The Pacific Northwest tags more fish and monitors fish movements more intensely than anywhere else in North America.  Tens of millions of dollars are spent on fish tagging and data retrieval each year.  However, there is no overriding analytical framework to the design and analyses of these studies to assure optimal performance or maximum benefits from the data.  The Columbia Basin Fish and Wildlife Program (CBFWP 1994-1995) recognized these needs in their call for a regional center for conducting analyses and the development of needed analytical tools and analyses (3.2F.1).  The plan also called for a coordinated information system on anadromous fish databases (3.3A.2) and the need to monitor populations, provide timely transfer of information, and develop needed monitoring technologies (4.3C.1).  Despite these overwhelming needs, given the high expenditures on tagging and data collection, no agreement on a coordinated information system for anadromous fish data, methods of their analyses, or timely dissemination of information to the public has been agreed upon.  This project exists to meet some of these pressing demands for sound statistical analysis of tagging data and the dissemination of data in a timely manner in order to evaluate and monitor mitigation activities in the Columbia Basin.  This program contributes to mitigation activities by providing sound information from historical or real-time tag detections in order to adapt management strategies to the most current information on salmonid migration and success.  This program also assists by resolving statistical and data analysis issues, so that managers can focus on biological or resource issues rather than on data analysis uncertainties.

Apart from its essential role in the fish and wildlife program, Bonneville Power Administration (BPA) needs and uses the support provided by this project to accountably perform and preserve its ability to independently make decisions related to operations of the Hydrosystem and commitment of fiscal and material federal resources for fish and wildlife mitigation programs.  This project also provides critical analysis for efficient use of limited mitigation funds on both a real-time and planning horizon basis.  This program also helps assist the BPA in the CBFWP (2000) requirement to produce an "annual report on flow augmentation" including "the benefits of flow augmentation for fish survival, and the precise attributes of flow they may make it beneficial."  
To support real-time monitoring and information dissemination, the program provides real-time analyses of PIT-tag data and smolt passage indices to predict outmigration timing at eight Snake and Columbia River dam sites.  This information is coupled with travel time information and barging projections in the Snake River Basin.  Daily updates are provided for yearling and subyearling chinook , steelhead, and sockeye from April through September of each year since 1994.  These predictions, along with supporting information, can be retrieved and independently analyzed using interactive software and graphs on the internet (http://www.cbr.washington.edu).

The second element this project provides is value-added analyses of historical tagging data by testing hypotheses, estimating parameters, and modeling interrelationships without the tremendous costs of additional field studies.  These analyses have provided information in the past to scientific groups like the Plan for Analyzing and Testing Hypotheses (PATH) forum.   An example is the analysis of adult returns from Priest Rapids hatchery in relationship to in-river conditions during juvenile outmigration (Skalski et al. 1996).  Other investigations have looked at the effects of pulsing river flows on smolt travel speeds (Perez-Comas and Skalski 1998a) and estimation of growth rates of outmigration smolt (Perez-Comas and Skalski 1998b).  This project has also examined 17 years of CWT data from coastal coho stocks and developed multivariate age-at-return models to investigate early ocean survival relationships.  Ryding (1998) and Ryding and Skalski (1998) have found that early ocean survival of coho to be correlated with sea surface temperature and upwelling.  They found a narrow range of optimal conditions that occur once in every 4-5 years that results in five-fold higher survival rates.  A follow-up study on coastal chinook salmon stocks in Washington and Oregon found the timing of spring transition and winter sea surface temperatures were major contributors to improved ocean survival and higher adult return rates (Burgess and Skalski 2001).

As provisions for handling salmonid stocks have become more restricted, the value of retrospective analyses of existing data sets becomes increasingly important.  These investigations also provide needed information to perform sample size calculations and to design field investigations more precisely. The specific analyses performed each year changes as BPA information needs shift to reflect the changing needs of the fisheries community.  A technical publication series of tag analyses is published under the contract to disseminate information learned from these added-value analyses.
A third element of this project is to develop the statistical tools and analyses to monitor and evaluate the status and trends of salmonid stocks and the hydrosystem.  Numerous performance measures and recovery goals have been established (e.g., BiOp. RM&E Hydro Working Group, Remand) to monitor the recovery of the Columbia Basin system.  Despite all the discussions of recovery and performance measures, no public source of readily available information exists.  In preparation for the 2005 "check in" and beyond, this project is compiling and summarizing the myriad of endpoints identified in recovery documents.  This project is also reporting the information and developing analysis tools to assist the public and agencies in independently assessing the status of recovery in the Basin.  The information and statistical tools to examine the status and trends of performance measures are being placed in the DART database managed by the University of Washington for public dissemination (http://www.cbr.washington.edu/dart/dart.html).  Collaborating with NOAA Fisheries and the action agencies will continue to refine and expand the capabilities to interpret recovery data.


