C05__Attachment_6_2002_Final_Klamath_Fish_Kill_Report.pdf

September 2002 Klamath River Fish-Kill:

Final Analysis of Contributing Factors and Impacts

July 2004

California Department of Fish and Game Northern California-North Coast Region

The Resources Agency State of California

II

Acknowledgements:

The California Department of Fish and Game would like to thank Dr. Gary L.

Hendrickson from Humboldt State University, Department of Fisheries Biology and Dr.

Douglas F. Markle from Oregon State University, Department of Fisheries and Wildlife for their constructive peer review of this report. We would also like to thank Dr. Walt Duffy for his help in coordinating that peer review. In addition, we want to thank the Yurok, Hoopa Valley and Karuk tribes, the California Department of Water Resources, the North Coast Regional Water Quality Control Board, the Oregon Department of Fish and Wildlife, the U.S. Fish and Wildlife Service, the U.S. Bureau of Reclamation, the U.S. Geological Survey, the NOAA National Climatic Data Center and the U.S. Forest Service, for their cooperation in providing data and/or for reviewing this document.

Thanks also to the Klamath Water Users Association for providing the legal Declaration of their biological consultant, which expressed several alternative hypotheses related to the fish-kill. The Northern California North Coast Region of DFG would like to recognize the efforts of its Habitat Conservation Program and Fisheries Program, as well as, the Native Anadromous Fish and Wildlife Branch, and Fish Health Laboratory in Sacramento, for their efforts in the compilation of data and statistical analyses in this report. Finally, we would like to recognize the DFG staff efforts of Steve Turek, Mike Rode, Bill Cox, George Heise, Wade Sinnen, Carl Reese, Sara Borok, Mark Hampton, and Calvin Chun in the preparation of this report.

III

Executive Summary:

This report presents the Department of Fish and Game’s (DFG) final evaluation of causative factors and impacts of the September 2002 Klamath River fish-kill and makes recommendations to minimize the occurrence of future fish-kills. This report finalizes and supercedes the January 2003 DFG report entitled: “September 2002 Klamath River Fish Kill: Preliminary Analysis of Contributing Factors”. This document addresses questions and concerns regarding the preliminary report. In addition, this report was peer reviewed by academia and contributing federal and state agencies, tribes and other stakeholders.

The September 2002 fish-kill was unprecedented in that it was the first major adult salmonid mortality event ever recorded in the Klamath River. Fall-run Chinook salmon were the primary species affected, but coho salmon, steelhead and other fish species were also lost. At least 33,000 adult salmonids died during mid to late September 2002 in the lower 36 miles of river. Although a larger number of Klamath River fall-run Chinook died, a greater proportion of the Trinity River run was impacted by the fish-kill, because the Trinity run is substantially smaller than the Klamath run on an annual basis and the peak of the Trinity run was present during the height of the fish-kill.

The primary cause of the fish-kill was a disease epizootic from the ubiquitous pathogens ich and columnaris. However, several factors contributed to stressful conditions for fish, which ultimately led to the epizootic. An above average number of Chinook salmon entered the Klamath River between the last week in August and the first week in September 2002. River flow and the volume of water in the fish-kill area, were atypically low. Combined with the above average run of salmon, these low-flows and river volumes, resulted in high fish densities. Fish passage may have been impeded by low-flow depths over certain riffles or a lack of cues for fish to migrate upstream. Warm water temperatures, which are not unusual in the Klamath River during September, created ideal conditions for pathogens to infect salmon. Presence of a high density of hosts and warm temperatures caused rapid amplification of the pathogens ich and columnaris, which resulted in a fish-kill of over 33,000 adult salmon and steelhead.

Flow is the only controllable factor and tool available in the Klamath Basin (Klamath and Trinity rivers) to manage risks against future epizootics and major adult fish-kills.

Increased flows when adult salmon are entering the Klamath River (particularly during low-flow years such as 2002) can improve water temperatures, increase water volume, increase water velocities, improve fish passage, provide migration cues, decrease fish densities and decrease pathogen transmission between fish.

The total fish-kill estimate of 34,056 fish, was conservative and DFG analyses indicate actual losses may have been more than double that number. If fish-kill numbers were substantially underestimated, more fall-run Chinook salmon could have been included in modeling efforts, for allocation to harvest allotments in ocean and in-river Klamath fisheries, during 2003. In addition, Klamath Basin tribal net and sport anglers may have lost the opportunity to harvest roughly 4,000 to 14,600 fall-run Chinook salmon in 2002, due to the fish-kill. This impact was more pronounced in the Trinity River than the Klamath River, because the fish-kill occurred below the confluence of the Trinity and Klamath, and precluded much of the harvest opportunity on the Trinity River.

IV

TABLE OF CONTENTS:

ACKNOWLEDGEMENTS:....................................................................................................................... II EXECUTIVE SUMMARY:.......................................................................................................................III TABLE OF CONTENTS:.......................................................................................................................... IV LIST OF TABLES...................................................................................................................................... VI LIST OF FIGURES.................................................................................................................................. VII

I. INTRODUCTION:

II. STUDY AREA:

II. A. GENERAL SETTING;

II. B. FISH RESOURCES;

III. FACTORS INVESTIGATED:

III. A. DISEASE;

III. A. 1. Introduction III. A. 2. Methods III. A. 3. Results III. A. 4. Findings

III. B. TOXIC SUBSTANCES;

III. B. 1. Introduction III. B. 2. Methods III. B. 3. Results III. B. 4. Findings

III. C. FLOW;

III. C. 1. Introduction III. C. 2. Methods;

III. C. 3. Results III. C. 4. Findings

III. D. TEMPERATURE;

III. D. 1. Introduction III. D. 2. Methods III. D. 3. Results

III. D. 3. a. Water Temperature and Flow Relationship III. D. 3. b. Water Temperature and Air Temperature Relationship III. D. 3. c. Actual and Predicted Water Temperature Data in the Fish-kill Area III. D. 3. d. Actual Water Temperature Data Upriver

