biologically rich. In some cases, conifers of one species
are replaced by others with faster initial growth; such is the case with longleaf
pine in the south, which is often converted to loblolly and slash pines (113).
There is no doubt that many species lose habitat in intensively managed forest
landscapes, though it should be borne in mind that habitat has never been a prior-
ity on the majority of industrial lands (there are notable exceptions) and has only
recently become so on public lands. To most foresters, trees past their peak of
growth are like savings bonds that no longer earn adequate interest; not cashing
in and replacing with higher-yielding young stands makes no economic sense.
The biological trade-offs involved in maximizing economic efficiency are now
known to be severe. Research beginning with the International Biological Pro-
gram (IBP) in the 1960s showed that forests past the peak of growth enter their
biological prime, a period characterized by uniquely rich habitat (7, 47, 98).
Riparian zones and wetlands are also biologically rich and functionally impor-
tant (56, 90, 162), and road building, logging, and grazing within riparian zones
have been major factors in the widespread decline of both migratory and resident
fish in the northwest (100). In some areas, such as the southern United States,
wetlands may be drained to plant commercial forests. A number of states and
provinces now restrict logging and roading in riparian zones and near wetlands.
The standard approach to conserving diversity in forests has been to establish
reserves, resulting in landscapes parceled out among two very different uses:
intensive forestry and no forestry. The importance of reserves to conservation is
clear. However, in most forested regions, the existing reserve network falls far
short of that needed to adequately protect regional diversity (113), and many
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

452 PERRY

scientists agree successful conservation will require viewing landscapes for

what they are—functional totalities in which both reserves and managed lands
play a role (45, 59, 65, 140). The implications of that view have been profound
for forestry, as maintaining habitat becomes a central issue in management
rather than something to be dealt with “elsewhere.” This strategy is not without
controversy. Most conservation ecologists agree that the amount of land likely
to be set aside will not, by itself, protect sensitive species. However, there are
concerns that unproven and possibly ineffectual changes in forestry practices
will be used to justify reduced commitment to reserves. The major scientific
challenges involve three levels of understanding: (a) the relationship between
managed forest structure and ecological function at the stand scale; (b) spatial
patterning of stand structures that meet diversity goals for a given bioregion; and
(c) the temporal dynamics of stand and landscape structures resulting from nat-
ural disturbance, anthropogenic disturbance, and interactions between natural
and anthropogenic disturbances (which, as I discuss later, can be significant).
Research on how forestry might be adapted to better protect biological di-
versity did not begin in earnest until the 1970s. Until that time, habitat concerns
on public forest lands focused largely on game animals, which were believed to
benefit from the open areas and edge created by patchwork clear-cutting. Even-
tually, concerns about the impacts of forestry on native diversity led to passage
of the National Forest Management Act of 1976 (NFMA), which directed the
USFS to seek scientific advice on how to maintain diversity of plants and an-
imals on lands under their jurisdiction. By the time NFMA was passed, US
Forest Service biologists had already begun to deal with the issue of nongame
bird habitat by hosting a National Symposium (150) followed by four regional
symposia. Similar symposia since have dealt with amphibians, reptiles, small
mammals, marten, fisher, wolverine, and lynx. In 1979, a group of USFS
and BLM wildlife biologists published what came to be known as the “Blue
Mountain Book” (because it dealt with the Blue Mountains of eastern Oregon
and Washington), a guide for maintaining terrestrial habitat in managed forests
(162). The first of its kind, the Blue Mountain Book was in demand through-
out the world, and similar guides for other regions soon followed. FEMAT, the
plan to protect old-growth associates and fish in the range of the northern spotted
owl, relied mostly on reserves (41) [wildlife biologists on the Forest Ecosys-
tem Management Assessment Team (FEMAT) were very wary of unproven
silvicultural techniques]. However, 10 large (37,000 to 140,000 ha) Adaptive
Management Areas (AMAs) were set aside to experiment with innovative man-
agement techniques.
Evidence from the northwest indicates that many species associated with old
growth also occur in younger stands that originate from natural disturbances,
presumably because of the rich structural legacies left behind in the form of large
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

SCIENTIFIC BASIS OF FORESTRY 453

dead wood, remnant green trees, and sprouting shrubs (58, 60). Large dead wood
(snags and logs) provides habitat for numerous terrestrial species, a fact that
almost certainly holds for all forest types (62). Logs are also critically important
in creating fish habitat because they dam streams to create pools (99, 139).
With the growing awareness of the needs of cavity-users in the late 1970s,
federal foresters began leaving scattered trees in clear-cuts (when permitted
by state safety inspectors), though dead wood on the ground was still rou-
tinely cleared (often because of concerns about wildfire). In the mid-1980s, re-
searchers and managers on the HJA and the Williamette National Forest installed
several trials using silvicultural techniques intended to capture the diversity of
naturally disturbed forests. Initially termed “new forestry,” and now “variable
retention harvest,” the basic idea is to leave a certain number of large, green trees
at harvest, either dispersed or aggregated, with the objective of providing at least
three functions not found in intensively managed forests: habitat continuity for
species requiring large, green trees; a diversity of age classes with concomitant
vertical and/or horizontal heterogeneity; and future sources of large dead wood
(46). At least two companies have since initiated operational retention harvests
(46, 158). Several replicated experiments are now installed in the United States
and Canada that manipulate retention levels and, in at least one case, the physi-
cal arrangement of retention (6, 46, 76). These incorporate a level of biological
detail far beyond anything in previous silvicultural experiments, including mea-
surements of small mammals, birds, invertebrates, plants, and microbes.
Results to date (including studies conducted in earlier trials) show that young
stands with remnant old trees support significantly greater abundance of some
old-growth associates than do young stands without old remnants, including
some not found at all in young stands without old trees (25, 60, 117, 137). Sim-
ulation predicts that bird richness will remain higher in mixtures of old and
young trees than in stands composed only of young trees for 140 years follow-
ing harvest (61). However, not all birds associated with old growth benefit from
leaving residual trees, and some that do benefit exhibit threshold responses in
which relatively small changes in the density of older trees have large effects
on abundance (60). Research on this issue is in its infancy, and it would be
surprising if new patterns did not emerge with time.
Resistance to Disturbances
A vast experiment is underway. Its unplanned and unwitting design is changing the spatial
and temporal structure of terrestrial ecosystem... R. L. Burgess and D. M. Sharp (24)

Insights gained with the emergence of landscape ecology and conservation bi-
ology as disciplines over the past two decades leave little doubt that the “vast
experiment” to which Burgess & Sharp refer to in the above quotation has sig-
nificantly perturbed the fabric of species relationships and ecological processes
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

