Changes in stream habitats are clearly related to decreases in fish populations in the South Fork Trinity River basin. However, there has been some debate about how great a role the deterioration of fresh water habitat has played in the decline of salmon and steelhead. Fisheries scientists performed field surveys before and after the 1964 flood and noted a 95% drop in spring chinook salmon runs (LaFaunce, 1967; Healy, 1963; Rogers, 1972; Dean, in press), but some habitats within the basin may have recovered considerably since that time.
Below is an assessment of the quality of fish habitat within the South Fork Trinity River basin, how it has changed over time, and how it might be limiting anadromous salmonid recovery. The physical processes of sediment changes in stream channels are reviewed in Chapters IV and X.
Recent "habitat typing" surveys, funded in part by the Trinity River Restoration Program, provide the most comprehensive data on which to base judgements (see Plate 1 for streams which have been "typed"). CDFG and USFS stream surveys also help to trace patterns of degradation or recovery over time. Comparisons with habitat typing information from other Trinity River and Klamath River tributaries (Wilcox and Johnson, in press) can provide insight into the current productivity of the South Fork Trinity River system. Finally, results from the South Fork Trinity River Watershed and Fisheries Monitoring Program (USFS 1990a, 1991a) are referenced with regard to temperature and fine sediment problems.
To properly manage fish populations, biologists need to know how much fish habitat is available and how fish use that habitat. Older stream survey reports provided useful observations, including assessments of fish abundance and habitat quality, but used a subjective rating system of "good, fair or poor." Consequently, comparing changes in habitat parameters or fish populations over time was not possible using these early surveys. Within the last decade, fisheries scientists have devised the more quantitative inventory technique called "habitat typing" to provide more specific information (Bisson et al., 1981; McCain et al., 1990).
Habitat Units
Even a casual observer of streams can discern between pools (deep areas with little current), runs (slow moving flat water in sections with less depth) and riffles (fast moving water with a broken surface). Bisson et al. (1981) divided these basic units into 21 specific habitat elements. The U.S. Forest Service (McCain et al., 1990) and California Fish and Game (Flosi and Reynolds, 1991) have refined this methodology for use in California. Habitat typing categorizes riffles according to the gradient of the stream, runs according to depth, and pools according to the elements that formed them. The length, width, and average and maximum depth of habitat units are measured. The amount and type of cover, stream gradient, the nature of substrate on the stream bottom, and the number of fish in various habitat types are also noted.
Fish Population Estimation
As habitat typing surveys are being conducted, 20% of all habitat units are surveyed by divers to estimate the number of fish. Since steelhead may spend several years in fresh water, the age of steelhead juveniles is also noted. Two or three divers swim upstream through a habitat unit, count the fish and then compare counts using a technique devised by Hankin and Reeves (1988). Dive counts are compared with estimates derived from electroshocking in some of the same locations, and then calibrated for accuracy.
While electroshocking can give a slightly more accurate count of fish in a specific location, direct observation by divers can be highly effective in clear northern California streams and causes no fish mortality. Hankin (1986) found that basin-wide fish population estimates using electroshocking were often in error by a substantial margin. While electroshocking counts in sample reaches might be very accurate, the sections sampled were not necessarily representative of the entire stream. Using dive counts can provide a more accurate basin wide estimate because much more of the stream is surveyed and fish abundance is correlated to habitat types
Channel Typing
Another habitat parameter which is included in habitat typing reports is Rosgen's channel type (1985). Channel types are classified on the basis of gradient, sinuosity, width, valley confinement, channel bottom substrate type, and bank stability. These various parameters often govern how a stream routes sediment, the abundance and distribution of stream habitats types, and whether treatment with habitat improvement structures is appropriate. Twenty five channel types are described by Rosgen (1985). There are four basic groupings: "A" channels are those which flow through incised valleys, often through hard rock types, "B" channels are moderately confined and of more moderate gradient, "C" channels are generally sinuous and flow through more open valleys, and "D" channels are braided and unconfined.
What Can Habitat Typing Tell Us?
