Ecological Objectives for Stream and Watershed Restoration along the Pacific Coast of North America

In British Columbia, Canada, over a century of forest harvesting has severely degraded many streams and as a consequence has contributed to the impairment of some other valuable forest resources, especially salmon. Recognition of the problem of land-use alteration on stream channels resulted in efforts to protect and restore stream channels to bring back stream and riparian ecosystems to their former states. Streams with fish populations, especially salmonids, receive riparian buffer strips which maintain some of the functions of riparian-stream ecosystems. British Columbia established a Watershed Restoration Program in 1994, but in specific terms most of the program’s activities are habitat enhancement for fish rather than strictly restoration. A number of techniques are used for restoration including instream placements, off-channel habitats, streambank planting and slope stabilization, road deactivation, and fertilization. There has been little evaluation of the success of these habitat creation projects beyond confirming the persistence of the structures. Most of the work has been done in mid-size streams where direct use by fish populations is possible, but this ignores the important influence of tributaries on downstream environments. This requires a true watershed perspective, rather than a local condition approach. We consider that measures of biodiversity and ecosystem processes, such as organic matter retention or primary productivity, should be assessed in addition to fish population dynamics. Considerations of wildlife use of streams and riparian areas remains peripheral to most of the current practices of stream restoration and protection. The technical development of stream restoration methods is far ahead of the conceptual framework that would specify what appropriate measures of success should be. Restoration might be more effective and efficient if ecological targets at the watershed scale were made explicit in the project objectives. KEY-WORDS: biodiversity, fish, forestry, habitat enhancement, headwaters, restoration, riparian areas, watersheds Ecosystem restoration is frequently applied to freshwater ecosystems around the world to remedy the impacts of human activities. In its strict definition, restoration accelerates the return of a system to its previous state of structure and rates of ecosystem processes. Other activities such as recovery, rehabilitation, replacement, enhancement, and protection are not considered restoration in the strict sense (Hunter 1996), but for the purposes here we will deal with all these activities which are popularly referred to as restoration. Restoration of aquatic ecosystems often takes place with a limited set of objectives or sometimes a single objective or target. Much of the stream restoration work in the past decades has been done with a focus on fish habitat, without consideration of other aspects of the stream ecosystem (e.g., Roper et al. 1997). Stream and riparian ecosystems are complex and the perturbation from human land use has broad effects, therefore the conservation goals of restoration should include more than a single species. Even when fish habitat restoration has been the primary goal there may have been too narrow a focus on the benefits to fish alone, or even just fish habitat, to properly measure the probability of success (Frissell and Nawa 1992). The west coast of North America (northern California to Alaska) is typified by a wet, temperate climate where winter rainfalls and snow can produce frequent and very flashy flooding. This same region if renowned for its timber and fisheries values, the latter of which is dominantly anadromous salmonids. Much of this region has had extensive modification to the land as a consequence of forest practices and there is a large movement to engineering river channels to restore fish habitat presumably lost due to forestry. In this paper we will consider the objectives of these river and watershed construction projects and whether these meet the definition of restoration. We will also provide an overview of restoration works underway in British Columbia, Canada. Among the important questions to ask are how effective these restoration activities are at meeting their objectives, and are there ways to make restoration more effective. We will consider these questions along with what lessons past experience provides for other temperate regions. Ecosystem targets of watershed restoration The recognition that streams are ultimately a product of their watersheds is a perspective of the late 20 century and previous to that the effects of land use on streams was poorly understood (Sedell et al. 1997). The riparian zone in particular is inextricably connected to a stream and provides a variety of functions critical to aquatic ecosystems. Streamside protection by leaving riparian buffer strips is an increasingly common