454 PERRY

within regions, with largely unpredictable consequences that may take decades
or centuries to unfold (85, 123, 163). Infestations of trees by a number of native
pathogens and defoliating insects are already outside of known historic ranges
in regions where humans have significantly altered forested landscapes, includ-
ing the western spruce budworm, Swiss needle cast, and several root rots in
western North America; fusiform rust and southern pine beetle in the south;
and eastern spruce budworm and ash yellows (a viral disease) in the northeast
United States and eastern Canada (26, 123, 127, 136, 173). Widespread changes
in stand structure have also increased susceptibility of forests throughout the
west to high-intensity crown fires (2).
Consistent with the resource concentration hypothesis of herbivore/
pathogen dynamics, epidemics of recent years are frequently related to the
spread of host trees. The spread of hosts stems in turn from various factors
that depend on locale but all involve changes in the historic disturbance regime.
In the south, agriculture, forestry, and fire exclusion have combined to sharply
reduce the extent of once widespread longleaf pine forests. Much of that area
has been planted with slash and loblolly pines, which have faster early growth
than longleaf (hence are favored by foresters) but are also more susceptible
to fusiform rust and southern pine beetle (127, 136). In the interior western
United States, southern British Columbia, northeast United States, and eastern
Canada, logging old growth coupled with fire exclusion allowed the spread of
Abies sp., principle hosts for spruce budworm (118, 173). An ongoing epidemic
of Swiss needle cast in the coastal mountains of Oregon is believed to have been
triggered by the large areas of young Douglas-fir plantations (G Filip, personal
communication). (Note, these are all native insects and pathogens—exotics are
another story.)
Depending on the particular situation, various other factors are believed to
have contributed to pest outbreaks. Roads have clearly accelerated the spread
of several diseases in the Pacific Northwest (123). Disruption of the natural
enemy complex through habitat loss and (in some cases) pesticides may be
involved in outbreaks of herbivorous insects (137, 165), though that is difficult
to demonstrate conclusively. The incidence of fusiform rust on southern pines
increases with management intensity, particularly weed control (181) and fer-
tilization (127). Eliminating broadleaved species from plantations also results
in greater incidence of root rots on Douglas-fir (146) and defoliators on white
spruce (52), phenomena that could stem from greater concentration of food
plants (conifers), disruption of natural enemies that depend on broadleaved
plants for habitat, or both.
Genetic and ecologic strategies have been employed to deal with herbivo-
rous insects and pathogens in forestry. Selection for resistance has significantly
reduced incidence of fusiform rust in southern pines (74); whether this is an
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

SCIENTIFIC BASIS OF FORESTRY 455

evolutionarily stable situation remains to be seen. Experiments on the stabil-

ity of resistance in slash pine indicate a highly dynamic relationship between
pathogen and host, with significant interactions between tree genotypes and
both environment (including the pathogen) and time (RA Schmidt, personal
communication). A single gene having major control of fusiform resistance
has been identified in loblolly pine (175), a situation in which rapid evolution
of the pathogen would be expected.
Ecologic strategies center on diversifying landscapes and promoting habitat
for natural enemies. Evidence from deliberate introductions of biotic control
agents strongly supports the importance of natural enemies as an effective,
and perhaps evolutionarily stable, source of control (77), and a great deal of
spatially explicit modeling points to the importance of landscape patterns in
host-pest dynamics (85). Ecologic approaches have so far been utilized mostly
in European forestry, though concepts of integrated pest management are gain-
ing a foothold in North American forestry.

THE FUTURE
There is one outstandingly important fact regarding Spaceship Earth, and that is that no
instruction book came with it. Buckminster Fuller (50)

The scientific challenges facing forestry are precisely those confronting basic
ecology: better understanding the relationships among structure, function, and
spatiotemporal dynamics of systems interconnected at many scales; and coping
with the “certain uncertainty” inherent in complex systems. Forest manage-
ment alters forest structure and thereby influences processes, some purposely,
as with productivity, some inadvertantly, as with erosion, population dynamics,
and susceptibility to disturbances. How much can the structure, hence pro-
cesses, of complex ecological systems be altered without compromising long-
term integrity, or, to turn that question on its head, how much structure must
be retained to maintain integrity? Can technology successfully substitute for
nature’s evolved controls? Implicit within these questions are numerous others
having to do with system states and measureable variables that tell something
useful about future system states.
During the 1990s a set of philosophies and general principles emerged under
the rubric of ecosystem management (EM). During its short life EM has gen-
erated numerous papers and symposia, most of which share common themes:
sustainability as the overall guiding principle; explicit goals; recognizing in-
terconnections; working across ownerships; taking a broad view spatially and
temporally; recognizing that ecological systems are dynamic but that some
changes have greater long-term consequences than others; accommodating
human needs within the constraints of ecological objectives (29, 57). The real
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

456 PERRY

challenges lie in translating abstract principles to what is (or is not) done on the
ground, and developing tools to assess whether goals are being achieved.
Several strategies have been or are being used to translate the general prin-
ciples of EM to specific practices. In most cases these were developed si-
multaneously over the past decade and are largely complementary rather than
competing. The approach exemplified by FEMAT (41) focuses largely on re-
serves and their interconnectance. In FEMAT, the spotted owl serves as an
umbrella species whose protection is hypothesized to de facto protect other
OG associates, and monitoring protocols are established to test that hypoth-
esis. FEMAT required relatively few modifications of forest practices in the
managed matrix (which under the plan composed a minority of the landscape
in many areas), though it took the ground-breaking step of devoting large areas
specifically to experimentation and learning, and it set the important precedent
of assembling multidisciplinary teams, along with protocols for synthesizing
sometimes widely divergent opinions about the probable outcomes of different
management scenarios.
A second approach, whose roots lie in the IBP-funded studies of OG conif-
erous forests and the emergence of landscape ecology as a discipline, deals
primarily with how lands outside of reserves are managed and uses historic pat-
terns of disturbance and recovery as templates for management. At the stand
level, this has led to techniques for protecting biological legacies, particularly
the variable retention harvests discussed earlier, and stimulated movement to-
ward longer rotations (35), the latter a strategy adopted by the Germans 30
years previously. At the landscape level, the concept of natural range of vari-
ability (NV) has received considerable attention. NV uses historic patterns of
disturbance to guide harvesting patterns, as, for example, identifying areas best
suited for longer rotations or small patch cuts and areas where shorter rotations
or larger cuts would be appropriate (30, 91). A third approach also focuses on
managed lands but uses functional models rather than historic. A good exam-
ple is Hansen et al (59), who used life histories to predict landscape patterns
required to maintain habitats for a suite of vertebrate species.
Each strategy has strengths and weaknesses. I discussed issues surround-
ing dependence on reserves earlier. NV has had the positive effect of making
foresters think about history (reliably determining history is another matter);
however, taken too literally the approach promotes the false belief that log-
ging can fit within a “natural” range of variability. Logging of any kind—
whether clear-cut or single-tree selection—is an unnatural event in the history
of a forest, and the relevant scientific question is not whether a practice fits
within NV, but how far management can depart from NV before compromis-
ing system integrity. Answering that question requires a functional approach
that deals with the suite of processes underpinning integrity, which in turn
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

