Ecosystem management
 Conservation biology and reserve-based management
 Late-successional and old-growth forest
 Aquatic ecosystems, salmon and riparian areas
 Social and economic environment
 Home

A primary focus of ecosystem management is managing to maintain the characteristics and processes of entire ecosystems that sustain component species, and human uses, over the long term (Grumbine 1994, Christensen et al. 1996). "Ecosystem management must include the following:

1. long-term sustainability as a fundamental value,

2. clear, operational goals,

3. sound ecological models and understanding,

4. understanding complexity and interconnectedness,

5. recognition of the dynamic character of ecosystems,

6. attention to context and scale,

7. acknowledgment of humans as ecosystem components, and

8. commitment to adaptability and accountability." (Christensen et al. 1996)
 

The attention to long time spans and multiple spatial scales requires management at larger spatial and temporal scales than has been common in the past: "Coordinated management at the landscape level, including across ownerships, is an essential component" (Society of American Foresters 1993). Since a particular ecosystem (e.g. prairie, wetland, forest) can occur in a range of different states and successional conditions, ecosystem management requires some guidance for determining the desired abundance and distribution of ecosystem types and conditions.
 

There are two general schools of thought for determining desired ecosystem conditions across landscapes (Cissel et al. 1994). In one, desired landscape patterns are determined based on those commodities and processes desired (e.g. wood fiber, habitat, water quality). This approach generally assumes that we have sufficient knowledge to produce those attributes we want. In the other approach, desired landscape patterns are based on historical patterns. This concept, which permeates much of the ecosystem management literature, recognizes that many species and processes are still unknown, so that maintaining or recreating the range of conditions that native species survived under for the last several thousand years should perpetuate them into the future. This goal of ecosystem management is often referred to as maintaining ecosystems within their "range of natural variability" (Swanson et al. 1993). The condition, abundance, and distribution of ecosystems before settlement by Euro-Americans (e.g. before ~1850 in western Oregon) is usually used as the baseline for determining the range of natural variability in North America (Wilson et al. 1991). Thus one primary question for adaptive management in the AMA is:
 

1. What is the natural range of variability of ecosystem abundance, distribution, and condition in the Oregon Coast Range?
 

Topography and climate help determine where different kinds of ecosystems occur, but disturbance (e.g. fire, wind, pests, pathogens, flood, landslide, logging) is perhaps the most important process determining the abundance and condition of ecosystem types across the landscape. Characterizing the size, intensity, and frequency of natural disturbances is a critical step for describing the range of natural variability of ecosystems. Native Americans no doubt had some effect on disturbances (e.g. fire on coastal headlands and valley savanna), but they are generally included in the range of natural variability, since their influence on native species was likely a factor for at least 10,000 years. The kinds of disturbances initiated by early Euro-American settlers (e.g. large clearing fires, mining, splash-damming, agriculture) and industrial-age people (e.g. road building, logging) are relatively new and usually qualitatively different from conditions that native species have adapted to for millennia, so these are not considered "natural" disturbances.
 

Determining the range of natural variability for the AMA will not be a simple task, in large part due to lack of evidence. Large areas of the Coast Range burned in catastrophic fires in the late 1840s (Morris 1934, Impara 1997), and those areas currently contain little evidence of conditions prior to 1850. Most of the areas that eventually became federal forest land burned in those fires (85% of USFS and 50% of BLM based on Teensma et al. 1991), and the rest of the landscape has been rather intensively managed, so there is little information left with which to reconstruct how the landscape functioned prior to settlement. Some very coarse evidence can be gained by analysis of early land surveys (Teensma et al. 1991). Analysis of fire history in areas similar to the AMA which have escaped recent catastrophic fire or heavy management, like the central Coast Range (Impara 1997), can provide valuable information. Analysis of abundance and periodicity of charcoal in lake sediments (Worona and Whitlock 1995) is also useful, although there appear to be relatively few suitable areas for such studies in the northern Coast Range (Cathy Whitlock, personal communication). Indeed, a landscape simulation model based on charcoal studies in the central Coast Range suggests that abundance of old-growth forest over the whole Oregon Coast Range varied between 38 and 60% over the last 3,000 years (Wimberly et al. submitted).
 

