Restoration Program

Program Administration

Session: Fish and Avian Habitat

Wednesday, February 7, 2007 - 2:00 to 3:00 PM

Trinity River Salmonid Life Histories: What we know, and what we need to know

Aaron Martin, Fisheries Biologist II, Yurok Tribal Fisheries Program, Weitchpec, CA
yurokfish@gmail.com, (707) 834 2595

Presentation [PPS - 2.5 mb]
Video [MOV - 9.3 mb]

To date there has been a considerable effort put forth into developing Habitat Suitability Criteria for Chinook salmon, coho salmon, and steelhead on the Trinity River . Salmonid fry on the mainstem Trinity River begin to emerge in late December (spring Chinook salmon) and continue through May/June for Steelhead. Therefore, the highest Concentrations of juvenile fish are seen from February through June in the areas where spawning is the heaviest. It is understood that salmonid fry utilize shallow low velocity habitats where structural cover is present, especially Chinook salmon and coho salmon fry. As they grow, coho salmon juveniles remain tight to or within the structural cover. Chinook juveniles move out into the slightly deeper eddies or edgewaters in search of food as they grow. Preferred habitats for both species include small woody debris, grasses, aquatic vegetation, exposed roots and young willows found along the edge of the water or in backwaters and side channels. This much is 'mostly' understood. However, many questions still exist; how long do naturally reared Chinook, coho and steelhead stay in the system (or the upper 40 miles)? What effects are hatchery or resident fish (residual steelhead, brown trout) having? Where do overwintering steelhead and coho salmon go? Does the current flow regime maximize natural production and survival of these fishes? Our efforts need to focus on these questions in order to improve our overall understanding of this dynamic system. The answer to these and other important questions will get us closer to recognizing the TRRP's goals.

Presentation notes:

60 % of spawning occurs in the first few miles of river below Lewiston Dam. Juveniles use slow water areas with cover; fry prefer depths < 3 feet and velocities < 1 fps. Salmon stay within 1 foot from escape cover. They like woody debris, grasses, exposed roots, aquatic vegetation, and young willows. They did not find fish at Hocker Flats as there is no escape cover. One predator is the large brown trout, also the hatchery steelhead (many do not leave until flows come up). Martin recommends fish need sufficient habitat velocity and escape cover. The Trinity River Flow Evaluation did not consider cover in their habitat suitability. He doesn't agree that flows 450 to 1500 cfs are necessarily bad for juveniles as it can offer access to additional cover. They snorkeled Lewiston at 800 cfs last year and found "happy" fish. There are high numbers of juvenile Chinook, coho and 0+ steelhead in upper 5-10 miles of river. Don't see 1+ or 2+ steelhead in upper river. Don't see over-wintering coho in mainstem. Don't see juvenile in deeper open water except for older Chinook.

Questions: Upper 5-10 miles are now the "headwaters"; fish need escape cover, predators are a problem. We don't know where the over-wintering coho and steelhead go. There hasn't been surveys of the tributaries and they have not done night surveys. Fish are not thought to be very active at night in winter. They haven't see too much stranding. Higher base flows in winters may be desirable, but how to accomplish? Can't carry over water from prior years. but some water is available from the high flow hydrograph.perhaps use a part of the 11,000 cfs to augment baseflow. These higher flows allows better access to cover than low flows. Why aren't steelhead seen more? Maybe they go downstream to Junction City , it is currently a mystery of where they go.

^ TOP

2-D Habitat Modeling: Predicted Versus Observed Habitat and Comparison with Fish Utilization

Dr. Thomas B. Hardy, Utah State University, Hardy@cc.usu.edu, (435) 797-2824
Dr. Ekaterina Saraeva, Utah State University, Saraeva@cc.usu.edu, (435) 797-8782

Presentation [PPS - 6 mb]