SCIENTIFIC BASIS OF FORESTRY 457

requires a measureable definition of integrity, testable hypotheses concerning

keystone processes and their links to system structure, and protocols for choos-
ing among hypotheses when the systems of interest are highly dynamic in
space and time and the information base is weak (e.g. 75). To date, ecological
processes such as disturbance, hydrology, and nutrient cycling have seldom
been integrated with habitat in EM models (the integrity of soils is virtually
never mentioned in the EM literature), though that is beginning to change
(115, 151).
The move toward ecosystem management by federal agencies may have lit-
tle relevance for forestry as a whole in the United States, where by far the
greatest amount of timber is produced on private lands. In 1991, for exam-
ple, 32% of timber harvested in the United States was from industry lands
and 51% from nonindustrial private lands (167). As in agriculture, maximiz-
ing profit in forestry is most often associated with simplifying systems rather
than creating or maintaining diversity (the degree to which the green certifi-
cation movement changes that equation remains to be seen). However, while
diversity may not be an explicit goal on most industrial lands, sustaining the
productive base is, which leads directly to questions concerning the functions
of diversity. A growing number of ecologists believe that complexity and sta-
bility of ecosystems are positively linked (where stability refers to ecological
functions and potentials rather than species composition on a given piece of
ground) (28, 36, 125); however, the nature of the linkages is poorly understood
and “very few models . . .incorporate biological complexity as a regulating
component of ecosystem function” (138). Maintaining genetic diversity within
expansive areas of high-yielding cultivars shifts what in the natural forest was
likely to have been a complex set of genetic, species, and landscape controls
(top down and bottom up) to a much greater reliance on genetics (bottom up),
an interesting and important experiment, unfortunately uncontrolled. With the
discovery of single gene control over susceptibility to fusiform rust in loblolly
pine (and probably slash pine), the stage is set for a classic, some might say
mythic, confrontation between the forces of nature and the ingenuity of hu-
mans. The rust is likely to evolve relatively quickly around a single gene. Can
biotechnologists put enough copies in the field to confound that process?
The forestry of the future seems likely to blend aspects of intensive man-
agement and EM, either on the same piece of ground, across landscapes, or
(more likely) some combination of the two. What Seymour & Hunter (140)
termed the triad approach—a mix of reserves, intensively managed lands, and
EM lands—provides a useful starting point. The central issue then becomes
one of assessing probabilities of outcomes associated with differing spatial
and temporal patterns of the three general land-use types. That requires, in turn,
improved understanding of structure-process-function at the scale of landscapes
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

458 PERRY

(e.g. habitat, disturbance) and sites (e.g. resilience following harvest or natural
disturbance).
Science is faced with the challenge of providing knowledge that helps soci-
ety achieve sustainability (96). Numerous scientists have concluded this can
only be accomplished if we begin to grapple successfully with complexity
(e.g. 34, 95, 119), a strategy that radically departs from the industrial approach
of simplifying systems to make them more predictable. But coming to grips
with complex reality requires more than change in management philosophies;
some fundamental approaches of science and the role of scientists in society
must be reassessed as well. The idea that ecological systems are quintessential
complex systems is old; however, much knowledge concerning the characteris-
tics of complex systems is new and growing. Though much of this knowledge
has come from physicochemical systems and their models, ecological sys-
tems share many characteristics with these: existence far from thermodynamic
equilibrium, interconnections through positive and negative feedbacks, self-
organizing dynamics, metastability and vulnerability to threshold transitions
(124, 166). Adding the social and economic dimensions defines the stage on
which forestry plays and greatly magnifies the complexity.
Assessing how much such systems can be altered without triggering un-
foreseen consequences is far from straightforward. Some changes are obvious,
others subtle; some set processes in motion that may not manifest for decades
or centuries. Some, perhaps many, changes in system structure alter probabil-
ities rather than being distinctly causal, as when susceptibility to herbivores,
pathogens, fire, or wind is increased or decreased. Whether stands so altered
actually burn down or are eaten up depends on factors such as climate, the
evolutionary interplay between hosts and pests, and diverse biotic controls with
complex spatial and temporal dependence. The logic of cause and effect has
limited utility in such a milieu.
Scientists have played a much expanded role in forestry in the 1990s, a trend
that will almost certainly continue. The traditional routes of experiment and
modeling remain vital, and, despite inadequate funding (at least in the United
States; 114), there are presently far more experiments dealing with links be-
tween structure and process in managed forests of the United States and Canada
than at any time in the past. Within the scope of traditional science, however,
issues of experimental analysis and data interpretation need to be dealt with, es-
pecially the obsession with “significance questing”, which, as Rothman pointed
out with regard to the medical sciences, and which is equally true for natural re-
sources sciences, “has become for some a clumsy substitute for thought” (134).
When the quest is for significance at very high levels, as is the norm, the burden
placed on establishing differences among management approaches effectively
favors implementing those with the highest economic payoff. Various scientists
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

SCIENTIFIC BASIS OF FORESTRY 459

have argued that the burden of proof needs to be shifted away from practices that
would change the management status quo and placed instead on practices that al-
ter systems most dramatically or have the greatest economic payoff (e.g. 34, 72).
Funtowicz & Ravetz (51) argue that situations in which both system uncer-
tainty and decision stakes are high require a “post-normal” or second-order
science, which Costanza (34) interprets as taking the scientific method into
new territory. “The scientific method,” Costanza writes, “does not, in its basic
form, imply anything about the precision of results achieved. It does imply a
forum of open and free inquiry without preconceived answers or agendas . . . .”
Whether one accepts the new arena as science or not, the fact is that scientists
are increasingly playing nontraditional roles in resource management. This has
taken various forms, including adaptive management, which effectively blurs
the boundary between management and experiment (172); various regional
scientific assessments (41, 72, 145, 151); expert testimony in courts and be-
fore legislative and regulatory bodies; and small teams of scientists that assess
the suitability of lands for green certification. “We need a new model,” wrote
Gordon & Lyons (55), “for linking science, management, and policy . . . .” The
circumstances of the past few years have thrust forestry scientists into a leading
role in developing that model.
ACKNOWLEDGMENTS
Tom Adams, Tim White, Sharon Freidman, Dan Binkley, Gordon Grant, Lee
Allen, Bob Wagner, and RA Schmidt were all generous with reprints and other
guidance. I’m grateful to NSF’s Long-Term Ecological Research Program and
my Department Head, Logan Norris, for support over the years, and to the fertile
learning environment I was provided as a member of the HJ Andrews LTER
Program and Oregon State University’s Sustainable Forestry Partnership.

