CHAPTER 3: RESEARCH AND LEARNING

 

Adaptive Management Areas...are specific areas dedicated primarily to the objective
of development and testing of new approaches....
 (Plan 1994, P. E-13, emphasis added)
   

BACK TO CONTENTS PAGE
 
 



 
 

What Are the Roles of Research and Learning in Adaptive Management?

Learning is what makes adaptive management adaptive; the formalized learning process in adaptive management is what distinguishes it from past forms of management. Any management approach with multiple objectives (for example, maintaining many different resources, species, and benefits to people) is of necessity an educated guess based on past experience, research, and theory. Land managers have tended to implement a few "best management practices" which were only adjusted when sufficient evidence of problems accumulated. In adaptive management, the goal is to try several different educated guesses in the beginning, monitor them systematically to see how they're doing, evaluate results and learn more about how the ecosystem works, and adapt our management plans to better meet our objectives. The AMAs in particular are also charged with testing the guidelines of the Northwest Forest Plan to see if better ways can be found to manage ecosystems and to avoid making new mistakes by following a narrow range of approaches.

Of course, forest management in the Oregon Coast Range isn't starting from scratch! We already know quite a bit about ecosystems in this area, and bringing that knowledge with us is important as we set out on a new course with different, broader objectives. Several assessments will help us do that:

Many terms and definitions have been used to describe mature and old-growth forest in the Pacific Northwest. One definition of late-successional forest is given in the introduction to this Guide. As used in the Northwest Forest Plan, the term refers to a range of forest conditions, beginning with stands in which tree crown expansion slows, openings between trees become larger and more stable, and large dead and fallen trees begin to accumulate. The definition includes older stands in which the oldest trees reach their maximum sizes, understory trees form multiple canopy layers, and large amounts of dead wood accumulate (ROD 1994, B-3). Late-successional characteristics typically begin to form in Douglas-fir forests between ages 80 and 140. The Forest Service definition of old growth applies to old late-successional stands that meet some minimum standards for numbers of large trees, range of tree sizes, multiple canopy layers, and abundance of snags and logs (Old Growth Definition Task Group 1986). The Bureau of Land Management classifies stands greater than 200 years of age as old growth, as a rule of thumb for analysis purposes. The Oregon State Department of Forestry is developing an "older forest" definition with structural categories similar to, but minimum standards different from, the Forest Service old growth definition.
 
 

What Do We Know?
 

Late-successional forest

As we pointed out in Chapter 2, there is much less late-successional forest in the Oregon Coast Range now than there was 150 years ago. In 1850, about 40% of the Coast Range between Astoria and Reedsport consisted of forests more than 200 years old, and 35% of the area had recently burned; thus 200-year-old forest may have covered about 60% of the area before the large fires that burned in the 1840s (Teensma et al. 1991, Ripple 1994). At least some, and maybe most, of the fires during the mid- and late-1800s were caused by people. This and other information suggests that coastal forests experienced catastrophic fire every 300 to 400 years, on average (Agee 1993, Ripple 1994, Siuslaw National Forest 1995). Estimates of currently existing older forest differ by method and geographic area: 0.8% of the Nestucca Watershed (200 acres) has forest older than 200 years (Nestucca Watershed Analysis 1994); 5.4% of the Siuslaw National Forest (33,800 acres) is old-growth forest (Bolsinger and Waddell 1993); and 10.0% of the federal lands in the Coast Range (140,500 acres) is multi-canopied forest with trees greater than 21 inches in diameter (FEMAT 1993). Note that federal lands occupy about one-third of the entire Oregon Coast Province (Siuslaw National Forest 1995). General classes of forest types in the AMA are shown in Map 4.

Late-successional forest is important habitat for many different species, including two species listed as threatened under the Endangered Species Act: the northern spotted owl and the marbled murrelet. The abundance of many other forest species seems to be associated more with the features that are found in natural and late-successional forests (for example, large snags and logs, multiple tree canopy layers, trees in a range of sizes), rather than with the age of the forest itself (Ruggiero et al. 1991, FEMAT 1993). We know that we can harvest some of the trees in mature Douglas-fir forests to accelerate the growth of the trees we leave and provide room for new trees to grow underneath to develop multiple canopy layers. We also know that we can increase the number of logs and snags in a forest by selectively killing trees and leaving them in place.
 
