BULL TROUT, SALVELINUS CONFLUENTUS, SPAWNING AREAS WITHIN THE YAKIMA RIVER BASIN, WASHINGTON BYSCOTT DAVID CRAIGAugust, 1997 This study was conducted in four reservoirs containing adfluv
Trang 1SPAWNING AREAS WITHIN THE YAKIMA RIVER BASIN, WASHINGTON
A ThesisPresented toThe Graduate FacultyCentral Washington University
In Partial Fulfillment
of the Requirements for the Degree
Master of ScienceBiology
byScott David CraigAugust, 1997
Trang 3iii
Trang 4BULL TROUT, SALVELINUS CONFLUENTUS,
SPAWNING AREAS WITHIN THE YAKIMA RIVER BASIN, WASHINGTON
BYSCOTT DAVID CRAIGAugust, 1997
This study was conducted in four reservoirs containing adfluvial bull trout,
Salvelinus confluentus, in the Yakima River Basin to quantify spawning habitat at
three spatial scales. The goals were to 1) build a Geographic Information System database containing georeferenced redd (nest) locations with habitat and redd
parameters; 2) quantify abiotic features within watersheds that may influence bull troutredd densities; 3) describe lotic stream reach features associated with spawners; 4) compare physical characteristics of redds within three levels of spawner density; and 5) estimate hatching and emergence timing of alevins
Redd distributions indicate that spawners utilize upper sections of available habitat more than lower areas, especially in larger watersheds. Watershed area and baseflow discharge can be used to predict adfluvial bull trout redd counts within watersheds while redd densities among stream reaches were significantly correlated with thalweg components of overhead cover and depth. Hatching and emergence periods varied considerably among watersheds because of differences in spawning timing and water temperature during incubation
iv
Trang 5of the trout, they live in the coldest, cleanest and most secluded waters.
No higher praise can be given to a Salmonid than to say, it is a charr.
(Jordan and Evermann, 1896)
This project was not possible without the help of many people. I would like to
thank my major advisor Dr. Paul W. James, for allowing me this opportunity to study this project and my committee members, Dr. Kristina Ernest and Dr. Douglas Reynolds for their help and guidance in writing the thesis. A grant from the Yakima Valley
Conservation Club was instrumental in providing me with funds to buy much needed waders as well as other field supplies, their generosity also allowed me to attend a bull trout conference in Oregon. Scott Hoffer and Tina Mayo from the U.S. Forest Service provided many thermographs as well as some critical winter transportation in retrieving them! Redd surveys were organized by Eric Anderson and Jim Cummans from the
Washington Department of Fish and Wildlife. Many thanks go to ARC/INFO guru,
Ken Rauscher, for help with the redd database coverage which I hope will be used for future research/enhancement projects of bull trout in the Yakima River Basin. Most of all, I wish to thank Laurie Connell for helping me in every aspect of my graduate
education. Your encouragement from beginning to end as well as countless hours in the field under extreme conditions made this thesis possible
Bull Trout Salvelinus confluentus
v
Trang 69Coal Creek
11Mineral Creek
11Gale Creek
11
vi
Trang 7Deep Creek 12
Bumping River 13
Rimrock Reservoir
13
Indian Creek 14
South Fork Tieton River 15
Bear Creek 15
Short & Dirty Creek 16
North Fork Tieton River 16
Clear Creek 17
III. METHODS and MATERIALS
18
vii
Trang 8Data collection
21
Data analysis
24
Table of Contents (continued)
IV. RESULTS
28
Redd counts and densities
28
Temporal redd distribution
31
Distribution of redds within watersheds
33
Watershed variables
39
Climatic conditions
44
Stream reach variables
46
Microhabitat (redd) characteristics
viii
Trang 9Estimated hatching/emergence periods
49
V. DISCUSSION
51
Redd counts and densities
51
Temporal redd distribution
52
Watershed variables
52
Stream reach variables
55
Microhabitat (redd) characteristics
57
Estimated hatching/emergence periods
59
Conclusions
60
VI. LITERATURE CITED
62
ix
Trang 10x
Trang 111. Sources of GIS coverages utilized or created for project 19
2. USFS level2 stream survey source and year of data collection 20
3. Variables measured at redd sites 23