method for providing for some of the ecosystem functions on which streams are dependent (e.g., Haycock et al. 1997). Riparian zone vegetation stabilizes the streambank by binding the alluvial soils, it supplies large woody debris (tree trunks and branches) which provide instream cover and habitat complexity, and provides inputs of organic matter. The riparian zone itself is a product of the moist conditions and fertile conditions provided by the aquatic environment. Protection of streams and riparian areas may avoid the need for costly restoration and is usually the least expensive and most ecologically sound approach to conservation. The maintenance of streamside buffers is not considered restoration, but is a potential consideration in the general plan of maintaining the functioning of stream ecosystems. Stream restoration is often considered as a means of creating artificial stream structures designed for fish habitat, however the broader context of the objectives of restoration are frequently overlooked (Frissell 1997). A narrow focus on fish, in particular salmonid fishes, diverts attention away from the necessary goals of restoration on the appropriate scales and from the broader ecological objectives. Current approaches also presuppose that a narrow range of physical habitats that can be designed by stream engineers is all that is needed to restore species considered to be of value. There have been many articles pointing out the failure of this approach to meet its targets and most emphasize the mismatch between spatial scales at which damage occurs and where the restoration work takes place (e.g., Frissell and Nawa 1992, Kauffman et al. 1997). Ecosystem restoration is distinct from single species enhancement, preservation, habitat creation, reclamation, mitigation, and replacement (Kauffman et al. 1997). Given sufficient time most ecosystems would recover many of their processes and structures, but the time frame is excessive, e.g., for instream, large woody debris (LWD) supply to recover following forest harvesting in riparian areas may take more than a century (Murphy and Koski 1989). Other processes like the recovery of slope stability may take >20 years (Sidle 1985) and sediment supply and transport rates may take considerably longer to “equilibrate”. The objective of restoration should be to re-establish ecosystem processes to create functional environments with most of their native structure (species) against which other objectives, e.g., fish population re-establishment, can be set as a secondary goal. There are a great many potential objectives for stream restoration referred to in the literature (Table 1), but most of these are rarely addressed. Many of these functions in the stream can be provided by intact riparian areas, but other goals require true watershed restoration. Most of these objectives depend upon a scale that should include watershed processes and not just local stream channel features. Targets at the watershed scale are hierarchical, i.e., local results are dependent upon larger-scale processes (Frissell and Nawa 1992). While most actions are to restructure the physical habitat it is also critical to consider the biological goals and results. Table 1: Some of the stated objectives for stream and watershed restoration activities in coastal streams of British Columbia. Targets Stream Riparian Area Water quality Reduced nutrients and sediment transport Buffers to retain sediments and nutrients Hydrology Moderated peak flows to reduce bed instability Moderated peak flows to reduce damage to banks and LWD Habitat complexity LWD, cover and channel complexity LWD; diversity of vegetation structures and ages Microclimate Temperature buffering, shading Temperature buffering, moist, shade Organic matter inputs Detrital inputs, primary productivity Detrital inputs, primary productivity Habitat stability Banks stable; bed more stable Floodplains resistant to rapid erosion Biological diversity Diverse habitats Diverse habitats Why is restoration needed in BC? In the last century and a half there has been large-scale land alteration by forestry in British Columbia, especially in the past 50 years. Older practices, including harvesting to the stream edge, have left many watersheds severely affected by harvesting (Slaney and Martin 1997). In addition, the sites logged first were those that were productive, valley-bottom sites, which are also among the most important salmon streams. In the past five years new regulations for forest practices in BC have been introduced in the Forest Practices Code Act and the hope is that these new approaches will reduce future damage. However, given the importance of other forest values, including fish and biodiversity in the broadest sense, it has been considered a high priority amongst legislators and the public to restore the previously damaged ecosystems to a functional state. In British Columbia most land is owned by the public, as opposed to private ownership. This