 

Fisheries and Watersheds

Anadromous fish populations in the Oregon Coast Range are a fraction of what they were 150 years ago. For example, the numbers of commercially caught salmon on the Siuslaw River (just to the south of the AMA) went from 87,500 to 7,000 for coho, and 11,000 to none for chinook between 1890 and 1960 (Sedell and Luchessa 1982). Accounts of rivers in the 1800s "choked" with salmon are common. Currently, 32 salmonid stocks in the AMA are considered at risk of extinction or of concern (FEMAT 1993). The National Marine Fisheries Service (NMFS) has grouped salmon stocks into "evolutionarily significant units"; the numbers and locations of these units are not yet available. The NMFS has proposed that coho salmon and steelhead be listed as threatened throughout Oregon.

Fish need clean water, in-stream gravel beds, and channel complexity (for example pools, riffles, and side channels) to complete their life cycles. Watersheds are dynamic systems that over time experience large fires, floods, and landslides. Natural stream systems maintain complex fish habitat through the periodic addition of gravels, boulders, and large conifer logs to streams from landslides and bank erosion. In many managed watersheds, however, streams are getting continual input of fine sediment, often from roads, but little coarse sediment and almost no wood. We know that we can install in-stream structures (mainly large logs) that increase channel complexity, but these approaches are inherently short term. Long-term solutions must consider the availability of large conifers that can fall or slide into streams, reduction of road density, and improved road design and location. Comprehensive solutions will also need to consider the entire upland, riparian, estuary, and ocean ecosystem on which fish populations depend, including the effects of commercial and sport fishing, agricultural activities along streams, and climate-driven changes in ocean currents.
 
 

People and communities

The communities in and around the north coast area are diverse, with a great variety of lifestyles and interests among residents. Communities tend to be somewhat isolated from each other because getting around in this mountainous landscape is not easy. Historically, the northern Coast Range has produced substantial natural resources, but these resources have mostly been exported to other areas for manufacturing and use in value-added products. Many people also value unmanaged forest landscapes (wilderness and roadless areas), which have become relatively rare in the northern Coast Range. The primary natural resources have been commercial fisheries on the coast, timber in the mountains, and agriculture in the river valleys. Tourism continues to be important, particularly near the ocean, and collecting special forest products like moss and mushrooms is a growing industry. Employment in commercial fisheries and canneries declined several decades ago, and jobs in forestry have dwindled more recently. There is a trend toward fewer, larger businesses in many economic sectors (for example, lumber mills and dairy farms). The total number of jobs in the area has increased, but most are in primarily non-manufacturing sectors, so average take-home wages are lower.
 
 

What Do We Need to Learn?

The primary objectives of this AMA represent an experiment to determine if we can provide for healthy populations of late-successional and aquatic species and provide value to local communities at the same time. Historical management in this area has not done this; current ideas of how to do it range from leaving the lands alone to intensively managing different parts of the ecosystem. Information is also needed on the basic habitat needs of a variety of rare species, concern for which is driving current management policies in the region. We know relatively little about how species, management, and natural disturbance interact at the watershed and landscape scales that are important for the long-term viability of many species and ecosystem functions.

No one knows whether people, through management actions, can develop stands that function as old-growth. Current old-growth forests originated and developed during the last several centuries. Some people believe that we don't know much about how those ecosystems developed, and even if we figure it out, we may not be able to reproduce those conditions under current climate, land-ownership patterns, competition from exotic plants and animals, and disturbance regimes--particularly if it's the large fires, floods, and wind-storms that were the critical factors. Some people believe that every young forest will develop naturally into old-growth without our intervention. Other people believe that, through management, we can imitate some kinds of natural disturbance and create stands that function as old growth.

