US Army Corps of Engineers Rogue Basin Dam Fisheries Evaluation Effects of Lost Creek Dam on the Distribution and Time of Chinoo

OREGO@ FISH & WILDLIFE REFERENCE LIBRARY

US Army Corps of Engineers %Fish & Wilcffilz; Portland District

Rogue Basin Dam

Fisheries Evaluation

Effects of Lost Creek Dam on the Distribution and Time of Chinook Salmon Spawning in the Rogue River Upstream of Gold Ray Dam

May 1991 EFFECTS OF LOST CREEK DAM ON THE DISTRIBUTION AND TIME OF CHINOOK SALMON SPAWNING IN THE ROGUE RTVPD 11P1ZTDr'A OF rnin DAv nAM

Rogue Basin Fisheries Evaluation Project Research and Development Section

Oregon Department of Fish and Wildlife

2501 SW First Street P.O. Box 59 Portland, OR 97207

May 1990

The United States Army Corps of Engineers funded this project under contract DACW57-77-C-0033. CONTENTS

Paqe

FOREWORD ...... ii

SUMMARY 1 ...... Spawning I Distribution ...... Spawning Ti 1 ...... RECOMMENDATIONS 2 ...... INTRODUCTION 3 ...... METHODS 4 ...... RESULTS 6 ...... Spawning Distribution 6 ...... Spawning Time 10 ...... DISCUSSION 17 ...... ACKNOWLEDGEMENTS 20 ...... REFERENCES 21 ......

APPENDIX A. Tables of Data Relating to Studies of Chinook Salmon 23 ...... APPENDIX B. Figures Showing Annual Relationship Between Location of Carcass Recovery and Date of Tagging 30 ...... APPENDIX C. Figures Showing Annual Relationship Between Date of Carcass Recovery and Date of Tagging 34 ...... APPENDIX D. Relation between Gregorian day, day-of-year and week-of-year 38 ...... FOREWORD

This report is based on 14 years of research funded by the U.S. Army Corps of Engineers. A study of this duration has necessarily involved the collective effort of many people since its inception in 1974. For this reason, it is being presented as a staff report of personnel on the Rogue Basin Fisheries Evaluation Project. The report was drafted by Thomas Satterthwaite who was largely responsible for the study design and for the analyses contained in the report. Barry McPherson supervised the project and critically reviewed the analyses, conclusions, and recommendations in the final document.

Research on the distribution and time of chinook salmon spawning above Gold Ray Dam was in response to concerns by fishery managers on the increased return of fall chinook salmon to spawning areas historically used by spring chinook salmon following the impoundment of water by Lost Creek Dam. The research was primarily an outgrowth of more intensive studies of chinook salmon and steelhead populations that began in the Rogue River in 1973. James Lichatowich was responsible for the original design and guidance of research on anadromous salmonids affected by the operation of Lost Creek Dam. These duties were subsequently assumed by Steven Cramer who served as program leader until 1985. Their leadership and insights on study designs were largely responsible for the ultimate success of research conducted by personnel in the Rogue Basin Fisheries Evaluation Project. The mainstem and tributaries of the Rogue River collectively produce the largest population of wild anadromous salmonids in Oregon. The Rogue River supports recreational and commercial fisheries of immense importance to Oregon citizens and is nationally renowned for its diversity and productivity. Authorizing documents for Lost Creek Dam stipulate that fisheries enhancement is to be an important benefit of the dam, mainly through improved temperature and flow. We hope our studies will ensure that these benefits are achieved for present and future generations of Oregon citizens.

Alan McGie Life History Studies Program Leader Research and Development Section Oregon Department of Fish and Wildlife Corvallis, Oregon

14 March 1990 SUMMARY

In this report, we evaluate the effect of the operation of Lost Creek Dam on the spawning distribution and the spawning time of chinook salmon Oncorhynchus tshawytscha in the Rogue River upstream of Gold Ray Dam. The Oregon Department of Fish and Wildlife conducted this study because managers became concerned about an increased run of fall chinook salmon to spawning areas historically used by spring chinook salmon. This change was not evident until juveniles produced during postimpoundment years matured and returned to spawn (Satterthwaite 1987). A summarv of our findinas fnllnw--

Spawning Distribution

1. Adults that migrated earliest spawned farthest upstream.

2. Spawning of fall chinook salmon and spring chinook salmon overlapped in the area between Trail Creek and the pool behind Gold Ray Dam. Few fall chinook salmon, but many spring chinook salmon, spawned upstream of Trail Creek.