4. Macrohabitat variables utilized in regression analysis with mean redd densities 25
5. Stream reach variables utilized in mesohabitat (stream reach) analysis 27
6. Watershed areas, available spawning habitat length and area with corresponding
longterm mean redd densities for all study areas 30
7. Best model (stepwise regression) in order of significance for correlation of
watershed variables with bull trout redd density by watershed area, wetted
habitat length and area 43
8. Best model (stepwise regression) in order of significance for correlating bull
trout redd density by watershed utilizing six highly correlated variables within
the Yakima River Basin 44
9. Best model (stepwise regression) in order of significance for correlating stream
reach variables with bull trout redd density by area and wetted habitat length in
all spawning reaches in the Yakima Basin 48
xi
Trang 121. Yakima River Basin and study watersheds 6
2. Elevation profile of the upper Yakima River main tributaries from confluence
of the Upper Yakima and Naches Rivers to lake reservoirs in study 7
3. Longterm mean redd counts (with 95% confidence intervals) of adfluvial bull
trout in six watersheds of the Yakima River Basin, WA (data from WDFW) 28
4. Temporal distribution of bull trout redd occurrence, indicating spawning activity,
with mean daily water temperature within six watersheds of the Yakima River
Basin in 1996 31
5. Bear Creek bull trout redd distribution in 1996 33
6. Box Canyon Creek bull trout redd distribution in 1996 34
7. Deep Creek bull trout redd distribution in 1996 35
8. Indian Creek bull trout redd distribution in 1996 36
9. Gold Creek bull trout redd distribution in 1996 37
10. South Fork Tieton River bull trout redd distribution in 1996 38
11. Relationship between longterm mean redd counts and watershed area (km2) 40
12. Relationship between longterm mean redd counts and available habitat
length (km) 40
xii
Trang 1314. Relationship between longterm mean redd counts and baseflow (m3 sec 1 km 2) 41
15. Relationship between longterm mean redd counts and maximum 7day mean summer water temperature (oC) 42
16. Relationship between longterm mean redd counts and maximum summer water temperature (oC) 42
17. Monthly air temperature at Keechelus Lake Dam in 1996 compared with longterm mean air temperature (data from USBR) 45
18. Monthly precipitation at Keechelus Dam in 1996 compared with 58year average and 95% confidence interval (data from USBR) 46
19. Mean daily water temperature (oC), peak spawning and estimated hatching
emergence times for bull trout in Yakima River Basin watersheds (199697) 50
xiii
Trang 141. Nominal and numeric redd densities by watershed stream reach 69
2. Baseflow, peakflow, minimum 7day summer water temperatures and mean values of physicalchemical water parameters from study watersheds in 1996 71
3. Road density, mean annual precipitation and percentage of watershed in three rainonsnow category areas 72
4. Drainage density, total channel length and stream order percentage of study watersheds 73
5. Percentage of USFS landtype associations in each watershed 74
xiv
Trang 15Bull trout, Salvelinus confluentus, is the only native charr species found in the
Yakima River Basin (Mongillo 1992). Recently, the U.S. Fish and Wildlife Service hasproposed to list the Klamath River population (southcentral Oregon) as endangered and the Columbia River population of bull trout as threatened (U.S. Fish & Wildlife Service 1995). The U.S. Department of Agriculture Forest Service (USFS) considers these charr to be "indicator species" for ecosystem management and they are
recognized as a "species of special concern" by the American Fisheries Society as well
as several State Management agencies (Batt 1996; Washington Department of Wildlife 1993). Charr are very susceptible to increased water temperature regimes (>15 oC), caused by global warming trends (Meisner 1990), increased fine sediments (Shepard et
al. 1984) and glacial silt or turbidity (Aass 1981). Bull trout were endemic to the entire Yakima River drainage prior to the creation of dams but increased water temperatures from dams throughout the system as well as logging in the mountain forests have greatly reduced their original habitat (Mongillo 1993).