III. D. 4. Findings III. D. 4. a. Water Temperature and Flow relationship III. D. 4. b. Water Temperature and Air Temperature Relationship III. D. 4. c. Water Temperature and Fish Stress in the Fish-kill Area III. D. 4. d. Water Temperature and Fish Stress Upriver

III. E. DISSOLVED OXYGEN;

III. E. 1. Introduction III. E. 2. Methods III. E. 3. Results III. E. 4. Findings

III. F. FISH PASSAGE AND RIVER GEOMORPHOLOGY;

III. F. 1. Introduction III. F. 2. Methods III. F. 3. Results III. F. 4. Findings

V

III. G. RUN TYPES AND CHINOOK SALMON RUN SIZE;

III. G. 1. Introduction III. G. 2. Methods III. G. 3. Results III. G. 4. Findings

III. H. CHINOOK SALMON RUN TIMING AND DENSITY;

III. H. 1. Introduction III. H. 2. Methods III. H. 3. Results III. H. 4. Findings

IV. FACTORS DISCUSSION:

IV. A. PRINCIPAL CAUSE;

IV. B. RELATED FACTORS;

IV. C. DIFFERENCES IN 2002 FACTORS AND OTHER LOW-FLOW YEARS;

IV. D. COMPARISON WITH OTHER FISH-KILLS;

IV. D. 1. Rogue River IV. D. 2. Butte Creek IV. D. 3. Babine River

V. IMPACTS OF THE KLAMATH RIVER FISH KILL:

V. A. TRINITY AND SALMON RIVERS;

V. A. 1. Introduction V. A. 2. Methods V. A. 3. Results V. A. 4. Findings

V. B. SCOTT AND SHASTA RIVER;

V. B. 1. Introduction V. B. 2. Methods V. B. 3. Results V. B. 4. Findings

V. C. IMPACTS TO NATURAL PRODUCTION;

V. C. 1. Introduction V. C. 2. Methods V. C. 3. Results V. C. 4. Findings

V. D. IMPACTS TO HATCHERY PRODUCTION;

V. D. 1. Introduction V. D. 2. Methods V. D. 3. Results V. D. 4. Findings

V. E. IMPACTS TO FISHERIES;

V. E. 1. Introduction V. E. 2. Methods V. E. 3. Results V. E. 4. Findings

VI. CONCLUSIONS:

VII. RECOMMENDATIONS:

VIII. REFERENCES:

IX. APPENDIX A:

VI

List of Tables:

Table 1. Fish species in the Klamath River downstream of Iron Gate Dam Table C1. Ten lowest average September flow years sorted by Klamath River gaging station. * = Flow falls in the lowest tenth percentile for the period of record (1951- 2002 for KNK, KAO, KSV, TRH and TRH+KAO and 1961-2002 for KIG). 2002 data is in bold-red and 2001 is in bold-blue

Table C2. Comparison of average September flows for low-flow years at various Klamath River and Trinty River Stations with means, maximums and minimums for the period of record of 1951-2002 at all stations except KIG. KIG period of record is 1961-2002

Table C3. Comparison of average September flow exceedence values (cfs) for modeled unimpared flows from Hardy and Addley 2001 with actual average September flows for the Klamath River below Iron Gate Dam, near Seiad Valley and at Orleans

Table G1. Documentation of the methods used to sample and estimate the 2002 Klamath River fall Chinook run size

Table G2. Age Composition of the 2002 Klamath River fall Chinook run as determined by the Klamath River Technical Advisory Team, with assistance from DFG's Klamath and Trinity River projects

Table G3. Summary of release and recovery data for coded-wire tags (CWTs) recovered during the 2002 lower Klamath River fish kill investigations

Table H1. Weekly percent composition of Trinity River and Iron Gate Hatchery coded-wire-tagged Chinook salmon recovered in the lower Klamath River sport creel harvest, 2002 and 1988-2001 average

Table V1. Comparison of fall Chinook runs in the Klamath basin in 1989, 1996, 2001, and 2002

Table V2. Production goals for anadromous salmonids at TRHat (DFG and USBR, 1996)

Table V3. Production goals established for anadromous salmonids at IGH (DFG and PacifiCorp, 1996)

Table V4. Sport and tribal net fishery quotas and estimated harvest of Klamath Basin, adult fall-run Chinook salmon, for the 2002 and 2001 seasons

Table V5. Modeled effects of underestimating the 2002 lower Klamath River fish-kill on 2003 fall-run Chinook salmon ocean abundance projections

VII

List of Figures:

Figure 1. Map of the Klamath River Basin showing pertinent features Figure C1. Average September flow values corrisponding to the percentage of years those average flows are exceeded for selected Klamath and Trinity River stations during the period of 1951 - 2002. Note that Iron Gate period of record is 1961 -

Figure C2. Mean daily flows for the Klamath River near Klamath gage (KNK) for the period from August 1 to October 15, 2001 and 2002

Figure C3. Mean daily flows for the Klamath River at Orleans plus Trinity River at Hoopa gages (KAO+TRH) for the period from August 1 to October 15, 2001 and

Figure C4. Mean daily flows for the Klamath River near Klamath gage (KNK) vs. the Klamath River at Orleans plus Trinity River at Hoopa gages (KAO+TRH) for the period from August 1 to October 15, 2002

Figure C5. Mean daily flows for the Klamath River below Iron Gate Dam from August 1 to October 15 for low-flow years

Figure C6. Mean daily flows for the Klamath River near Seiad Valley from August 1 to October 15 for low-flow years

Figure C7. Mean daily flows for the Klamath River at Orleans from August 1 to October 15 for low-flow years

Figure C8. Mean daily flows for the Trinity River at Hoopa from August 1 to October 15 for low-flow years identified on the Klamath River

Figure C9. Combined mean daily flows for the Klamath River at Orleans and the Trinity River at Hoopa from August 1 to October 15 for low-flow years

