Carnation Creek Project
 

Geographic Location of Project

Interactive map that details several points of interest relating to the Carnation Creek Watershed Experiment.

Project Summary:

Introduction

The effects of forestry practices on watershed processes and salmon populations have been studied for over 35 years at Carnation Creek on the west coast of Vancouver Island, British Columbia (B.C.). This single-watershed, intensive case study has generated the longest series of continuous data on fish-forestry interactions anywhere. Studies at the Carnation Creek Experimental Watershed were initiated in 1970 by the federal government agency now known as Fisheries and Oceans Canada (DFO) and by MacMillan Bloedel Ltd. (now Weyerhaeuser Company), and soon expanded into a multi-agency, multidisciplinary program on the effects of forest harvesting on a coastal watershed and its salmon and trout populations. The Research Branch of the B.C. Ministry of Forests is currently the lead agency continuing this investigation of long-term, forestry-related effects and watershed recovery processes in partnership with other provincial government agencies, universities, and the forest industry.

Purpose and Objectives

In the 1960s, resource managers and land-use planners needed to base judgements on the effects of logging on fish populations from studies conducted elsewhere in North America; for example, in Oregon, Alaska, and as far away as New Hampshire. Both the forest industry and government resource agencies were concerned that these extrapolations might not lead to the most appropriate planning decisions for areas on the west coast of British Columbia. Therefore, the Carnation Creek study was initiated in order to provide fish-forestry interaction information on at least one type of drainage basin in coastal British Columbia.

Initially, the three principal objectives were:

1. to provide an understanding of the physical and biological processes operating within a coastal watershed;

2. to reveal how the forest harvesting practices employed in the 1970s and early 1980s changed these processes; and

3. to apply the results of the study to make reasonable and useful decisions concerning land-use management, and fish and aquatic habitat protection.

The project has achieved these initial goals despite the typical limitations of studies conducted only in a single watershed. Over 200 publications have been produced from Carnation Creek research. The results from this project have made major contributions to the British Columbia Coastal Fisheries-Forestry Guidelines (CFFG) implemented in 1987, the legally binding provisions for aquatic habitat protection within the British Columbia Forest Practices Code (FPC), which replaced the CFFG in 1995, and the current Forest and Range Practices Act, which is supplanting the FPC.

Research and monitoring is currently focused on determining the mechanisms, rates, and levels of natural resource recovery in a harvested coastal drainage by quantifying the long-term changes in biological and physical watershed processes as the second forest grows. Research is presently documenting the 25- to 30-year post-harvest responses to logging practices from the condition and attributes of the hydrologic regime, hillslopes, stream channel network, riparian forest (canopy closure), aquatic habitats (mainstream, tributary, and off-channel network), water temperatures, and fish populations. Key project components continue to include studies of salmon species abundance, growth, age structure, survival, and smolt production at Carnation Creek.

There is a continuing need for information on the effects of forestry practices from intensive, watershed-scale, process-oriented studies conducted over sufficient periods of time that also incorporate the pre-harvest, during-harvest, and post-harvest data necessary to determine long-term impacts of forestry practices and post-harvest recovery processes. Carnation Creek is one site where these relationships can be thoroughly studied and understood. This study is responsible for much of our present understanding of how small Pacific coastal watersheds function, and how forestry practices affect these functions.

The basin-scale, multidisciplinary approach continues to be employed to explain changes to stream channels and fish habitats attributable respectively to: (1) alternative riparian management treatments (riparian buffer strips of known dimensions versus clearcut harvesting); (2) hillslope processes (landslides and debris flows); and (3) the overall level of harvesting within the basin (hydrologic effects).

These linked study components are intended to support the improved effectiveness of British Columbia forestry policies, regulations, and guidelines concerned with: sustaining water quality, water quantity, and timing of flow; riparian management; hillslope and stream channel integrity; aquatic habitats; and fish.

Study Site

Carnation Creek is located ~20 km northeast of Bamfield on the south shore of Barkley Sound in southwestern Vancouver Island (49°N, 125°W).