Bull trout, cutthroat trout, Oncorhynchus clarki, and mountain whitefish, Prospium williamsoni, probably colonized the Yakima River drainage prior to the arrival of
Trang 161 anadromous salmon (Meehan and Bjornn 1991). Bull trout coexist with other
Salmonidae by occupying a different niche and tend to be more piscivorous than other trout and salmon. Their preferred habitats include cool waters (<15 oC) of headwater streams, rivers and lakes where they can move from natal streams into larger rivers or lakes depending upon the population type (Bonneau and Scarnecchia 1996; Fraley and Shepard 1989; Goetz 1989; Pratt 1984; Rieman and McIntyre 1993; Thiesfeld et al. 1996).
There are four population types, or life history scenarios, of bull trout based upon available habitat. 1) Adfluvial bull trout live from one to four years in their natal stream before migrating to large lakes (>5 km2 in the Yakima River Basin) where they mature and take advantage of more diverse and abundant food supplies. Examples of these populations include the charr of Keechelus, Kachess, Rimrock and Bumping Lakes. 2) Fluvial bull trout utilize natal streams for rearing much like their adfluvial counterparts, but reside in large streams and rivers after maturation. Some of the larger streams in the Yakima River Basin such as North Fork Teanaway River, Rattlesnake Creek, Ahtanum Creek and upper Naches River have fluvial populations. Only two bull trout have been recorded in the mainstem Yakima River below Keechelus Lake since 1993. One fish was observed in a fish weir 100 m above the mouth of Swauk Creek in 1993 and another was illegally harvested from Lake Easton in 1996 (Anderson1997). This small reservoir (<1 km2) is located at the junction of Kachess and Yakima Rivers. 3) Resident bull trout are found in small tributaries at high elevation and live intheir natal streams for their entire life cycle. They are normally located upstream of fish barrier falls and anadromous fish populations. 4) Anadromous bull trout rear in
Trang 17Identifying lotic characteristics that are relevant for sustaining bull trout persistence
is important because habitat loss leading to fragmentation is one of the key problems associated with bull trout declines (Leary et al. 1993; Rieman and McIntyre 1993) and this trend is clearly evident in the disjunct populations of the Yakima River Basin(Mongillo 1993). Studies testing island biogeography theory have shown that aquatic organisms are restricted by the spatial boundaries and patterns of their environment(Broenmark et al. 1984; Browne 1981; March and Bass 1995; Rieman and McIntyre 1995; Shaffer and Samson 1985; Villa et al. 1992). Natural features such as geology, geomorphology and climate also provide the physical framework for species
recruitment and sustained residence (Resh et al. 1988). Physical watershed
characteristics such as elevation, gradient, drainage area and stream order have been significantly correlated with the number of fish species collected in the Cascade
Mountains of Washington State (Beecher 1988). Geologic bedrock type has been shown to influence trout densities in the Flathead River Basin, Montana (Fraley and Graham 1981). Geology can also be a factor involved in increased clay and sand bedloads in salmonid spawning areas (Ducan and Ward1985). Timber harvest and roads have also been tied to degradation of trout spawning areas in Washington
(Cederholm and Salo 1979; Reid 1981), Montana (Shepard et al. 1984) and Idaho(Cross and Everest 1994). Physical parameters such as geology, hydrology and microclimate of watersheds have yet to be correlated with bull trout spawning areas within the state of Washington. The diverse climate and geomorphology of the Eastern Cascades of central Washington provide a plethora of conditions from which to seek abiotic features impacting spawning adfluvial bull trout. These conditions could also
Trang 18in making decisions by providing a baseline record for future management activities. Utilizing salmonid spawner redd counts as a descriptor of current and future population size has been acknowledged (Beland 1996) and this premise will be the basis for providing a unit of bull trout density as well as denoting the presence or absence of the species in the study watersheds. This density indicator will then be utilized to ascertain
if relationships exist between abiotic watershed features and bull trout abundance. The objectives of this study were to 1) build a Geographic Information System database containing georeferenced redd (nest) locations with habitat and redd
parameters; 2) quantify abiotic features within watersheds that may influence bull troutredd densities; 3) describe lotic stream reach features associated with spawners; 4) compare physical characteristics of redds within three levels of spawner density; and 5)estimate hatching and emergence timing of alevins
Trang 19in the Yakima River Basin. Fourteen watersheds larger than 13.5 km2 that have present
or potential access by adfluvial bull trout spawners were examined (Figure 1).