Project Goal

The project goal is to provide state-of-the-art analysis and interpretation of life-cycle information for monitoring and evaluation of salmonid stocks.


Project Objectives

1. Provide real-time predictions of smolt runtime through the hydrosystem to better manage water resource in the Columbia Basin.

2. Provide integrated analysis of historical tagging  data to improve the understanding of salmonid life history information and the effects of hydrosystem operations on salmon recovery.

3. Develop state-of-the-art statistical methods for monitoring and evaluating the status and trends of the Columbia Basin hydrosystem and associated salmonid stocks.

4. Provide the BPA with technical assistance on topics, proposals, and technical reports of interest to them on an "as needed" basis, and to extend this statistical service to other agencies.

5. Provide ready access to information on the status and trends of performance measures related to salmonid recovery in the Columbia Basin.


Project Accomplishments

Integration, Coordination, and Information Transfer

This program, besides fulfilling essential BPA mandated responsibilities, also addresses essential issues in the Columbia River Basin Fish and Wildlife Program (CRBFWP 2000).  This program, through data compilation, analysis, and dissemination, is consistent with the Council's scientific principles of ecological management (i.e., Principle 7).  This program strives to understand the dynamics of the aquatic systems and anthropogenic effects by analyzing historical tagging data and providing real-time access to the monitoring data to the public, fisheries managers, and the TMT.  Under the Hydrosystem Passage and Operations section of the CRBFWP, this program has made the following contributions:  (a) developed, measured, and monitored juvenile fish passage; (b) monitored and evaluated adult passage, and provided technical support to the installing adult PIT-tag counters; (c) evaluated the effects of flow and spill strategies on outmigration success; (d) assist the BPA in preparing an annual report on flow augmentation; (e) provided mitigation information and specialized data summaries to the TMT; (f) investigated biological measures of mainstem operations; and (g) investigated key uncertainties associated with mainstem migration and transportation.  Under the Council's Ocean Condition section, this program has made the following coordinated contributions:  (a) helped better understand conditions salmon face in the ocean, (b) identify factors critical to ocean survival, (c) help distinguish ocean effects from other effects on adult returns, (d) separate the effects of ocean-related mortality from that caused in the freshwater part of the life cycle.

Equally important, this program has contributed to the Council's call for "Research, monitoring and evaluation" in all three critical areas they have identified:  (a) identify and resolve key uncertainties, (b) monitor and evaluate, (c) make information readily available.  This program provides statistical support, software development, and publicly accessible interfaces for data interpretation as part of the internet data capabilities of the Columbia River Data Access in Real Time (DART).  


Biological/Ecological Results to Date

Since 1991, this project has evaluated historical tagging studies to help design future research and determine the reliability of existing information.  The project has explored possible relationships between juvenile survival and adult returns with river flows, abundance of hatchery stocks, pulsing, ocean conditions, and ambient river conditions.  This project has also developed new approaches and statistical tools which generate on-line internet-based information including real-time in-season predictions of migration status.  This information, along with flow predictions and transportation levels, provides fish managers with the opportunity to integrate the status of the smolt migration with river management options.  Run-timing predictions are available for the Snake, Mid-Columbia, and Lower Columbia hydroprojects for NMFS evolutionary significant units (ESUs) as well as river runs as a whole for spring and fall chinook salmon, steelhead, and sockeye.

Specific biological results of the project to date include the following:

1.   Evaluation of travel time information contained in freeze-brand data.

2.   Evaluation of the annual pattern of adult salmon returns to the upper Snake River from PIT-tag data.

3.   Recommendations for complete life-cycle survival studies using combined PIT/CWT releases.

4.   Evaluation of the relationship between Priest Rapids hatchery fall chinook adult returns and river conditions.

5.   Evaluation of effects of river pulsing on smolt travel times.

6.   Evaluation of the benefits of smolt transportation at Priest Rapids and McNary dams.

7.   Evaluation of coho ocean survival relationships.

8.   Evaluation of chinook salmon ocean survival relationships.

9.   Evaluation of the internal consistency of hydroacoustic, PIT-tag, and balloon-tag studies in explaining project survival at Lower Granite Dam.