The watershed occurs within the Coastal Western Hemlock Biogeoclimatic Zone, which spans the west coast of North America from the Oregon Coast Range and High Cascades to the Queen Charlotte Islands and southeast Alaska. The forests consist primarily of western hemlock (Tsuga heterophylla), amabilis fir (Abies amabilis), and western redcedar (Thuja plicata), with some Sitka spruce (Picea sitchensis) on the valley bottom, Douglas-fir (Pseudotsuga menziesii) on the ridges, and western white pine (Pinus monticola) at high elevations. Prior to logging, bigleaf maple (Acer macrophyllum) and red alder (Alnus rubra) provided a substantial proportion of the tree cover in riparian areas.

The watershed area is 11 km² and contains rugged terrain between 0 and 800 m elevation. The valley walls have gradients up to 80%. The coarse, well-drained soils, forest cover, hydrology, and heavy annual precipitation (varying from 210 to over 500 cm/yr) are typical of western Vancouver Island and many other areas of coastal British Columbia. About 95% of the annual precipitation falls as rain, primarily during autumn and winter. High variations in seasonal rainfall cause stream discharge to range from 0.03 m³/s in summer to 64 m³/s in winter. Streamflows may increase 200 fold within 48 hours because of rapid runoff from rainstorms that can deliver up to 26 cm of precipitation within that period.

The mainstream channel is ~7.8 km long. The lowermost 3.1 km, extending from the stream mouth to a steep-gradient canyon, is inhabited by several species of anadromous salmonids, including coho (Oncorhynchus kisutch) and chum (O. keta) salmon, and steelhead (O. mykiss) and cutthroat (O. clarki) trout.  Coastrange (Cottus aleuticus) and prickly (C. asper) sculpins also inhabit this lowermost stream reach, which flows through a valley bottom of about 55 ha that is 50–200 m wide. The valley flats in this area contain both perennial and seasonally flooded tributaries, ponds, and depressions used as winter habitat by juvenile salmonids.  The canyon is a barrier to anadromous salmonids; however, an isolated, relict population of land-locked (non-migratory) cutthroat trout inhabit about 1 km of Carnation Creek immediately upstream of this steep reach.

Annual numbers of adult salmon spawners returning to the watershed from the Pacific Ocean have historically included from 23–4,186 chum salmon, 74–639 coho salmon, 12 or fewer steelhead trout, and 9 or fewer cutthroat trout. Occasionally, other species, such as pink (O. gorbuscha), chinook (O. tshawytscha), and sockeye (O. nerka) salmon enter the stream in small numbers as stray adult spawners in autumn.

Study Design

The Carnation Creek project is a single-watershed, intensive case study incorporating pre-harvest, during-harvest, and post-harvest observations. Spatial controls are available for a variety of study components in Tributary C, a sub-basin mainly unharvested for the duration of the investigation. The study design and the methods employed for monitoring physical variables, fish populations, and biological processes before, during, and after forest harvesting have been described by Hartman and Scrivener (1990) and summarized by Tschaplinski (2000) and Tschaplinski et al. (2004).

This study was designed initially to examine the effects of progressive clearcutting and three different types of streamside forest harvest treatments on stream channels and fish populations. The three riparian treatments applied within the lowermost 3.1 km of the stream were:

1. a leave-strip buffer 1–70 m wide from the estuary to 1300 m upstream;

2. “intensive” clearcutting along 900 m of stream channel immediately upstream from the leave-strip treatment (no riparian trees were left, some trees were felled and yarded across the stream, and commercially valuable windthrown trees were recovered from the stream channel); and

3. “careful” clearcutting along 900 m of stream immediately upstream from the intensive treatment area (no activity was permitted in the stream; vegetation on the streambank, such as salmonberry [Rubus spectabilis], was untouched but red alders were removed).

The study was carried out in three phases:

1. pre-harvest period, 1970–1975;

2. during-harvest period, 1976–1981, when about 41% of the watershed (including almost all of the valley bottom) was harvested; and

3. post-harvest period, currently from 1982 to 2006. Within the post-harvest phase (1987–1995), about 21% more of the basin was logged in headwater areas.

