BPA Fish and Wildlife FY 1997 Proposal

Section 1. Administrative
Section 2. Narrative
Section 3. Budget

see CBFWA and BPA funding recommendations

Section 1. Administrative

Title of project
Fish Passage Video Monitoring

BPA project number   8812005

Business name of agency, institution or organization requesting funding
Yakama Indian Nation

Sponsor type   WA-Tribe

Proposal contact person or principal investigator
 NameMel Sampson
 Mailing addressYakima Indian Nation
Fisheries Resource Management
P.O. Box 151
Toppenish, WA 98948

BPA technical contact   Kelly Kittel, EWN 503/230-5199

Biological opinion ID   

NWPPC Program number   7.4K.1

Short description
This project is a component of the monitoring/evaluation program for the YKFP. In addition, the project provides fish passage information for the fish managers in the Yakima Basin. Data collected includes: enumeration of fish passage by species, identification of external marks, weekly fish passage counts, run timing and fish length data. Information is also recorded pertaining to fish viewing conditions at each of the ladders.

Project start year   1993    End year   

Start of operation and/or maintenance   

Project development phase   Implementation.

Section 2. Narrative

Related projects
-Training & Education: Project #8812004 Training Assistance for Personnel for YKFP needs.
-Fish Technicians: Project #8812008 Technical personnel support for field data collection, task
activities/assignments and special fishery projects consistent with YKFP objectives.
-Quantitative Production Objectives, Sh/FC: Project #8812009 Refining objectives for Steelhead/Fall Chinook
and assist in selection of appropriate treatment strategies, risk analysis and monitoring plans.
-Yakima Hatchery Construction: Project #88-115 Construction of Spring Chinook Supplementation Facilities
for YKFP.
-Yakima Hatchery Final Design: Project #90-69 Design of Spring Chinook Facilities for YKFP
-YKFP Scientific & Mgt Services (Master To): Projects Contracted to Co-Managers, Washington State

1. YSIS (Yakima Species Interaction Studies) |
2. Marking Studies | All of these Tasks represent an aspect of the Goals and
3.Genetic Management Framework | Objectives of the YKFP
4. Policy/Management |
5. NIT Design; (New Innovative treatment) one of two methods to be tested in YKFP.

Project history
The YKFP has been in the Fish and Wildlife plan since 1982. The Yakima phase of the Project initially included an all stock initiative. There still remains an all stock initiative but not all at once. The Policy Group's preferred an all stock initiative. There still remains an all stock initiative but it will be implemented through a phase in process. The Policy Group's preferred alternative that's included within the FEIS is to implement the supplementation aspects for Spring Chinook and develop and implement a monitoring plan for the now released coho in the basin. The other stocks will be implemented through a tiering process. The Klickitat aspect of the Project currently is doing fishery surveys, population monitoring, habitat inventory, and engineering surveys of passage barriers in the Klickitat River watershed. This is consistent with the Preliminary Design Report.

Biological results achieved
The use of video as a methodology to monitor adult fish passage in the Yakima Basin began in 1988. Beginning in 1993 with the initiation of the YKFP, renewed effort was made towards solving some of the problems encountered with this methodology that had been previously identified. The focus has been at Prosser Dam, to address issues that negatively affect the fish viewing environment (e.g. light system, window and backboard cleaning issues), and to make improvements to the data retrieval and management processes. Also, beginning in 1993 QC issues were addressed. These specifically dealt with ranking the magnitude of each “problem” affecting fish image quality recorded on the video tape, and to determine our level of accuracy pertaining to species identification, total fish counts, etc. This information resulting from this QC work is currently being used to make improvements to the fish viewing sites at Prosser Dam.

Annual reports and technical papers
Fast, D. E. et al. 1989. Yakima/Klickitat Natural Production and Enhancement Program. Prepared for Bonneville Power Administration. Project No. 88-120. Grant DE-A179-88BP93203.

Hubble J. D. et al. 1990. Yakima/Klickitat Natural Production and Enhancement Program. Prepared for Bonneville Power Administration. Project No. 88-120. Grant DE-A179-88BP93203.