The Yakima River drainage (15,997 km2) contains nearly ten percent of the state area, the largest basin wholly located within the state of Washington. The drainage has two major branches, the lower gradient Upper Yakima River (5,476 km2) and the higher gradient Naches River (2,901 km2) (Figure 2). Both headwaters begin along the high peaks of the Eastern Cascade Mountain range and flow south and east before converging at the city of Yakima, WA. The mouth of the Yakima River converges withthe Columbia River approximately 17 km north of the Snake River confluence.
Trang 205
Trang 21Figure 1. Yakima River Basin and study watersheds.
Trang 23oligotrophic lake has a spillway elevation of 767 m above sea level, an area of 10.5
km2, a volume of 308,500,000 m3 and a maximum depth of 108 m (Dion et al. 1976; Goodwin and Westley 1967)
No fish passage facilities were constructed for the anadromous or migratory
salmonids that utilized the Keechelus Basin for spawning, rearing or adult habitat prior
to dam construction. Natural populations of spring chinook salmon, Oncorhynchus tshawytscha, coho salmon, O. kisutch, sockeye salmon, O. nerka, and steelhead trout,
O. mykiss, have been extirpated from this lentic system (Tuck 1995). Today the main game fish is kokanee salmon, O. nerka, but limited numbers of rainbow trout, O. mykiss, cutthroat trout and brook trout, Salvelinus fontinalis, are also harvested. The burbot, Lota lota, a large piscivorous fish, competes with bull trout as the toplevel
consumer within the lake (personal observation)
Gold Creek
Gold Creek is a thirdorder stream that covers 35.2 km2 and is the only recorded spawning tributary for the Keechelus bull trout population. The upper Gold Creek Basin is located in the Alpine Lakes Wilderness Area but the lower portion is highly degraded by Keechelus Reservoir inundation, past timber harvest practices and gravel excavations resulting in formations of several ponds or small lakes. Gravel for the interstate (I90) road bed construction in the late 1970s was excavated from the flood
Trang 24A fish barrier falls is located approximately 11.4 river km upstream from the Interstate I90 bridges, but a bedrock cascade at 8.5 km may impede spawning
migration. The highest densities of redds are normally found in a springfed channel near 6.6 km and in the immediate vicinity of the ephemeral Silver Creek located at 7.3
km (Wissmar and Craig 1997).
Meadow Creek
Meadow Creek is a secondorder stream that covers 21.8 km2 before draining into the southwest corner of Keechelus Lake near the dam. No bull trout or redds have beenrecorded in Meadow Creek. Recent surveys in 199596 detected no salmonids of any kind above a series of barrier falls located 6.4 km above the lake (Unpublished data, USFS, Wenatchee, WA). Timber harvest, including clearcutting throughout the watershed, has opened up many riparian areas to increased light levels.
Coal Creek
Coal Creek is located within the Snoqualmie Pass summit trough and this
geomorphic feature produced the lowest mean basin elevation of all the watersheds in this study. This thirdorder tributary (14.2 km2) has clearly been modified
Trang 25observation).
Kachess LakeThe adfluvial population of bull trout in Kachess Lake is ranked as "High Risk" by the WDFW (Mongillo 1993) due to the low number of spawning adults (<5 redds year 1). Kachess Lake actually consists of two natural basins, the large main lake and Little Kachess Lake, situated directly north of the main lake. The main lake was impounded in 1912 to increase the water supply for irrigation in the lower Yakima Valley. Today, at full capacity, the spillway elevation is 687 m above sea level and the volume is 9,854,926,000 m3 (Goodwin and Westley 1967). The main lake has an area
of 18.4 km2 and maximum depth of 144 m. Little Kachess has an area of 2.9 km2 and maximum depth of 91 m. Both basins are considered to be oligotrophic (Goodwin and Westley 1967). The main lake basin is located 6 km due east of the Keechelus Dam.