Historical Data Collections

Data collected historically included comprehensive information on climate, stream temperature and discharge, groundwater levels (from piezometers), precipitation (rain) chemistry, water chemistry, terrain (hillslope) stability, stream channel morphology, large woody debris (LWD) abundance and distribution, streambed particle-size composition (frozen-core methods), suspended sediment generation and transport during high flows, streambed scour and deposition, forestry-related ground disturbance, post-harvest revegetation, biomass of aquatic algae (periphyton), abundance and distribution of benthic macroinvertebrates, and fish habitats. Fish population studies were focused primarily on salmonids and included research on: spawning and migration patterns; abundance, seasonal movements, rearing, growth, and survival of juvenile salmonids; chum egg incubation; egg survival; fry emergence; and fecundity determinations for female chum and coho for estimates of annual egg-to-fry survival.

Current Research and Monitoring Components

Many of the historic study components continue today, but several new components have been added in recent years. Routine studies on suspended sediment, primary (algal) production, water chemistry, and groundwater levels were discontinued in the mid to late 1980s. Principal study components currently include: climate; watershed hydrology and hydrologic recovery; forest regeneration and growth; hillslope processes; stream channel morphology; streambed erosion and sedimentation processes; and fish populations, migrations, and habitats.

Intensive riparian treament.		Riparian buffer treatment area.		Stream temperatures and discharge.

A basin-scale, integrated approach is used: to describe and model the sequential linkages among hillslope, drainage network, and floodplain processes; to develop watershed sediment and woody debris budgets; and to document and predict the ultimate consequences of these processes for the evolution of stream channel morphology, aquatic habitats, and fish production. Core study components are:

1. Hydrological and meteorological data collections

• Data are collected at permanent hydrology weirs and climate stations equipped with continuous-operation sensors linked to dataloggers where information is collected and archived electronically.
• Water temperature, depth, and discharge are monitored at hydrology weirs located on the mainstream (B weir) and on two principal tributaries (C control sub-basin; H clearcut basin). Our project partners from Water Survey of Canada (Environment Canada) monitor B weir as part of their provincial hydrology network.

• Climate stations located at several sites and elevations monitor air temperature, solar radiation, precipitation, relative humidity, wind speed, and wind direction.

2. Hydrologic recovery

• Hydrologic recovery studies involve forest microclimatology and canopy rainfall interception and runoff.
• Rain gauges and sampling troughs in both mature forest and second-growth stands are used to determine the proportion of rainfall that is intercepted by the vegetation and channelled down tree trunks (stem-flow monitoring) to the ground.
• Hydrologic recovery models are developed and continually validated on the basis of these field data.

3. Channel morphology and sediment dynamics

• Surveys for monitoring LWD, streambed textures, channel bedload movements, and sediment budgets involve several subcomponents.
• Annual ground-based surveys are conducted during summer in eight representative study sections (60–120 m long) distributed in both clearcut and riparian-buffered stream areas.
• All pieces of LWD are mapped and identified with numbered, metal tags to follow changes in distribution and abundance.
• Textural distributions of surface sediments are described visually with grid samplers and mapped.
• Cross-sectional and longitudinal profiles are obtained to describe the geomorphic characteristics of each study section.
• Annual channel scour and deposition is quantified with scour-and-fill monitors installed within the channel.
• Sediment (bedload) transport is determined by following the distribution of large populations of magnetically tagged, colour-coded stones representing natural size distributions and placed into the stream in previous years. Particles are located with magnetometers and laser survey equipment to calculate distances and depths moved over time.

4. Annual ground-based and aerial photographic surveys
m of riparian areas and hillslopes

• Ground-based techniques with angular canopy densiometers are supplemented with aerial surveys of the entire creek channel to monitor the rates of second-growth canopy closure over the creek, its tributaries, and off-channel refugia. Mid-summer canopy closure and forest growth in riparian buffer versus clearcut riparian areas are related to water temperature (peak daily and mean monthly) and fish growth/survival changes by study section, by riparian treatment, and in the watershed overall.
• Stereo aerial photographs are used to determine changes in channel structure in areas between study sections as part of cumulative effects monitoring and to generate a full inventory of fish habitats.
• Air-photo inventories in steep terrain are used to document the location, frequency, and sizes of sediment and debris sources, including open hillslope and gully failures.