Figure C10. Mean daily flows for the Klamath River near Klamath from August 1 to October 15 for low-flow years

Figure D1. Regression of maximum daily September water temperatures in the lower Klamath River (1998 and 1999 at Omagaar and 2001 and 2002 at Terwer) against the combined flows of TRH+KAO. Slopes are not significantly different between 1998 and 1999 (t = 1.95, p > 0.05) or 2001 and 2002 (t = 1.04, p > 0.05)

Figure D2. Regression of maximum daily September water temperatures in the lower Klamath River (1999 at Omagaar and 2001-2002 at Terwer) against the combined flows of TRH + KAO. Slopes are significantly different between 1999 and 2001- 2002 (t = 4.68, p < 0.05)

Figure D3. Regression of maximum daily September water temperatures in the lower Klamath River (1998 and 1999 at Omagaar and 2001 and 2002 at Terwer) against maximum daily air temperatures at Terwer

Figure D4. Regression of maximum daily September water temperatures in the lower Klamath River (1998 and 1999 at Omagaar and 2001 and 2002 at Terwer) against minimum daily air temperatures at Terwer

Figure D5. Regression of maximum daily September water temperatures in the lower Klamath River (1998 and 1999 at Omagaar and 2001 and 2002 at Terwer) against maximum daily air temperatures at Orleans. Slopes are not significantly different between 1998 and 1999 (t = 1.93, p > 0.05) or 2001 and 2002 (t = 0.25, p > 0.05)

VIII

Figure D6. Regression of maximum daily September water temperatures in the lower Klamath River (1998 and 1999 at Omagaar and 2001 and 2002 at Terwer) against minimum daily air temperatures at Orleans. Slopes are significantly different between 1998 and 1999 (t = 2.04, p < 0.05) but not significantly different between 2001 and 2002 (t = 1.54, p > 0.05)

Figure D7. Regression of maximum daily September water temperatures in the lower Klamath River (1998 at Omagaar and 2001-2002 at Terwer) against maximum daily September water temperatures at Orleans. Slopes are significantly different between 1998 and 2001-2002 (t = 2.87, p < 0.05)

Figure D8. Regression of maximum daily September water temperatures in the lower Klamath River (1998 and 1999 at Omagaar and 2001 and 2002 at Terwer) against 7-day average maximum daily air temperatures at Orleans. Slopes are significantly different between 1998 and 1999 (t = 2.90, p < 0.05) and not significantly different between 2001 and 2002 (t = 0.36, p > 0.05)

Figure D9. Regression of maximum daily September water temperatures in the lower Klamath River (1998 at Omagaar and 2001-2002 at Terwer) against 7-day average maximum September water temperatures at Orleans. Slopes are significantly different between 1998 and 2001-2002 (t = 8.53, p < 0.05)

Figure D10. Regression of maximum daily September water temperatures in the lower Klamath River (1998 and 1999 at Omagaar and 2001 and 2002 at Terwer) against 7-day average minimum daily air temperatures at Orleans. Slopes are significantly different between 1998 and 1999 (t = 2.47, p < 0.05) and not significantly different between 2001 and 2002 (t = 0.90, p > 0.05)

Figure D11. Regression of maximum daily September water temperatures in the lower Klamath River (1998 and 1999 at Omagaar and 2001-2002 at Terwer) against 7-day average minimum September water temperatures at Orleans. Slopes are significantly different between 1998 and 1999 (t = 2.47, p < 0.05) and 1998 and 2001-2002 ( t = 2.78, p < 0.05), but not significantly different between 1999 and 2001-2002 (t = 0.32, p > 0.05)

Figure D12. Comparison of actual Terwer daily max with predicted max water temperatures from regression analysis for 2001-2002. Regression uses 7-day average max and min air temperatures at Orleans as the predictor of max daily water temps at Terwer

Figure D13. Actual and predicted max daily water temperatures at Terwer from 7-day running averages for max and min air temperatures at Orleans. Daily Max Wat T = 0.2516x(7-DRA Max Air T Orleans)+0.2010x(7-DRA Min Air T Orleans)+35.48, (r2=0.53 p<0.001)

Figure D14. Comparison of actual and predicted 7-day running average of September maximum daily water temperatures at Terwer

Figure D15. Daily maximum September water temperatures in the lower Klamath River at Terwer (2001 and 2002) and Omagaar (1998 and 1999)

Figure D16. 7-day running average of daily maximim water temperatures for September at Terwer (2001 and 2002) and Omagaar (1998 and 1999)

Figure D17. Comparison of September 2002 daily maximum water temperatures for various Klamath River stations

Figure D18. Comparison of September 2002 7-day running averages of daily maximun water temperatures for various Klamath River stations

IX

Figure D19. Regression of maximum daily September water temperatures in the lower Klamath River for 2001-2002 at Terwer against average daily September flows at TRH+KAO </= 1,900 cfs and > 1,900cfs. Slopes are significantly different between </= 1,900 cfs and > 1,900 cfs plots (t = 2.13, p < 0.05)

Figure D20. Average daily September flows for TRH+KAO during 1998, 1999, 2001 and 2002

Figure E1. Dissolved oxygen concentrations from Klamath River at Terwer on a 30 minute recording interval from September 12, 2002 at 11:00 am to September 19, 2002 at 4:00 pm. Provisional data provided by the Yurok Tribe

Figure E2. Minimum and average daily dissolved oxygen concentrations from the Klamath River at Terwer for August 19 - September 27 of 2001 and 2002.