2. Are there effective approaches to maintaining ecosystems that do not require imitating natural dynamics?
 

It is possible that approaches to ecosystem management which do not rely exclusively on the range of natural variability can provide for native species and ecosystems. For example, disturbances of moderate intensity and frequency can enhance biodiversity compared to disturbances of high or low intensity and frequency in some ecosystems (Connell 1978). Therefore it may be possible to use carefully planned management disturbances which aren't "native" to a particular area (e.g. partial cutting similar to the effects of moderate-intensity fire in moist coastal forests) to enhance habitats for species of concern. In addition, the range of natural variability may have been a function of disturbances and succession operating at the landscape scale (e.g. large fires and large blocks of old-growth) which could not be recreated in current mixed-ownership landscapes if only a portion of the landscape will be managed to sustain native biodiversity. (Approximately 1/3 of the lands in the Oregon Coast Range are in federal ownership [USDA Forest Service 1995]). Evaluating the question would require implementing different landscape management strategies over relatively large areas (several thousand acres) for long time spans (at least 100 years) with periodic multi-species inventories and monitoring.
 

The basic approach of conservation biology as it has been applied in most settings, and in the NWFP, is to place the best remaining habitat into reserves from which most management is excluded and attempt to provide dispersal habitat between reserves, with the surrounding landscape being managed relatively intensively (Harris 1984, FEMAT 1993). Indeed, the NWFP allows the harvest of remaining stands of old-growth forest outside of reserves under the assumption that young forest areas within "late-successional reserves" will grow into high quality habitat in large blocks. This conservation strategy is intended to maintain organisms that require large areas of habitat (e.g. the northern spotted owl), as well as organisms with smaller home ranges that require the same habitat.
 

Species that are rare or occur in only a few places (e.g. Columbia torrent salamander, Rhyacotitron kezeri) may not be protected by reserves, however (FEMAT 1993). Many species of plants, lichens, arthropods, and amphibians require features of old-growth forest ecosystems like large trees and logs, but disperse slowly or over short distances. These species may not be able to disperse between large blocks of late-successional reserves and could become genetically isolated, increasing their risk of extinction. In the NWFP, retention of live trees and woody debris in harvest units in the non-reserve "matrix" lands, protection of "riparian reserves", and "survey and manage" protocols for selected species are intended to maintain species not well served by late-successional reserves, at least in the short term. Even if federal lands in the northern Oregon Coast Range and some portion of the Tillamook State Forest provide late-successional habitat within the next 100 years, they would probably be quite isolated from habitat in the rest of the region (assuming that the privately-owned forests in southern Washington and in the Highway 20 corridor will be managed on 40-60 yr rotations)--the distance between federal lands in the AMA and those to the south is at least 23 km (14 mi). Other reserve-based approaches argue that long-term maintenance of functional ecosystems requires areas large enough to allow natural disturbance and recovery processes to occur (i.e. large fires) and to support all native biota, including wide-ranging carnivores (e.g. grizzly bear, Ursus arctos, and wolves, Canis lupus) (Noss 1983). Such a proposal for the Oregon Coast Range might allocate 50% of the landscape in reserves that exclude most resource extraction and roads (Noss 1993).
 

3. Are there effective approaches to maintaining native late-successional species that do not require the NWFP reliance on reserves?
 

Other approaches to maintaining biodiversity do not require extensive systems of reserves, and instead rely on silvicultural treatments to create the appropriate distribution of ecosystem types and conditions, while allowing commodity extraction at the same time (Oliver 1992, Carey and Curtis 1996). This kind of management would shift different forest successional stages across multi-owner landscapes using a variety of long-rotation, thinning, and partial-harvest techniques which would also maintain some old-growth legacies (old trees, snags, and logs) in most stands. A multi-disciplinary modeling effort for the western Olympic Peninsula suggests that such an approach could maintain most biodiversity (including late-successional species), and could produce 82% of maximum net present value and greater long-term economic returns than short-rotation intensive management (Carey et al. 1996). The proportions of different forest age classes could be determined by a variety of sustainable harvest schedules (Henderson 1994). One potential advantage to this approach is that most portions of the landscape would pass through all successional stages and be connected over time, which could maintain populations of rare or low-mobility species which would not survive extended periods in early-successional forests. There is still uncertainty, however, about whether all organisms and ecosystem processes can be maintained in managed forest stands.
 

The immediate interest in late-successional and old-growth forest under the NWFP is for providing habitat for two species listed under the Endangered Species Act: the northern spotted owl and the marbled murrelet (Brachyramphus marmoratus). In addition, a total of 1,098 terrestrial species (not counting arthropods) are believed to be closely associated with late-successional forests on federal lands within the range of the northern spotted owl, including 527 fungi, 106 bryophytes, 157 lichens, 124 vascular plants, 102 mollusks, 18 amphibians, 38 birds, and 26 mammals (FEMAT 1993). Of those species, 9 of the amphibians, 28 of the birds, and 19 of the mammals are believed to occur in the AMA (USDA Forest Service and USDI Bureau of Land Management 1997). Many aquatic species also depend directly and indirectly on late-successional forests (see next section).
 