No fish passage facilities were constructed for the anadromous or migratory
salmonids that utilized the Kachess basin for spawning, rearing or adult habitat prior to dam construction. Natural populations of spring chinook, coho and sockeye salmon and steelhead trout have been extirpated from the Kachess Lakes (Tuck 1995). Today the main game fish is kokanee salmon, but limited numbers of rainbow trout, cutthroat
trout and brook trout are also harvested. Lake trout, Salvelinus namaycush, were
Trang 26Box Canyon Creek
Box Canyon Creek is a thirdorder tributary (31.6 km2) and the only known
spawning tributary for the adfluvial bull trout of the Kachess Lakes. Very few
spawners return to Box Canyon Creek (<5 redds year 1). The upper basin is located in the Alpine Lakes Wilderness Area, but spawning bull trout are restricted to the lower 2.5 km of stream channel due to a natural barrier falls. The bull trout of Box Canyon Creek are incredible vertical navigators. Several threemeter vertical jumps are
required to reach spawning habitat (personal observation)
Subsurface channels in the exposed Little Kachess lake bed have been recorded practically every year in which precipitation is belownormal and 1996 was no
exception
Mineral Creek
The Mineral Creek watershed (28.5 km2), a thirdorder tributary, is defined as all channels located above Little Kachess Lake. This includes the Kachess River which is actually a small tributary to Mineral Creek. The creek contains 3.4 km of available stream habitat but much of the area below the Kachess River confluence contains degraded stream habitat similar to lower Gold Creek. Subsurface channels in this reachare common. In 1996 subsurface channels were first observed on July 23 and persisted until the last visit on October 21 (personal observation)
Trang 27Gale Creek, a thirdorder tributary (16.8 km2), contains over 2.7 km of available spawning habitat. A 1996 snorkel survey yielded no bull trout, but numerous cutthroat trout and brook trout were observed (personal observation). During this survey, no fishbarriers were encountered that would present an obstacle to a large bull trout. Spawner access to Gale Creek would have to be accomplished before the lake recedes because low baseflow discharge in August and September would prevent fish from ascending the exposed lake bed near the Kachess campground.
Bumping LakeThe adfluvial population of bull trout in Bumping Lake was listed as "Unknown" in
This natural lake was impounded in 1910 to increase the water supply for irrigation
in the lower Yakima Valley. Today, at full capacity, the spillway elevation is 1044 m above sea level with an area of 5.17 km2. Bumping Lake is located approximately 90 river km upstream from the YakimaNaches confluence
No fish passage facilities were constructed for the anadromous or migratory
salmonids that utilized the Bumping River Basin for spawning, rearing or adult habitat
Trang 28harvested. Bull trout are the toplevel consumer within this lake.
Deep Creek
Deep Creek is a thirdorder tributary that has 5.6 km of spawning habitat up to a natural barrier falls. In 1996 spawners were restricted to only 2.0 km of the stream due
to dewatered channels. Recent escapement surveys by the WDFW have indicated a fluctuating number of redds (12100). Several stream segments become dewatered during periods of belowaverage precipitation (Anderson 1997)
Bumping River
Bumping River is a fourthorder tributary that has only 1.5 km of habitat available for spawners because of a natural barrier falls. No spawning activity was observed in
1996, (personal observation) although USFS personnel have reported redds in the past. The vast majority of the basin is located in the William O. Douglas Wilderness Area and therefore contains no roads. A trail crosses the river directly above the barrier falls and parallels the upper river
Rimrock ReservoirThe adfluvial population of bull trout in Rimrock Reservoir is rated as "Low Risk"
by the WDFW (Mongillo 1993) due to the high number of spawning adults (>100 redds year 1). Rimrock Reservoir was created in 1925 to increase the water supply for irrigation in the lower Yakima Valley. Today at full capacity, the reservoir
Trang 29No fish passage facilities were constructed for the anadromous or migratory