Provisional data provided by the Yurok Tribe

Figure E3. Dissolved oxygen concentrations for Klamath River at Martins Ferry recorded at 30 minute intervals for August - September 2001 and 2002. Provisional data provided by the Yurok Tribe

Figure E4. Dissolved oxygen concentrations for Trinity River near the mouth recorded at 30 minute intervals for August - September 2001 and 2002. Provisional data provided by the Yurok Tribe

Figure E5. Dissolved oxygen concentrations for Klamath River at Weitchepec recorded at 30 minute intervals for August - September 2001 and 2002. Provisional data provided by the Yurok Tribe

Figure F1. River stage for the Klamath River near Klamath during September of low-flow years. Data were not available for 1973 and 1981

Figure F2. River stage for the Klamath River near Klamath during September of non-low-flow years since 1993

Figure G1. Total in-river run-size estimates for fall-run Chinook salmon in the Klamath Basin since 1978

Figure G2. Total in-river run-size estimates for fall-run Chinook salmon in the Klamath Basin for low-flow years

Figure H1. Run-timing for Chinook salmon in the Klamath River Estuary as the number of fish collected by beach seine during 1977, 1979, 1981, and 1984 - 1990

Figure H2. Run-timing for Chinook salmon in the Klamath River Estuary as a weekly average of the number of fish collected by beach seine during 1977, 1979, 1981, and 1984 - 1990

Figure H3. Average weekly expanded coded-wire tag returns from lower Klamath River sport creel surveys, 1988-2001, for Trinity River Hatchery spring-run and fall-run Chinook and Iron Gate Hatchery fall-run Chinook salmon

Figure H4. Average weekly expanded coded-wire tag returns from lower Klamath River sport creel surveys, 1988-2001 average versus 2002, for Trinity River Hatchery spring-run and fall-run Chinook and Iron Gate Hatchery fall-run Chinook salmon

Figure H5. Average weekly expanded coded-wire tag returns for Iron Gate Hatchery fall Chinook salmon from Klamath River sport creel surveys, 1988-2002 and average

Figure H6. Average weekly expanded coded-wire tag returns for Trinity River Hatchery fall Chinook salmon from lower Klamath River sport creel surveys, 1988-2002 and average

Figure H7. Average weekly expanded Iron Gate and Trinity River Hatchery fall Chinook salmon CWT recoveries for low-flow years (1988, 1991, 1992, 1994 and 2001) from the lower Klamath River sport creel census. Data for 1973 and 1981 were not available

X

Figure H8. Average weekly expanded Iron Gate and Trinity River Hatchery fall-run Chinook salmon CWT recoveries for non low-flow years (1989,1990, 1993, 1995 - 2000) from the lower Klamath River sport creel census

Figure H9. Average weekly catch per unit effort of adult Chinook salmon in the lower Klamath River sport creel survey during low-flow years (1988, 1991, 1992, 1994, 2001 and 2002)

Figure H10. Comparison of average weekly catch per unit effort of adult Chinook salmon in the lower Klamath River sport creel survey and Yurok Tribal Klamath Estuary net harvest, 1994 - 2001 average compared to 2002

Figure V1. Trinity River spring-run and fall-run Chinook salmon run-size estimates. No spring-run estimates were available for 1983 and 1995. 2002 estimates are provisional and may be subject to change

Figure V2. Regression of spring-run vs. fall-run Chinook salmon population estimates for the Trinity River since 1978. Point estimate and 95% CI is indicated for 2002 fall-run. Y fall-run = 1.7 (X spring-run) + 13,969, (r2 = 0.57, p < 0.01)

Figure V3. Salmon River spring-run and fall-run Chinook salmon run-size estimates since 1980

Figure V4. Regression of spring-run vs. fall-Run Chinook salmon population estimates in the Salmon River. Point estimate and 95% CIs are indicated for 2002 fall-run. Y fall-run = 2.6 (X spring-run) + 1,403, ( r2 = 0.48, p<0.01)

Figure V5. Timing of Path 1 Chinook salmon carcass recoveries (Path 1 = fresh carcasses with clear eyes and/or firm flesh) in the Scott River spawning ground surveys for 2001 and 2002

Figure V6. Run timing of Chinook salmon observed at the Shasta River Fish Counting Facility in 2001 and 2002

I. Introduction:

A substantial fish-kill occurred on the lower Klamath River, California (Figure 1), in September 2002. The kill was first reported to the California Department of Fish and Game (DFG) on September 19, 2002. A cooperative effort, including DFG, U.S. Fish and Wildlife Service (USFWS), U. S. Forest Service (USFS) and the Yurok, Hoopa, and Karuk tribes, was implemented to evaluate the numbers of fish killed. Cooperators conducted surveys on September 20, 24, and 27, 2002. Results of the fish-kill surveys were summarized and reported by USFWS (USFWS 2003a).

The fish-kill took place on the lower 36 miles of the Klamath River, extending from the river’s mouth upstream to Coon Creek Falls and entirely within the Yurok Indian Reservation. Actual beginning and ending dates of the fish-kill are unknown; however, dead fish were observed between at least September 18 and October 1, 2002.

The USFWS estimated over 34,000 adult fish died (USFWS 2003a). This total included:

32,553 fall-run Chinook salmon (Oncorhynchus tshawytscha), 344 coho salmon (O.

kisutch), 629 steelhead (O. mykiss), 311 Klamath smallscale sucker (Catostomus rimiculus), 87 sculpin (Cottus sp.), nine speckled dace (Rhinichthys osculus), one coastal cutthroat trout (O. clarki), one American shad (Alosa sapidissima), one green sturgeon (Acipenser medirostris) and 120 unidentified fish (USFWS 2003a). Of the 32,553 Chinook salmon lost, a little over 7,000 or about 22 % of those fish, were of hatchery origin, with the remainder being naturally produced fish (USFWS 2003a). Estimates were considered conservative due to the nature of conducting fish-kill investigations (American Fisheries Society [AFS] 1992), observations of biologists conducting the fish-kill survey (Bairrington 2002, personal communication), and past experience by DFG in conducting surveys for carcasses of spawned-out anadromous salmonids in the Klamath Basin (DFG 2000a and 2002c).

DFG began to identify and evaluate potential causative factors and impacts of the fish-kill on September 19, 2002. This report summarizes the findings of that effort. The purposes for preparing this report were: to document conditions under which the fish-kill occurred; to assess and identify primary and underlying factors leading to the 2002 Klamath River fish-kill; to identify measures that if implemented would reduce the potential for future fish-kills; and to allow better management of the fishery resources in the Klamath Basin in the future.