5. Salmonid migration assessments (main fish fence)

• Adult coho and chum salmon, and cutthroat and steelhead trout spawners returning to Carnation Creek (September to mid-December) are enumerated at the main fish weir located near the mouth of the stream.
• Each spawner is identified to species and sex. Age is determined from scale samples. Lengths and weights are obtained.
• Chum salmon spawning downstream of the main fence are enumerated each day by observers on foot. Their seasonal abundance is determined from "area-under-the-curve" techniques.
• Juvenile salmonids (fry and smolts) and sculpins migrating seaward in spring (mid-February to mid-June) are enumerated and identified to species, and large numbers (up to 50 individuals per species per day) are sampled for scales, measured for length, and weighed.
• Coded-wire tags are applied to all coho salmon smolts to determine marine survival.

6. Salmonids rearing in fresh water

• Multiple fish surveys are performed each summer to describe the abundance, distribution, age structure, growth, and survival of juvenile salmonids rearing in the watershed.
• Three surveys are conducted in 10 representative study sections (60–120 m long) distributed in clearcut and riparian-buffered stream areas between mid-June and late September. Eight of these study sites are the same as, or adjacent to, those used for channel morphology research.
• Abundance of each species by study site is determined by the Seber-LeCren two-catch removal method with electrofishing, seining, and barrier nets. Lengths, weights, and scales for age analysis are taken.
• The total abundance of fish in Carnation Creek is determined by extending the numbers captured in the survey sections to the total length of stream inhabited by each species.
• Aquatic habitat inventories for fish rearing in Carnation Creek are obtained simultaneously with the fish inventories.  Habitats are described for periods of summer base flows within each study section–Resource Information Standards Committee (RISC) standard methods are used to quantify channel, pool, glide, and riffle habitat units for fish density determinations.
• Seasonal and interannual changes in numbers and fish size are determined to track fish species survival and growth by riparian treatment area, by channel impact state, and in the watershed overall.

7. Overwinter habitats and survival

• Juvenile salmonid movement between the main channel and tributary or –off-channel– overwintering refugia are quantified by two-way fish fences operated in autumn and spring.
• Baited Gee traps and snorkel counts are used to assess the proportion of juvenile populations overwintering in the main channel versus tributary/off-channel habitats.
• Post-harvest habitat assessment, and use of these sites by rearing fish, are determined annually to assess the importance of low-order streams for salmon survival and production.
• Total overwinter survival is calculated from the difference between the population abundance determined in late summer / autumn and the number of smolts migrating seaward from Carnation Creek in spring (plus residual fish rearing in the stream in spring).

Core studies at Carnation Creek collect and employ empirical data:

1. to determine current hillslope, stream channel, and fish habitat conditions, and the status of salmonid populations within the watershed;

2. to compare current channel and aquatic habitat conditions in clearcut and buffered riparian areas;

3. to relate these conditions to the ecological implications of past and current forestry practices within the watershed, and quantify the state of recovery;

4. to quantify and explain annual and seasonal trends in fish abundance, growth, distribution, and habitat use; and

5. to compare the relative effects on stream channels and fish habitats made by alternative riparian forestry treatments versus those delivered to the stream network by hillslope processes (landslides and debris flows); that is, the –downstream cumulative effects.–

In addition to the core studies, Carnation Creek data support process studies and model development/refinement by our research partners to describe and predict post-harvest alterations or recovery to hydrologic regime, hillslopes, stream channels, aquatic habitats, fish populations, and habitat capability. These spatially linked studies include:

• additional rainfall interception and runoff models (UBC);
• modelling water quantity and quality of surface flows (UBC);
• predictive models of landslide frequency and location (UBC);
• hillslope/channel sediment-and-debris budgets (UBC partners);
• channel sediment mobilization and channel condition studies (UBC and associates);
• dissolved organic carbon (DOC) shielding of UV light versus fish survival (Fisheries and Oceans Canada and Environment Canada); and
• export of water, DOC, particulate organic carbon, and nitrogen from Carnation Creek and its clearcut and unlogged sub-basins to the ocean, and the consequences for coastal marine alga blooms and juvenile salmon survival (University of Western Ontario).