Much of our effort in the AMA will be concentrated near existing late-successional forest, with the hope that the plants and animals living in them will move into the surrounding younger stands as they develop. How many of these species are necessary for a functioning old-growth ecosystem, and how readily or rapidly they move between stands, are unknown. We do know, for example, that several species of mycorrhizal fungi associated with tree roots are essential to tree nutrient transfer and tree development. What if a forest can't develop into old growth without some of them? We don't know what the roles of different fungal species are, nor of countless microorganisms, snails, insects, spiders, lichens, and so on--many of which have never been identified. Nevertheless, we believe that applying what we do know about old-growth ecosystems by testing changes in structure and composition of maturing stands is the likeliest approach to reaching the goal of sustainable old-growth forests in the Northwest.

Some research needs--for example, studies of spotted owl feeding habits and monitoring approaches for "survey and manage" species--will be addressed through regional efforts (although sampling may be done within our AMA). Other learning needs can be met best through management experiments within the AMA itself, research directly linked to management questions, and independent research. The following list of general questions is meant to introduce some major issues; more specific items are contained later in this document, in watershed analyses, and in the forthcoming research and learning assessment.
 
 

Late-successional forest

In the long-term, what is the most effective way to maintain late-successional forest ecosystems and provide economic value to local communities at the same time? Although forests have been grown on long (150-300 year) rotations in a few countries, we have little experience with such practices, and no one has tried to manage a functioning late-successional ecosystem. Whether appropriate tree species and animal habitats can be maintained under different rates of selective harvest of trees or other uses of the forest over time is not known, nor do we know whether some kind of long rotation may be most appropriate.

What are the characteristics of late-successional forests in the different ecological zones within the AMA? We do not have a good idea of the composition and structure of old-growth forests in the AMA because of their rarity and a lack of studies. Current ecological descriptions of old growth were primarily developed from Douglas-fir forests in the Cascade Range. Even those descriptions focus on a small part of the ecosystem; we are still ignorant about the role of most insects, fungi, plants, and soil organisms in old-growth forests. Studies of existing late-successional stands, stand reconstruction from stump measurements in clearcuts, and examination of historical records would help clarify what the "desired future condition" of late-successional forests in the AMA should be.

Can desired late-successional forest characteristics be promoted, and what prescriptions would be most effective in the different ecological zones? Although general responses of coastal forests to silvicultural manipulation are known, the specific responses in terms of habitats and ecosystem function are not. For example, we would like to know the most effective stand age at which to begin density treatments, the intensity of the efforts, the number of entries, and the specific stand attributes to provide. Repeating a set of different prescriptions across the AMA would help answer these questions.
 
 

Fisheries and watersheds

Can riparian reserve boundaries and management guidelines be modified to allow more management options while meeting conservation objectives? The interim riparian reserves under the Northwest Forest Plan cover at least 75% of the federal land area in the Oregon Coast Range. Many activities, including most timber harvest, are not allowed in the reserves unless they can be clearly shown to benefit aquatic habitat conditions. Guidelines allow for modification of reserve boundaries after watershed analysis, but the effects of many different management activities on stream habitats are not well known.

What is the most effective method for increasing structure in streams for fish habitat? Large conifer logs are the longest lasting and best providers of structure in streams, but most streams in the AMA are dominated by hardwoods. Several projects to reintroduce a component of conifers in riparian zones have been initiated, but we don't know how successful they will be in the long term, whether focusing our efforts on tributary headwalls or stream-sides would be more effective, or whether hardwood-dependent wildlife (for example, neotropical migratory songbirds) will be negatively affected by the introduction of conifers.

What is the role of coastal estuaries and large rivers in maintaining salmon populations? Although most federal lands in the AMA are in upper portions of watersheds, fish travel through and live in large rivers and estuaries during part of their life cycle. The importance of different parts of the river network to fish survival is not well understood. Studies examining the abundance and timing of fish use of different parts of a watershed, particularly where wetlands have been restored (for example, the Salmon River estuary) would aid in determining the importance of estuaries. Such studies may also identify methods and opportunities for collaboration among landowners.
 