10. Evaluation of the relationship between turbine passage survival and turbine operating efficiency, and the efficiency of the "1% turbine efficiency rule."

Together, these biological results are helping to evaluate past management decisions and improve future efforts to enhance wild salmonid stocks.


Nonbiological Results to Date

The most visible product of the project is the annual forecasts of smolt outmigration time each spring and summer.  As many as 34 salmonid stocks, ESUs, and composite groups are monitored each year to produce the "percentage of passage by date" and the "date to specific passage percentiles" at hydroprojects throughout the Columbia Basin.  Each year, these run-timing predictions are evaluated post-season to report on predictive accuracy and precision.

The project has also focused on the added value analyses of historical tagging data.  Using information collected over time and enumerable investigations, this project has sought to understand salmonid life history and hydroproject effects at scales larger than the extent of individual investigations.  Such investigations have examined the effects of ocean conditions on early ocean survival of fall chinook salmon and coho.  Other investigations have been used to provide information to better design future tagging studies such as the proposed summer spill study in 2004.

More recent results have focused on compiling and summarizing performance measures as stated in the BiOp and Remand as the fisheries community approaches the 2005 check-in.  These efforts have included evaluating statistical methods to assess status and trend data, and to summarize recovery status.


Specific accomplishments by year include:

1991:  Report on historical brand release data for the Snake River.  Report identified serious bias in travel time estimation using freeze-brand data and the difficulties establishing travel time relationship with existing data.

1992:  Report on a "strawman" smolt monitoring design for Snake/Columbia River systems.  This strawman became the conceptual design for the existing PIT-tag detection system for the Snake River.
1993:  Report on adult PIT-tag returns.  Results showed annual patterns of adult returns differed across years even for closely related river systems.

1994:  Developed and implemented statistical software called Program RealTime PIT Forecaster using pattern recognition to predict the outmigration timing of spring runs of wild Snake River spring/summer chinook at Lower Granite Dam real-time.

1995:  Refinement of Program RealTime PIT Forecaster and initial testing of Program RealTime Passage Index Forecaster to predict outmigration timing of summer/fall runs of juvenile subyearling chinook at Lower Granite Dam based on Fish Passage Center (FPC) passage indices.  Report on transportation benefit analysis methods showed estimation differences among investigators may be related to analysis methods they selected.  Recommendations on proper statistical models provided.

1996:  RealTime prediction algorithm of this project linked with CRiSP model of Project 89-108 to extend predictions of migration timing to include Lower Snake River hydroprojects and McNary Dam.  Investigated extending real-time outmigration prediction to Mid-Columbia hydroprojects. Expansion of Internet access to cover historical run timing and flow data at all major hydroprojects.  Expansion of on-line interactive internet-based information to include in-season time series of PIT detections for NMFS Snake River Basin ESUs [i.e. ESA-listed juvenile spring/summer chinook, fall chinook and sockeye salmon] as they pass through the FCRPS.  Report on Priest Rapids hatchery returns of fall chinook versus river conditions found upstream-downstream study designs to be of limited quantitative reliability.


1997:  RealTime projections continued along with the addition of Redfish Lake sockeye salmon and steelhead.  Investigations included the study of the effects of river pulsing on smolt travel times, the estimation of fish guidance efficiency (FGE) and spill effectiveness (SE) from PIT-tag data and the estimation of season-wide survival rates and smolt growth rates.

1998:  Completed study of the relationships between coho age-at-return CWT data and early ocean survival as a function of ocean conditions.  Assessed internal consistency of hydroacoustics, PIT-tag, and balloon-tag studies in explaining Lower Granite Dam project survival of outmigrating smolts.

1999:  Assessed the validity of using Rice Island PIT-tag recoveries in estimating smolt survival through the Bonneville project.   Assessed the feasibility of using hatchery rearing information to assess smolt quality and subsequent downriver migration performance in cooperation with the US Geological Survey.

2000:  Completed study of relationships between fall chinook age-at-return CWT data and early ocean survival as a function of ocean conditions.  Developed optimal weighting schemes to improve the predictive performance of smolt outmigration timing for Program RealTime.  Completed synthesis and comparison of smolt survival estimates as generated with different technologies and experimental protocols.  Assessed the effects of the detection efficiency of adult PIT-tag detection facilities at Bonneville Dam on the performance of transportation evaluation studies and estimation of ocean survival.