Habitat typing surveys are typically conducted during low flow conditions in summer, to reveal how fish are using habitat at that time. Survey data entered into personal computers and data base management programs allow comparison between reaches within a stream, different streams or within the same stream over time. For example, changes in pool depth could indicate filling due to sedimentation or recovery from past floods. Comparing streams with similar watershed areas and similar geology, but different land use, can reveal changes in stream habitat in response to varying land management. Fisheries biologists within the USFS can also use habitat typing information in interdisciplinary team meetings to evaluate the potential for impacting fisheries resources associated with proposed land use activities.
The Trinity River Restoration Program has funded habitat typing to provide a basic inventory, but also to guide what and where in-stream habitat improvement structures might be used to increase fish populations (Hampton, 1988). The use of habitat typing as a diagnostic tool for stream restoration operates under the assumption that stream conditions during summer low flows are limiting salmon and steelhead production. Data are analyzed to determine which habitat units or features are preferred by juvenile salmonids. Various techniques have been employed to increase favored habitat types. Depending on the channel configuration and stream gradient, pools can be created, cover elements increased, or spawning areas supplemented (see Chapter X).
Problems With Habitat Typing Information and Its Use
While habitat typing surveys provide much more specific information than former stream surveys, some subjectivity can enter into the interpretation of habitat types. Fisheries technicians often work under time constraints; therefore, habitat units may sometimes be lumped together to save time. A short plunge pool segment between two runs may become one long habitat unit called a step run. However, even if this occurs, a considerable amount of useful information is still acquired.
Habitat typing information alone can be misleading. It must be interpreted in a broad context if it is to be used as a tool for prescribing restoration activities. Little fisheries information has been collected in streams during times of high flows or severe cold. Some research now indicates that over-winter survival may be more limiting to production of juvenile salmonids than summer low-flow conditions (Platts and Nelson, 1988). Habitat typing also has little predictive value with regard to changes in stream habitat caused by upslope erosion. If structures are placed in a stream to provide cover or to create pools, any subsequent problems with high sediment production rates or high bedload movement in the channel can negate any possible benefits. Factors such as streamflow and water quality must also be integrated into any limiting factor analysis. For example, it makes little sense to add cover logs to enhance rearing habitat for juvenile salmonids during low flow conditions if stream temperatures exceed the lethal limit for these fish.
Habitat types in a stream may change with changing flow. During low flows in summer, a habitat unit might be characterized as a run due to shallow depth but, during higher flows, the same feature might be interpreted as part of an adjacent pool. Unfortunately, measuring streamflows during habitat typing surveys is not yet standard practice. The use of electrofishing to measure relative accuracy of dive counts also enhances the reliability of comparisons based on habitat typing data. Comparisons between habitat typing surveys on the same stream reaches in different years may be complicated by lack of streamflow data.
Direct observation counts of fish populations are not always checked, using electroshocking techniques, because of the additional time required. However, dive counts may have substantial bias, particularly in small streams. For example, young of the year steelhead (0+) often favor low gradient riffles, but divers have difficulty spotting these fish in water shallower than six inches. An apparent lack of 0+ steelhead in low gradient riffles of cold streams could be the result of sampling error, and not truly reflective of habitat preference.
Fish population estimates acquired during habitat typing surveys are only a snapshot. Low populations can reflect lack of access for spawning adults during drought years or over-harvest in the ocean, rather than lack of spawning or rearing habitat. The number of fish present can influence how fish use habitat, so habitat preference may change in years of varying fish abundance (Chapman, 1966). Therefore, one year of data on fish and their use of habitat is insufficient for drawing conclusions about how to enhance habitat. Knowledge of the life history of the species being restored is also necessary. There may be no need to improve habitat for young of the year steelhead in larger streams if they generally stay in smaller tributaries during their first year of life.
South Fork Trinity River Basin Habitat Typing Reports
Most large sub-basins and reaches of the main stem of the South Fork Trinity River have been habitat typed using the McCain et al. (1990) methodology (Plate 1). The information collected yields an interesting ecological picture of the streams surveyed. When considered in combination with older file reports and other information sources, underlying problems limiting salmon and steelhead production become clearer (Plate 2). Recommendations for restoration measures that were included in the reports can be found in the chapter on Habitat Restoration (Chapter X).