Visit the Annual Reviews home page at

http://www.AnnualReviews.org

Literature Cited

1. Aber JD, Melilo JM. 1982. FORTNITE: Kress LW, King JS. 1998. Leaf area
a computer model of organic matter and and above- and belowground growth re-
nitrogen dynamics in forest ecosystems. sponses of loblolly pine to nutrient and
Univ. Wisc. Res. Bull. R3130. Madison, water additions. For. Sci. In press
WI. 49 pp. 4. Alig R, Adams D, Chemelik J, Bettinger
2. Agee JK. 1990. The historical role of fire P. 1998. Private forest investment and
in Pacific Northwest forests. In Natural long-run sustainable harvest volumes.
and Prescribed Fire in Pacific Northwest In National Research Council 1997.
Forests, ed. JD Walstad, SR Radosovich, Forested Landscapes in Perspective.
DV Sandberg, pp. 25–38. Corvallis, OR: Washington, DC. National Acad. Press.
Oregon State Univ. Press In press
3. Albaugh TJ, Allen HL, Dougherty PM, 5. Amaranthus MA, Perry DA. 1994. The
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

460 PERRY

functioning of ectomycorrhiza in the and Process in a Forested Ecosystem.

field: linkages in space and time. Plant New York: Springer-Verlag. 253 pp.
Soil 159:133–40 20. Bosworth B, Studer D. 1991. Compar-
6. Amaranthus MP, Darbyshire R, Bormann isons of tree height growth on broad-
B. 1998. Long-term ecosystem productiv- cast burned, bulldozer-piled, and non-pre-
ity integrated research sites. Agron. Soc. pared sites 15 to 25 years after clear-cut
New Zealand. In press logging. Proceedings—Management and
7. Amaranthus MP, Trappe JM, Bednar L, Productivity of Western Montane Soils.
Arthur D. 1994. Hypogeous fungal pro- USFS Gen. Tech. Rep. INT-280, ed. AE
duction in mature Douglas-fir forest frag- Harvey, LF Neunshwander, pp. 197–200.
ments and surrounding plantation and its Washington, DC: USGPO
relation to coarse woody debris and an- 21. Boyle JR. 1975. Nutrients in relation to in-
imal mycophagy. Can. J. For. Res. 24: tensive culture of forest crops. Iowa State
2157–65 J. Res. 49:297–303
8. Amaranthus MP, Li CY, Perry DA. 1990. 22. Brocke RH, O’Pezio JP, Gustafson KAA.
Influence of vegetation type and madrone 1989. A forest management scheme
soil inoculum on associative nitrogen fix- for mitigating impact of road networks
ation in Douglas-fir rhizospheres. Can. J. on sensitive wildlife species. Is Forest
For. Res. 20:368–71 Fragmentation A Management Issue in
9. Angermeier PL, Karr JR. 1994. Biolog- the Northeast? USFS NE Forest Exp.
ical integrity versus biological diversity Sta. Gen. Tech. Rep. NE-140, pp. 7–12.
as policy directives. BioScience 44:690– Washington, DC: USGPO
97 23. Burger JA, Pritchett WL. 1988. Site
10. Baker FS. 1950. The Principles of Silvi- preparation effects on soil moisture and
culture. New York: McGraw-Hill available nutrients in a pine plantation in
11. Ballard R, Gessel SP, eds. 1983. IUFRO the Florida flatwoods. For. Sci. 34:77–87
Symposium on Forest Site and Continuous 24. Burgess RL, Sharp DM. 1981. Forest
Productivity. USDA For. Serv. Gen. Tech. Island Dynamics in Man-Dominated
Rep. PNW-163. Portland, OR Landscapes. New York: Springer-Verlag
12. Behan M. 1997. Scarcity, simplicity, sep- 25. Carey AB. 1995. Sciurids in Pacific
aratism, science—and systems. See Ref. Northwest managed and old-growth for-
88, pp. 411–18 ests. Ecol. Appl. 5:648–61
13. Beuter JH, Johnson KN. 1989. Economic 26. Castello JD, Leopold DJ, Smallidge PJ.
perspectives on maintaining the long- 1995. Pathogens, patterns, and processes
term productivity of forest ecosystems. in forest ecosystems. BioScience 45:16–
See Ref. 120, pp. 221–29 24
14. Binkley D. 1992. Mixtures of nitrogen- 27. Chang SX, Preston CM, Weetman GF.
fixing and non-nitrogen-fixing tree 1995. Soil microbial biomass and micro-
species. In The Ecology of Mixed Species bial and mineralizable N in a clear-cut
Stands of Trees, ed. MGR Cannell, DC chronosequence on northern Vancouver
Malcome, PA Robertson, pp. 99–123. Island, British Columbia. Can. J. For. Res.
London: Blackwell Sci. 25:1595–607
15. Binkley D. 1986. Forest Nutrition. 28. Chapin FS III, Sala OE, Burke IC, Grime
Chicester, NY: Wiley & Sons. 290 pp. JP, Hooper DU, et al. 1998. Ecosystem
16. Bormann FH, Likens GE, Sicama TG, consequences of changing biodiversity.
Pierce RS, Eaton JS. 1974. The ex- BioScience 48:45–52
port of nutrients and recovery of sta- 29. Christensen NL, Bartuska AM, Brown
ble conditions following deforestation at JH, Carpenter S, D’Antonio C, et al. 1996.
Hubbard Brook. Ecol. Monogr. 44:255– The report of the Ecological Society of
77 America committee on the scientific ba-
17. Bormann BT, Bormann FH, Bowden WB, sis of ecosystem management. Ecol. Appl.
Pierce RS, Hamburg SP, et al. 1993. Rapid 6:665–91
N2 fixation in pines, alder, and locust: ev- 30. Cissel JH, Swanson FJ. 1997. The present
idence from the sandbox ecosystem study. landscape-scale application of the range-
Ecology 74:583–98 of-historical-variability concept in the
18. Bormann BT, Wand D, Bormann FH, Oregon Cascades. Abstracts of the 1997
Benoit G, April R, Snyder MC. 1998. Meeting of the Ecological Society of
Rapid, plant-induced weathering in an ag- America, p. 9
grading experimental ecosystem. Biogeo- 31. Clayton JL. 1979. Nutrient supply to soil
chemistry Submitted by rock weathering. See Ref. 136, pp. 75–
19. Bormann FH, Likens GE. 1979. Pattern 96
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