Hubble, J. D. et al. 1991. Yakima/Klickitat Natural Production and Enhancement Program. Prepared for Bonneville Power Administration. Project No. 88-120. Grant DE-A179-88BP93203.

Hubble, J. D. 1992. Yakima/Klickitat Natural Production and Enhancement Program. Prepared for Bonneville Power Administration. Project No. 88-120. Grant DE-A179-88BP93203.

Hubble, J. D. 1993. Yakima/Klickitat Natural Production and Enhancement Program. Prepared for Bonneville Power Administration. Project No. 88-120. Grant DE-A179-88BP93203.

Management implications
The use of video has proven to be a successful method to monitor adult fish passage in the Yakima Basin. There is a solid commitment by the YIN and the BOR, which maintains these monitoring facilities, to address the current problems that affect video monitoring, in order to make them state-of-the-art. The adult monitoring facilities at Prosser and Roza dams are key components, along with the Chandler and Roza juvenile facilities, to measure smolt-to-adult and adult-to-adult survival rates for both the wild and hatchery populations for the YKFP. Adult escapement counts at these two facilities also provide a means to estimate spawner escapement into subbasins without adult counting facilities. These monitoring sites provide a location where marked hatchery adults from various YKFP experimental treatment groups can be non-intrusively examined in order to measure survival attributes (e.g. smolt-to-adult). Of equal importance is the use of this data for the fish managers.

Specific measureable objectives
Future objectives will be directed in the following areas, 1) to upgrade the physical facilities at Prosser and Roza dams pertaining to fish monitoring, 2) improve the fish image acquisition procedures, 3) improve the data processing and archival procedures, and 4) incorporate the quality control data into deriving better estimates of fish passage.

Testable hypothesis
That current improvements being made to the physical fish viewing facilities will result in improved accuracy in fish enumeration and species identification (Specific questions to be addressed in 1996 are delineated in the Methodology section).

That external marking methods being devised by the newly formed monitoring/evaluation TAC of the YKFP can be successfully integrated by the video monitoring system.

That turbidity significantly affects species identification and enumeration of fish comprising the fall and spring runs past Prosser Dam.

Underlying assumptions or critical constraints
An inherent constraint to a video based monitoring system in the Yakima River at Prosser is the level of turbidity that degrades the resultant fish image. The underlying assumption is that amount of time (days) impacted by high turbidity levels does not significantly affect fish identification and enumeration accuracy.

It is assumed that improvements to the physical facilities will achieve a yet to be determined level of accuracy that will satisfy monitoring/evaluation criteria for the YKFP.

The video based adult monitoring system consists of a analog video camera placed in front of the fish passage viewing window. The video signal is fed into a time-lapse video recorder, that typically compresses 24 hr of real time monitoring onto a 2 hr VHS tape. This is accomplished by incorporating a pause in between each exposure taken by the video camera of less than one second. The VHS tapes are replaced each day. The fish passage data is recorded by playing the VHS tape back on a video editor unit which allows for variable speed movement of the tape through the unit. Initially, the data is recorded onto a hard copy as the tape is being played back, and then is keyed into a database on the PC. After the tapes have been played back they are achieved. Currently, there are improvements being made to the monitoring system.

For the past four years the focus of our research has been two-fold, 1) to quantify the level of accuracy using a video based monitoring system and to understand the reasons for reduced accuracy, and 2) to make improvements to the lighting system and in the fish image acquisition and data management procedures. In the past year other known problems affecting fish viewing have begun to be addressed, such as maintaining a clean viewing window and back board. Initiation of research item two was a result of knowing what needed the most immediate attention to improve fish image quality.

I. For the past four years approximately 20% of the video tapes from the spring and fall runs have been double read. This is accomplished by having reader one read reader two’s tapes and visa-versa. The tapes are selected randomly, but weighted in proportion to the weekly percent of fish passage relative to the entire run. Evaluated from these tapes is the level of agreement between the two readers with regards to total fish per tape, species identification and fish lengths. For fish pairs where there is a difference in species identification, the data set from reader one is sorted against reader two to determine, by ranking, what the most frequent “disagreements” were and what they were most frequently identified as by the second reader. Where significant differences in the parameters being evaluated between the readers are seen, then reasons for this occurrence is examined.