From a historic perspective, the September 2002 fish-kill represented a one-time, unprecedented event on the Klamath River. The focus of this report was to analyze potential factors contributing to the fish-kill, particularly from the perspective of what made 2002 different. Factors were considered individually and collectively for their role in causing the fish-kill. Factors considered included: disease; toxic substances; flow; air and water temperature; dissolved oxygen; fish passage and river geomorphology; run-type, timing and density; and run-size. Where possible, the report identified competing hypotheses and analyzed their consistency with available data. This report also compared the Klamath River fish-kill with other fish-kills, evaluated the potential impacts of the fish-kill on fishery resources in the Klamath River Basin1 and discussed potential fishery management implications. The report did not attempt to develop predictive models for future fish-kills, given that at least to date, this was a one-time event.

1 Throughout this report, frequent references to the Klamath River Basin or Klamath River System are made. These references are meant to be inclusive of all features within the basin (Figure 1). These features include the Klamath, Shasta, Scott, Salmon and Trinity rivers in the lower basin and the Klamath, Sycan, Williamson, Sprague and Lost rivers as well as Upper Klamath Lake in the upper Basin.

II. Study Area:

II. A. General Setting;

The Klamath River originates in south-central Oregon, east of the Cascade Mountain Range. The 263-mile river flows in a general southwesterly direction, bisecting the Cascade Range as it courses through Oregon into California. Once in California, the river continues flowing southwesterly, before turning northwesterly near its confluence with the Trinity River, thence continuing to the Pacific Ocean. The Klamath River enters the Pacific Ocean about 15 miles south of Crescent City, California (Figure 1). The mainstem Klamath River drains about 5,000 square miles in Oregon and 10,000 square miles in California. The Klamath River Basin is California’s second largest river system (DFG 2002a), and the state’s second most important salmon-producing river (California Advisory Committee on Salmon and Steelhead Trout 1988).

The Wood, Williamson, Sprague, and Sycan rivers are significant headwater tributaries.

These rivers join to form Upper Klamath Lake. Water flows from Upper Klamath Lake into Link River (approximately 1.2 miles long), and thence into Lake Ewauna near Klamath Falls, Oregon (FishPro 2000). The Klamath River officially begins at the lower end of Lake Ewauna. Significant tributaries do not enter the Klamath River until the Shasta River confluence at river mile (RM) 177 in California (Pacific Southwest Interagency Committee 1973). The Scott, Salmon, and Trinity rivers enter at RM 143, 66, and 44, respectively, as the Klamath River flows through California.

A number of smaller tributaries enter the Klamath River in California within the fish-kill area, downstream of Coon Creek Falls (RM 36). These include Ah Pah (RM 17), Blue (RM 16), Tarup (RM 8), McGarvey (RM 6), Terwer (RM 5), Waukell (RM 4), Hoppaw (RM 3), and Hunter (RM 1) creeks. These small tributaries often have very little flow during summer. However, Blue Creek accretions can range from 30 to 60 cfs during summer months.

PacifiCorp operates six hydroelectric facility dams on the mainstem system, beginning with Link River Dam (RM 253) at the outlet of Upper Klamath Lake and ending at Iron Gate Dam (RM 190). The total generating capacity of the six facilities is 153.8 megawatts (PacficiCorp 2000). Iron Gate Dam, in coordination with upriver facilities, is used to re-regulate flow in the Klamath River downstream of the hydroelectric facilities.

There are no major mainstem water projects downstream of Iron Gate Dam.

Water diversions by the U. S. Bureau of Reclamation (USBR) Klamath Project, and private diversions from the Klamath River Basin in Oregon and the Trinity River in California, result in notable changes in flow of the Klamath River. Similarly, in-basin diversions from the Scott and Shasta rivers, also reduce contributory flows to the Klamath River during summer and early fall.

The Klamath River Basin encompasses three major geologic provinces: the Cascade Mountains in its headwaters region, the Klamath Mountains in the middle, and the North Coast Ranges as the river nears the Pacific Ocean (Helley and LaMarche 1973). The general topography along the upper reaches of the river ranges from relatively flat to almost vertical canyon walls. Landforms along these reaches are mostly volcanic in origin, and include transition forms between the Cascade, Klamath and North Coast provinces. Basaltic and andesitic volcanic deposits are common. Drainages within the Klamath Mountains are deeply incised, and old land surfaces are commonly exposed along river channels. Granitic and ultramafic rocks, intrude into the highly metamorphosed volcanic and sedimentary rocks that lie beneath the mountains. A combination of sheared rocks, shallow soil profile development, and steep slopes are common in the North Coast Ranges.

Downstream of Iron Gate Dam, the river generally has a cobble-bed and pool-riffle channel form. Occasionally, granitic bedrock outcroppings occur. Downstream of Blue Creek, the substrate generally includes more gravel-cobble than further upstream.

River gradient is variable and generally decreases downstream of Iron Gate Dam.

Gradient averages about 17 ft-per-mile between Iron Gate Dam and the Scott River, 14 ft-per-mile between the Scott and Salmon rivers, 13 ft-per-mile between the Salmon and Trinity rivers, and 7.5 ft-per-mile between the Trinity River and Pacific Ocean. The river's gradient in its last five miles below Terwer Gage, drops only about six feet (1.2 ft-per-mile) before entering the Pacific Ocean.

Flows upstream of Keno Reservoir are largely dependent upon flows from the Oregon headwater tributaries and losses of water privately diverted for off-stream purposes.

Significant water is stored and then diverted for agricultural purposes during the spring-summer growing season, by private diverters and the USBR Klamath Project (Project).