Definitions

As used in the Northwest Forest Plan, "late-successional forest" 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. This includes older stands in which the oldest trees reach their maximum sizes, understory trees form multiple canopy layers, and dead wood accumulates to high levels (NWFP1994, 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 older late-successional stands which meet some minimum standards of numbers of large trees, range of tree sizes, occurrence of multiple canopy layers, and abundance of snags and logs (Table 1). More recent interim Forest Service guidelines for the western hemlock zone are similar, but the minimum diameters for large trees vary by site class (a measure of an area's productivity) and stands must be at least 200 years in age (USDA Forest Service 1992). The Bureau of Land Management classifies stands greater than 200 years of age as old-growth. 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 definitions. Some definitions distinguish between unmanaged, virgin "old-growth" stands and "fully functional" stands developed through silviculture that would provide some or most (depending on assumptions) of the same ecosystem properties (FEMAT 1993, Carey et al. 1996).
 

Table 1: Minimum structural criteria for Douglas-fir old-growth forest in the western hemlock zone (Old Growth Definition Task Group 1986):

• 20 Douglas-fir per ha (8/ac) >80 cm (32") DBH¹ or >200 yrs old

• 30 or more shade-tolerant trees per ha (12/ac) >40 cm (16") DBH (includes hemlock, cedar, grand fir, silver fir, and big-leaf maple)

• deep, multilayered canopy

• 10 or more conifer snags per ha (4/ac) >50 cm (20") DBH and > 4.5 m (15') tall

• 33.6 or more metric tons per ha (15 tons/ac) of logs, with 10 or more pieces per ha (4/ac) >60 cm (24") in diameter and >15 m (50') long

Abundance

There is much less old-growth forest in the Oregon Coast Range now than there was prior to Euro-American settlement. In 1850, approximately 40% of the Coast Range between Astoria and Reedsport consisted of forests more than 200 yrs old, and 35% of the area had recently burned (Teensma et al. 1991); applying the proportion of 200 yr old forest in the unburned area to the entire area suggests that forests more than 200 yrs old may have covered about 60% of the area before the large fires that burned in the late 1840s (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, USDA Forest Service 1995). As mentioned on page 11, most of the lands that came into federal ownership in the AMA were burned in the 1850's (the lands of the other large public landowner in the area, the State of Oregon, were burned in the Tillamook fires after 1933). Most of the old-growth forest left after the fires was logged over the last several decades.

Current estimates of older forest vary by method and geographic area: 33,800 acres of old-growth on (or 5.4% of) the Siuslaw National Forest (Bolsinger and Waddell 1993); 200 acres of forest greater than 200 yrs old in (or 0.8% of) the Nestucca Watershed (USDA Forest Service and USDI Bureau of Land Management 1994a); and 140,500 acres of multi-storied forest of trees greater than 53 cm (21 inches) diameter on (or 10.0% of) federal lands in the Coast Range (FEMAT 1993) (federal lands occupy about 1/3 of the land base). The rapid change in forest composition that occurred in the Pacific Northwest in the last century is illustrated by the fact that timber cruisers for King County, Washington in 1902 considered trees less than 70 cm (28 inches) diameter "unmerchantable" and didn't count them (Estella Leopold, personal communication). Classification of satellite imagery suggests that closed-canopy forest and larger trees are more abundant on federal lands than on other ownerships in the northern Coast Range (Figure 5). Forest Service and BLM inventories suggest that about one third of the federal lands in the AMA are occupied by forests over 80 years old (Figure 6-older forest types were not classified).

Restoration

Old-growth ecosystems are commonly believed to be a predictable stage of forest development (Oliver 1980, Franklin et al. 1981). Thus it is possible that the passage of time alone, without intervention by management or major disturbance, is sufficient for the development of old-growth conditions from younger forest stands (Figure 7). Some people caution, however, that current old-growth stands originated and developed over the last several centuries, that we know little about how they developed, and that we may not be able to reproduce those processes under current climate, land-ownership patterns, impacts from exotic plants and animals, and disturbance regimes (FEMAT IV:31-32). Other people point out that most of the younger (i.e. 20-80 yr-old) forests in the region today are relatively dense stands of slow-growing trees that are developing on a different trajectory from that of many old-growth stands, which appear to have developed at low densities (Tappeiner et al. 1997). Many existing old-growth stands did not develop without disturbance, but experienced significant tree mortality from fire, wind, or flood (Spies and Franklin 1988, Morrison and Swanson 1990). It is possible that by reducing stand density through harvest (i.e. "thinning"), trees in such stands could develop the size and structure of old-growth forest more readily (Newton and Cole 1987, Oliver 1992, Bailey 1996). Combined with other treatments, such as killing trees to create standing snags or down logs, damaging tree tops to create cavities, introducing wood-rotting fungi, and distributing tree harvest patchily within stands, harvest to reduce tree density could provide habitat for many late-successional species earlier than if stands were left alone (Hansen et al. 1991, McComb et al. 1993, Carey et al. 1996). Regardless of whether harvest is required in young stands to ensure development of late-successional communities, there is interest in the potential of thinning harvests to accelerate forest development and produce economic returns from harvested timber.
 