The next step (for future proposed work) is to use this information in a statistical manner to adjust the fish counts by species.

II. Improvements to the Prosser right bank monitoring site will be evaluated in spring and fall of 1996. The following questions will be evaluated:

1. Fish behavior. Is there are difference in passage behavior, specifically the incidence of “fall back” in the viewing window, at night between incandescent light and infra-red light?

Experimental Approach. Beginning at twilight alternate between the two light sources at 2 to 4 hour intervals. This could be done through out most (50%+) of the fish run. The difference in the mean number of “fall back” events throughout the run between the two light sources would be the evaluation parameter. A t-test will be used to test for a significant difference.

2. Species identification accuracy. Is there a difference in species identification between the two light sources, as defined by the ratio of positive to unknown species identification?

Experimental Approach. Within each respective light source treatment compare the ratio of positive to unknown species, using a t-test to evaluate this. Positive identification means the reader made an identification with regards to “wild vs hatchery”, “adult vs jack” and/or the species itself (ie. chinook, steelhead, etc). An unknown identification means the reader could not determine the aforementioned categories of identification.

3. Image quality. Is there a difference in fish image quality between the two light source treatments?

Experimental Approach. To some extent questions 2 and 3 are linked together. To put this into context to the readers, it’s typically image fuzziness or lost of image details (too dark) that leads to species mis-identification or the inability to make a positive species identification. The approach is to rank each image (derive a scale) based on image sharpness (perhaps this could be measured in some fashion from a digitized image) and level of key detail features visible in the image (ie. spotting, reflexivity, fins). In the end I would like to be able to state in a quantitative sense what the difference in image quality is between the two light source treatments.

4. Pre and post image quality. A. Overall at the right bank is there a difference in fish image quality comparing a random sample of fish images from the past three years (pre) to a random sample of fish images from the 1996 spring run (post)? B. What were the causes historically affecting image quality and what are they now, and to what magnitude were they affecting image quality relative to each other?

Experimental Approach. The purpose of this question is to answer the question how much better (I’m assuming they are better) are the post images to the pre images? The same protocol described for question three would be used to evaluate this question. To address the second part of this question (B) I would rely on data that is kept for each tape describing the various problems affecting image quality. This data records the relative magnitude of the problem (ie. fair, bad, v. bad) and the duration of the problem. Lost of recording time would be part of this analysis, since no images are recording if for example the electricity is off or the record button wasn’t pressed. The same data will be recorded for the spring 1996 run. My approach then for question B will be to rank the top five problems in the pre and post treatments and describe how they impacted image quality and to discuss improvements made in the post relative to the pre. Finally, to summarize for the post facility/environmental conditions how severe is each problem and make recommendations for improvements (if possible).

5. Image quality between windows (rt vs. ce vs. lf). Is there a difference in image quality between the three ladders?

Experimental Approach. Image quality would be evaluated following the protocol described for question four. However, to kept light source treatments the same only fish images recorded under the incandescent light source treatment would be use in the evaluation, since this is the only light source type at the center and left ladders. As in question four, the same evaluation procedure would be used to identify and rank problems affecting image quality.

6. Fish fork length accuracy. What is the difference in length between the actual fish length and of the estimated fish length derived from the fish image?

Experimental Approach. This test will require allowing a live fish of known species and length to swim in front of the viewing window for 10 seconds or so to record several images from which the length from the fish image can be measured. This will require construction of a “cage” that can be lowered down into the passage slot. Fish could be collected at the right bank denil ladder. On days when fish were being collected we would not be able to conduct the aforementioned tests. The difference in fork length would be the evaluation parameter.

Brief schedule of activities
The emphasis for FY1997 will be to continue to monitor the adult spring and fall chinook, coho and steelhead runs, as in past years. Secondly, continued improvements will be made at Prosser and Roza dam fish viewing sites (4 ladders) and evaluation of these improvements towards improvement in fish image quality1. Thirdly, improvements will be made in the fish image acquisition and data processing/storage procedures. Fourthly, use of the quality control information to derive better estimates of fish passage and species identification will be initiated. This task will involve the services of a statistician.