 

People and communities

Can the AMA provide significant economic support for local communities? Timber harvest in the AMA is likely to produce a relatively low volume and will consist primarily of thinning treatments, at least for the next decade. Other opportunities to provide income include tourism, special forest products harvest, recreation, and possibly an eventual resurgence in sport and commercial fishing. How these products could be sustainably provided to meet the needs and interests of different local communities is not clear. Involving local residents in evaluating opportunities will be critical to answering these questions.

What are the best operational techniques for achieving AMA objectives, and what skills are needed for the personnel who will carry them out? Potential activities in the AMA include surveying, marking, cutting, removing trees, planting trees and shrubs, measuring plants, sampling water-quality, and keeping records. There is much to learn about relative ecological and economic costs and benefits of different methods (for example, roads and skyline logging, fewer roads and helicopter logging, wider or narrower rights-of-way, trails and horse logging) and the availability of expertise to carry out these actions. We may wish to try different techniques for getting work done, for example, contracting the management of stands or subwatersheds.

How can the management agencies fund the landscape treatments needed? Current funding and contracting regulations limit the ability of AMA managers to "reinvest" receipts from sale of commodities, such as harvest of timber or sale of special forest products, in land and resource improvements. Greater ability is needed to fund restoration projects; a number of potential ways have been discussed, but none are implementable under current law and policy.

How does management by different landowners (federal, private, state, and tribal) affect each other's management and the function of the ecosystem as a whole? Federal lands comprise only a part of the ecosystems within the AMA. For example, streams that provide fish habitat flow through, and are affected by, many different ownerships before they reach the ocean. We need to explore how activities on different ownerships affect the behavior of the whole system and whether ways can be developed to ensure that activities on different ownerships complement, rather than work against, each other.

What can we do to promote more effective community involvement in the Northern Coast Range AMA, including both communities of interest and communities of place? The direction for management of the AMAs provides a clear mandate to increase agency interaction and collaboration with persons and organizations in local communities. Indeed, the social objectives of the AMA cannot be met without meaningful contacts with a variety of people and groups. The potential benefits of such involvement are significant:

What could be done to increase the educational opportunities relating to natural resources in the AMA? One of the principal goals of the Adaptive Management Area network is, not only to increase the rate of learning from management activities, but to explore new and more effective ways of sharing that information--not just with other agency offices, but with local schools, colleges, and communities. We need to identify and develop better-structured ways of initiating and maintaining connections with local educational institutions.

What can we do to creat more local employment opportunities? In order for the AMA to fully meet the goal of increasing economic support to local communities, ways must be found and developed to direct a greater share of employment to local companies and workers. A variety of tools to accomplish this will be needed, as well as possibly some changes in existing agency procurement regulations.
 
 

How Do We Go about Learning?

Different kinds and intensities of efforts can lead to learning opportunities to further adaptive management (Bormann et al. in preparation). At one end of the spectrum is traditional research, where the importance of a few factors is intensively studied and confounding variation is controlled. At the other end is the application of single management prescriptions that are well documented and monitored to detect future change. In between are research linked to management questions, and management experiments that have several attributes of experimental design (such as comparison, randomization, and replication). Most activities in our AMA will probably fall under these in-between categories, which--in practice--may be difficult to distinguish because most activities will involve collaboration of managers and scientists to some extent. Scientists will help managers to define learning objectives, review experimental designs, predict outcomes, and design systems for long-term record keeping, monitoring, and evaluation. The major distinction, however, is that scientists will have primary responsibility for the linked research and managers for the management experiments. Independent research projects will continue to provide valuable new information about Coast Range ecosystems.

Applying some basic principles of experimental design to management projects will maximize learning for a given amount of effort and minimize confounding factors that throw results into question. Study designs that ensure learning are based on comparisons, randomization, and replication. By applying solid learning designs, we will arrive at the truth more quickly than by trial and error, and be able to predict more accurately the benefits and outcomes of future activities.
 