 

Figure 7: The natural processes by which young forests with relatively simple structure
and composition (left panel) develop into old-growth forests with complex structure and
composition (right panel), and the ability of vegetation management to accelerate those
processes, are not well known.
 
 
 
 
 
 
 

4. What are the advantages and disadvantages of different methods of restoring and maintaining late-successional forest ecosystems?
 

We do not know how to restore functioning late-successional forest ecosystems from a landscape dominated by young stands--it's never been tried. In addition to the uncertainty about how current old-growth forests developed, and incomplete knowledge of the species and ecosystem functions found within them, we are not sure how forests will develop following different kinds of management. Although general responses of coastal forests to thinning are known, especially growth of overstory trees in response to commercial thinning (Drew and Flewelling 1979, Curtis and Marshall 1993, Bailey 1996), long-term responses of species habitat and ecosystem function are not. For example, development of multi-storied canopy structure in mature forests is not a guaranteed process because of limited seed dispersal and particular microsite requirements for establishment of shade-tolerant conifers like western hemlock (Gray and Spies 1997, Schrader 1998). There is general agreement on the primary goals of thinning to promote late-successional structure: large overstory trees in an uneven, patchy distribution, and large woody debris (logs and snags). There are many different silvicultural approaches that could be taken to attain these goals, and the ramifications of each are unclear. Several interacting variables of interest include:

• number of harvest entries into a stand (and implications for maintaining a road network),

• tree density remaining after harvest (and consequences for blowdown),

• spatial heterogeneity of harvest (uniform vs. patchy),

• amount of wood of different sizes that can be removed without compromising habitat goals for coarse woody debris, and

• treatments to enhance decadence, including topping, girdling, felling and leaving, and introducing rot fungi into trees.
 

A fundamental assumption behind these stand-level manipulations is that plants, animals, fungi, and microorganisms characteristic of late-successional forests will arrive and prosper as the treated younger stands develop. Nevertheless, we know little about the roles of most of these different species, how many are necessary for a functioning old-growth ecosystem, or how readily or rapidly they move between stands.

   Return to top of page
 

The primary interest in aquatic and riparian ecosystems under the NWFP is the need to restore habitat for fish populations, particularly anadromous salmonids (Oncorhynchus spp.). These ecosystems are important habitat for many other species as well, including mollusks, lichens, mosses, vascular plants, amphibians, and mammals (FEMAT 1993).
 

Characteristics

Aquatic ecosystems include the waters of intermittent and perennial streams, rivers, and lakes, and the organisms living in them. Riparian zones are the interface between terrestrial and aquatic ecosystems, and include the land areas that affect and are affected by aquatic ecosystems (e.g. forests casting shade on streams, and wetlands).
 

Aquatic and riparian ecosystems in forested mountain landscapes are very dynamic on time scales from days to centuries. Habitat structure in streams is primarily determined by floods interacting with sediment from landslides, and by live vegetation and large woody debris from forests on headwalls, streamsides, and floodplains (Gregory et al. 1991, Sedell and Beschta 1991). Large woody debris is a key structural element that helps retain sediments (e.g. cobbles and gravel), and creates--by diverting flood waters and sediments--complex habitat features (pools and side channels) in streams and floodplains, provides cover, and is also occupied by organisms. Wood may fall directly into streams from riparian forests, or be transported by landslides from smaller (non-fish-bearing) tributaries or hillslopes. The abundance of sediment and large wood in a given stream is linked to forest successional stages throughout a watershed (Benda 1994, Reeves et al. 1995). Under the pre-Euro-American disturbance regime, infrequent wildfires occurred over large areas, killing many old and large trees. Landslides were common on recent burns, adding large amounts of sediment and wood to streams (Swanson et al. 1987), and often destroying and simplifying fish habitat. As hillslopes reforested and floods continued to move sediments downstream, complex stream habitat would form, with gravel-rich conditions possibly persisting between 50-300 years after wildfire. After long periods without disturbance, most of the sediments could be flushed downstream, leaving poor, bedrock-dominated habitats. Because wildfires were apparently patchy in time and space, however, there would always be a large number of watersheds with good habitat conditions at any one time (Reeves et al. 1995). In addition to providing and retaining coarse woody debris, riparian forests and streamside vegetation affect physical habitat by shading streams (affecting light levels and stream temperatures). They also greatly influence aquatic foodwebs and water quality via inputs of organic litter and regulation of inputs of inorganic nutrients and fine sediments (Gregory et al. 1991, Sedell and Beschta 1991).
 