ownership affects who holds the responsibility for restoration of lands. In BC the responsibility rests with the government and so regulations around forest harvest and grazing attempt to manage the effects on the land, but any damage as a result of an inappropriate regulatory framework may have to be restored at public expense. The time frame over which extensive land-use alterations have taken place in British Columbia is relatively short in the view of world history. This latter fact means that we have an opportunity lacking in many parts of the world and that is that we have undisturbed watersheds which can serve as templates to direct the actions of restoration works. The other advantage this provides is that many species that may have been locally extirpated from a watershed may be able to recolonize from nearby, “pristine” watersheds which retain most of the native ecosystem. What gets done in British Columbia? Is it effective? In British Columbia restoration of watersheds is largely stream restoration, and probably more correctly fish habitat enhancement. British Columbia started instream fish enhancement in the 1970s and in 1994 began the Watershed Restoration Program. Maintenance of streamside buffer strips during forest harvesting is a practice that only began to be broadly applied in the early 1990s and is now law for certain classes of streams. Even the maintenance of buffers will not be sufficient to rapidly recover the characteristics of streams (e.g., supply of LWD) and so there are numerous activities planned and ongoing, including construction of instream habitat, creation of off-channel habitat, riparian planting and streambank stabilization, stream fertilization, and road deactivation. The focus in BC is slightly more towards development of off-channel areas than appears to be the case in the nearby US. There are still boulder and log placements in larger streams of BC, and in some smaller tributaries, but there are many more artificial channels which are dependent on surface water or groundwater flows. Instream habitats: The creation of instream habitat has become a major industry in BC, as in the adjacent US states (Frissell and Nawa 1992, Newbury and Gaboury 1993, Slaney and Zoldokas 1997). The simplification of channel form results from the loss of large woody debris inputs and higher peak hydrographs affecting streambanks and sediment loads which fills in pools. Most of the instream habitat enhancements involve placement of boulder arrays or large woody debris in an attempt to modify local geomorphic units to recreate complex habitats for fish, primarily salmonids. Replanting of riparian areas with fast-growing trees (e.g., poplar, alder) may accelerate the supply of LWD, but these fast-growing species also tend to have small stems and high rates of decay. Since it may take up to 250 for pre-logging rates of channel LWD to re-establish, placement of instream structures, especially LWD, has been intended to compensate for the current deficit of woody debris in many salmon streams. In most cases LWD placements are strongly anchored to resist peak flows. Boulders (some weighing up to 1500 kg) are also placed as clusters to create geomorphic variation or as anchors for the LWD (Slaney and Zoldakas 1997). The instream structures are designed to provide several of the conditions needed by some lotic fishes, e.g., deep pools (especially needed during low-flow periods), overhead cover from predators, and flow refuge. In some streams the effects of forest harvesting or road building are to channelize the stream. In the latter case it may be possible to create pool-riffle sequences using placements of boulder ramps across the channel at intervals of about six channel widths (Newbury and Gaboury 1993). In smaller streams (up to about 4 order) creation of channel complexity this way has had reasonable success, in part because the rock bars are able to adjust to stream flow conditions during peak flows rather than being fixed in place. Off-channel habitats: Technological solutions to declines in commercially valuable fishes resulted in creation of spawning and rearing channels for salmonids. As an outgrowth of the early efforts to create additional habitat for spawning and rearing, off-channel habitats have been built in many areas. In general, these off-channel areas are artificial channels supplied with surface water from a nearby stream or groundwater inputs. The channels are built with many of the features critical to fish such as deep pools, interspersed with riffle sections, and lots of large woody debris to create cover (Figure 1). These channels are not subject to the hydrologic flow variation of nearby natural streams. There is some evidence that these artificial channels produce a greater number of salmonid smolts than equivalent areas of natural streams. These side channels supplied with groundwater may provide juvenile salmonids with better than natural growing conditions since relatively warm