SCIENTIFIC BASIS OF FORESTRY 461

32. Cole DW. 1995. Soil nutrient supply in variable retention harvest systems. See
natural and managed forests. Plant Soil. Ref. 88, pp. 111–39
168–69:43–53 47. Franklin JF, Cromack K Jr, Denison W,
33. Compton J, Cole DW. 1990. Impact of McKee A, Maser C, et al. 1981. Ecolo-
harvest intensity on growth and nutrition gical characteristics of old-growth Doug-
of second rotation Douglas-fir. See Ref. las-fir forests. USDA For. Serv. Gen. Tech.
59, pp. 151–62 Rep. PNW-118. Portland, OR
34. Costanza R. 1993. Developing ecological 48. Freidman ST, Foster GS. 1997. Forest
research that is relevant for achieving sus- genetics on federal lands in the United
tainability. Ecol. Appl. 3:579–81 States: public concerns and policy re-
35. Curtis RO. 1997. The role of extended ro- sponses. Can. J. For. Res. 27:401–8
tations. See Ref. 88, pp. 165–70 49. Fried JS, Boyle JR, Tappeiner JC II, Cro-
36. Ehrlich PC. 1994. Biodiversity and eco- mack K Jr. 1990. Effects of bigleaf maple
system function: need we know more? on soils in Douglas-fir forests. Can. J. For.
In Biodiversity and Ecosystem Function, Res. 20:259–66
ed. E-D Schulze, HA Mooney, pp. vii–xi. 50. Fuller B. Quoted in Explorations. Spring/
New York: Springer Summer 1998, p. 6
37. El-Kassaby YA, Ritland K. 1996. Impact 51. Funtowicz SO, Ravetz JR. 1991. A new
of selection and breeding on the genetic scientific methodology for global envi-
diversity in Douglas-fir. Biodiversity Con- ronmental problems. In Ecological Eco-
servation 5:795–813 nomics: The Science and Management of
38. El-Kassaby YA, Namkoong G. 1995. Sustainability, ed. R Costanza, pp. 137–
Genetic diversity of forest tree planta- 52. New York: Columbia Univ. Press
tions: consequences of domestication. In 52. Gagnon RP, Chabot M. 1991. Preven-
Consequences of changes in biodiversity, tion des pertes de bois attribuables a
IUFRO World Congress 2:218–28. Tam- la torduese des bourgeons de l’epinette.
pere, Finland Governement du Quebec, Ministere des
39. Eriksson G, Namkoong G, Roberds JH. Forets, Montreal
1993. Dynamic gene conservation for un- 53. Gessel SP, Lacate DS, Weetman GF, Pow-
certain futures. For. Ecol. Manage. 62:15– ers RF, eds. 1990. Sustained Productiv-
37 ity of Forest Soils. Proceedings of the 7th
40. Farnum P, Timmis R Jr, Kulp K. 1983. North American For. Soils Conf.; 1988
Biotechnology of forest yield. Science Jul; University of British Columbia. Fac-
219:694–702 ulty of Forestry, Univ. Br. Columbia. Van-
41. FEMAT. 1993. Forest ecosystem manage- couver, Canada V6T 1W5
ment: an ecological, economic, and social 54. Glover GR, Creighton JL, Gjerstad DH.
assessment. Rep. For. Ecosystem Manage. 1989. Herbaceous weed control increases
Assessment Team (FEMAT). 1996-793- loblolly pine growth for twelve years. J.
071. Washington, DC: USGPO For. 87:47–50
42. Foster DR, Orwig DA, McLachlan JS. 55. Gordon JC, Lyons J. 1997. The emerging
1996. Ecological and conservation in- role of science and scientists in ecosystem
sights from reconstructive studies of tem- management. See Ref. 88, pp. 447–53
perate old-growth forests. Trends Ecol. 56. Gregory SV, Lamberti GA, Erman DC,
Evol. (TREE) 11:419–24 Koski KV, Murphy ML, Sedell JR. 1987.
43. Fox TR, Morris LA, Maimone RA. 1989. Influence of forest practices on aquatic
The impact of windrowing on the produc- production. In Streamside Management:
tivity of rotation age loblolly pine. Proc. Forestry and Fisheries Interactions, ed.
5th Biennial Southern Silviculture Conf. EO Salo, TW Cundy, pp. 233–55. Contri-
Ashville, NC: USDA For. Serv. bution No. 57, Inst. For. Resources, Univ.
44. Franklin JF. 1993. The fundamentals of Washington, Seattle
ecosystem management with applications 57. Grumbine RE. 1994. What is ecosystem
in the Pacific Northwest. In Defining Sus- management? Conserv. Bio. 8:27–38
tainable Forestry, ed. GH Aplet, N John- 58. Hansen AJ, Spies TA, Swanson FJ, Oh-
son, JT Olson, VA Sample, pp. 127–44. mann JL. 1991. Conserving biological di-
Washington, DC: Island Press versity in managed forests. BioScience
45. Franklin JF. 1993. Preserving biodiver- 41(6):382–92
sity: species, ecosystems or landscapes. 59. Hansen AJ, Garman SL, Marks B. 1993.
Ecol. Appl. 3:202–5 An approach for managing vertebrate di-
46. Franklin JF, Berg DR, Thornburg DA, versity across multiple-use landscapes.
Tappeiner JC. 1997. Alternative Silvi- Ecol. Appl. 3:481–96
cultural approaches to timber harvesting: 60. Hansen AJ, McComb C, Vega R, Raphael
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

462 PERRY

MG, Hunter M. 1995a. Bird habitat rela- Oregon and Washington. Bethesda, MD:
tionships in natural and managed forests The Wildlife Society
in the west Cascades of Oregon. Ecol. 73. Hodge GR, White TL, Powell GL, de
Appl. 5:555–69 Souza SM. 1989. Predicted genetic gains
61. Hansen AJ, Garman SL, Weigand JF, Ur- from one generation of slash pine tree im-
ban DL, McComb WC, Raphael MG. provement. So. J. Appl. For. 13:51–56
1995b. Alternative silvicultural regimes 74. Hodge GR, White TL, Schmidt RA, Allen
in the Pacific Northwest: simulations of JE. 1993. Stability of rust infection ra-
ecological and economic effects. Ecol. tios for resistant and susceptible slash and
Appl. 5:535–54 loblolly pines across rust hazard levels.
62. Harmon ME, Franklin JF, Swanson FJ, So. J. Applied For. 17:188–92
Sollins P, Gregory SV, et al. 1986. Ecol- 75. Holling CS. 1992. Cross-scale morphol-
ogy of coarse woody debris in temperate ogy, geometry, and dynamics of ecosys-
ecosystems. Adv. Ecol. Res. 15:133–302 tems. Ecol. Monogr. 62:447–502
63. Harr RD, Harper WC, Krygier JT, Hsieh 76. Hollstedt C, Vyse A. 1997. Sicamous
FS. 1975. Changes in storm hydrographs Creek Silvicultural Systems Project. Br.
after road building and clear-cutting in Columbia Min. For. Res. Prog. Work. Pap.
Oregon Coast Range. Water Resour. Res. 24. Victoria, BC
11:436–44 77. Holt RD, Hochberg ME. 1997. When
64. Harr RD. 1976. Hydrology of small for- is biological control evolutionarily stable
est streams in western Oregon. Gen. Tech. (or is it)? Ecology 78:1673–83
Rep. PNW-55. USFS Pacific Northwest 78. Hornbeck JW, Swank WT. 1992. Wa-
Exp. Station. Portland, OR tershed ecosystem analysis as a basis
65. Harris LD. 1984. The Fragmented Forest. for multiple-use management of eastern
Chicago, IL: Univ. Chicago Press forests. Ecol. Appl. 2:238–47
66. Hart SC, Binkley D, Campbell RG. 1986. 79. Hutten M. 1998. Performance of pon-
Predicting loblolly pine current growth derosa pine seedlings outplanted in for-
and growth response to fertilization. Soil est, patch-cuts, and clear-cuts in central
Sci. Soc. Am. J. 50:230–33 Oregon. MS thesis. Oregon State Univ.,
67. Hart SC, Binkley D, Perry DA. 1997. In- Corvallis
fluence of red alder on soil nitrogen trans- 80. Ingestad T, Aronsson A, Agren GI. 1981.
formations in two conifer forests of con- Nutrient flux density model of mineral nu-
trasting productivity. Soil Biol. Biochem. trition in conifer ecosystems. Studia Fore-
29:1111–23 stalia Suecica 160:61–71
68. Harvey AE, Neuenschwander LF, eds. 81. Johnson DW. 1992. Effects of forest man-
1991. Proceedings—Management and agement on soil carbon storage. Water,
Productivity of Western-Montane Forest Air, Soil Pollution 64:83–120
Soils. USFS Gen. Tech. Rep. INT-280. 82. Jones JA, Grant GE. 1996. Peak flow re-
254 pp. sponses to clear-cutting and roads in small
69. Harvey AE, Larsen MJ, Jurgensen MF. and large basins, western Cascades. Ore-
1980. Clear-cut harvesting and ectomyc- gon Water Resour. Res. 32:959–74
orrhizae: survival of activity on residual 83. Jongmans AG, van Breemen N, Lund-
roots and influence on a bordering stand strom U, van Hees PAW, Finlay RD,
in western Montana. Can. J. For. Res. 10: et al. 1997. Rock-eating fungi. Nature
300–3 389:683–84
70. Harvey AE, Meurisse RT, Geist JM, 84. Jurgensen MF, Larsen MJ, Graham RT,
Jurgensen MF, McDonald GI, et al. Harvey AE. 1987. Nitrogen fixation in
1989. Managing long-term productivity woody residue of northern Rocky Moun-
in northwest-mixed conifers and pines. tain conifer forests. Can. J. For. Res. 17:
See Ref. 120, pp. 164–84 1283–88
71. Harvey AE, Jurgensen MF, Larsen MJ, 85. Kareiva P, Wennergen U. 1995. Connect-
Graham RT. 1987. Relationships among ing landscape patterns to ecosystem and
soil microsite, ectomycorrhizae, and nat- population processes. Nature 373:299–
ural conifer regeneration of old growth 302
forests of western Montana. Can. J. For. 86. Feller MC, Kimmins JP, Scoullar KA.
Res. 17:58–62 1983. FORCYTE 10: calibration data and
72. Henjum MG, Karr JR, Bottom DL, simulation of potential long-term effects
Perry DA, Bednarz JC, et al. 1994. In- of intensive forest management on site
terim Protection for Late-Successional productivity, economic performance, and
Forests, Fisheries, and Watersheds: Na- energy benefit/cost ratio. See Ref. 82, pp.
tional Forests East of the Cascade Crest, 179–200
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