The Project supplies water to approximately 240,000 acres within the upper Klamath Basin. USBR also supplies water to the Basin's Lower Klamath and Tule Lake National Wildlife Refuges. The actual acreage served by the Klamath Project from upper Klamath Lake, is about 176,000 acres. There are an additional 26,000 acres served from the Lost River watershed and 30,000 acres of non-farmed acreage in the national wildlife refuges.

USBR's flow regulation at Upper Klamath Lake, typically results in higher and earlier peak flows in the Klamath River, decreased summer flows, and greater annual flow variability (Balance Hydrologics, Incorporated 1996). However, in USBR comments received on the draft of this report, they suggest any early peak flows would probably be influenced by water diverted from the Lost River to the Klamath River, to avoid flooding in the Tule Lake area. Ecological effects of these flow modifications are commingled with downstream hydroelectric generation diversions, storage, and releases.

The Southern Oregon/Northern California Coasts (SONCC) Coho salmon, was listed as threatened in 1997 and the Lost River sucker (Deltistes luxatus), and shortnose sucker (Chasmistes brevirostris) as endangered in 1988, pursuant to the federal Endangered Species Act. California had listed the two sucker species as endangered in 1974 and now considers coho salmon as a candidate species. All three species occur in the Klamath

Basin. Prior to the federal listings, flow regimes from Upper Klamath Lake were altered by PacifiCorp to shape water releases, in order to optimize hydroelectric energy production. This shaping for energy production has altered flows and the ability to understand historic flow regimes on the Klamath River since installation of Link River Dam in 1921.

There is ample evidence to suggest that the historic flow regime of the upper Klamath River exhibited greater flow in the late spring and summer months and had a rather “smooth” hydrograph. This historic hydrograph was attributed to hydraulic buffering from a large storage capacity of natural wetlands in and around Tule Lake, and Upper and Lower Klamath lakes (Balance Hydrologics, Inc. 1996, Hardy and Addley 2001).

Although the Klamath Project now stores water in Upper Klamath Lake for spring and summer irrigation and for use on wildlife refuges, this stored volume may actually be considerably less than the historic natural storage capacity. Irrigation diversions and hydroelectric operations, contribute to altered flow regimes downstream of each impoundment and diversion facility. From 1960 to 2000, average impaired monthly flows at Iron Gate Dam ranged from 791 cfs in July to 3,733 cfs in March. Prior to the hydroelectric and irrigation projects, median monthly flows (50% exceedence values) at Iron Gate ranged from 3,640 cfs in April to 1,361 cfs in September (Hardy and Addley 2001). These flows were derived using daily flow records for the 1905-1912 period (which represented an above normal precipitation period) at the Keno gage and were corrected downward to represent a “normal year”.

Substantial water diversion and water use occurs in other areas of the Klamath River Basin. The California Department of Water Resources (CDWR 1997) estimated that current annual agricultural water use in the Shasta and Scott River basins totals 110,000 acre-feet and 71,800 acre-feet, respectively. In comparison, average annual irrigation and urban water use above Keno Dam in Oregon totals 503,700 acre-feet (CDWR 1997).

Water use and diversions have significantly changed in the Trinity River over the past 50 years. The USBR Trinity River Diversion (TRD) is a feature of the Central Valley Project. The TRD was completed in 1964 and diverts a significant proportion of the Trinity River to the Sacramento River Basin, for agriculture and hydroelectric generation.

From 1964 to 1986, an annual average of 1,146,800 acre-feet of Trinity River water was diverted out of the basin to the Sacramento River. From 1986 through 2000, those diversions were reduced to an average of 732,400 acre-feet annually (CH2M Hill 2000, USFWS and Hoopa Valley Tribe 1999). Prior to the TRD, summer and early fall flows in the Trinity River near Lewiston, ranged from less than 100 cfs in dry years to 300 cfs in wet years (CH2M Hill 2000, USFWS and Hoopa Valley Tribe 1999). Summer and early fall flows in the Trinity River near Lewiston, were held at 300 cfs from 1978 through the 1990s. Releases at Lewiston Dam during the 2002 fish-kill were 450 cfs.

This flow was established for dry-year conditions in the Trinity River Mainstem Fishery Restoration Environmental Impact Statement/Report (CH2M Hill 2000) and by court order.

Climate in the Klamath River Drainage is generally characterized by damp, mild winters, and dry hot summers. Temperatures become more moderate and precipitation increases, as the river approaches the Pacific Ocean. Mean annual precipitation ranges from about

13.5 inches near Klamath Falls in the upper basin, to about 110 inches near Blue Creek (RM 16). The upper drainage receives most of its precipitation as winter snowfall.

Rainfall dominates the lower reaches of the drainage.

Vegetation along the river reflects precipitation patterns. Upstream of Keno Reservoir, vegetation typifies river/wetland complexes and agricultural development. Down-river in the steep sided canyon, pine, oak and shrub predominate on the steep, rocky slopes.

Juniper woodland communities occasionally occur. Limited stands of Douglas fir infrequently occur below upper canyon rims. Proceeding down-river, there is a slow transition from primarily mixed pine-oak-juniper-shrub stands, to conifer forest-oak woodlands, to primarily conifer forests. Upslope vegetation downstream of Blue Creek primarily is conifer forest, interspersed with oak woodlands.

II. B. Fish Resources;

Historically, much of the Klamath River was home to abundant runs of anadromous salmon, steelhead, Pacific lamprey, and other species, as they migrated to various tributaries of the river and Upper Klamath Lake (Oregon Department of Fish and Wildlife [ODFW] 1997). Fall and spring Chinook salmon are believed to have spawned within the Sprague River System of the Upper Klamath Basin (Klamath River Basin Fisheries Task Force 1992). Runs of Chinook were thought to go as far up the Sprague River as Beatty, Oregon, and spawning was reported in the north and south forks of the Sprague. These fish runs were halted as early as 1910 by the construction of Copco I Dam, which permanently blocked fish passage (City of Klamath Falls 1986). Today, Iron Gate Dam forms the upstream limit of anadromy, because this dam does not have fish passage facilities. This discussion of fish resources is limited to the river downstream of Iron Gate Dam.