 

Structured management experiments

Comparison: Let's say for example that we do some type of thinning (a "treatment") in a forest to create old-growth habitat; we come back in 10 or 20 years and find that some older-forest features are present, but others are missing. The problem is, we don't know if the change in the forest was caused by the cutting we did, or by the forest growing older on its own. And we never will know, unless we have something to compare it with. In a thinning example, the comparison might be a nearby similar patch of forest that we leave in its present condition. It doesn't have to be a place where you do nothing--it could be a place where you do something the old way to compare with doing something the new way; it depends on what you want to learn. A solid design consists of different treatments (including controls) applied to similar pieces of ground (or stream, or road, or watershed). Moreover, not every test needs to achieve the intended result--our greatest learning often occurs from legitimate treatments that fail. If treatments don't differ much, we usually don't learn much from comparing them.

Randomization: Another important design element is random assignment of treatments on the ground, which protects a study from bias in selecting sites from among the unavoidable variability in nature. In a thinning project with two treatments, for example, we might find several nearby stands (or parts of one stand) that appear "similar". However, there will always be some differences in tree density or size, aspect and slope steepness, or shrub density (the fewer the better!). When we randomly assign treatments to different pieces of forest, we ensure that we won't consciously or unconsciously bias a particular treatment towards a particular kind of site.

Replication: The third element of experimental design is replication. That is, we ideally want to do several applications of each kind of activity. The more examples we have of each kind of treatment, the more confident we can be that the results we see from an activity aren't caused by something unusual about a particular site--for example, the wind exposure, soil composition, or growth history of a particular patch of forest. Replication and randomization are especially important if we know that our sites are not very similar to begin with.

The design concepts of controls, randomization, and replication can be applied at any scale, from trees to watersheds. For example, if we want to look at ways of developing large limbs on trees for marbled murrelet nest platforms, we could compare different ways of removing the top or pruning some limbs to encouraging the growth of other limbs on individual trees. Or, if we want to look at ways of partial cutting to increase forest diversity, we could compare different thinning rates on 40-acre patches of trees. If we want to look at ways of increasing fish populations, we could compare different management approaches relating to road densities, harvest rates, or stream-side plantings in nearby watersheds. The treatments don't need to be simple to provide valid comparisons, however: for example, some thinning treatments could have a range of densities or different sizes of patchy openings in them.
 
 

Keeping records

Probably the most important thing we can do, regardless of what kind of activities we pursue, is to keep good records. This may sound too obvious to bother mentioning, but it takes real dedication to keep activities well documented. Decisions or observations that were vivid in the woods easily turn into hazy memories in the office four months later. Even records that are well organized and carefully stored may not make sense to someone who looks them up ten years later. The best approach is to decide what we're going to record, how we're going to measure it, and how we're going to store that information even before we actually do anything. This process becomes a key part of our monitoring plan. As in all activities, we need to keep the process simple. When we undertake a project, we should ask ourselves: "If I were someone else coming back to this project 20 or 30 years from now, what would I want to know to evaluate the effect of this action?"
 
 

Following through

Once we decide to learn something, we will have to follow through with treatments and measurements long enough to answer our question. "Policies to learn must persist for times of biological significance, and they must affect human action on the scale of ecosystems" (Lee 1991, p. 161). Many of the changes we hope to see in late-successional habitats and healthy watersheds might not show up for several decades. A commitment to learning and to repeated monitoring will allow us to evaluate how we're doing. Even if we do not have a lot of resources (i.e., personnel and funds) to measure changes in the forest now, setting up valid comparisons--by using thoughtful design and carefully documenting what we did and why--will provide future managers with greater options to learn about future pressing issues.
 
 

Non-federal lands

Many of the answers we're looking for about management of forests and watersheds are needed by more than just managers of federal lands. Collaboration among federal, state, tribal, and private landowners could provide us with better answers sooner. Some treatments that are presently incompatible with federal land designations could be tested on private lands. Conversely, other landowners may wish to try something on federal lands that they do not have the resources or leeway to do on their own lands. Cooperation between landowners may also allow us to schedule activities so that we reduce needs for road maintenance and enhance conditions across watersheds. We may also learn much about ecosystem response to management over large areas by comparing management of forests on state, private, tribal, and federal lands.
 
 

What Are We Doing to Learn So Far?