Results - Forestry Effects on Fish and Fish Habitat

Carnation Creek research has identified three broad and interrelated categories of forestry-related effects upon fish and aquatic habitats:
(1) physical habitat-structure alterations; (2) temperature-related shifts; and (3) trophic responses. These categories, separately and in combinations, have had different effects on fish depending upon species, life stage, and distribution in fresh water. Population and habitat responses to harvest practices are thus complex.  Moreover, long-term trends in anadromous fish abundance are often difficult to interpret because they may be confounded by changes in climate, ocean conditions, and fisheries management strategies.

Chum and coho salmon have responded differently to forest harvesting. Chum salmon populations have exhibited the most marked changes over the past 36 years although this species uses freshwater habitats much less extensively than do coho salmon.

Mean numbers of chum spawners returning to the creek have fallen 3.3 fold from pre-harvest levels.

Numbers of adult coho (excluding precocious young males called jacks) returning to spawn have also declined during the post-harvest period.

However, the difference between pre-harvest and post-harvest mean returns is not statistically significant. Mean post-harvest returns were significantly lower than pre-harvest levels until the mid-1990s when strict reductions in commercial and recreational coho fisheries, followed by complete closure between 1998 and 2003, resulted in four of the five highest returns observed since 1971, including the 34-year peak in 2003. This pattern demonstrates that fishing mortality is an important determinant of the abundance of coho spawners returning to Carnation Creek.

Project scientists in the late 1980s used a series of sequential life history–based regression models to determine the relative effects of fishing, forest harvesting, and climate change on the returns of coho and chum to Carnation Creek. They concluded that most of the post-harvest decline in both species was due to factors other than forestry practices. About 26% of the reduction in the numbers of chum adults, and < 10% of the decline in coho during the 1980s was attributable to forestry operations. Most of the decline for both species was caused by climatic shifts affecting both marine and freshwater environments. In spite of the prominence of climatic factors, forestry has been an important cause of the observed declines in chum and coho abundance at Carnation Creek.

Structural Habitat Alterations

The main effects of forest harvesting on the stream were: (1) bank erosion causing the channel to become wider and shallower; (2) loss of large woody material and the accumulation of woody debris into recently formed logjams; and (3) movement of sand and pea-sized gravel to the lower reaches of the stream and estuary.

Some of these changes, affecting one-half or more of the lower 3070 m of Carnation Creek, were a consequence of riparian clearcutting and were initiated shortly after harvesting concluded. However, longer-term, basin-wide processes reflecting critical linkages between steep hillslopes and the stream channel network have overwhelmed the effects of the riparian treatments. These longer-term changes to the stream channel resulted from increased frequencies of landslides and debris torrents after logging. Over 80 small landslides and three major debris torrents have been observed, all in the logged portions of the watershed. The overall volume of landslide material has increased 12 fold after logging. Large logjams and associated sediments deposited by debris torrents in 1984 have moved progressively downstream through the clearcut riparian treatments to the leave-strip zone, and continue to cause major channel alterations and fish habitat loss more than two decades after their initiation.

The post-logging widening of the channel, accelerated scour and deposition, and loss of stable LWD have been largely due to the stream moving around these logjams and sediment deposits, and redistributing materials downstream. This complex of short-term and long-term habitat changes has affected all salmonids in the watershed.

Two-thirds of the effect on chum salmon related to forestry practices was explained by impaired embryo development and reductions in egg and alevin survival due to elevated sedimentation of spawning gravel in Carnation Creek. Between 68 and 100% of all chum salmon spawn in a channel segment about 100 m long in the upper estuary and immediately upstream where fine sediments are deposited. Egg-to-fry survival declined there from 20.3% before logging to 10.9% after logging. The accumulation of fine material reduced intra-gravel water flow and oxygen delivery to the eggs. Sand and pea gravel deposition also buried alevins and prevented their emergence.