We 3. did not detect any change in the spawning distribution of fall chinook salmon, possibly because we sampled few adults that originated from preimpoundment broods.

4. Spring chinook salmon broods produced after full operation began at Lost Creek Dam spawned farther downstream compared with preimpoundment broods.

A 5. decrease in the relative abundance of early migrating adults, compared with late migrating adults, was responsible for the downstream shift in the spawning distribution of spring chinook salmon.

6. Increased water temperature during the period eggs and alevins incubated in the gravel, or increased harvest rate within the sport fishery upstream of Gold Ray Dam, may have decreased the relative abundance of early migrants among wild spring chinook salmon that returned to the Rogue River.

Spawning Time

1. Adults that migrated earliest spawned earliest.

2. Fall chinook salmon excavated few redds of spring chinook salmon, but probably interbred with the spring race. In the area where spawning of fall and spring races overlapped, spawning time differed little between races.

3. Spring chinook salmon broods produced after full operation began at Lost Creek Dam spawned later compared with preimpoundment broods. Time of spawning did not change among cohorts of hatchery origin. 4. The change to later spawning was most pronounced for early migrating spring chinook salmon. Late migrating adults were less affected.

I 5. Spawning time correlated with water temperature when eggs and alevins incubated in the gravel. Adults spawned later when broods were exposed to increased incubation temperatures. Spawning time was not correlated with river physical parameters during spawning. 6. Later spawning of wild spring chinook salmon that originated from postimpoundment broods was probably the result of decreased survival rate among progeny of early spawning adults. Increased harvest rate within the sport fishery upstream of Gold Ray Dam may have also decreased the relative abundance of early spawning adults.

RECOMMENDATIONS

1. The evaluation of modified strategies of water temperature released from Lost Creek Dam should be completed. Decreased outflow temperature during autumn may increase the abundance of wild spring chinook salmon that migrate (and spawn) early in the season. Decreased outflow temperature may also decrease the proportion of fall chinook salmon among wild chinook salmon that spawn upstream of Gold Ray Dam. 2. Spawned carcasses should be surveyed annually in the mainstem between Cole M. Rivers Hatchery and Shady Cove, and in Big Butte Creek, as part of a long-term management program for spring chinook salmon in the Rogue River basin. Any significant changes in spawning time should be further evaluated by replicating the study described in this report.

Also, data from these surveys can be used to estimate the spawning escapement of wild spring chinook salmon. The estimated run of adults at Gold Ray Dam is not a reliable estimate of spawning escapement of wild fish because (1) unmarked hatchery fish compose a large portion of the run and (2) harvest and prespawning mortality of wild fish above Gold Ray Dam is not accurately estimated.

3. The migration time of wild and hatchery spring chinook salmon should be estimated annually at Gold Ray Dam as part of a long-term management program for spring chinook salmon in the Rogue River basin. Estimates of migration timing can be developed by scale analysis or by an expanded program of marking juveniles at Cole M. Rivers Hatchery. Analysis of scales would probably require extensive sampling at Gold Ray Dam, possibly as many as 200 adults every 2 weeks. A better option might be to mark a constant proportion of the release groups with fin clips. Based on the findings of Hankin (1982), we recommend a minimum mark rate of 25%. The marking rate should be constant between years and between release groups. Also, the selected mark should be visible as adults pass the counting station at Gold Ray Dam. 4. Efforts to enhance the production of spring chinook salmon should concentrate on the restoration of the early migrating component of the wild stock. Early migrants contribute to the river fisheries at a higher rate than late migrants. Wild fish also contribute at a higher rate than hatchery fish to the fishery upstream of Gold Ray Dam because wild fish remain in the river rather than enter Cole M. Rivers Hatchery.

2 INTRODUCTION

The Rogue River is the largest producer of spring chinook salmon Oncorhynchus tshawytscha among river basins south of the Columbia River. The stock contributes to commercial fisheries off the coast of northern California and Oregon (Jones 1988). Satterthwaite (1987) estimated that ocean landings of the 1971-76 brood years averaged about 60,000 wild fish annually. The stock also supports important recreational fisheries in the Rogue River. Cramer et al. (1985) estimated that freshwater harvest averaged about 7,000 adults annually during 1964-81. This estimate did not include jacks smaller than 60 cm (24 inches).