Research has been conducted in parts of the AMA for many decades. The Cascade Head Experimental Forest (managed by the Forest Service), north of Lincoln City, was designated in 1934. Early studies described the natural history, growth, and yield of native trees, and many of the original permanent plots are still being measured. Studies in the 1950s and 1960s included a variety of harvest techniques, including staggered-setting clearcuts, progressive strip clearcuts, shelterwood, and several types of thinning. Current activities include studies of forest development, ecology of wildlife and plant species, nutrient cycling, salt marsh restoration, and applied forestry studies of harvest techniques and various species mixtures (Greene and Blinn 1991). Two state-owned experimental forests also lie within the AMA watershed boundary. Some of the earliest research on forest thinning techniques was done at Black Rock Forest Management Research Area (managed by the Oregon Department of Forestry), just west of Falls City. Forestry research is also being done at McDonald-Dunn Forest (managed by the College of Forestry, Oregon State University), near Corvallis. Several Research Natural Areas (on both Forest Service and BLM-administered lands) also exist in the AMA (see appendix C); RNAs are mostly pristine, unmanaged areas, which provide opportunities for a variety of studies of natural ecosystems.

Many unplanned "experiments" are already in place; for example, adjacent forests that we know were the same once but have been managed differently, or places where managers have tried innovative plantings, thinnings, or road construction methods. These places can provide valuable information while we wait for more definitive results from newly installed studies. Some opportunities for retrospective studies have already been identified (Thomas et al. 1993); others no doubt could be found if we were to look carefully at current conditions. The sooner such areas are identified, the better the opportunities we will have to capture known information regarding how they came to develop the way they did.

Several research projects associated with the Coastal Oregon Productivity Enhancement Program (COPE) are being conducted on a variety of ownerships in the AMA, including studies on forest thinning and tree underplanting, changes in riparian forests, road-building techniques, and fish and wildlife abundance and habitat. The Coastal Landscape Analysis and Modeling Study (CLAMS) has an interdisciplinary team of scientists who are developing tools to understand changes over large landscapes and analyzing the economic and ecological consequences of different forest policies at the scale of the entire Coast Range. Scientists contributing to these studies work for many organizations, including the Biological Services Division of the USGS (formerly National Biological Service), the Forest Service's PNW Research Station, Oregon State University, and the University of Oregon.

Three sites in our AMA are being used for the Density Management Study conducted by the Biological Services Division of the USGS and the BLM. This study is evaluating several thinning methods in mid-aged and mature stands (40 to 70 years old) to achieve late-successional stand characteristics as quickly as possible, and the effects of those methods on wildlife (small mammals, amphibians, fish), understory plant species, and microclimate. Installations at three different sites (Callahan Creek, Sand Lake, and Elk Creek) will consist of four basic treatments:

Different types of unthinned buffers along streams are also planned.

The goal of the Adaptive Management Pilot Project being conducted by the Siuslaw National Forest and the Pacific Northwest Research Station is to continue developing the theory of adaptive management by implementing the concepts in local projects such as the the Hebo Restoration Study and designing other large-scale management experiments. The Hebo Restoration Study is an attempt to accelerate development of late-successional stand characteristics in a 1910 plantation of nonlocal tree stock in which tree growth appears to be reduced compared to natural stands. We do not know to what extent these poorly adapted trees will ever develop effective late-successional structure, so the idea of "starting over" with locally adapted tree stocks is being compared with various ways of managing the existing trees and adding components of local stocks. Soil productivity on the Hebo plantation may also be degraded from past fires. The four treatments, replicated three times each, consist of:

Treatment effects on forest structure, woody debris, soil nutrients, microbial weathering of soil, and soil invertebrates will be evaluated.

Learning is also supported by managers in the AMA. A great deal of effort is going into watershed analysis, which provides valuable site-specific information that will help improve future decisions. Watershed analysis is used to quantify past and current ecosystem conditions, characterize the natural range of ecosystem conditions in the area, identify management opportunities and restoration strategies that will sustain long-term productivity and meet societal needs, and designate potential indicators for monitoring. Some watershed analyses have been completed, and several others are in various stages of completion (see Map 4 and the  discussion in Chapter 5). Several implementation and monitoring projects are under way in the AMA, using various methods of placing logs in streams and planting conifers in streamside areas. Several timber sales that include alternative stand treatments are also planned (see chapter 5). These projects focus on different degrees of thinning, different types of variable thinning, and planting with different tree species on a variety of forest types around the AMA.
 