water in winter allows these fish to grow when their kin in natural channels are having almost no growth in water temperatures near 0 C (Hinch unpublished data). As with other engineering work described above, whether these efforts constitute restoration or habitat creation is debatable. Figure 1: One example of an off-channel stream created using diversion of surface water from a natural channel Borden Creek, near Chilliwack, British Columbia. Note the placement of large woody debris and boulders, as well as the stabilization of the “toe” slopes with rock. Some of these channels can be several kilometers long. Along with the promotion of artificial, off-channel habitats as a habitat enhancement technique, fisheries biologists increasingly recognize the importance of natural off-channel habitat as a refuge from flooding. Small tributary streams may themselves provide refuge from either floods or high temperatures in bigger streams. Unfortunately road-building practices for logging roads frequently produce barriers to fish movement into tributaries. As part of BC’s watershed restoration practices road crossings over streams are being made more accessible to fish (Slaney and Martin 1997). Streambank and watershed stabilization: In many areas forest harvest to the streambank or cattle grazing can lead to destabilization of stream banks and loss of LWD inputs. The logical solution in the short term is to plant riparian areas with fast-growing trees and shrubs to regain the shading and stabilization afforded by streamside vegetation. In serious cases the streambank may be armored with rock or vegetative mats (Slaney and Zaldokas 1997). The use of rip-rap in streams in BC is generally limited to urban and highway developments. Work further from the stream channel may be performed to reduce problems higher in the watershed. Roads are frequently deactivated to reduce the erosion and slope failures associated with road surfaces. Deactivation may include planting seedlings on the roads, opening up drainage channels (ditches), or hydroseeding to produce vegetative cover. The access to forestry roads may be cut off by removal of bridges or digging deep ditches across roads. Hillslope failures that result in mass wasting to a stream are treated by hydroseeding to establish vegetation to cover the failure. Gullies, defined as steep channels with sidewalls >40% gradient, are a major source of debris flows in coastal BC, and these are treated by cleaning out excess logging debris to reduce the risk of debris flow initiation (Slaney and Martin 1997). Fertilization: Many of BC’s streams are nutrient poor as a result of high precipitation and relatively insoluble bedrock. Historically, many of the salmonid spawning streams would have had a major input of nutrients as the carcasses of spawned-out salmon decomposed (Bilby et al. 1996). The loss of such a major source of nutrient input as a result of fish harvesting, and stock depletion in some cases, has resulted in the frequent use of nutrient additions to mitigate this potential loss of productivity. Several studies have confirmed the loss of productivity and the promise of nutrient additions to boost stream productivity (summary in Richardson 1993, Perrin and Richardson 1997). There is some evidence that nutrient additions can stimulate the food webs leading to juvenile salmonid fish with the result that some fish experience enhanced growth rates (e.g., Johnston et al. 1990). Other problems and goals: Although temperate regions are cooler on average it is the extremes that pose problems for many of our native species. Is shading a necessary target in temperate waters? Even in coastal British Columbia stream temperatures in watersheds that have been clearcut may exceed 20 C, and therefore can be close to lethal conditions for some species such as salmonids or tailed frogs. In BC and the US Pacific Northwest summer temperatures often coincide with low rainfall periods which may result in excessive heating of streams. One solution is to maintain shade, but how much? Coastal streams may not require shading of the entire drainage to maintain temperatures but there has been very little research on how much is enough. Full shade, as in closedcanopy, second-growth forests, may be as detrimental as full sunlight if it retards the rate of primary productivity of the stream. There may be some advantages to creating reaches of stream with no shading. A hundred kilometers or so from the coast in BC, Washington, and Oregon, the forests become more open and the areas are used for cattle grazing. In these areas forestry and ranching combine to present a potential source of stream degradation. Buffer zones are required when forestry activities take place alongside fish-bearing streams, but there has been very little effort to restrict damage from cattle. In a limited number of demonstration projects cattle have been kept away from stream edges using fences until riparian vegetation is established to the