In recent years, fishery managers have expressed concern about the wild component of the run. Production (ocean catch plus freshwater escapement) broods 4 of produced after the first years of full operation of Lost Creek Dam at river kilometer (RK) 253 averaged 53% of the production from preimpoundment broods (Satterthwaite 1987). Also, the relative abundance of fall chinook salmon increased among wild adults that entered areas upstream of the counting station at Gold Ray Dam (RK 202)(Satterthwaite 1987).

The shift in race composition concerned fishery managers because the earlier-spawning spring race spawns in only a small portion of the basin, whereas the later-spawning fall race spawns in a large portion of the basin (Cramer et al. 1985). Some replacement of the spring race would occur if fall chinook salmon disturbed redds excavated by spring chinook salmon (McNeil 1964).

Surveys of spawned carcasses in the area upstream of the mouth of Elk Creek (RK 245) suggested that wild females that originated from postimpoundment broods spawned later than wild females that originated from preimpoundment broods (Satterthwaite 1987). Prior to the start of full operation at Lost Creek Dam, spring chinook salmon spawned most intensively in this area (Cramer et al. 1985). The later time of spawning may be the result of (1) fall chinook salmon spawning in the area or (2) later spawning by spring chinook salmon. Seemingly, there should be less concern if the spring race was spawning later.

A later spawning time among spring chinook salmon has management implications. Late spawners do not contribute as well as early spawners to the river fisheries. Satterthwaite (1987) reported a correlation between migration time and spawning time among chinook salmon produced before full operation began at Lost Creek Dam. The relationship indicated that late spawners Rogue migrated into the River later than early spawners. Consequently, later spawning by spring chinook salmon may indicate a later time of migration. Later migration would probably decrease freshwater harvest. Cramer et al. (1985) concluded that early migrating adults contributed to the river fisheries at a higher rate than late migrating adults.

To evaluate these concerns, the Oregon Department of Fish and (ODFW) Wildlife developed and conducted a study funded by the United States Army Corps of Engineers (USACE). The study was designed to (1) compare the spawning

3 distribution and spawning time of spring chinook and fall chinook salmon and (2) assess the effect of Lost Creek Dam on the spawning distribution and the spawning time of chinook salmon.

Lost Creek Dam was completed in 1976, but the reservoir only partially filled during 1977 because of drought. Storage capacity of the reservoir is 465,000 acre-feet, of which about 180,000 acre-feet is released during the summer and fall. A water intake structure allows for withdrawal from 5 levels when the reservoir is full. Selective withdrawal permits some manipulation of water temperature at release, particularly when the reservoir is stratified (USACE 1983). Flexibility in release flow and water temperature provides the opportunity to use release strategies designed to optimize production and harvest of salmon and steelhead 0. mykiss in downstream areas.

METHODS

We tagged chinook salmon that passed Gold Ray Dam with individually numbered tags. During 1974-78 and 1987 we used Floy t-bar tags and in 1986 we used aluminum jaw tags. We also tagged adults during 1979-81, but we chose not to analyze these data because the tags were not individually numbered and were color-coded by 2-week intervals. Chinook salmon that passed Gold Ray Dam prior to 16 August were classified as spring chinook salmon. Later migrants were classified as fall chinook salmon based on tagging studies at the time of river entry (Oregon Department of Fish and Wildlife, unpublished data). We recovered tags during weekly surveys for spawned carcasses in areas upstream of Gold Ray Dam (Table 1). Upon recovery of a tag, surveyors recorded the date and the river kilometer of the recovery site. We assumed that tagged fish spawned in close proximity to the recovery site.

Each week, surveyors walked 2 km up Big Butte Creek from its confluence with the Rogue River at RK 250. Few chinook salmon spawn farther upstream because of a waterfall that usually blocks fish migration during the early

Table 1. Areas surveyed for spawned carcasses of chinook salmon that migrated upstream of Gold Ray Dam, 1974-87.

River Survey Area kilometer period Years

Ray Dam Gold pool - Tou Velle Park 205-212 09/24-11/18 1974-81, 1986-87 Tou Velle Park - Dodge Bridge 212-223 09/24-11/18 1974-81, 1986-87 Dodge Bridge - Shady Cove 223-235 09/24-11/18 1974-81, 1986-87 Shady Cove - Rogue-Elk Park 235-245 09/17-11/18 1974-81, 1986-87 Rogue-Elk Park - Hatchery 245-252 09/17-11/18 1974-87

Big Butte Creeka 0-2 09/17-11/18 1974-81, 1986-87

a Enters the Rogue River at RK 250.