These same processes have affected juvenile coho salmon. The survival of coho eggs has also declined by about one-half after forest harvesting–from 28.8 to 15.6%. This decline was partly caused by increases in the amount of streambed sand and pea-sized gravel downstream of the clearcut portions of the creek. However, most of this reduction was caused by increased streambed scour and deposition in clearcut areas during freshets after logging. Streambank erosion and loss of LWD caused further impacts on juvenile coho salmon.

Changes in channel morphology and cover also reduced the quality and amount of main-channel rearing habitat available for coho, and thus reduced the capacity of the stream to support populations of coho fry and yearlings during summer. Habitat complexity in both the careful and intensive clearcut treatments decreased after logging due to reductions in the amount, size, and stability of LWD within the stream channel. Volumes of LWD declined by at least 50% in sections of stream adjacent to the clearcut treatments. Pools became shallower due to bedload deposition. Long stretches of channel were filled with large sediment deposits upstream of new logjams, thus creating ephemeral channels and reducing stream wetted area and salmonid rearing habitat in the main channel. The freshwater habitats in Carnation Creek that supported approximately 12 000 coho juveniles in late summer each year before harvesting have sustained between one-half and two-thirds of these numbers afterwards.

The stream structure and fish habitat characteristics in the lower reaches of the creek have been partially buffered from riparian forestry practices for decades by the leave-strip treatment, and have remained similar to those observed in pre-logging years. However, this zone is presently being degraded as excess sediment and debris move downstream from the intensive- and careful-treatment sections. This process will likely cause further reductions in the rearing capacity for coho in the future. Therefore, the full extent of the harmful effects of logging on the stream channel, fish habitats, and juvenile coho abundance are yet to be observed two decades after most harvesting ended.

Water Temperature Alterations

Riparian clearcutting has increased stream temperatures in all seasons. Mean monthly temperatures in small, logged tributaries have increased by as much as 4°C in summer. Mean temperatures in the main creek have increased by as much as 3.2°C in August, 0.7°C in December, and 2°C in April. Increases during autumn and winter were relatively small, but had the most profound effects on salmonid populations.

Warmer stream temperatures in fall and winter accelerated egg and alevin development rates that caused sequential changes in emergence timing, growth, survival, and seaward migration timing in both chum and coho salmon. Chum fry emerged and migrated seaward earlier in spring, and were also smaller than they were in pre-logging years. Both reduced fry size and earlier seaward migration have been strongly correlated with reduced ocean survival. Links between freshwater and marine life history processes were thus identified. Increased mortality of chum fry early in their ocean life history was attributed to increased susceptibility to predation due to small size and early-season entry into near-shore waters during winter-like conditions of relatively low salinity and biological productivity.

Post-harvest temperature increases had more complex effects on coho salmon life histories.  During and after logging, coho fry emerged from the streambed up to 6 weeks earlier in spring, permitting 6 more weeks for summer growth than had been available to fry in pre-logging years. The lower numbers of fry rearing in Carnation Creek in most years after logging also resulted in increased growth rates due to density-dependent reductions in competition for food. Coho fry consequently grew 11 mm longer on average by the end of their first summer after logging compared with sizes during pre-logging years (trout fry also increased in mean length by 16 mm after logging). This larger body size was positively associated with improved overwinter survival after logging. Larger coho have apparently been better able to survive winter conditions that include frequent scouring freshets. Increased overwinter survival was responsible for the elevated numbers of smolts produced at Carnation Creek after logging.

Although the creek sustained fewer fry after logging, smolt production increased during and after harvesting 1.6 fold.

Smolt size and biomass also increased. These counter-intuitive relationships are examples of several temperature-related effects upon coho juveniles at least partially attributable to logging.  Increased water temperatures also radically changed the age structure of coho smolt populations.