salmonids that utilized the upper Tieton River Basin for spawning, rearing or adult habitat prior to dam construction. Natural populations of spring chinook and coho salmon as well as steelhead trout have been extirpated from their original lotic habitat(Tuck 1995). Today the main game fish is kokanee salmon, but limited numbers of rainbow trout, cutthroat trout and brook trout are harvested. Bull trout are the toplevel consumer within the lake
Clear Lake dam located immediately above Rimrock contained no fish passage facilities until 1994. The absence of fish passage over the last 60 years has probably eliminated or drastically reduced bull trout from both the North Fork Tieton and Clear Creek. The new fish ladder has not been monitored and thermal barriers may deter bulltrout from utilizing the structure. Recruitment of large numbers of spawners into the Clear Lake Basin appears unlikely, because recent tagging projects at Indian Creek and South Fork Tieton have shown no straying by spawning adults between these
watersheds (Unpublished data, Paul James, Central Washington University, 1996)
Indian Creek
Indian Creek, a thirdorder tributary (50.0 km2), contains the largest number of bull trout spawners by watershed area in the entire Yakima River Basin. The upper basin is located in the William O. Douglas Wilderness Area but the lower watershed is
composed of private ownership and a large USFS campground. Spawners are limited
to 8.1 km of habitat below a natural falls. Spawning substrate is extremely limited above the falls and this may be one indication of why no redds have been recorded
Trang 30channel. This channel has consistently contained high redd numbers, even after the
1996 flood devastation. This main spring is adjacent to a road that contains the
trailhead and primitive camping area and the mouth of this channel is located about 3.5
km upstream of the lake. Two other spring channels are utilized as spawning habitat and they are located on the opposite bank from the main spring. The southern spring mouth is located at 4.3 km while the much longer northern spring is located 5.2 km upstream from the lake.
South Fork Tieton River
South Fork Tieton River is a large fourthorder tributary (189.1 km2) that flows intothe southeast end of Rimrock Reservoir. The South Fork contains over 27 km of habitat up to a large natural barrier falls but spawning is normally restricted to the upper
17 km. During reservoir drawdown, a nonnatural fish barrier falls forms at the TietonRoad bridge. It appears not to affect bull trout spawning because the fish migrate before the falls develops in late September when the water is released from the
reservoir. Earlier water releases or belownormal pool elevation could hinder access to spawning areas. No postspawning mortalities have been observed from immigrating fish in the last two years (personal observation)
South Fork Tieton River originates in several glacial fields of Old Snowy Mountain within the Goat Rocks Wilderness Area. Although the South Fork is a pristine river,
Trang 31Bear Creek
Bear Creek, a small thirdorder tributary (16.9 km2) of South Fork Tieton, is an extremely high redd density area within the South Fork Tieton River Basin. Spawning
in Bear Creek is restricted to 0.93 km of habitat below a natural barrier falls. Steep gradients and low flows often constrict spawners to less than 0.3 km of habitat because
North Fork Tieton River
North Fork Tieton River is a large thirdorder (105.1 km2) tributary that flows into the western end of Rimrock Reservoir via the Clear Lake Reservoir (0.93 km2) which was built prior to the Rimrock dam. The North Fork, like the South Fork Tieton,
Trang 32temperatures reduce glacial runoff. The North Fork contains over 16 km of accessible spawning habitat. A natural barrier is located approximately 16.4 km above Clear Lakebut migration could be curtailed by a marshyswamp stream environment located 3 km below the barrier falls
One adult bull trout was observed 0.2 km below the confluence with Scatter Creek during a snorkel survey in the Fall of 1996 (personal observation). The last observation
of bull trout in the North Fork was in 1973 (Anderson 1997).
Clear Creek
Clear Creek is a thirdorder tributary (45.4 km2) of Clear Lake. This stream
experienced channel inundation by alluvial sediments from the 1996 floods. A large spring located 3.0 km above the lake provides the majority of discharge during baseflow periods. Completely dewatered channels upstream of the spring were observed in the Fall of 1996, suggesting that spawning areas would be restricted to the habitat below the spring.