Although young salmon feed on organisms in mountain stream foodwebs before heading out to sea, returning adult salmon can be a major food and nutrient source for riparian and forest ecosystems; indeed, it may be the only way (other than geologic uplift) that nutrients move from the ocean to the mountains. Where runs in the Northwest are still healthy, live and dead salmon provide nutrition for a wide array of animals (including mammals, birds, inverterbrates, and even young salmon) and stream-side plants. For example, the amount of ocean-derived nitrogen in different aquatic and riparian organisms in healthy salmon spawning streams can range between 17-100% (Kline et al. 1990, Bilby et al. 1996).
 

Current conditions

Salmon populations in the Oregon Coast Range are a fraction of what they were before Euro-American settlers arrived about 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). Coastwide, chinook (O. tshawytscha) are at 50-70% of the estimated 3-600,000 adults present in 1900, and coho (O. kisutch) from Cape Blanco north to the Columbia are at 4% of the estimated 1.7 million adults present in 1900 (Kostow 1997). Accounts from the 1800s of Oregon rivers "choked" with salmon are common. Currently, 32 salmonid stocks in the AMA are considered at risk of extinction or of concern (from FEMAT 1993).
 

The causes of salmon population declines include over-harvesting in commercial and recreational fisheries, migratory impediments such as dams, and loss of genetic integrity due to the effects of hatchery practices and introduction of non-local stocks, but the most common cause is the loss and degradation of freshwater and estuarine habitats (Nehlsen et al. 1991, FEMAT 1993). Habitat impacts arise from agriculture, timber harvest and associated activities, road construction, livestock grazing, water withdrawal and diversion, and dams. These serious human impacts exist on top of natural factors like predation by marine mammals and cycles in marine food supplies and ocean currents.
 

Logging and road construction have degraded mountain streams in the Northwest and the AMA by destabilizing steep slopes, causing repeated landslides through time (Swanson et al. 1982, Swanson et al. 1987, USDA Forest Service and USDI Bureau of Land Management 1997). Often the areas that slide have no trees or small trees, and streams are already low in wood due to salvage and "stream cleaning" activities from the 1960s and 70s (Sedell and Beschta 1991), so there are few large logs available to retain sediments and structure fish habitat. Forests along streams and many headwalls are composed primarily of hardwoods, which do not provide long-lasting woody debris, as most conifers do. Roads in particular deliver large quantities of sediment to streams via landsliding, surface erosion, and stream channel diversions (Beschta 1978). Because harvest and road-building activities have been dispersed throughout the landscape, sediment input is a chronic disturbance and there is no period for any one stream to recover, unlike conditions under episodic wildfire disturbances (Swanson et al. 1982, Reeves et al. 1995). Fine sediment can reduce survival of fish eggs and developing alevins, decrease ecosystem productivity and food availability, and fill in pools (FEMAT 1993). The low frequency of pools and woody debris has been documented for Coast Range and AMA streams (USDA Forest Service 1995). Paradoxically, most Coast Range streams are sediment-poor and bedrock-constrained, probably because although recent management has triggered many landslides, there was little wood present to trap it in the system and it flushed out of the drainages (Gordon Reeves and Gordon Grant, personal communication). Increased sedimentation rates in recent decades has been documented for Devils Lake on the central coast (Eilers et al. 1996).
 

Many salmon species spend extended periods of time in the stream network before heading out to sea, and large rivers and estuaries may be particularly important habitat. Yet there are relatively few estuaries on the Oregon Coast compared to the rest of the region (FEMAT 1993), and most of the land along estuaries in the AMA are on private property and have been drained, diked, channelized, or heavily grazed, providing little habitat for fish; some are also impacted by nutrient pollution from agricultural runoff.

Restoration

Several options have been proposed for restoring aquatic and riparian ecosystems to sustain fish and other species; some are short-term fixes, some are long-term, and additional information is needed on implementing and choosing amongst them. The initial emphasis of restoration is reducing or eliminating activities that damage stream habitat within areas that currently contain the best habitat and healthiest fish runs ("key watersheds" in the NWFP), and working to restore watersheds that are of moderate quality (i.e. those most likely to respond to treatment, and that will be available to replace key watersheds degraded by future catastrophic disturbance) (Reeves et al. 1995, Reeves and Sedell 1992). Several aquatic assessments have identified those watersheds in the AMA (Figure 8).