Sixteen species of freshwater fishes are native to the Klamath River downstream of Iron Gate Dam, and there are at least ten introduced species (Table 1). Many of these native fishes once supported significant in-river fisheries. The Klamath River once supported a highly productive fishery for Chinook salmon, coho salmon and steelhead trout. For example, from 1918 to 1930, the annual Chinook salmon catch ranged from 11,500 to 61,500 fish annually (Snyder 1931), although the fishery had already been greatly reduced at that point. Today, Chinook salmon, coho salmon and steelhead trout are less abundant. Coho salmon are listed as threatened pursuant to the federal Endangered Species Act, and DFG was directed by the California Fish and Game Commission on February 4, 2004 to prepare a rules package for listing coho pursuant to California's Endangered Species Act. Pacific lamprey once supported a significant Native American fishery. Today, lamprey continues to be a sought-after species, but the fishery is greatly reduced.

Table 1. Fish species in the Klamath River downstream of Iron Gate Dam.

Scientific name Common name Anadromous/

Resident Species Status

Native species Lampetra tridentata Pacific lamprey A Proposed 1/ Lampetra ayresi river lamprey R Proposed 1/ Lampetra pacifica 2/ Pacific brook lamprey R Lampetra similis 2/ Klamath River lamprey R Lampetra richardsoni western brook lamprey R Proposed 1/ Oncorhynchus tshawytscha Chinook salmon A Oncorhynchus kisutch coho salmon A Threatened 3/ Oncorhynchus mykiss steelhead A Oncorhynchus clarki cutthroat trout A Rhinichthys osculus speckled dace R Catostomus rimiculus Klamath smallscale sucker R Catostomus snyderi Klamath largescale sucker R Cottus aleuticus coastrange sculpin R Cottus asper prickly sculpin R Cottus klamathensis marbled sculpin R Acipenser medirostris green sturgeon A Acipenser transmontanous white sturgeon A Thaleichthys pacificus eulachon A Gasterosteus aculeatus threespine stickleback A Introduced species Notemigonus crysoleucas golden shiner R Pimephales promelas fathead minnow R Ictalurus nebulosus brown bullhead R Perca flavescens yellow perch R Micropterus salmoides largemouth bass R Lepomis cyanellus green sunfish R Pomoxis nigromaculatus black crappie R Archoplites interruptus Sacramento perch R Alosa sapidissima American shad A Salmo trutta Brown trout R

1. Petitioned for listing as threatened in 2000 pursuant to the U.S. Endangered Species Act.

2. The distribution of Pacific brook lamprey and Klamath River lamprey in the Klamath

River system is unclear.

3. Listed as threatened pursuant to the U.S. Endangered Species Act. Currently considered warranted for listing by the California Fish and Game Commission pursuant to the California Endangered Species Act.

III. Factors Investigated:

III. A. Disease;

III. A. 1. Introduction

Large adult fish-kills caused by pathogens and disease, are relatively rare in the wild and much more common in hatcheries. The 2002 fish-kill was unprecedented and represented the first recorded massive mortality of adult anadromous fish on the Klamath River. Traxler et al. (1998) reported high pre-spawning mortality during 1994 and 1995, due to ichthyophthiriasis in sockeye salmon in the Babine River (Skeena River System), British Columbia. Until that time Ichthyopthirius multifilis (ich), the pathogen responsible for ichthyophthiriasis, had been reported in numerous species of freshwater anadromous fish in Canada, but no epizootics had occurred in wild salmonids. Traxler et

al. (1998) identified high densities of fish as the primary factor for the epizootic. In this case, sockeye salmon were held in high densities below weirs for several weeks, before they were allowed to migrate into artificial spawning channels. High densities stressed the fish and made them more susceptible to ich infection and eventual death (Traxler et

al. 1998). Although ambient water temperature has been identified as an important factor in the rate of development of ich (Warren 1991), water temperatures during the epizootics in 1994 and 1995 were similar to prior years, when significant pre-spawning mortality was not evident (Traxler et al. 1998).

Similar fish-kills occurred at two widely separated spawning channels on the Frasier River, British Columbia in 1995 (Traxler et al. 1998). Losses were also attributed to ich and affected wild adult sockeye and Chinook salmon. The combination of warm water and high fish density were identified as factors related to the severe ich outbreaks (Higgins 2002, personal communication).

The Rogue River in southern Oregon has experienced extensive pre-spawning mortality of both spring and fall-run Chinook salmon (ODFW 2000). Prior to the start of operation of Lost Creek Dam in 1978, ODFW indicated at least three years in which large pre-spawning mortalities of spring and fall Chinook salmon occurred. In addition, pre-spawning mortality was mentioned in 13 of 30 years, in which reports were written (ODFW 1992). Annual rates of pre-spawning mortality were as high as 70.2% for spring Chinook (ODFW 2000) and 81% for fall Chinook (ODFW 1992). The three largest recent spring-run Chinook fish-kills were documented in 1987, 1992 and 1994, when losses were 31,579, 13,684 and 20,134 fish, respectively (ODFW 2000). Disease was implicated as the probable immediate cause of both spring and fall-run Chinook pre-spawning mortality, but death could not be attributed to any one organism (ODFW 1992, 2000). The bacterium Flavobacterium columnare (columnaris) was the disease pathogen most often found in dead and dying fall-run Chinook. Fall-run Chinook mortality rates were positively related to water temperatures in the Rogue River Canyon (RM 34.2 – 68.4) during late summer. As water temperatures increased, the mortality rate increased.

Regression analysis, indicated mortality rates of <1% at 66.2 ºF (19 ºC), 15% at 68.0 ºF

(20 ºC) and 89% at 69.8 ºF (21 ºC) (ODFW 1992). The effect of flow on mortality was not analyzed, but flow was closely correlated to water temperature (ODFW 1992). Fish density had less affect on mortality than water temperature (ODFW 1992).