4 fall. We used driftboats to survey salmon carcasses in the Rogue River. We surveyed alternate banks of the channel on alternate weeks, except we surveyed every week in areas where carcasses tended to concentrate. We cut all retrieved carcasses in half to (1) verify the sex of the fish, and (2) prevent them from being counted during succeeding surveys. Surveyors also checked each carcass for tags or clipped fins.

We estimated the spawning time of wild female spring chinook salmon from weekly counts of carcasses without fin clips. We assumed that unmarked females of hatchery origin did hot bias the data. This assumptinn appeared reasonable because hatchery fish usually accounted for less than 10% of the carcasses recovered during surveys of spawned carcasses (Appendix Tables A-1 and A-2).

We also assumed that carcasses were counted 2 weeks after spawning occurred. This assumption was based on a postspawning longevity of 9 days (van den Berghe and Gross 1986) and a postmortem period of 5 days until discovery by a surveyor. A postmortem period of 3.5 days would have been appropriate if all areas were surveyed weekly. However, because some areas were surveyed on alternate weeks, we selected 5 days as an approximation of the postmortem period.

We grouped data from unmarked carcasses of spawned females into preimpoundment and postimpoundment broods. Because almost all wild females mature at age 4 or age 5 (Oregon Department of Fish and Wildlife, unpublished data), females counted during 1974-79 were designated as preimpoundment broods and females counted during 1981-87 were designated as postimpoundment broods. We grouped data from females that spawned during 1980 with preimpoundment broods because age 5 adults were twice as abundant as age 4 adults (Cramer et al. 1985).

We used the spawning time of females that entered Cole M. Rivers Hatchery (RK 252) just below Lost Creek Dam as a statistical control. Most of the females that entered the hatchery during 1975-77 were wild fish. By 1979, returning females were mostly of hatchery origin. During all years, hatchery personnel randomly selected adults to spawn. Consequently, spawning time at the hatchery should accurately reflect spawning time had the adults been allowed to spawn naturally (telephone conversation on 20 February 1990 with Michael Evenson, ODFW, Cole M. Rivers Hatchery, Trail, Oregon).

We made no attempt to estimate the quantity or quality of physical habitat available to either juvenile or adult spring chinook salmon. We assumed that temporal variations in channel morphology and physical characteristics of the substrate had no influence on either spawning time or spawning distribution during the time period of the study. We recognize that the habitat of spring chinook salmon is a dynamic, rather than static, variable. However, other than water quality, we lacked the data necessary to evaluate temporal changes in habitat.

Flow and water temperature of the Rogue River was estimated from data recorded at the automated gage near McLeod (RK 248) that was operated by the United States Geological Survey. We used data on maximum water temperature in analyses because mean water temperature was not estimated prior to 1979.

5 We chose P < 0.05 as a criteria for significance. We used analysis of variance to test for differences between means. To identify relationships between variables, we used correlation analysis and assumed data were random pairs of observations from a bivariate normal distribution. To quantify relationships between variables, we used regression analysis. Independent variables were assumed to be measured without error. We used analysis of covariance to compare slopes and elevations of regressions. Elevations of regressions were not compared when there were significant differences between regression slopes. Chi-square statistics, with corrections for continuity where needed, were calculated to test for differences between distributions. We referred to Snedecor and Cochran (1967) and Zar (1984) for analytical procedures.

RESULTS

We recovered 561 tags from chinook salmon that spawned upstream of Gold Ray Dam. Annual recoveries of tags ranged from a low of 11 in 1977 to a high of 171 in 1987. We recovered 260 tags from spring chinook salmon that originated from broods produced prior to the start of Lost Creek Dam operation. We also recovered 213 tags from spring chinook salmon produced during the postimpoundment years. However, tag recoveries from fall chinook salmon totaled only 19 from broods reared in preimpoundment years and 69 from broods reared in postimpoundment years. A summary of tag recoveries is presented in Appendix Tables A-3 and A-4.

Spawning Distribution The spawning distribution of fall chinook salmon overlapped the spawning distribution of spring chinook salmon in the Rogue River upstream of Gold Ray Dam. Tag recoveries indicated that the overlap was greatest in the area downstream of Trail Creek (RK 240). Few fall chinook salmon spawned in areas farther upstream (Figure 1). However, overlap in spawning distribution between spring and the fall races increased when adults that originated from postimpoundment broods returned to spawn. Tag recoveries from spring chinook salmon indicated that postimpoundment broods spawned farther downstream than preimpoundment broods. During 1974-78, 48% of the tag recoveries came from the area closest to Lost Creek Dam (RK 241-252). During 1986-87, only 31% of the tag recoveries came from this area. The difference in distribution of tag recoveries was significant (P < 0.001). However, there was no significant change in the percentage of fish that spawned in Big Butte Creek (Figure 1).