 

What Do We Need to Do next?
 
We are still conducting several assessments and will continue to consult with interested partners in the AMA effort. This information will help us set priorities for what we want to learn and how we're going to learn it. We expect that these priorities will continue to shift as we learn. In the short term, the following issues and approaches will guide much of the management in the AMA. Most of the implementation and basic monitoring will be carried out by land managers and their staffs, but participation by other interested persons and groups will be encouraged. Researchers will be able to assist with design of some projects and will help to provide detailed information on specific issues; other projects may be initiated to address specific research questions. Although forest, stream, social, and landscape issues are addressed separately below, actual prescriptions will probably address a combination of issues. Although many of the activities described in this Guide focus on the effects of tree harvest, current and future activities will likely examine the effects of exotic species, insect outbreaks, windthrow, fire, floods, landslides, and other influences on late-successional forests and streams.
 
 

Development of late-successional habitat

Some people believe that most forest stands will reach the old-growth stage on their own, without intervention or disturbance (Oliver 1981). There are many examples of old-growth forests that developed at low density or with repeated disturbance (for example, wind and fire), however, which suggest that we may be able to accelerate development of densely stocked stands by treating them in ways that mimic natural disturbances, through thinning, underplanting, and other management activities. Densely stocked plantations might not develop optimum old-growth characteristics on their own, however, and repeated thinning of older stands may provide both timber output and improved late-successional habitat.

When designing thinning projects, we should consider that commercial thinning differs from natural tree mortality caused by competition, windthrow, insects, or fungi because wood is removed from the forest, trees are taken all at once, and stumps are left in the ground (compared to the root wads and churned soil created by windthrow). These differences may be important to some animal and plant species. The density of trees remaining after thinning affects their growth as well as the density, growth, and composition of understory trees and other plants. The same total number of trees in a thinning unit can also be arranged uniformly or in variable density, which may also affect growth of overstory and understory vegetation. Although spatial variation is a characteristic of late-successional forest, uniform thinning in younger stands may create more large trees for future timber or woody debris; variation could be produced by later thinning or allowed to develop naturally. These issues could be addressed with the following types of management projects:

Because forests vary substantially across the AMA (in climate, composition, slope position, and so on) the same thinning study should be repeated in a wide range of forest types across the AMA. Although general vegetation response would be monitored in all projects, selected projects should also monitor the response of a range of species, including understory plants, small mammals, birds, and fungi. Map 4 displays major vegetation classes in the AMA.
 
 

Restoration of aquatic habitats

The focus of most aquatic restoration strategies is to reduce or eliminate effects from timber harvest and roads (fine sediment erosion, channel constraints, landslides) and to place or promote input of large conifer logs into stream channels. Trees fall into streams naturally from stream banks and from landslides in unstable headwalls. Closure or modification of roads may have short-term impacts on streams as well as long-term benefits. Thinning to promote conifer growth, or conifer planting in hardwood areas (with or without thinning), may also affect stream temperature, erosion, and riparian habitat in the short term. How much buffering from different types of management, some of which may be beneficial, is needed in riparian zones? What is the relative importance of riparian habitat in smaller and larger streams within a watershed? These issues could be addressed with the following types of management projects:

Similar to the forest topics above, the same project should be repeated in a wide range of streams across the AMA. Although the response of fish habitat (or stream structure) would be monitored in all projects, selected projects should also monitor fish abundance, physical factors (water temperature, sediment movement), the response of a range of species (stream invertebrates, riparian vegetation), and the use of riparian buffers as corridors for movement of terrestrial species.
 