point where cattle cannot reach the stream edge. In these areas the stream gradients are sufficiently low and the peak flows moderate so that recovery of riparian vegetation, either by planting or by cattle exclusion, is often sufficient to speed the recovery of those streams. What are appropriate measures of success? In many projects the success of restoration is judged simply to be the continued presence of physical structures (Frissell and Nawa 1992). In the majority of cases there is only a minor amount of effort at even determining if fish populations responded positively, and in some cases fish show no significant change (e.g., Gowan and Fausch 1996). In the original proposal for BC’s Watershed Restoration Program there were several measures proposed for evaluation of the success of their activities, but these were not carried through. A similar example exists for water pollution abatement programs in the US where over $500US billion was spent from the 1970s to the early 90s, but <0.2% of the money was spent on monitoring to determine if the programs had been successful (Hart 1994). There are many predictions from our theoretical framework for stream ecosystems that might make good measures for success. For instance, the influence of large woody debris on the retention of organic matter and nutrients have been shown for streams (e.g., Bilby 1981, Richardson 1991), but to date there are few studies where either of these components of the stream ecosystem have been measured in response to restoration activities. Wallace et al. (1995) found large increases in organic matter retention and minor increases in solute storage when log placements were made in small streams. The diversity and productivity of macroinvertebrates also appears to increase as a result of log placements through the creation of more complex and retentive habitats (Wallace et al. 1995, Hilderbrand et al. 1997). The point of reference for the success of stream and watershed restoration should be the recovery of species populations and process rates to the natural range found in undisturbed watersheds (Dolédec and Statzner 1994, Ralph et al. 1994). One of the measures should be fish populations given that we have extensive data on these species and a great deal of public interest, however fish may not be the best indication of restoration. Given that there are many other water quality and biodiversity objectives it seems appropriate that other measures for the evaluation of the success of restoration should be included in projects (e.g., Ebersole et al. 1997). What traits would go into a good indicator of the success of restoration? Ideally these indicators would be important components or processes in the stream ecosystem, perhaps key species or rates of organic matter transport or water quality measures. The measures also need to be relatively simple to obtain and reliable indicators of successful restoration. In British Columbia salmonid densities are often advocated but these species are largely anadromous and therefore transient as far as the stream and watershed are concerned. Catastrophic impairment of the stream will be reflected in salmonid numbers, but more subtle changes in the ecosystem may be difficult to measure by population numbers. Stream invertebrate assemblages are being used by several jurisdictions to evaluate stream condition (e.g., Wright 1995) and these same indicators could be used to assess restoration. In BC the fish fauna is too depauperate (largely a result of glacial history in the region) to provide a good indicator although an index of biotic integrity has been used elsewhere based on fish assemblages. Measures of wildlife and other terrestrial components of the riparian ecosystem would be in order as an indication of the effectiveness of restoration actions. Many of the riparian management guidelines and proposals for watershed restoration make mention of wildlife values (in the broad sense including all wild organisms), but seldom evaluate responses of these species. Terrestrial invertebrates would be a group with lots of species on which to evaluate restoration activities but these taxa are poorly known and require a specialist to identify. Two good indicators of riparian condition are amphibians or small, herbaceous plants. Amphibians are a relatively sensitive group with many species, and occur at very high densities in good habitat (sometimes exceeding the biomass per unit area of all other vertebrates). Can we make restoration more effective? Often a small portion of a stream channel is treated with stream restoration without removing sources of degradation which may be higher in watershed (Kauffman et al. 1997). This “postage stamp” effect of restoration of a small reach may be unsuccessful if upstream environments remain degraded and influence downstream reaches where stream works have taken place. Small streams typically do not receive protection from forestry activities, or other land-use activities, and so