A two-dimensional hydrodynamic model was utilized to estimate the depth and velocity characteristics associated with pre- and post-restoration activities at Hocker Flat, Trinity River. Pre-restoration channel topography was derived from a combination of aerial photogrammetry, LIDAR, and surveyed cross sections. Aerial photographs were also used to digitize both substrate and vegetation polygons that were used to assign spatially variable roughness for the hydrodynamic model and substrate/vegetation for use in the habitat modeling of target anadromous species and life stages. Post-restoration topography was derived from measured cross section geometries and design plans for the restoration site. Calibration data for both pre- and post topographies consisted of three-discharge and water surface elevation profile surveys. The efficacy of the hydrodynamic modeling under both the pre- and post-restoration conditions was assessed by a comparison of the predicted versus observed water surface profiles through the study reach. Fish observation data were provided for the study site for both pre- and post-restoration conditions at several flow rates. Habitat suitability criteria for use in computing physical habitat were derived using binary relationships for depth, velocity, and substrate/cover developed for use as part of an Expert Habitat Mapping (EHM) component of the technical restoration studies. Chinook, coho, and steelhead fry, chinook young-of-year, adult anadromous holding, and adult anadromous spawning life stages were evaluated for both pre- and post-restoration conditions. Fry and juvenile habitat modeling relied upon depth, velocity, and escape cover, while adult holding used only depth and velocity and spawning used depth, velocity, and substrate. Application of the coupled two-dimensional hydrodynamic model and physical habitat model was effective in modeling the spatial distribution and quantity of habitat under both pre- and post-restoration conditions over a range of simulated flows. Modeling results for all life stages compared favorably with fish observation data under both pre- and post-restoration conditions. Significant increases in post-restoration habitat for fry and juveniles were associated with estimated vegetation colonization expected two/three years post site construction. Modeling results strongly support that restoration objectives are being met through the implemented channel changes.

Presentation notes:

They predict fish habitat for pre- and post-restoration using a 2-dimension model. They required pre- and post-construction topography. Roughness values will provide velocities. They observed some differences at Hocker Flat in pre- and post-restoration for depths and velocities; differences are greater at higher flows. Spawning habitat increased but fry rearing decreased. Lack of rearing habitat is due to lack of escape cover or vegetation.

Questions: Pre-dam photos show wide cobble bars with little cover along the river, what did the fish do? These photos were likely further downstream and that may have been more Chinook habitat that do not require rearing cover. The coho and steelhead were traditionally in the upstream areas.

^ TOP

Predicting changes in bird abundance caused by removal and succession of riparian habitat

Sherri L. Miller, U.S. Forest Service, Redwood Sciences Laboratory, smiller02@fs.fed.us, (707) 825-2949
C. John Ralph, U.S. Forest Service, Redwood Sciences Laboratory, cralph@humboldt1.com , (707) 825-2992
and Pablo Herrera, U.S. Forest Service, Redwood Sciences Laboratory, Pablo_Herrera@fs.fed.us, (707) 825-2994

Presentation [PPS - 20 mb]

Riparian birds are one component of monitoring response of wildlife following implementation of the Program. Riparian bird abundance is expected to increase or be maintained in the 40-mile treatment reach as the complexity of riparian habitat from the water's edge to the upland boundary of the floodplain increases. A set of target riparian birds have been selected for monitoring because they are good integrators of riparian habitat conditions, are of some management concern and can respond rapidly (1 to 2 years) to changes in habitat. From point count surveys conducted from 2002 to 2006, we have developed statistical models to predict abundance of target species from a set riparian habitat characteristics, including: spatial arrangement, structural complexity, vegetative characteristics and species composition. By monitoring changes in habitat conditions and bird abundance we can validate the predictive models, identify trends, and estimate population size for the set of target riparian bird species.

The Yellow-breasted Chat is one riparian nesting species of concern that has been selected for monitoring. We have found Chat abundance as high as 4 birds within 150 m of a station with higher numbers tending to be downstream of Steiner Flat. About half of our stations had Chat detections. Our predictive model suggests that higher abundance of Chats is associated with larger riparian habitat patches, presence of wet soil plants, and Himalaya blackberry within 100 m of the count station. Lower abundance is associated with higher percentages of tree cover, more bedrock and medial bars within 100 m of the station, and arroyo willow at the count station. In general, Chats require larger patches of habitat with a mix of riparian plant species and structural complexity. We will be able to use the parameter estimates from the models to predict the number of Chats that will be present at the station, reach, or 40 mile scale as restoration and alluvial processes change the riparian habitat over time.

Presentation notes:

Actual bird presence data at Hocker Flat (counts) agree well with statistical model for count data over the entire 40 miles of river. Therefore, it seems reasonable that we can predict changes in presence due to restoration (changes in vegetation) at single sites.

Questions: There are plans to validate the model on more sites than Hocker Flat. They were aware of autocorrelation among variables and avoided this.