Figure 8

The emphasis for reducing existing and ongoing damage to aquatic ecosystems is on roads. The need is to identify the specific causes of road-related runoff and sediment production and then prevent them by either decommissioning or upgrading the road (e.g. outsloping roadbed, placing larger culverts, removing unstable fills) (FEMAT 1993 Appendix V-J). In watersheds with relatively few roads and relatively good habitat, small expenditures to upgrade and remove roads can reap large benefits (Harr and Nichols 1993). Removing or moving roads and structures (e.g. dikes) that constrain streams may be crucial to restoring valley-floor features like logjams and side channels (Sedell and Beschta 1991).
 

5. What are the most effective approaches for increasing fish habitat in streams?
 

The most important elements of watershed restoration are restoration of instream habitat complexity and restoration of riparian vegetation condition. Short-term habitat improvement measures include placing wood and other structures in streams to create pools and side channels. However, this work is expensive and only a limited number of stream segments can be treated in this way. The most effective, long-term solution to creating and maintaining aquatic habitat is thought to be restoring vegetation in riparian zones and headwalls (Sedell and Beschta 1991, FEMAT 1993), particularly by introducing conifers where they are deficient and accelerating the growth of conifers where they occur. Rather than establishing fixed standards as benchmarks for stream and forest conditions (e.g. a fixed number of pools or large conifers per kilometer), many feel that we should model forest management on natural dynamics of disturbance and succession in order to maintain a diversity of habitat types (e.g. for spawning, rearing, and overwintering) and functional food webs and aquatic ecosystems (Reeves et al. 1995, Bisson et al. 1997). While current management is a series of chronic low to moderate disturbances dispersed throughout the landscape, natural disturbance tends to be short pulses of intense disturbance followed by long periods of recovery. Planning on watersheds of 4th-6th order may be necessary to match goals to salmon life cycles (e.g. maintaining summer and winter habitats and migration corridors). The most promising area for future land management will be integration of upslope and riparian management at basin and landscape scales (Gregory 1997).
 

6. Can riparian reserve boundaries and guidelines be modified and still meet conservation objectives?
 

The interim riparian reserve designation in the NWFP applies to 75% or more of the federal land area in the Oregon Coast Range (USDA Forest Service 1995). That level of reserve designation may not be necessary, depending on upslope management and the specific issue in question. For example, much of the coarse woody debris found in fish-bearing portions of streams may derive from relatively few headwalls in a drainage, because landslides on some tributaries may never reach those portions, and landslides on lower portions where streams are more powerful may be flushed out of the drainage without contributing to habitat (Gordon Reeves, personal communication). Evaluation of riparian reserve boundaries should probably focus on their specific intended goals and consider their intended role as habitat for terrestrial species as well. A related issue is that silvicultural activities are allowed in young stands in riparian zones if they are believed to accelerate riparian restoration. However, agency personnel are often concerned about potential short-term damage from increasing bank-side erosion and light levels into streams. While light and temperature effects from riparian silviculture can (and are being) readily evaluated, characterizing fine sediment impacts of stand-level management is notoriously difficult and unreliable (Hall et al. 1987, Gordon Grant, personal communication).
 

One troubling aspect for restoration in the Coast Range and AMA is that the natural abundance of conifers along streams is not clear (Nierenberg 1996). Getting conifers to grow in stands dominated by hardwoods and Rubus spectabilis (salmonberry) can be difficult, and may naturally occur only rarely. Many of the riparian hardwood stands in the AMA are close to the maximum life span for Alnus rubra (red alder), and may be replaced by salmonberry brush fields and (eventually?) by conifers; either scenario could seriously impact rare or declining hardwood-associated terrestrial species like the white-footed vole (Phenacomys albipus) and several migratory songbird species (USDA Forest Service and USDI Bureau of Land Management 1997). Maintaining hardwoods could also be important for the aquatic ecosystem, since a mixture of conifer and hardwood litter is optimal for the aquatic food chain.

The primary social and economic interest in AMAs under the NWFP is experimentation with policies and management to become "prototypes of how forest communities might be sustained" (NWFP D-4). This goal is to be reached collaboratively, by providing opportunities for all interested parties to work together to develop innovative approaches. Most policy direction for federal agencies comes from national or regional bureaucrats and elected officials. One goal of adaptive management is that interested parties have a more direct influence on local policies by helping define issues and goals and evaluate success at achieving them (FEMAT 1993, Bormann et al. 1994, Stankey and Shindler 1997). The relevant issues for this topic are public involvement, sustainable economic systems, and management systems.
 

Characteristics and current conditions

There are many challenges, beyond simply technical issues, to developing an adaptive management system that explicitly treats management policies as experiments. "Policies to learn must persist for times of biological significance, and they must affect human action on the scale of ecosystems" (Lee 1993, p. 161). The track record for past attempts at adaptive management, however, is not good. Problems identified by Carl Walters, author of Adaptive Management of Renewable Resources (1986) include:

• poor design: scale too small, treatments too weak, lack of replication, lack of baseline data, limited tolerance to truly test options,

• inadequate monitoring,

• lack of long-term commitment,

• inadequate funding, and

• "management difficulties": risk aversion, inability to admit failure, resistance due to perceived threats to existing interests (from a 1995 talk, as summarized by Stankey and Shindler 1997).
 