Trang 33Geomorphic, climatic, anthropomorphic (roads and trails) and stream habitat indices were computed for each watershed by the use of GIS coverages and stream survey reports. The information assembled for this aspect of the project was compiled from a variety of sources. Locations of the data sources are presented with descriptions of methodologies
Existing data sourcesMap coverages or GIS themes were obtained from the United States Forest Service (USFS) Wenatchee National Forest Supervisor's Office located in Wenatchee, WA and the Central Washington University (CWU) GIS laboratory (Table 1). Map coverages are described by the following attributes and features. Roads and trails are similar to features observed in USFS roads and trails maps as well as United States Geological Survey
(USGS) 7.5minute quadrangle maps. Channel lengths displayed in USGS 7.5minute maps are similar to those within the stream coverage. Stream order classification was inferred by the methods of Strahler (1957) and edited into the stream coverage. The precipitation coverage depicts the estimated yearly water equivalent of rain and snow(Miller et al. 1973). Peak rainonsnow zones represent general snow accumulation zones that are based on average amounts of snow on the ground in early January, relative to the amount that could be melted during a model storm event. Five snow zones (based on climate, elevation, latitude and vegetation) are defined within this coverage. The
Trang 34regolith/edaphic composition and characteristics (McColley 1976; Snyder and Wade 1973).Watersheds were digitized by delineated tributary areas from USGS 7.5minute quadrangle maps. Redd locations were spatially located by the use of a Garmin45 global positioning satellite (GPS) and entered into GIS as coverages. Spatial resolution was + 35 meters.Leveltwo and levelthree stream survey reports were obtained from the USFS district offices to ascertain wetted channel area or spawning habitat area. Surveys were obtained from the North Bend, Naches and Cle Elum USFS District Ranger Stations (Table 2)
Table 1. Sources of GIS coverages utilized or created for project
Map coverage theme Coverage origin Publication or source
Precipitation CWUGIS laboratory (Miller et al. 1973)
RainonSnow CWUGIS laboratory DNR Geology Division
Landtype association USFS Wenatchee, WA USFS Wenatchee, WA
Trang 35The U.S. Bureau of Reclamation (USBR) Lake Keechelus weather station (elevation
754 m) was utilized to determine if air temperature and/or precipitation anomalies occurred during the 1996 study period. The data sets contain monthly air and precipitation (water equivalent) values from 1944 to 1994. These data were obtained from the USBR hydromet database
Data collection
Trang 36on three types, "Definite", "Probable" and "Possible" (Washington Department of Wildlife 1993). "Definite" redds are distinguished as those structures that are unquestionably viable redds which have either been positively identified with actively spawningdigging fish or contain the appropriate redd dimensions and substrate compositions. "Probable" redds are defined as those structures that contain two or more of the appropriate of
pit/tailspill/substrate compositions but have not been positively identified with actively spawning/digging fish. "Possible" describes structures that could be redds, but are created
by factors such as stream hydraulics or testdigs from spawners that contain nonviable or
no eggs within the redd. Only definite and probable redds are tallied for the final WDFW count. Redds were marked by placing surveyor's flagging near the spawning location. Three or more surveys were conducted on streams that were identified as bull trout
spawning "index" areas by the WDFW. The remaining study streams were either surveyed once or not at all because of time constraints. Streams not surveyed for redds were
snorkeled in the summer of 1996 to determine the presence/absence of juvenile bull trout. Streams in this category were Meadow Creek, Gale Creek and North Fork Tieton River. Discharge, pH, turbidity and conductivity were measured bimonthly from June to October, 1996 within a twoday period among all watersheds. One sampling station was chosen for each watershed. Locations were selected to obtain maximum discharge per unit area. Sites were generally characterized by a single constricted stream channel containing impermeable substrates such as bedrock or other coarse materials. If stream channels did not conform to the above criteria, discharge locations were selected in areas not associated with late summer downwelling or dewatering of stream channels. Discharge was
measured by the transect method with a MarshMcBirney model flow meter and calculated
Trang 37of the gauge height to the calculated discharge was used to approximate discharge when theflow meter method was not available. Water parameters of hydrogen ion concentration (pH) and conductivity, recorded in units of milivolts DC (mV), were measured in the field
by the use of a Hanna Instruments Incorporation water analyzer that was calibrated
monthly. Turbidity, measured in nephelometric turbidity units (NTU), was determined using a LaMotte model 200B meter
Water temperature was collected at each discharge location by the use of Onset digital recording type thermographs. The units were preset to record at evenhour intervals (12 readings per day) and were placed in streams during March of 1996 and removed in AprilMay of 1997.