^ TOP

Temporal and spatial distribution of salmon spawning from Lewiston Dam to Cedar Flat River Access: 2002-2006

Charles Chamberlain, US Fish and Wildlife Service, Arcata Fish and Wildlife Office, 1655 Heindon Road, Arcata, CA, charles_chamberlain@fws.gov

Presentation [PPS - 4.1 mb]

Redd surveys in the mainstem Trinity River from Lewiston Dam to the Cedar Flat River Access have been collaboratively conducted by the California Department of Fish and Game , U.S. Forest Service , U.S. Fish and Wildlife Service, and the Yurok Tribal Fisheries Department since 2001. Beginning in 2002, river kilometer (distance from the river's mouth as interpreted from a centerline drawn on U.S. Geological Survey USGS 1:24,000 topographical maps) was recorded with redd data to achieve a spatial representation finer than the reach-scale resolution of prior years. This finer resolution facilitates analysis of spatial and temporal distribution of Chinook salmon spawning activity, and geomorphic unit (meander bend) scales. Our data do not suggest strong temporal or spatial segregation between spring and fall Chinook salmon. Annual spatial distributions of 2002-2006 redds correlate poorly with total yearly redd counts. In all years, spawning intensities are highest near Lewiston Dam/Trinity River Hatchery. A mean of 50% of the annual mainstem spawning distribution for the survey periods in 2002-2006 occurred within the first 5 km downstream of Lewiston Dam (from about Sawmill Side Channel upstream; range 37 to 61%), a mean of 75% of the spawning occurred within the first 24 km (from about Biggers Road Bridge upstream; range 67 to 79%), and a mean 90% of the spawning occurred within the first 49 km (from about Sheridan Creek upstream; range 85 to 92%). Lastly, the spatial resolution of our data allow interpretation of spatial response to site-specific restoration actions such as the Hocker Flat rehabilitation site.

Presentation notes:

Redd surveys were performed by floating and during carcass surveys. Most spawning occurs in Lewiston area; there are consistent high spawning areas such as Poker Bar, Vitzthum Gulch, Sheridan Creek , Del Loma. At Hocker Flat, there seemed to see some increased spawning in 2006. Redd counts ranged from 3276 to 7085 in 2001 to 2006. Seem to be consistencies in timing from year to year.

Questions: Superimposition does occur in the upper reach. Any trends of spawning with flow releases are not clear-there is not enough variation in flow at this time of year to tell. Did not compared redd counts to weir counts at Junction City . Redds do correlate with female carcasses recovered.

^ TOP

Where are Summer Coho?

Patrick Garrison, California Department of Fish and Game, pgarrison@dfg.ca.gov, (530) 623-9328

Presentation [PPS - 1.7 mb]

In 2006, the upper Trinity River from Lewiston Dam (Rkm 180.1) to Steelbridge (Rkm 158.8) was snorkeled during the summer flow period (450 cfs) to enumerate juvenile coho salmon, Oncorhynchus kisutch, and measure habitat suitability criteria. During the course of the survey, we enumerated 2,768 young of the year (0+) coho salmon; no yearling coho (1+) were observed. Geographically, the greatest numbers of juvenile coho were found in proximity to the BLM/Rush Creek fishing access, Cemetery side channel complex and immediately upstream of Salt Flat.

Juvenile coho were predominantly found in four meso-habitat types: side-channels (38.73%), glides (28.72%), backwaters (16.58%), and runs (10.55%). Juvenile coho were predominantly found in close proximity to the bank (mean=5.6 feet) and to both object (mean=1.08 feet) and overhead cover (mean=2.16 feet). The dominant object cover type utilized by juvenile coho salmon was non-emergent rooted aquatic vegetation (55.74% of observations), while the second most popular object cover type was small woody debris (26.12% of observations). The mean water column velocity selected by juvenile coho was 0.20 feet per second, while the mean water depth was 2.4 feet. Water temperature at juvenile coho sightings averaged 10.4 degrees Celsius and ranged from nine to 16 degrees Celsius; although above 12 degrees Celsius juvenile coho were rarely observed.

Inter-species interaction between juvenile coho and other fish was also examined. In the presence of non-predatory fish species, juvenile coho appeared to mimic those species schooling behavior. When with stickleback, Gasterosteus spp , juvenile coho readily intermixed with their schools, and were less likely to utilize cover. In the presence of juvenile Chinook salmon, O ncorhynchus tshawytscha , juvenile coho tended to select faster, deeper habitats and positioned themselves further from object and overhead cover. Brown trout, Salmo trutta, appeared to have a negative affect on preferential juvenile coho habitat utilization. When brown trout were present, juvenile coho were absent from highly preferable habitat. This was primarily observed in habitats including both small woody debris and lower water velocities.