2001:  Evaluated the statistical behavior of proposed statistical tests for recovery in the NMFS BiOp based on annual estimates of smolt survival.  Proposed methods were found to have low-to-moderate chance of identifying the correct status of recovery.  Alternative statistical methods were developed to assess recovery that had much higher probabilities of correctly identifying the true status of recovery.  Examined the statistical bias of treatment/control returns for assessing transportation benefits under the conditions of low adult return numbers and developed bias-corrected estimators.

2002:  Compiled estimates and associated standard errors on smolt-to-adult ratios (SARs) using CWT recoveries for 90 Columbia Basin hatcheries from the 1970s to the present.  Currently compiling inriver survival and travel time information using PIT-tag detections for approximately 20 Columbia Basin hatcheries and screw traps from the 1990s to the present.  Results of the CWT and PIT-tag analyses are available in DART for examination and analysis with ambient river data, power operations, and ocean conditions.  Assisted NMFS in developing multidimensional decision rules for 2005 and 2008 progress "check-in," and 2010 compliance with BiOp smolt survival performance measures.

2003:  Completed compilation of smolt travel time and survival estimates from PIT-tag releases, mid-1990s to 2002 (http://www.cbr.washington.edu/pitSurv/).  Also updated PIT-tag survival and travel times for the spring outmigration in 2003, and awaiting completion of the summer outmigration to complete the 2003 summaries.  Provided assistance to NOAA Fisheries on better statistical decision rules for recovery.  Compiled adult escapement data and associated NOAA interim recovery goals for 24 stocks and tributaries in the Columbia Basin.  A new website with adult escapement trends and goals will be established in January 2004.

2004:  Compiled smolt survivals and travel times through the hydrosystem based on 21 wild and hatchery PIT-tagged stocks in 2003.  Updated historical data on survival and travel times of summer 2004 migrants.  Developed new web page on adult fallback rates of salmonids at Bonneville Dam using PIT-tag information.  Compiled historical information on annual flows (since 1950s) and dissolved gas levels (since mid-1980s) and performance compliance at 10 Columbia Basin dams.  Developed graphical tools to summarize status of performance measures used in the recovery process.


Project Accomplishments:  1 January 2004 - 31 December 2004


Major milestones accomplished this last year include the following:

1. Program REALTIME monitored the time of smolt outmigration of 34 stocks, ESUs, and composite groups through 10 different hydroprojects in 2004.

2. In collaborating with tribes, agencies, and PUDs, smolt migration timing was monitored at smolt traps throughout Mid-Columbia tributaries for the first time.

3. Interactive assessment capabilities were added to DART monitoring and evaluation databases to permit users to examine annual trends and compare with performance standards.

4. Annual monitoring and evaluation data were complied and added to DART to allow examination of the following:

a. Adult return counts
b. Hanford redd counts
c. Dissolved gas percentages versus standards
d. Annual flows versus flow targets
e. Water temperatures versus EPA standards

5. Adult fallback statistics at Bonneville Dam based on PIT-tag detections were added to DART.  This prototype is expected to be extended to the other hydroprojects in 2005.

6. Analysis of interannual variation in smolt survival to assist in the design of a system-wide or project-specific summer spill evaluation study.

7. Development of a prototype web pages to summarize the status and trends of performance measures during the 2005 check-in.  These prototypes were used to demonstrate the ability to summarize the myriad of performance measures in a comprehensive manner for the public and fishery managers.

8. Assisted NOAA Fisheries, US Army Corps of Engineers, Northwest Power and Conservation Council, Washington Department of Fish and Wildlife, Mid-Columbia public utility districts, US Geological Survey, Nez Perce Tribe, and the Confederated Tribes of the Umatilla Indian Reservation in the design and analysis of monitoring and evaluation data and studies.  

9. The following technical reports and publications were completed in the last year as a consequence of this BPA project.

a. Burgess, C., and J. R. Skalski.  2003.  Evaluation of the 2003 predictions of the run-timing of wild and hatchery-reared salmon and steelhead smolt to Lower Granite, Rock Island, McNary, and John Day dams using Program RealTime.  Volume XI in the Design and Analysis of Salmonid Tagging Studies in the Columbia Basin.  Bonneville Power Administration, Portland, OR.  
b. Skalski, J. R.  2004.  Overview of design options for a summer spill study.  Prepared for BPA.  
c. Skalski, J. R.  2004,  Responses to ISRP questions concerning the Bonneville summer spill study.  Prepared for BPA.
d. Skalski, J. R.  2004.  Response to state, federal and tribal agencies joint technical staff review of summer spill study.  Prepared for BPA.