Although AMAs were created with the mandate of cross-agency and multi-stakeholder participation, there was little demand or involvement from agencies and the public for the creation of most AMAs. Without a local sense of ownership and legitimacy, public interest could be weak and agency managers could perceive AMAs as yet another unfunded obligation during a time of shrinking budgets and staff (Stankey and Shindler 1997). Initial public involvement efforts in the Northern Coast Range AMA generated little interest, but recent efforts appear to be generating more interest and discussion (Warren Tausch, personal communication). The level of community cooperation and vitality varies across the area; many residents acknowledge tension within their communities between new and long-term residents (Leonard 1997).
 

The geographic communities in and around the Northern Coast Range AMA are diverse, with a great variety of lifestyles and interests among residents. Approximately 60,000 people live in the rural and coastal areas in the AMA. They are primarily Caucasian and about one third of the adults are high school graduates (Leonard 1997). Two major demographic trends in coastal counties are population growth from migration (mostly from other areas in the Northwest), and a proportional increase in older age groups (e.g. for retirement). The north coast region attracts more than 2 million visitors each year, 60% of whom engage in outdoor activities, primarily along the coast itself. Demand for camping and dispersed recreation in the interior appears to have increased over the last several years, however, and large numbers of people hunt large game, fowl, and fish. Other commodities extracted from public lands include special forest products (e.g. moss, Christmas trees, mushrooms), firewood, timber, rocks and minerals, grazing leases, and water (Leonard 1997).
 

Employment in the retail and service industries has increased while employment in manufacturing and the forest industry has decreased over the last 35 years. Commercial fishing has declined over the last several decades due to reductions in fish abundance as well as a decline in prices. Within the last decade, several timber-processing mills and two hospitals have closed and two large casinos have opened. Economic growth is frequently related to tourism or residential development of the coastal strip. Wages for a significant portion of the workforce have declined due to a scarcity of full-time jobs paying more than minimum wage, while the cost of rental housing has climbed. About 14% of the population in the AMA is below the federal poverty level (Leonard 1997).
 

Sustainable economies
 

7. How can the AMA provide significant economic support for local communities?
 

Economic value to local communities can be derived from federal lands in the AMA in a number of ways. Hunting, fishing, gathering of edible foods, and collecting firewood provides some subsistence to many local residents. While fish should become more abundant in the long term in response to habitat restoration, some game animals that use early successional stages for part of their life cycle (e.g. Cervus elaphus, Roosevelt elk) may become less abundant until more late-successional forests, which often provide better forage (Happe et al. 1990), become available. Tourists and visitors engaged in hunting or fishing generate local value by purchasing gas, food, lodging, and souvenirs in local communities. Tourism along the coastal strip is expected to continue to increase, and could be encouraged within forest lands by developing additional trails and recreational facilities. Increased tourism may not be desirable, however: access may become more limited as fewer roads are maintained, not all local residents approve of the increased traffic and commercialization associated with tourism, and recreation may negatively affect species that are of concern. Economic value can also be derived through extraction of commodities for resale or manufacturing, including timber, special forest products (e.g. moss, boughs, mushrooms, native plants), firewood, and minerals.
 

Timber has been the most important commodity, in terms of value and employment, for Coast Range communities over the last few decades. Projections for annual timber harvest from thinning in young and mature stands on federal lands in the AMA are for 23,500-47,000 m³ (10-20 MMBF) cut from 320-640 ha (800-1600 ac) (USDI Bureau of Land Management and USDA Forest Service 1997). The value of thinning sales tends to fluctuate greatly with regional and international markets; interest and bid amounts for recent federal thinning sales has been low (Warren Tausch, personal communication). Given the emphasis of private landowners on growing fiber, there may be potential for federal forest lands to provide comparable value from high-quality wood by growing and harvesting older, larger trees using long-rotation or selection-system silviculture (Weigand et al. 1994, Hansen et al. 1995, Curtis and Carey 1996). Of course, large live and dead trees are desired for many purposes, including forest habitat, stream habitat, and forest products. The long-term question that needs to be addressed for this AMA is:
 

8. Can wood products be removed from the forest in appropriate amounts and/or locations on the landscape and still maintain vital late-successional ecosystems?
 

Communities also derive values from forests that are primarily non-economic in nature, including clean water, clean air (pollution abatement), moderation of floods, and amenity values (e.g. visual, spiritual, historical). Choosing forest management policies that meet competing economic demands and values should consider local capabilities and regional economic trends (Niemi and Whitelaw 1997), and is probably most effectively done with local participation.
 