Presentation notes:

Two divers did not work to count fish as they stirred up water so only one diver was used with raft recorder. Counted naturally produced coho (but some may be hatchery fish). Only saw three juvenile coho in the thalweg. Did not see any coho at gravel augmentation project but this may be due to poor visibility. Side channels were vacant at the steelbridge. Cemetery side channel had no coho at low flow. Found more at coho at broad bends in the river, but had to wait between schools for water to clear from mud. Coho used side channels but many side channels may have been too deep or too fast. Only 4 % of coho were observed in pools-may be too deep as coho seem to seek edges. Coho like cover and low velocity, found near areas of high spawning, don't like brown trout. Recommendations: quantify fall winter rearing and brown trout predation.

Questions: Is non-emergent vegetation chosen because there is so little wood? They seemed to like vegetation as brown trout were observed searching wood debris.

^ TOP

Restoration Actions and Juvenile Coho Rearing Habitat: Research Design and One Year of Data

John Hannon, US Bureau of Reclamation, jhannon@mp.usbr.gov, (916) 978-5524
Mark Bowen, US Bureau of Reclamation, mbowen@do.usbr.gov, (303) 445-2222

Presentation [PPS - 3.4 mb]

The objective of this project is to determine effects of specific rehabilitation actions on juvenile coho salmon by comparing coho densities, site fidelity, habitat use, growth, and food availability between treatment (rehabilitated) and reference (unmodified) sites.

The study sites we established in June 2006 are:

• Upstream pair
• Old Lewiston Bridge reference site (~river mile 110)
• Lewiston Cemetery side channel treatment site (~river mile 109)
• Downstream pair
• Lime Point reference site (~river mile 74.5)
• Pear Tree constructed alcove treatment site (~river mile 73)

The assessment of coho response to rehabilitation actions is being measured by mark and recapture of juvenile coho salmon at the study sites. We tagged 1,200 coho from Trinity Hatchery and 800 wild coho at the study sites with passive integrated transponder (PIT) tags and visually marked them with elastomer. We captured the coho primarily using baited minnow traps and occasionally by using a seining/snorkeling combination. The minnow traps successfully captured coho in most areas where we sighted them by snorkeling. The recaptures all occurred at the upstream pair of sites and at the rotary screw traps at Pear Tree and Willow Creek. We captured or recaptured no coho at the downstream pair of sites (Pear Tree and Lime Point). Coho density was highest at the Old Bridge during the summer and highest at the cemetery side channel in winter.

Site fidelity for wild coho occurred at the two upstream study sites as evidenced by recaptures and snorkel survey sightings of marked coho. Site fidelity for hatchery coho released into study sites was low. Hatchery coho mostly left the study site where they were released within a few days. Based on recaptures at downstream rotary screw traps and absence of recaptures in study sites, the hatchery fish released in both June and January seemed to head downstream rather than staying near the release locations.

We sampled food availability using primarily drift nets. Although the samples have not yet been analyzed, the impression of the samplers was that the presence of food in the drift is low (samples in June and September) but invertebrate densities on the substrate are high. Stomach content analysis in 2007 will show what the juvenile coho are eating. We plan to use the food availability results combined with coho growth data to model the bioenergetics and compare habitat quality between study sites.

We hope to establish a new treatment and control site in the Dark Gulch/Bucktail area in 2007 if resources allow. This will provide four sites with coho densities high enough to more effectively evaluate fish response to habitat projects.

Presentation notes:

Chose two paired reference and treatment sites to monitor. Cemetery Hole was treated in the 1980s and Pear Tree was treated more recently. No evidence of growth differences for juveniles between treatment versus control in upper site. Growth could not be assessed at the at Pear Tree site as no recaptures were made. Only 2.7 % of the marked fish were recaptured at Cemetery Hole. No hatchery fish were recaptured. Lewiston old bridge had only 3.5 % recaptures. Lack of recaptures at downstream sites may be due to higher temperatures causing coho to move out downstream. There doesn't appear to be enough food in the drift for fish.

Questions: Nighttime snorkel surveys have not been tried, however, minnow traps were set overnight.