In 2002 and 2003, Butte Creek, a tributary to the Sacramento River, experienced large pre-spawn mortalities of federally and state listed spring-run Chinook salmon. DFG estimated fish losses of 3,431 salmon (21%), out of a population of 16,028 adults in 2002 and 11,231 salmon (65 %) out if an estimated population of 17,294 adults in 2003 (DFG 2004). The unique life history of spring-run Chinook includes a protracted period of adult fresh water residency, which in Butte Creek is up to seven months prior to spawning, and thus exposes those fish to a higher risk of pre-spawn mortalities.

Additionally, water temperatures in the holding/spawning reach of Butte Creek, are generally higher during the months of July and August than is optimal for Chinook salmon, with average daily water temperatures during the period generally exceeding 62° F. In both 2002 and 2003 peak mortalities occurred from late July until early September (96% in 2002 and 99% in 2003) following abnormally high water temperatures. During 2002, water temperatures at a key site in the holding reach of Butte Creek peaked at 69.4° F during mid-July, while in 2003, the peak was 69.7° F in late July. While the entire holding spawning reach of Butte Creek is affected by a Pacific Gas and Electric Company (PG&E) hydropower project, DFG concluded the pre-spawning losses were primarily due to large numbers of fish concentrated in limited holding pools, high water temperatures, and an outbreak of two pathogens, columnaris and ich (DFG 2004). DFG also concluded that operation of the hydropower project provided a net benefit to salmon holding and spawning, but also identified the need to further evaluate operational changes that would potentially decrease water temperatures.

All the aforementioned wild epizootics appear to relate to at least one of four main factors. Those factors including; restricted fish passage, high densities of fish, warm water temperatures, and low-flow, can act either individually or in concert to create conditions favorable for an epizootic.

Warren (1991) stated: “Disease can occur when susceptible fish encounter virulent pathogens and adverse environmental conditions stress the fish.” Many fish pathogens are present in the Klamath River System. The major fish pathogens documented in recent years include the myxozoan parasite Ceratomyxa shasta (ceratomyxa), the trematode Nanophyetes salmincola, and the bacterium columnaris (Foott et al. 2002).

When environmental conditions are not stressful to fish, disease may not occur, even though fish are infected or in contact with pathogens. However, disease can occur when environmental conditions degrade, such as with increased water temperature, decreased flow, and increased fish density (Warren 1991, Lasee 1995). Poor environmental conditions are stressful to fish and result in compromised immune function, making fish more susceptible to disease (Wedemeyer 1970). Certain protozoan parasites having direct life cycles (i.e. no intermediate host is required), can multiply to high numbers quickly, when water temperatures and flows are favorable (Noga 1996). Consequences of disease by any of the serious pathogens are lethargy, debilitation, weight loss, and mortality of the host fish.

The purpose of this section is to address the null-hypothesis; the September 2002 Klamath River fish-kill was not caused by disease. The alternative hypothesis is; the Klamath River fish-kill was caused by disease. To address these hypotheses, this section will review pathology reports prepared by DFG and USFWS on the fish-kill, and determine if disease was a major factor in the death of fish in the lower Klamath River during September 2002.

III. A. 2. Methods

On September 25, 2002, the DFG Fish Health Laboratory in Rancho Cordova was notified of a large fish-kill on the lower Klamath River, and was requested to investigate disease as a possible cause of mortality. Pathological evaluation of freshly dead and live Chinook, coho and steelhead (caught by anglers), was made on September 26 near Blake’s Riffle (approximately RM 8) and September 27 near Blue Creek (RM 16).

Standard necropsy procedures for evaluation of fish were followed (Thoesen 1994). Fish were observed externally for gross pathological lesions on the body surface, fins, gills and eyes. Samples were collected by Tresa Veek, DFG Associate Fish Pathologist.

Tissue squash preparations were made from lesion areas of the skin and gills, on microscope slides using 22 mm² cover slips. Microscope slides were placed in ziplock plastic bags and stored on ice for approximately two hours prior to reading. Direct observations were made using phase contrast microscopy with a Nikon Labophot-2 compound microscope at 40x and 400x magnification. The microscope was set up at the Terwer launch site using electrical power supplied by an inverter connected to a car battery. Internal pathological lesions were recorded and squash preparations of intestinal lesions were examined as described above.

III. A. 3. Results

Concurrent investigations by DFG and USFWS, CA-NV Fish Health Center, were conducted and independent reports of the fish-kill prepared. The DFG and USFWS pathology reports are provided in Appendix A. At Blake’s Riffle, heavy ich infestations of gill tissue were the predominant finding. Lesions typical of columnaris, were not seen, nor was the bacterium isolated. Intestinal lesions, typical of ceratomyxa, were not observed in any of twenty dead Chinook salmon, nor in any of four live specimens (one steelhead, two Chinook and one coho).

At Blue Creek, three freshly dead (two Chinook and one coho) and four moribund Chinook salmon were examined. Heavy ich infestations were observed in five of seven samples, and columnaris lesions and bacteria were present in seven of eight samples.

Ceratomyxa was also noted in one moribund Chinook salmon of eight fish sampled.

Motile protozoan parasites, such as ich, either leave fish hosts or degrade soon after death of the fish host. Only one Chinook salmon of three dead fish had ich, while four of four live Chinook were infested.

III. A. 4. Findings

The USFWS (2002), CA-NV Fish Health Center concluded: “Gross clinical signs of swollen gills containing ich and the high incidence of gill rot (columnaris) provide strong evidence that disease, induced by these two contagious pathogens, was the immediate cause of the fish-kill.” Fish examined by DFG’s Fish Health Laboratory on September 26 and 27, found ich and columnaris to be the principal causes of disease and death. Fish caught by anglers were primarily infected with ich. Moribund and fresh dead fish were heavily infected with both ich and columnaris. Fish dead for several hours or more were not examined for ich, because tissues were too degraded for analysis.

Fish entering the lower Klamath River in…