SCIENTIFIC BASIS OF FORESTRY 463

87. Kimmins JP. 1977. Evaluation of the con- Wissmar RC, Clarke SE, et al. 1993. Man-
sequences for future tree productivity of agement history of eastside ecosystems:
the loss of nutrients in whole-tree harvest- changes in fish habitat over 50 years,
ing. For. Ecol. Manage. 1:169–83 1935–1992. In Eastside Forest Ecosys-
88. Kohm KA, Franklin JF, eds. 1997. Creat- tem Health Assessment, ed. PF Hessburg,
ing a Forestry for the 21st Century. Wash- Vol. III. Washington, DC: USDA For.
ington, DC: Island Press Serv.
89. Korzukin MD, Ter-Mikaelian T, Wagner 101. McKeand SE, Crook RP, Allen HL. 1998.
RG. 1996. Process versus empirical mod- Genotypic stability effects on predicted
els: which approach for forest ecosystem family responses to silvicultural treat-
management? Can. J. For. Res. 26:879– ments in loblolly pine. So. J. Appl. For.
87 In press
90. Kuenzler EJ. 1989. Value of forested wet- 102. McKeand S, Svensson J. 1997. Loblolly
lands as filters for sediments and nu- pine: sustainable management of genetic
trients. In Proc. Symposium: The Forested resources. J. For. March:4–9
Wetlands of the Southern United States, 103. Miller JH, Zutter B, Zedaker SM, Cain M,
ed. DD Hook and R Lea, pp. 85–96. Edwards MB, et al. 1986. A region-wide
USDA For. Serv. Gen. Tech. Rep. SE-50. study of loblolly pine growth relative to
Asheville, NC four competition levels after four growing
91. Landres P, Morgan P, Swanson F. 1997. seasons. In Proc. 4th Biennial Res. Conf.,
Evaluating the usefulness of natural range ed. DR Phillips, pp. 581–91. Asheville,
of variability in managing ecological sys- NC: USDA For. Serv. Gen. Tech. Rep. SE-
tems. Abstr. 1997 Meet. Ecol. Soc. Am., 42
p. 20 104. Miller RE, Barker PR, Peterson C, Web-
92. Langston N. 1994. Forest Dreams, Forest ster SR. 1986. Using nitrogen fertiliz-
Nightmares. Seattle: Univ. Wash. Press ers in management of coast Douglas-fir.
93. Ledig FT. 1986. Heterozygosity, hetero- In Douglas-fir: Stand Management for
sis, and fitness in outbreeding plants. In the Future, ed. CD Oliver, DP Hanley,
Conservation Biology: The Science of JA Johnson, pp. 290–303. Inst. For. Res.
Scarcity and Diversity, ed. ME Soule, pp. Contribution No. 55. Seattle: Univ. Wash.
77–104. Sunderland, MA: Sinauer Coll. For. Resources
94. Li CY, Hung LL. 1987. Nitrogen-fixing 105. Moldenke A. 1990. One hundred twenty
(acetylene-reducing) bacteria associated thousand little legs. Wings Summer:11–
with ectomycorrhizae of Douglas-fir. 14
Plant Soil 98:425–28 106. Morris LA. 1989. Long-term productiv-
95. Lubchenco J, Olson AM, Brubaker LB, ity research in the US southeast: expe-
Carpenter SR, Holland MM, et al. rience and future directions. In Research
1991. The sustainable biosphere initia- Strategies For Long-term Site Productiv-
tive: an ecological research agenda. Ecol- ity, ed. WJ Dyck, CJ Mees, FRI Bull.
ogy 72:371–412 152, pp. 221–35. Rotorua, NZ: For. Res.
96. Lubchenco J. 1998. Entering the cen- Inst.
tury of the environment: the need for a 107. Morris LA, Pritchett WL. 1983. Effects
new social contract for science. Science of site preparation on Pinus elliotii-Pinus
279:491–97 palustris flatwoods forest soil properties.
97. Mann LK, Johnson DW, West DC, Cole See Ref. 17, pp. 243–51
DW, Hornbeck JW, et al. 1988. Effects of 108. Morris LA, Pritchett WL, Swindel BF.
whole-tree and stem-only clear-cutting on 1983. Displacement of nutrients into
post-harvest hydrologic losses, nutrient windrows during site preparation in a flat-
capital, and regrowth. For. Sci. 34:412– wood forest. Soil Sci. Soc. Am. J. 47:591–
28 94
98. Marcot BG. 1997. Biodiversity of old 109. Morrison PH, Swanson FJ. 1990. Fire
forests of the west: a lesson from our el- History and Pattern in a Cascade Moun-
ders. See Ref. 130, pp. 87–105 tain Landscape. USDA For. Serv. Gen.
99. McArthur JV. 1989. Aquatic and terres- Tech. Rep. PNW-GTR-254. Portland, OR
trial linkages: flood plain functions. In 110. Namkoong G. 1984. A control concept
Proc. Symposium: The Forested Wetlands for gene conservation. Silvae Genetica
of the Southern United States, ed. DD 33:160–33
Hook, R Lea, pp. 107–16. Asheville, 111. Namkoong G, Bishir. 1987. The fre-
NC: USDA For. Serv. Gen. Tech. Rep. quency of lethal alleles in forest tree pop-
SE-50 ulations. Evolution 41:1123–27
100. McIntosh BA, Sedell JR, Smith JE, 112. NC State Forest Nutrition Cooperative.
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