Each stream or tributary occupying more than 1,350 ha was stratified into one or more reach breaks. Reach breaks were delimited by changes in gradient, geomorphology or tributary junctions. Variables computed for stream reaches were channel gradient,
sinuosity and random densiometer readings (overhead cover). Wetted habitat area and length were calculated for each reach so that 1996 redd counts could be converted to densities.
Up to 30 redds per stream reach were sampled for five mesohabitat (channel) and nine microhabitat (redd characteristics) variables. Mesohabitat variables measured at redd locations included bankfull and wetfull channel widths, thalweg depth and velocity as well
as thalweg densiometer (% overhead cover) readings. Microhabitat variables include pit and tailspill dimensions, overhead cover, tailspill velocity (0.6 depth), pit and tailspill depths and redd substrate composition. The substrate composition index was a
measurement of the mean crosssectional diameter of randomly selected redd particles
Trang 38Channel variable measurements were based on USFS, Stream Inventory Handbook (1993) protocols. Measurements of all redd variables are similar to the methodology of Thurow and King (1994). Pit and tailspill lengths were measured parallel with the stream flow. Pit length was the distance between the upstream edge of the pit and the upstream edge of the tailspill. Tailspill length was the distance between the upstream edge of the tailspill and the downstream terminus of disturbed gravels. A single measurement of mean pit and tailspill widths were measured perpendicular to stream flow. In all, fourteen variables were measured at redd sites (Table 3).
Table 3. Variables measured at redd sites
Factor Variable Measurement unit Method or description
Meso Thalweg densiometer % overhead cover Spherical densiometer
Meso Thalweg velocity cm sec 1 0.6 depth, Flow meter
Micro Redd densiometer % overhead cover Spherical densiometer
Micro Tailspill velocity cm sec 1 0.6 depth, Flow meter
Trang 39Spreadsheet organization and statistical analysis were performed with Excel 4.0 and
Statview 4.0 software
The longterm redd counts for each watershed were averaged from WDFW data for the macrohabitat analysis. Historical surveys utilized the same protocols and covered the same areas as the 1996 redd surveys.
Mean redd counts for the macrohabitat analysis were converted to three units of
density: 1) redds per watershed area (km2), 2) redds per available habitat length (km) and 3) redds per available habitat area (ha). Available habitat was defined by all stream
channels below the first fish barrier, even if some stream reaches were dewatered or not utilized by spawners during the 1996 redd surveys
Redd densities utilized for the mesohabitat analysis were calculated by dividing the number of redds identified during the 1996 surveys into a unit of length (km) or area (ha) for each stream reach. Nominal redd densities (high, medium and low) were also
calculated to determine differences among spawning reaches within watersheds. Reaches were ranked within watersheds so that each watershed contained all nominal density levels. Watersheds that contained fewer than three spawning reaches (Bear and Deep Creeks) had their highest density reach classified as high, with the next reach classified as medium. Box Canyon Creek was not included in this analysis because only 8 redds were observed and a high nominal rating did not fit well when compared with other watersheds.
In 1996, multiple redd counts were obtained for each of the watersheds used as index areas by the WDFW. Temporal distributions on the majority of redds (some flagging dates
Trang 40Simple linear regression analysis was utilized to determine which macrohabitat
variables were significantly correlated with the longterm mean redd densities. Sixteen variables were utilized for this aspect of the macrohabitat analysis (Table 4).
Table 4. Macrohabitat variables utilized in regression analysis with mean redd densities
unit
Method of calculationRoad density km km 2 Total road length/watershed area
Drainage density km km 2 Total stream length/watershed area
Stream order density % total length Stream length by order/stream length
Water temperature* oC Maximum mean weekly water temperatureWater temperature* oC Maximum 2hour summer water
temperatureBaseflow m3 sec 1 km 2 Lowest discharge per unit area (Jun. Oct.)PeakFlow m3 sec 1 km 2 Highest discharge per unit area (Jun. Oct.)
length
km Wetted habitat length to barrier fallsAvailable habitat area hectares Wetted habitat area to barrier falls
* Mean weekly=7day average water temperature, 12 readings per day
** Two readings per month, all watersheds measured within a twoday period