Public participation and AMA management

The primary agency purposes for public involvement usually include building public support for decisions and reaching better decisions (Lawrence and Daniels 1996). The stated objectives for public involvement in the Northern Coast Range AMA are to educate the public about issues, learn from the public about issues and local knowledge, create better plans, avoid conflicts, create opportunities for the public to participate in getting work done, and work towards gaining the public's trust (USDI Bureau of Land Management and USDA Forest Service 1997). These objectives differ somewhat from what citizens want from public involvement, which includes chances to discuss issues and provide input, be listened to openly and respectfully, and see evidence that their input is actually used (Shindler and Cheek 1997).
 

9. How can more effective community involvement be promoted in the AMA?
 

One challenge for the North Coast AMA is in developing a constituency of interested citizens and organizations for an entity that was created with no local demand and includes communities with a wide range of social and economic interests. The traditional agency constituency of forest commodity users generally does not support the policies of the NWFP and the stated goals of the AMA (personal observation), and environmental and recreation constituencies have generally paid much less attention to Coast Range forests than Cascade forests, where montane parklands and unmanaged landscapes are more abundant. The AMA will therefore have to cultivate public interest and involvement.
 

The AMA as a whole may be too large and too conceptual to galvanize interest, however. Local "watershed councils" have been formed as part of the state government's Oregon Coastal Salmon Restoration Initiative. Although some are dominated by government employees, most also contain individuals representing a range of local interests,. These may be a more effective vehicle for public involvement in the AMA, where salmon restoration is also a major concern. Councils, which have funding to support planning, restoration, and monitoring, have been formed for the Yamhill, Tillamook, mid-Coast (Lincoln City), and Rickreall basins. Two of these have formed subcommittees to work with the AMA coordinator on AMA issues. The Tillamook Bay National Estuary Project (TBNEP) may be another vehicle for public involvement in AMA activities. The TBNEP is part of a national Environmental Protection Agency effort to develop partnerships with government agencies and communities to protect and restore the health of estuaries while supporting economic and recreational activities. Projects include information brochures, monitoring and stream restoration by volunteers, social and environmental assessments and studies, and demonstration projects.
 

Characteristics of successful citizen/agency interactions include (Shindler and Cheek 1997):

• inclusive with wide variety of forums and mechanisms,

• interactive and sincere with personal contact and cooperation,

• innovative and flexible by using a variety of methods,

• early and continuous involvement throughout a process, and

• designed with basic strategies and purposes spelled out.
 

The depth and type of involvement that is desired or possible for the AMA probably depends on who wants to be involved locally and the specific issues under consideration. Honest, open exchanges of knowledge help democratize decision-making by making technical knowledge available for public scrutiny and providing detailed knowledge of local conditions for analysis (Stankey and Shindler 1997). Yet agency planning efforts usually involve a substantial amount of time spent considering a wide range of multi-disciplinary information; many people may not have enough free time or background knowledge to participate in such planning. In addition, how to incorporate non-technical ("experiential" or "personal") knowledge, which may provide insights about changes in the environment or in cultural values, into learning and planning is not clear. Developing a shared vision and public involvement will need to be an ongoing process, including broad representation and honoring the legitimacy of the range of concerns and knowledge present. To succeed, the future of AMAs must be linked to the wider social and economic concerns within the communities (Stankey and Shindler 1997). Fostering local initiatives, commitments, and knowledge will help maintain sustainability by insulating ecosystem management from the political cycles that affect large organizations (Naiman et al. 1997).
 

There are few models and no expertise for "how to do" adaptive management, putting front-line managers in unknown terrain and working against the "can-do" spirit common in the land-management agencies (Stankey and Shindler 1997). An analysis of institutional barriers to ecosystem management (Cortner et al. 1996) identified five basic areas needing more study:

• how existing laws, regulations, and policies constrain or aid ecosystem management,

• potential mechanisms for managing across jurisdictions,

• need for organizational change and new relations between agencies and public,

• philosophical underpinnings of ecosystem management in relation to existing management and government institutions, and

• new methods for studying and integrating natural and social sciences.
 

A few of the existing regulations that represent constraints in this AMA include the Federal Advisory Committee Act, which places strict requirements on non-governmental involvement in decision-making, and the reserve land allocations and other guidelines that are overlaid on the AMA designation. Implementing adaptive management may require a major shift in management culture to one that encourages risk, accepts that "failures" (which can be the most productive form of learning) will occur, and support innovation in the face of uncertainty. It may require a shift for researchers as well, because in management situations control of important variables is often low and noise is high (Stankey and Shindler 1997).