large parts of a watershed may be disrupted. Unfortunately, it is often the downstream, larger reaches where socially important fish species reside and these are the sites that receive restoration action. Small tributaries need to be considered before downstream reaches can be restored and this is a good example of where the watershed perspective is critical. Many of the effects of watershed perturbation by land use are small, incremental changes. There has been little study of cumulative effects of changes in forested watersheds. In other settings environmental scientists have been pursuing the study of cumulative effects of non-point-source pollutants on water quality. Some of the effects of forest harvest are small, but incremental, and there is no framework for evaluating the significance of these, and therefore restoration of these sources of disturbance have not begun to be considered. Among the sites where these kind of effects may propagate within a watershed are the many small, headwater streams where increased erosion or increased transport or organic matter starts. Many of the small gullies carry water only during high precipitation periods but in combination with all gullies in a watershed these may be major sources of materials. Buffers or restoration of these areas by cleaning out logging debris may not be sufficient. Restoration efforts remain focused on fish habitat and rehabilitation of the most obviously impaired reaches of streams. In the future, attention to headwater areas will come about as we develop a stronger conceptual framework for ecosystem restoration. There may be alternatives to the long, thin buffers that are typically left following logging or other land use. These thin buffers are highly prone to windthrow because they have grown up in a continuous forest and are not resistant to wind exposure. Moreover, strips of trees no more than 30 m wide on each side of a stream may be so influenced by edge effects that there is no “interior” forest habitat left. In temperate areas where shading for temperature protection may not be as critical as in other areas, it may be possible to set up reserve patches along a stream rather than a continuous buffer. If one was to select sites for these reserve patches, one suggestion would be to put approximately circular reserves around the confluences of streams. The confluences of streams are often geomorphically dynamic and as a consequence also have high structural diversity which is an important element leading to the high biodiversity associated with riparian areas. A further benefit of such an approach is that fish frequently seek out small tributaries (or move from tributaries to larger streams) as temperature or flow refuges when conditions change during the year. Some alternative designs for stream protection and restoration should be tested as part of the long-term development of tactics for managing stream and riparian ecosystems. Are there lessons for other temperate regions? Buffers are widely advocated as a way to protect streams from degradation from land use of various sorts (e.g., Haycock et al. [1997] and references therein) . However, buffers are often deployed primarily on larger streams where their influence might actually be less useful to conservation. Small streams potentially heat more quickly and have a larger impact from the forest (higher edge to area ratio). The usefulness of riparian buffers along streams has not been evaluated other than as a source of LWD. Given that riparian buffers are rarely greater than 30 m wide, the riparian area may be so modified by edge effects from the conditions beyond that it lacks the natural functions attributed to riparian areas. In a large, integrated experiment in coastal BC we are evaluating buffer widths of various widths versus streamside clearcuts and forested controls to test the effectiveness of buffers at maintaining key components and processes of small stream and riparian systems (Richardson and Hinch unpublished). The effectiveness of riparian buffers for meeting the full suite of goals sometimes attributed to them awaits further tests. Stream restoration frequently is motivated by socio-political objectives of providing the appearance of environmental stewardship. These reasons for restoration may be justified but there are times where the goal is to create jobs rather than simply to restore environments. Public participation in some restoration activities provides a connection for personal stewardship of local streams which further supports the call for continued restoration work at the political level. Community involvement usually is critical to support the continuation of publicly-funded restoration programs like that in British Columbia. Many stream and streamside projects are discussed as restoration but the list of objectives may makeclear what is really at stake. The term “restoration” is widely misused when applied to instreamhabitat creation. If the objective is simply