464 PERRY

1998. HTTP://WWW2.ncsu.edu/unity/ 129. Powers RF, Alban DH, Miller RE, Tiarks
lockers/project/ncsfnchpg AE, Wells CG, et al. 1990. Sustaining site
113. Noss RF, LaRoe ET III, Scott JM. 1995. productivity in North American forests:
Endangered Ecosystems of the United problems and prospects. See Ref. 53, pp.
States: a Preliminary Assessment of Loss 49–79
and Degradation. Biol. Rep. 28. Wash- 130. Radosovich SR, Ghersa CM. 1992.
ington, DC: USDI Nat. Biol. Service. Weeds, crops, and herbicides: a modern-
58 pp. day “neckriddle.” Weed Tech. 6:788–95
114. National Research Council. 1990. Fores- 131. Read DJ. 1993. Plant-microbe mutu-
try Research: A Mandate for Change. alisms and community structure. In Bio-
Washington, DC. Nat. Acad. Press. 84 pp. diversity and Ecosystem Function, ed. ED
115. Pastor J. 1993. Modelling the effects of Schulze, HA Mooney, pp. 181–236. New
timber management on populations dy- York: Springer-Verlag
namics, diversity, and ecosystem pro- 132. Rehfeld GE. 1992. Breeding strategies for
cesses. In Modelling Sustainable Forest Larix occodentalis: adaptations to the bi-
Ecosystems, ed. DC LeMasetr. Washing- otic and abiotic environment in relation to
ton, DC: Soc. Am. Foresters improving growth. Can. J. For. Res. 22:5–
116. Pastor J, Post WM. 1986. Influence of cli- 13
mate, soil moisture, and succession on 133. Reiter ML, Beschta RL. 1995. Effects
forest carbon and nitrogen cycles. Biogeo- of forest practices on water. In Cumula-
chemistry 2:3–27 tive Effects of Forest Practices in Oregon.
117. Peck JE, McCune BM. 1997. Remnant Rep. for Oregon Dept. For. Ch. 7, RL
trees and canopy lichen communities Beschta, JR Boyle, CC Chambers, WP
in western Oregon: a retrospective ap- Gibson, SV Gregory, et al, Salem, OR
proach. Ecol. Appl. 7:1181–87 134. Rothman KJ. 1986. Significance quest-
118. Perry DA. 1994. Forest Ecosystems. Bal- ing. Ann. Int. Med. 105:445–47
timore, MD: Johns Hopkins Univ. Press 135. Savolainen O, Karkkainen K. 1992. Ef-
119. Perry DA, Maghembe J. 1989. Ecosystem fect of forest management on gene pools.
concepts and current trends in forest man- New For. 6:329–45
agement time for reappraisal. For. Ecol. 136. Schowalter TD, Turchin P. 1993. South-
Manage. 26:123–40 ern pine beetle infestation development:
120. Perry DA, Meurisse R, Thomas B, Miller interaction between pine and hardwood
R, Boyle J, et al, eds. 1989. Maintaining basal areas. For. Sci. 39:201–10
Long Term Productivity of Pacific North- 137. Schowalter TD. 1995. Canopy arthropod
west Forests. Portland, OR: Timber community response to forest age and al-
121. Perry DA, Amaranthus MP, Borchers ternative harvest practices in western Ore-
JG, Borchers SL, Brainerd RE. 1989. gon. For. Ecol. Manage. 78:115–25
Bootstrapping in ecosystems. BioScience 138. Schulze E-D, Mooney HA. 1994. Ecosys-
39:230–37 tem function of biodiversity: a summary.
122. Perry DA, Rose SL. 1983. Soil biology In Biodiversity and Ecosystem Function,
and forest productivity: opportunities and ed. E-D Schulze, HA Mooney, pp. 497–
constraints. See Ref. 17, pp. 229–38 510. New York: Springer-Verlag
123. Perry DA. 1988. Landscape patterns and 139. Sedell JR, Bison PA, Swanson FJ, Gre-
forest pests. Northwest Environ. J. 4:213– gory SV, eds. 1988. What we know about
28 large trees that fall into streams and
124. Perry DA. 1995. Self-organizing systems rivers. Gen. Tech. Rep. PNW-GTR-229.
across scales. TREE 10:241–44 Portland, OR: USDA For. Serv.
125. Perry DA, Amaranthus MP. 1997. Distur- 140. Seymour R, Hunter ML Jr. 1992. Prin-
bance, recovery, and stability. See Ref. 88, ciples and Applications of New Forestry
pp. 31–56 in Spruce-Fir Forests of Eastern North
126. Plochmann R. 1989. The forests of cen- America. Publication 716. Orono, Maine
tral Europe: a changing view. In Oregon’s Ag. Exp. Sta.
Forestry Outlook: An Uncertain Future. 141. Shainsky LJ, Radosovich SR. 1992.
The 1989 Starker Lectures, pp. 1–9. Cor- Mechanisms of competition between
vallis, OR: Coll. Forestry Douglas-fir and red alder seedlings. Ecol-
127. Powers HR, Miller T, Belanger RP. 1993. ogy 73:30–45
Management strategies to reduce losses 142. Shainsky LJ, Rose CL. 1995. Effects of
from fusiform rust. So. J. Appl. For. 17: competition on the foliar chemistry of
146–49 young Douglas-fir in monoculture and
128. Powers RF. 1988. Unpublished data given mixed stands with red alder. Can. J. For.
in Ref. 120. Res. 25:1969–77
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