^ TOP

Evaluating the effectiveness of rehabilitation actions in creating fish habitat in the Trinity River

Darcy Pickard, ESSA Technologies Ltd., dpickard@essa.com, (604) 733-2996

Presentation [PPS - 1.5 mb]

Evaluating the effectiveness of different river habitat rehabilitation actions on salmonid populations is difficult. This is especially true within a single large watershed restoration effort like the Trinity River Restoration Program, due to lack of replication and independence as well as effects occurring outside the watershed.

We developed a simple, spatially explicit model of the loss and creation of Chinook fry-rearing habitat in the Trinity River. The model describes the expected relationship between the rehabilitation action, covariates (flow levels, region, gravel, etc.) and the amount and sustainability of habitat created. Cost was used as a simple measure of the complexity of a single rehabilitation action. Since the model can be updated as new knowledge is gained and can utilize information provided by more detailed models it is a useful adaptive management tool.

We used this simple model to compare implementation designs in their ability to distinguish between different types/intensities of mechanical river rehabilitation actions. We considered alternative choices in how to distribute such actions across space (river locations) and time (years). We also considered several climate scenarios of interest (e.g. droughts, varying frequencies of floods), and their effects on the ability to detect changes in habitat from different treatments.

A 3x6 Latin square design (with blocking variables: year and region) had comparable power to a design where only two regions were treated in each year, suggesting that if the number of rehabilitation sites per year is limited the more convenient of the two implementation designs may be used. High cost actions must be much more effective than low & moderate cost actions to make a difference in the mean habitat created after 15 years. Low and moderate cost actions along with a flood every 5 years and/or cheap repairs after droughts may be more effective than one-time very costly actions.

Presentation notes:

Used a simple model to estimate effectiveness of restoration chose project cost as a surrogate for complexity. Lower cost actions may be better than high costs actions as they could be just as effective. Drought conditions may require some additional management as willows could easily establish on cleared floodplains and reduce effectiveness. She recommended that the TRRP consider alternative flow schedules in drought years if peak flows are not achieving the desired goal of scouring willows. This water could be used for something else.

Questions: How was model parameterized? Pickard used expert opinions so model is more theoretical. The model results were not compared to real river data or results.

^ TOP

Discussion Session: Fish and Avian Habitat

Panelists: Aaron Martin, Thomas Hardy, Sheri Miller, Patrick Garrison, Charles Chamberlain, Darcy Pickard, John Hannon and Mark Bowen

How were reference sites chosen by Hannon and Bowen? Why was a side channel being compared to a main channel site? Bowen noted that a side channels might be thought of as a treatment in that it could be added in to a mainstem reach as a treatment.

Has hatchery to natural production been examined as a function of distance below dam? Carcass data has been looked at. Hatchery proportion is indeed higher near the dam, but sample sizes were small. Regarding spawning distribution: any thoughts about geomorphic relations in choice of redd sites? Need to get away from the straight u-shaped channel. Spawning habitat is not limiting on the upper river. Spawning looked better in wider floodplains and perhaps areas of lower slopes. A few areas get lots of superimposition, but most other areas are not superimposed. Not sure why areas that look good do not have spawners. May be linked to hyporheic flows, or upwelling that may keep gravels clean. Fish seem to use same spawning areas year after year. Backwater areas during winter may be very important. Indeed these surveys want to assess how well coho are using habitats.

Why were the bird count data at Hocker Flat so close to the prediction model results? Riparian habitat is not that variable spatially, so the model results, which are essentially means for all sites, should be close to most individual sites. Sites were originally picked in early 90's during an earlier study.

Since the system seems to be dramatically changed: cut off headwaters, more predators, warmer waters, hatchery fish. is it realistic to expect recover coho in the mainstem? Rush and Brown's Creek are two areas that may be good areas to start on restoration. But generally the tributaries may not be good habitat.

The early literature shows similar results regarding spawning-high spawning near the hatchery, so some of these results are not novel. Are there any thoughts of restoration that might spread out spawning. Can you think strategically? Current work, such as Hannon and Bowen, is different from the earlier work in that it is focused on changes due to restoration.

Himalayan blackberry is important cover for yellow warblers. There is not as much plans to plant actively as to let plants come in.

How well coordinated are the studies? Somewhat coordinated.

Issues such as engineering side channels, importance of cover, role of predators-are these inconsistent with the ROD? Will this be followed up? IAP is organized to examine question or hypothesis of the ROD. Maybe predator control is more important than engineering channels. Is the program only designed to test the ROD or is it an adaptive management? It is both, since ROD says to follow adaptive management. As we learn more we can prioritize hypotheses.

^ TOP