to enhance fish habitat or restore water quality then theproject may not be restoration in the strict sense. A watershed perspective is essential for mostrestoration work since local modifications will not be successful if broader impairment (e.g.,sediment loading) of the stream system has not been remedied (Frissell 1997). Clear definition of thegoals for stream activities would make evaluation of success more obvious and at the same time makeclear that there are often narrowly defined reasons for the activity. Watershed restoration is still earlyin the development of a conceptual framework and admission of that would make progress easier toachieve. REFERENCESBilby,R.E. 1981. Role of organic debris dams in regulating the export of dissolved and particulate matter from aforested watershed. Ecology 62:1234-1243.Bilby, R.E., B.R. Fransen, and P.A. Bisson. 1996. Incorporation of nitrogen and carbon from spawning cohosalmon into the trophic system of small streams: evidence from stable isotopes. Can. J. Fish. Aquat. Sci.53:164-173.Dolédec,S and B.Statzner (1994): Theoretical habitat templets, species traits, and species richness: 548 plant andanimal species in the upper Rhône River and its floodplain. Freshwat. Biol. 31:523-538.Ebersole, J.L., W.J. Liss, and C.A. Frissell. 1997. Restoration of stream habitats in the western United States:restoration as reexpression of habitat capacity. Environ. Manage. 21:1-14.Frissell, C.A. 1997. Ecological principles, pp. 96-115 In: J.E. Williams, C.A. Wood, and M.P. Dombeck (Eds.).Watershed Restoration: principles and practices. American Fisheries Society, Bethesda, Maryland.Frissell, C.A. and R.K. Nawa. 1992. Incidence and causes of physical failure of artificial habitat structures instreams of western Oregon and Washington. N. Am. J. of Fish. Manage. 12:182-197.Gowan,C. and K.D. Fausch (1996): Long-term demographic responses of trout populations to habitatmanipulation in six Colorado streams. Ecol. Applic. 6:931-946.Hart, D.D. 1994. Building a stronger partnership between ecological research and biological monitoring. J. N.Am. Benthol. Soc. 13:110-116.Haycock, N., T. Burt, K. Goulding, and G. Pinay (Eds.). 1997. Buffer Zones: their processes and potential inwater protection. Quest International, UK, 322 p.Hilderbrand, R.H., A.D. Lemly, C.A. Dolloff, and K.L. Harpster. 1997. Effects of large woody debris placementon stream channels and benthic macroinvertebrates. Can. J. Fish. Aquat. Sci. 54:931-939.Hunter, M.L., Jr. 1996. Fundamentals of Conservation Biology. Blackwell, MA.Johnston, N.T., et al. 1990. Increased juvenile salmonid growth by whole-river fertilization. Can. J. Fish. Aquat.Sci. 47:862-872.Kauffman, J.G., R.L. Beschta, N. Otting, and D. Lytjen. 1997. An ecological perspective of riparian and streamrestoration in the western United States. Fisheries 22(5):12-24.Murphy,M.L. and K.V.Koski. 1989. Input and depletion of woody debris in Alaska streams and implications forstreamside management. N. Am. J. Fish. Manage.9:427-436.Newbury,R.W. and M.N.Gaboury. (1993): Stream analysis and fish habitat design: a field manual. ManitobaFisheries Branch, Canada.Perrin, C.J. and J.S. Richardson. 1997. N and P limitation of benthos abundance in the Nechako River, BritishColumbia. Canadian Journal of Fisheries and Aquatic Sciences 54:2574-2583.Ralph,S.C. et al. 1994. Stream channel morphology and woody debris in logged and unlogged basins of westernWashington. Can. J. Fish. Aquat. Sci. 51:37-51.Richardson, J.S. 1991. Seasonal food limitation of detritivores in a montane stream: an experimental test.Ecology 72:873887.Richardson, J.S. 1993. Limits to productivity in streams: evidence from studies of macroinvertebrates, p.9-15. InR.J. Gibson and R.E. Cutting [Ed.] Production of juvenile Atlantic salmon, Salmo salar, in natural waters.Canadian Special Publication of Fisheries and Aquatic Sciences 118.Roper, B.B., J.J. Dose, and J.E. Williams. 1997. Stream restoration: is fisheries biology enough? Fisheries22(5): 6-11.Sedell, J.R., G.H. Reeves, and P.A. Bisson. 1997. Habitat policy for salmon in the Pacific northwest, pp. 375-387 In: D.J. Stouder, P.A. Bisson, and R.J. Naiman (Eds.). Pacific Salmon & Their Ecosystems: statusand future options. Chapman & Hall.Sidle, R. 1985. Hillslope Stability & Land Use. American Geological Union.Slaney,P.A. and A.D.Martin 1997. The watershed restoration program of British Columbia: accelerating naturalrecovery processes. Water Qual. Res. J. Canada 32:325-346.Slaney, P.A. and D. Zaldokas (Eds.). 1997. Fish Habitat Rehabilitation Procedures. Watershed RestorationTechnical Circular No. 9, BC Ministry of Environment, Lands & Parks, Victoria, BC, Canada.Wallace, J.B., J.R. Webster, and J.L. Meyer. 1995. Influence of log additions on physical and bioticcharacteristics of a mountain stream. Can. J. Fish. Aquat. Sci. 52:2120-2137.Wright, J.F. 1995. Development and use of a system for predicting the macroinvertebrate fauna in flowingwaters. Australian J. Ecol. 20:181-197.