SCIENTIFIC BASIS OF FORESTRY 465

143. Sharitz RR, Boring LR, Van Lear DH, ern United States: SRIEG satellite work-
Pinder JE III. 1992. Integrating ecological shop summary remarks. Can. J. For. Res.
concepts with natural resource manage- 27:442–46
ment of southern forests. Ecol. Appl. 2: 156. Stoehr MU, El-Kassaby YA. 1997. Levels
226–37 of genetic diversity at different stages of
144. Shumway JS, Chappell HN. 1995. Preli- the domestication cycle of interior spruce
minary DRIS norms for coastal Douglas- in British Columbia. Theor. Appl. Genet.
fir soils in Washington and Oregon. Can. 94:83–90
J. For. Res. 25:208–14 157. Stonecypher RW, Piesch RF, Helland GG,
145. SICB. 1996. Status of the Interior Chapman JG, Reno HJ. 1996. Results
Columbia Basin: Summary of Scientific from genetic tests of selected parents
Findings. USFS Gen. Tech. Rep. PNW- of Douglas-fir (Pseudotsuga meneziessi
GTR-385. Portland, OR. 144 pp. Mirb. Franco) in an applied tree improve-
146. Simard S, Vyse A. 1994. Paper birch: ment program. For. Sci. Monogr. 32:1–
weed or crop tree in the interior cedar- 35
hemlock forests of south British Colum- 158. Stuart-Smith K, Hebert D. 1996. Putting
bia. In Interior Cedar-Hemlock-White sustainable forestry into practice at
Pine Forests: Ecology and Management. Alberta-Pacific. CPPA Woodlands Pap.
pp. 311–16. Pullman, WA: Wash. State April/May:57–60
Univ. Dept. Nat. Resource Sci. 159. Swanson FJ, Clayton JL, Megahan WF,
147. Simard SW, Perry DA, Jones MD, My- Bush G. 1989. Erosional processes and
rold DD, Durall DM, Molina R. 1997. long-term site productivity. See Ref. 120,
Net transfer of carbon between ectomy- pp. 67–81
corrhizal tree species in the field. Nature 160. Tappenier JC, Huffman D, Marshall D,
388:579–82 Spies TA, Bailey JD. 1997. Density,
148. Small MF, Hunter ML. 1988. Forest frag- ages, and growth rates in old-growth and
mentation and avian nest predation in young-growth Douglas-fir in coastal Ore-
forested landscapes. Oecologia 76:62–64 gon. Can. J. For. Res. 27:638–48
149. Smith CT Jr, McCormack ML Jr, Horn- 161. Taylor SP, Alfaro RI, DeLong C, Rankin
beck JW, Martin CW. 1986. Nutrient and L. 1996. The effects of overstory shad-
biomass removals from a red spruce— ing on white pine weevil damage to white
balsam fir whole-tree harvest. Can. J. For. spruce and its effects on spruce growth
Res. 16:381–88 rates. Can. J. For. Res. 26:306–12
150. Smith DR (Technical Coordinator). 1975. 162. Thomas JW, ed. 1979. Wildlife Habitats
Proceedings of the Symposium on Man- in Managed Forests of the Blue Moun-
agement of Forest and Range Habitats for tains of Oregon and Washington. Agricul-
Nongame Birds. USFS Gen. Tech. Rep. ture Handbook No. 553. Washington, DC:
WO-1. Washington, DC: USGPO. 343 pp. USDA For. Serv.
151. SNEP. 1996. Status of the Sierra Nevada. 163. Tilman D, May RM, Lehman CL, Nowak
Sierra Nevada Ecosystem Project, Final MA. 1994. Habitat destruction and the ex-
Report to Congress. Wildland Rd Res. tinction debt. Nature 371:65–66
Center Rep. No. 36. Davis, CA: Univ. 164. Todd D, Pait J, Hodges J. 1995. The im-
Calif. Centers for Water and Wildland pact and value of tree improvement in the
Resources south. Proc. 23rd So. Forest Tree Improve-
152. Deleted in proof ment Conf. 7–15
153. South DB, Zwolinski JB, Allen HL. 165. Torgerson TR, Mason RR, Campbell RW.
1995. Economic returns from enhancing 1990. Predation by birds and ants on two
loblolly pine establishment on two upland forest insect pests in the Pacific North-
sites: effects of seedling grade, fertiliza- west. Stud. Avian Biol. 13:14–19
tion, hexazinone, and intensive soil culti- 166. Turner MG, Romme WH, Gardner RH,
vation. New For. 10:239–56 O’Neill RV, Kratz TK. 1993. A revised
154. Spies TA, Franklin JF. 1991. The structure concept of landscape equilibrium: distur-
of natural young, mature, and old-growth bance and stability on scaled landscapes.
Douglas-fir stands in western Oregon and Landscape Ecol. 8:213–27
Washington. In Wildlife and Vegetation of 167. USDA Forest Service. 1995. The 1993
Unmanaged Douglas-Fir Forests, ed LF RPA Timber Assessment Update. General
Ruggerio, KB Aubrey, AB Carey, MH Tech. Rep. RM-GTR-259. Fort Collins,
Huff. USDA For. Serv. Gen. Tech. Rep. CO. 66 pp.
PNW-GTR-285. Portland, OR 168. Velazquez-Martinez A, Perry DA, Bell
155. Stelzer HE, Goldfarb B. 1997. Imple- TE. 1992. Response of aboveground bio-
menting clonal forestry in the southeast- mass increment, growth efficiency, and
P1: KKK/mbg P2: ARS/ary QC: ARS/kba T1: ARS
September 17, 1998 14:48 Annual Reviews AR067-16

466 PERRY

foliar nutrients to thinning, fertilization, PhD thesis, North Carolina State Univ.
and pruning in young Douglas-fir planta- Raleigh, NC. 125 pp.
tions in the central Oregon Cascades. Can. 176. Williams CG, Hamrick JL, Lewis PO.
J. For. Res. 22:1278–89 1995. Multiple population versus hierar-
169. Vitousek PM, Matson PA. 1985. Distur- chical breeding populations: a compar-
bance, nitrogen availability, and nitrogen ison of genetic diversity levels. Theor.
losses in an intensively managed loblolly Appl. Genet. 90:584–94
pine plantation. Ecology 66:1360–76 177. Williams CG, Hamrick JL. 1996. Elite
170. Wagner RG. 1993. Research directions populations for conifer breeding and gene
to advance forest vegetation management conservation. Can. J. For. Res. 26:453–61
in North America. Can. J. For. Res. 23: 178. Yarie J. 1989. A comparison of the nu-
2317–27 tritional consequences of intensive forest
171. Walstad JD, Kuch PJ, ed. 1987. Forest harvesting in Alaskan taiga forests as pre-
Vegetation Management for Conifer Pro- dicted by FORCYTE-10 and LINKAGES
duction. New York: Wiley & Sons 2. IEA/BE Project A6 Rep. No. 1. Forest
172. Walters C. 1986. Adaptive Management Soils Lab. Univ. Alaska, Fairbanks. 59 pp.
of Renewable Resources. New York: 179. Zeimer RR, Lewis J, Rice RM, Lisle TE.
MacMillan 1991. Modelling the cumulative water-
173. Wickman BE, Mason RP, Swetnam TW. shed effects of forest management strate-
1993. Searching for long-term patterns of gies. J. Environ. Qual. 20:36–42
forest insect outbreaks. In Proceedings: 180. Zeimer RR, Lewis J, Lisle TE, Rice RM.
Individuals, Populations, and Patterns, 1991. Long-term sedimentation effects of
September 7–10. Norwich, England different patterns of timber harvesting.
174. Wiersum KF. 1995. 200 years of sustain- Int. Assoc. Hydrol. Sci. 203:143–50
ability in forestry: lessons from history. 181. Zutter BR, Gjerstad DH, Glover GR.
Environ. Manage. 19:321–29 1987. Fusiform rust incidence and sever-
175. Wilcox PL. 1995. Use of DNA markers ity in loblolly pine plantations follow-
for the genetic dissection of breeding of ing herbaceous weed control. For. Sci.
fusiform rust resistance in loblolly pine. 33:790–800
View publication stats