Lower South Fork Trinity River (Grouse to Madden Creek)
Two habitat typing surveys have been carried out on the lower South Fork. During June through August, 1990, Six Rivers National Forest habitat typed the main South Fork from Grouse Creek to Madden Creek (Dale, 1990). Shasta Trinity National Forest personnel surveyed from Coon Creek to Eltapom Creek in August and September, 1991 (Gilroy et al., 1992a). This stream reach over-lapped somewhat with the Six Rivers habitat typing work.
Six Rivers National Forest Survey (1990)
The six mile reach from Madden Creek to Ammon Creek was comprised of 55% main channel pools, 16% lateral scour bedrock pools, and 9% runs, by length. This survey was conducted by raft from August 6-8, 1990. It varied from standard techniques in that widths and lengths of habitat units were visually estimated. Juvenile steelhead of different age classes and a few juvenile chinook inhabited main channel and lateral scour bedrock pools immediately below the mouths of cold water tributaries. Many juvenile salmonids were found to take refuge in the lowest reaches of tributary streams (see Lower South Fork tributary reports). The surface water temperatures in 1990, during the course of the survey, ranged from 74 to 81°F. Six Rivers National Forest has determined through this survey that high stream temperatures are the major limiting factor for fisheries in the lower South Fork Trinity River.
Shasta-Trinity National Forest Survey (1991)
Approximately eight miles of the lower South Fork was covered in this survey which was conducted from 8/20/91 to 9/6/91. About half of the reach (from Coon Creek to Grouse Creek) over-lapped with that covered by Six Rivers N.F. (Dale, 1990). This survey report characterized most of the reach as a Rosgen B-2 channel type, moderately entrenched and confined. A short segment below Eltapom Creek was in a broader valley (Rosgen C-3). In 1991, stream temperatures averaged 69.4°F and ranged from 62-75°F.
Shasta Trinity N.F. staff found main channel pools (28%) to be the most prevalent habitat type by length, with step runs (16%), runs (15%), low gradient riffles (10%), and high gradient riffles (9%) also frequent (Gilroy et al., 1992a). Main channel pools constituted 57% of stream volume, while step runs made up 16%. The deepest pools recorded in any reach of the South Fork were found here. Average depth of main channel pools was 2.1 m (6.8 ft.) and one pool was 9 m (29.1 ft.) deep. Substrate composition in pool habitats averaged 32% sand and fine sediment. Spawning gravels accounted for 13% of all substrate but had extremely high levels of fine sediments (44%) based on visual estimates.
Only nine chinook salmon juveniles were observed during dives, with seven yearlings and just two 0+ noted. Steelhead juveniles were predominantly 2+ (89%), with few 1+ (7%) and 0+ (4%) counted. Low gradient riffles and step runs showed by far the highest concentration of all age classes of steelhead. Deeper pool units and high gradient riffles had substantial numbers of 2+ steelhead as well. The chinook juveniles were found in high gradient riffles and step runs.
Gilroy et al. (1992a) determined habitat preference for salmonids using an electivity index (Jacobs, 1974) which compares the use of habitat type relative to their frequency of occurrence. Only low gradient riffles, high gradient riffles, and step pools showed positive electivities. High water temperatures and low dissolved oxygen levels may have forced salmonids to congregate in turbulent areas where dissolved oxygen would be somewhat elevated. The concentration of fish in so few habitat types may lead to competition and predation, which would account for the extremely low number of 0+ and 1+ steelhead present in the survey reach. The findings of this survey substantiate those of the Six Rivers work;high stream temperature appears to be the principal limiting factor in this reach of the South Fork Trinity River.
High levels of fine sediment in pools noted in Gilroy et al. (1992a) substantiate other evidence that excessive sediment may also be a major limiting factor in this reach. Substantial pulses of sediment pass through this reach even in years of low flow. Dean (personal communication) reported that ten feet of sediment was deposited in one pool during the winter of 1991-92. However, the steep, confined channel of the lower South Fork prevent pools from being completely filled so they remain the dominant habitat category. Numerous personal accounts indicate existing pools are currently much shallower than they were prior to the 1964 flood. Restricted depth prevents formation of cold water strata in most pools, therefore limiting the availability of the pool to serve as summer refuge for juvenile salmonids. High bedload movement may also greatly reduce survival of salmon eggs in the lower South Fork (CDWR, 1982). Chapter 3 continued.