Prior to logging, approximately one-third to one-half of all Carnation Creek coho required 2 years to grow large enough to transform into smolts and migrate seaward. Increased seasonal growth due to earlier emergence during and after riparian logging has resulted in most coho reaching smolt size and emigrating seaward after just 1 year in fresh water. Age-2 smolts have become relatively rare after logging.

These temperature-related effects upon coho salmon that began in 1976 still persist. Because much of the Carnation Creek channel is more than twice its pre-logging width, these effects will likely continue for several years until a new riparian forest canopy is established over the stream, and both water temperatures and fish growth decline toward pre-harvest levels.

Despite increased smolt abundance and size of 1-year-olds, adult coho returns to Carnation Creek have clearly not increased after logging. Much of this trend appears to be due to reduced marine survival in coho resulting from long-term shifts in marine climate, decreased ocean productivity, and increased predator abundance. However, as for chum fry, marine survival variations are also linked with the temperature-related effects of forest harvesting. Seasonal increases in water temperatures after logging have also shifted the timing of the spring seaward migration of coho smolts from the stream to about 10 days earlier than during the pre-logging period. While this shift appears trivial, most mortality in salmonids in marine environments is known to occur soon after they enter the ocean.  Migration timing and ocean conditions in late winter and spring appear to be critical in determining how many chum fry or coho smolts survive to be adults.

Trophic Shifts

Forest harvesting, including riparian clearcutting, post-harvest slashburning, and ground preparation, has had a variety of effects on aquatic communities. However, some relationships were unclear while other interactions were either beneficial or harmful, particularly to periphyton and aquatic macro-invertebrates. Riparian harvesting has increased both water temperatures and the amount of solar radiation reaching the stream. These changes, together with short-term increases in nutrient availability due to ground disturbance and slashburning, allowed aquatic primary production to increase briefly at Carnation Creek within the during-logging phase of the study. However, this increase was limited by stream channel alterations, which increased streambank erosion, streambed mobility, and sediment transport, and consequently scoured periphyton from the streambed during freshets.

Riparian canopy removal also changed the composition and seasonal pattern of organic litter entering the stream by virtually eliminating the supply of conifer needles between May and December and increasing the supply of deciduous leaf litter from red alders and salmonberry between September and November. The consequences of these changes have been unclear. Project scientists have concluded that benthic macro-invertebrate populations were reduced in the during-harvest phase of this study in areas where riparian clearcutting occurred because of reduced leaf litter input and retention, and increased erosion, transport, and deposition of sand in the benthos. However, these results are not consistent with those from some other studies. Also, the harvest-associated trophic changes at Carnation Creek have not reduced the size or abundance of coho smolts produced from the watershed. The wide variation evident in the literature on the responses of aquatic communities to riparian clearcutting suggests the strong influence of site-specific conditions, including stream size, hydraulic characteristics, gradient, channel and canopy type, and other circumstances.

Species-specific Variations

Fish response to the effects of forestry practices is also influenced strongly by life history strategy. Multiple life history strategies for salmonids during their juvenile stages in fresh water may provide population stability for a species. With one life history strategy each, chum salmon and steelhead trout are influenced exclusively by habitat conditions in the main channel of the creek. Such species have been more strongly affected by forest harvesting than coho salmon, which have employed three life history strategies, including main channel habitation, estuary use, and seasonal use of refuge habitats in the floodplain. Coho may have persisted more successfully at Carnation Creek than the other species because they are less sensitive to main channel habitat loss and alteration.

In spite of the application of harvest practices, which long pre-date those of the Forest Practices Code and the Forest and Range Practices Act, both steelhead and cutthroat trout continue to survive at Carnation Creek in numbers generally comparable to those occurring prior to forest harvesting. Although steelhead smolt production from the stream has fallen in the post-harvest period, perhaps reflecting sensitivity to main channel habitat alterations, their small numbers and high interannual variation render their long-term patterns difficult to interpret.

Application of Carnation Creek Results

The Carnation Creek study is responsible for much of our present understanding of how small Pacific coastal watersheds function, and how forestry practices affect these functions. Over 200 publications have been generated from studies of fish-forestry interactions and watershed processes over the past 36 years at Carnation Creek. The long-term, continuous data sets generated for several biologic, climatic, and watershed physical parameters have allowed the determination of the relative impacts of forest harvesting, climate change, variation in ocean conditions, and fisheries management on its salmonid populations.

