Skip Links

USGS - science for a changing world

Open-File Report 2012-1120

Prepared in cooperation with the U.S. Army Corps of Engineers

A Markov Chain Analysis of the Movements of Juvenile Salmonids, Including Sockeye Salmon, in the Forebay of McNary Dam, Washington and Oregon, 2006–09

By Noah S. Adams and Tyson W. Hatton

Abstract

Passage and survival data were collected at McNary Dam between 2006 and 2009. These data have provided critical information for resource managers to implement structural and operational changes designed to improve the survival of juvenile salmonids as they migrate past the dam. Much of the valuable information collected at McNary Dam was in the form of three-dimensional (hereafter referred to as 3-D) tracks of fish movements in the forebay. These data depicted the behavior of multiple species (in three dimensions) during different diel periods, spill conditions, powerhouse operations, and testing of the surface bypass structures (temporary spillway weirs; TSWs). One of the challenges in reporting 3-D results is presenting the information in a manner that allows interested parties to summarize the behavior of many fish over many different conditions across multiple years. To accomplish this, we used a Markov chain analysis to characterize fish movement patterns in the forebay of McNary Dam. The Markov chain analysis allowed us to numerically summarize the behavior of fish in the forebay.

This report is the second report published in 2012 that uses this analytical method. The first report included only fish released as part of the annual studies conducted at McNary Dam. This second report includes sockeye salmon that were released as part of studies conducted by the Chelan and Grant County Public Utility Districts at mid-Columbia River dams. The studies conducted in the mid-Columbia used the same transmitters as were used for McNary Dam studies, but transmitter pulse width was different between studies. Additionally, no passive integrated transponder tags were implanted in sockeye salmon. Differences in transmitter pulse width resulted in lower detection probabilities for sockeye salmon at McNary Dam. The absence of passive integrated transponder tags prevented us from determining if fish passed the powerhouse through the juvenile bypass system (JBS) or turbines. To facilitate comparison among species in this report, we combined JBS and turbine passage for yearling Chinook salmon, steelhead, and subyearling Chinook salmon even though we were able to differentiate between passage through the JBS or turbines for these three species. Information on passage proportions through the JBS and turbines can be found in the first report.

Numerically summarizing the behavior of juvenile salmonids in the forebay of McNary Dam using the Markov chain analysis allowed us to confirm what had been previously summarized using visualization software. For example, within the powerhouse region, passage proportions among the three powerhouse areas were often greater in the southern and middle areas of the powerhouse compared to the northern area of the powerhouse for yearling and subyearling Chinook salmon. The opposite generally was observed for steelhead. The results of this analysis also allowed us to confirm and quantify the extent of milling behavior that was observed for steelhead. For fish that were first detected in the powerhouse region, less than 0.10 of the steelhead, on average, passed within each of the powerhouse areas. Instead, steelhead transitioned to adjoining areas in the spillway before passing the dam. In comparison, greater than 0.20 of the Chinook salmon passed within each of the powerhouse areas. Less milling behavior was observed for all species for fish that first approached the spillway. Compared to the powerhouse areas, a higher proportion of fish, regardless of species, passed the spillway areas and fewer transitioned to adjoining areas in the powerhouse.

In addition to quantifying what had been previously speculated about the behavior of fish in the forebay of McNary Dam, the Markov chain analysis refined our understanding of how fish behavior and passage can be influenced by changes to the operations and structure of McNary Dam. For example, the addition of TSWs to the spillway area clearly influenced the passage of fish. Previous results have been reported showing that TSWs increased passage through non-turbine routes and the fish-track videos indicated, in general, how fish behaved before passing the TSWs. However, the analysis presented in this report allowed us to better understand how fish transitioned across the face of the dam before passing the TSWs and resulted in a quantitative way to measure the effect of moving the location of the TSWs from year to year. Installation of the TSWs in bays 22 and 20 clearly increased passage proportions through the southern one-third of the spillway area for all species, most significantly for steelhead. When the TSWs were moved to bays 19 and 20 in 2008, overall passage through the southern one-third of the spillway remained higher than 2006, but decreased from what was observed in 2007. Shifting the TSWs to the north decreased the proportion of fish passing through the TSWs and increased the number of fish that transitioned to adjoining areas before passing the dam.

Perhaps the most interesting new information to come out of the two-step Markov chain analysis relates to how the performance of the TSWs was influenced by their proximity to the powerhouse. During 2007, the highest proportion of fish passing through TSW 22 was for fish that transitioned from the powerhouse area. In contrast, a relatively low proportion of fish passed through TSW 20 after coming from the powerhouse area. Instead, the proportion of fish that passed TSW 20 after coming from the northern part of the spillway was twice as high as the proportion of fish that passed through TSW 20 after coming from the powerhouse. During 2008, the TSW in bay 22 was moved to bay 19, leaving the TSW in bay 20 as the one closest to the powerhouse. As was the case when a TSW was located in bay 22, the proportion of fish passing through TSW 20 after coming from the powerhouse was higher than the proportion of fish passing TSW 20 after coming from the northern part of the spillway. Passage proportions for fish passing through TSW 19, the farthest north of the two TSWs during 2008, was higher for fish that came from the northern part of the spillway compared to the proportion of fish that passed through TSW 19 after coming from the powerhouse.

The Markov chain analysis provided a mathematical way to characterize fish behavior in the forebay of McNary Dam and helped refine our understanding of how fish movements were influenced by operational and structural changes at the dam. The numerical information used to quantify the behavior of fish also can be used to construct simulations to examine how proposed fish passage structures might influence passage of juvenile salmonids. To demonstrate this, we used the results of the Markov chain analysis to examine how a virtual fish collector located in the center of the powerhouse might influence passage of juvenile salmonids at McNary Dam.

First posted May 31, 2012

For additional information contact:
Director, Western Fisheries Research Center
U.S. Geological Survey
6505 NE 65th Street
Seattle, Washington 98115
http://wfrc.usgs.gov/

Part or all of this report is presented in Portable Document Format (PDF); the latest version of Adobe Reader or similar software is required to view it. Download the latest version of Adobe Reader, free of charge.


Suggested citation:

Adams, N.S., and Hatton, T.W., 2012, A Markov chain analysis of the movements of juvenile salmonids, including sockeye salmon, in the forebay of McNary Dam, Washington and Oregon, 2006–09: U.S. Geological Survey Open-File Report 2012-1120, 104 p.



Contents

Abstract

Introduction

Markov Chain Methods

Environmental and Biological Setting

Results

Discussion

Acknowledgments

References Cited

Glossary

Appendix A. Locations of Hydrophones in the McNary Dam Forebay, 2006–09

Appendix B: Results from the 2006-2009 one-step Markov chain analysis for all fish regardless of where they first approached the dam during the day and night periods

Appendix C: Results from the 2006-2009 two-step Markov chain analysis for all fish regardless of where they first approached the dam during the day and night periods

Appendix D: Results from the 2006-2009 two-step Markov chain analysis for all fish that first approached the powerhouse during the day and night periods

Appendix E: Results from the 2006-2009 two-step Markov chain analysis for all fish that first approached the spillway during the day and night periods

 


Accessibility FOIA Privacy Policies and Notices

Take Pride in America logo USA.gov logo U.S. Department of the Interior | U.S. Geological Survey
URL: http://pubsdata.usgs.gov/pubs/of/2012/1120/index.html
Page Contact Information: GS Pubs Web Contact
Page Last Modified: Saturday, 12-Jan-2013 15:15:05 EST