For FY1998-2001, it’s envision that the task of applying the QC data to derive better fish passage estimates, etc. will be completed by the end of FY1998 (assuming it’s initiated in FY1997). Because of the uncertainty of the BOR’s funding abilities, it’s difficult to assess when Prosser and Roza dam fish monitoring facilities will be completed2. It’s anticipated that most of this work will be completed by FY2000. In the next two FY’s attention will focus on continued improvements to the fish image acquisition and data processing/storage issue. Specifically, to investigate using a digital format and using of an automatic tape (or CD) editor to reduce playback time. It’s anticipated that various marking techniques (that will be used to mark hatchery treatment groups) will be evaluated to determine how well they can intrusively interrogated by the video system.

1 Improvements to the fish viewing sites is presently the responsibility of the BOR. Therefore, completion of all desired improvements to theses sites is dependent upon their funding ability.
2 Proposed budgets reflect estimated costs to complete capital improvements in light of the uncertainty with BOR funding.

Biological need
The YKFP is in essence an experiment to test hypothesis that supplementation is a successful method to rebuild natural salmon populations, with adaptive management as the underlying experimental/management approach. This requires that monitoring/evaluation be conducted in order to assess the progress and direction to take with regards to supplementation strategies towards rebuilding of the salmon populations. The Prosser and Roza adult fish monitoring sites provide a key component with regards to monitoring/evaluation of the YKFP, namely estimating run size by species, recording the number of marked from various treatment groups, and the collection of other life history data. In addition, this data is valuable to the fish managers in the basin, who deal with issues outside the scope of the YKFP that help to enhance the success of the project (e.g. habitat protection and enhancement).

Critical uncertainties

Summary of expected outcome
1. That improvements to the fish viewing sites will result in the best quality of fish image possible for the Yakima River.

2. That the fish image acquisition and data processing/storage procedures be devised which result in reduced tape processing time, better reader accuracy, better data management/storage protocol, and which takes advantage of current digital technology.

3. That statistical procedures be devised to incorporated the QC data to derive the best estimate of fish passage for each species.

4. That specific marking techniques are identified that will allow for intrusive interrogation using a video based monitoring system.

Dependencies/opportunities for cooperation
As aforementioned, physical improvements to the fish viewing sites currently is the BOR’s responsibility (however this issue is not clearly resolved at this time), and thus the time table for all the needed improvements is currently dictated by their ability to fund improvements. As stated earlier costs for these improvements have been incorporated into the projected FY1997-2001 budgets.


Monitoring activity
Outcomes 1-3 will be measured by the YIN adult fish passage video monitoring staff. In general, it will be up to the staff to determine if it’s felt that the best quality fish image has been obtained. This will be measured in terms of how well all the known problems which negatively affect the image have been addressed, and by how well improvements to the fish image acquisition and data processing/storage procedures are working (are they meeting the staff’s expectations). Outcome 4 will largely be monitored through the monitoring/evaluation TAC. Success of this outcome will be determined by being able to interrogate a pre-determined level (e.g. 95%) of marked fish passing through the fish viewing sites.

Section 3. Budget

Data shown are the total of expense and capital obligations by fiscal year. Obligations for any given year may not equal actual expenditures or accruals within the year, due to carryover, pre-funding, capitalization and difference between operating year and BPA fiscal year.

Historic costsFY 1996 budget data*Current and future funding needs
1993: 86,358
1994: 78,526
1995: 90,870
1996: 92,949
Obligation: 92,949
Authorized: 0
Planned: 92,949
1997: 215,000
1998: 215,000
1999: 180,000
2000: 155,000
2001: 150,000

* For most projects, Authorized is the amount recommended by CBFWA and the Council. Planned is amount currently allocated. Contracted is the amount obligated to date of printout.

Funding recommendations

CBFWA funding review group   Bonneville Dam - Priest Rapids Dam

Recommendation    Tier 1 - fund

Recommended funding level   $215,000

BPA 1997 authorized budget (approved start-of-year budget)   $215,000