Carnation Creek results have been widely used to develop forest management practices, regulations, and guidelines in British Columbia and elsewhere in the Pacific Northwest. The British Columbia Coastal Fisheries Forestry Guidelines (CFFG), which were first implemented in 1987 and guided forestry practices in coastal forest districts until 1994, were based primarily upon research information from Carnation Creek and the Queen Charlotte Islands.

The Carnation Creek study subsequently made substantial contributions toward the development of the hillslope, stream, and riparian management provisions of the Forest Practices Code of British Columbia, which replaced the CFFG in 1995. The FPC
was applied to the entire province, and contained legally binding regulations together with guidelines and recommended best management practices. These provisions remain as default environmental standards under the new Forest and Range Practices Act.

Future Directions

Current research and monitoring at Carnation Creek continues to generate information relevant for validating current forest practices and to support forest policy, regulatory, and guideline development. The ultimate goal is to ensure the sustainable use of forest resources by protecting watershed processes and aquatic values. Forestry practices of high management interest to be validated or changed over time include those that relate to area of harvest (rate-of-cut, ECA, other cumulative-effects parameters), gully and steep slope management, and riparian practices, including retention along small tributaries both in steep non–fish-bearing headwaters and in low-gradient, fish-bearing areas in valley flats.

Studies and results that relate directly to sustainability include:
time to hydrologic recovery of second-growth stands; peak flow response to level of watershed harvesting; landslide risk and prediction for different forestry practices; stream channel morphology and aquatic habitat responses to riparian and upslope harvesting (sediment and debris delivery, channel aggradation and degradation, LWD budgets, and habitat complexity); water temperature changes due to riparian canopy removal and regrowth; and fish population responses to habitat, water quality, and water temperature shifts.

The watershed role of small, low-order streams is an important element of the Carnation Creek study. These channels do not receive mandatory riparian reserves or tree retention in regulation.  They include: (1) steep streams implicated in debris flows and cumulative downstream impacts; and (2) low-gradient valley tributaries that provide critical fish refuges and winter habitats in both the main floodplain and at higher elevations. Presently, with the deterioration of mainstream habitats, low-gradient small streams provide winter habitats that support up to 70% of all coho smolts generated from Carnation Creek.

Our unprecedented 36-year data sets provide the framework required to address difficult management questions: what is the cumulative effect of harvesting beyond specific levels (i.e., 65% basin harvest); what is the effect on both large and small stream channels when the riparian vegetation is removed or modified; what biological changes result from altered stream habitats; and how long do these changes persist?  Achieving sustainability depends on advanced understanding of these watershed relationships connected with forestry impacts and recovery.

Carnation Creek's 36 years of observations have shown that some forestry-related changes, such as water temperature increases and related shifts in salmon growth and survival, happen rapidly, while others may take decades to fully develop. Some effects of riparian harvesting at Carnation Creek, such as long-term reductions in LWD supply to stream channels from harvested riparian zones, will not be fully observable until at least 60 years from now. However, 25 years after harvesting concluded, key changes are occurring rapidly now in spite of tree regeneration and growth in both riparian and upland areas. These prolonged effects need to be described and quantified.

Further forestry-related declines in the quantity and quality of aquatic habitats are anticipated within the next several years due to: (1) increased deciduous canopy closure over the stream channel (resulting in lowered salmonid growth, survival, and smolt output due to suboptimally lowered water temperatures and aquatic primary production); and (2) downstream progression of cumulative impacts of landslide-generated sediment and debris into the part of Carnation Creek, historically protected by riparian buffers, which contains the last remaining high-quality anadromous fish habitats. This demonstrates the long-time course of some harvest-related alterations, and the vulnerability of streams to hillslope failures and debris flows regardless of riparian reserves.

Publications

Contacts

Please direct questions regarding webpage to For.Prodres@gov.bc.ca
Updated March 2008