 

People and communities

The range of values that could be provided to people while also providing late-successional habitat is not known. Many of the questions build on the issues raised above. For example, different thinning intensities not only provide different amounts of harvested wood, but may affect abundance and growth of plants that provide special forest products, condition of fish habitat, and suitability for recreation. Even activities like hiking or recreational driving, generally thought to be relatively benign, may lead to harmful effects on some species and habitats. The ways economic use is organized also have important implications to people, including the types of regulation, harvest, manufacture, and distribution. The techniques that are adopted will depend greatly on the interests, values, and abilities of the people participating. These issues could be addressed with the following types of management projects:

Although landowner costs and the amount and value of harvested products will be monitored in all projects, selected projects or areas should also monitor people's movements, activities, and perceptions about the AMA.
 
 

Landscape design

Many of the issues surrounding late-successional forest habitat and fisheries must be addressed at a larger scale than we have focused on in the past. For example, the effect of the spatial distribution of different forest ages on wildlife, and the effect of road density and placement on stream flows and fish habitat are issues that must be addressed at the watershed or river-basin scale. An adaptive management approach for these issues might attempt different types of management strategies on different watersheds (or groups of watersheds).

To assign different strategies to different geographic units--such as watersheds--a landscape plan will be needed. This plan will need to be based, not only on experimental design considerations, but also on existing resource conditions, including past impacts, current management allocations, and sensitivity of the resources present in each area.

An important step to developing a landscape plan will be to set priorities for what we would like to learn by treating the landscape in different ways. For example, most ideas about the role of people in conservation of natural resources fall along a spectrum between active landscape management, which proposes that most of the landscape should be managed (often with modified kinds of long rotations and intensive efforts for specific objectives), and reserve-based management, which maintains that large portions of the landscape should be allowed to develop without intervention, usually with most commodity-based activities excluded. We might want to choose two or more different approaches along the active-passive spectrum and try them out on different landscape blocks. Interested citizens would help design some of the different strategies. Instead of being separate efforts, projects to address stand-scale questions could be nested within particular landscape strategies.

Because changes over large landscape areas take place gradually, evaluating changes in watershed condition will require consistent strategies over periods of several decades or centuries. It would be useful to identify those areas in which future management might change to test newly identified strategies, and those areas in which strategies would be continued over a century or more.
 
 

Data management

The different landowners in the AMA, as well as state and local governments, already collect a great deal of information about resources and people's activities within the AMA. However, the objectives for collecting the information are often not clear, the standards used by different people are different, and the information often is not easily retrieved or analyzed. To set adaptive management on a sure footing, it will be important to:

Efforts to standardize data collection techniques among federal agencies are being pursued at regional and national scales, but we will also encourage our partners in the AMA to work with us on ways of translating existing information from different inventories or of adding measurements to existing inventories so that the comparable information can be examined across the AMA.
 
 

Research database and information map

Sometimes we aren't aware of things that have already been learned that could help us out. The AMA Research and Learning Assessment will assess past and current research studies in the AMA and similar portions of the Coast Range Province. Objectives, methodologies, and results of different studies will be summarized into a readily accessible and searchable data-base. Locations of study sites and raw data will be documented. The locations of research, monitoring, and retrospective sites could be compiled into GIS data layers readily available to all.
 
 

Research Coordination and Logistics

The lead scientist for the AMA is currently responsible for keeping track of existing learning projects in the AMA and helping coordinate the work of other scientists there. Information from research projects will be centrally stored for easy access and sharing among the partners in the AMA. The lead scientist also helps design management and monitoring activities to address key questions and aids in procuring funding for research activities in the AMA. An ad-hoc group of agency and university scientists provides advice and assistance in these efforts. Current funding for research is not sufficient to tackle all of the questions identified in a short period of time, but it is expected that researchers with external sources of funding will find opportunities to use proposed or existing projects within the AMA as settings in which to conduct their studies.

 Approval of proposed studies is ultimately the land manager's responsibility. Many projects will be developed through the collaboration of managers, scientists, and the public. Other projects, especially those requiring coordination of resources and selection of specific sites in the forest, should be submitted to the AMA management team for evaluation and approval well in advance of proposed activities. Research studies planned specifically for Cascade Head Experimental Forest and Scenic Research Area must be approved by the Pacific Northwest Research Station Director.