We did not detect a significant change in the spawning distribution of fall chinook salmon. Adults of preimpoundment origin and adults of postimpoundment origin spawned in the mainstem primarily between RK 205 and RK 217 (Figure 1). In this area, we found 68% of the total tags recovered during 1974-78 and 55% during 1986-87. The difference between the samples was 69 not significant (P = 0.58). Small samples sizes (19 in 1974-78 and in 1986-87) decreased the sensitivity of this analysis.

6 so

SPRING CHINOOK SALMON

EZ;1 PREIMPOUNDMENT BROODS IIIIIIIII POSIIMPOUNDMENT BROODS 60 -

40 -

20 -

I-0 L- 0 1--

z - LLI 80 C-) Of FALL CHINOOK SALMON LLJ IL

60 -

U2 PREIMPOLINDMENT BROODS IIIIIIIII POSTIMPOUNDMENT BROODS

40 -

20 -

0 - 205-217 217-229 229-241 241-253 BIG BUrrE CREEK RIVER KILOMETER ON ROGUE RIVER

Figure 1. Distribution of tags recovered from spawned carcasses of chinook salmon. Preimpoundment broods were represented by fish tagged and recovered during 1974-78. Postimpoundment broods were represented by fish tagged and recovered during 1986-87. Data were grouped so that areas were of similar distance.

7 We found that the location of spawning was related to the time of migration at Gold Ray Dam. Early migrants tended to spawn farther upstream than late migrants (APPENDIX B). The negative relationship between the two variables was significant for each year except 1974 (Table 2). During 1974, we were unable to recover any tags from late migrants that passed Gold Ray Dam. Correlation coefficients ranged between -0.47 and -0.53 for those years when we recovered at least 10 tags from spawners that passed Gold Ray Dam after 15 August (APPENDIX B).

Table 2. Regressions of river kilometer of tag recovery (Y) on day of tagging (X) for chi-nook salmon tagged at Gold Ray Dam and recovered as spawned carcasses. Day-of-year calendar is in APPENDIX D.

Year Regression SE of slope N r P

1974 Y 50 = 246.2 - 0.055X 0.159 -0.05 0.729 1975 Y 27 - 303.4 - 0.394X 0.111 -0.58 0.002 1976 Y = 294.7 - 0.342X 0.059 45 -0.66 <0.001 1977 Y = 292.7 - 0.27OX 0.060 9 -0.86 0.003 1978 Y = 279.2 - 0.235X 0.036 110 -0.53 <0.001

1986 Y 0.18OX - 266.9 - 0.030 98 -0.53 <0.001 1987 Y = 264.9 - 0.175X 0.026 160 -0.47 <0.001

Regressions of spawning location on migration timing did not differ significantly (P = 0.22 for slopes; P = 0.16 for elevations) among brood years produced prior to the start of Lost Creek Dam operation. Also, regressions of spawning location on migration timing did not differ significantly (P = 0.90 P for slopes; = 0.59 for elevations) among postimpoundment broods. Consequently, we used pooled data from 1974-78 to represent preimpoundment broods and used pooled data from 1986-87 to represent postimpoundment broods. We found that the relationship between spawning location and migration time did not change after Lost Creek Dam began full operation. Elevations and slopes of pooled regressions did not differ between preimpoundment and postimpoundment broods (Table 3). We also examined the mean kilometer of May spawning of fish tagged in and found no significant difference (P = 0.73) between preimpoundment and postimpoundment broods. Because there was no change in the spawning location of early migrating adults, we concluded that a decrease in relative abundance of early migrating adults was responsible for the downstream shift in spawning distribution among all spring chinook salmon. Analysis of the spawning distribution of wild female chinook salmon also indicated that the spawning distribution of spring chinook salmon shifted downstream after the juveniles produced after full operation began at Lost Creek Dam returned to spawn. Among wild females that originated from preimpoundment broods, an average of 54% spawned upstream of Elk Creek. Among counterparts that originated from postimpoundment broods, an average of 40%

8 spawned upstream of Elk Creek. The difference in means was significant (P = 0.015) and agreed with results of analysis of tagged fish.

Table 3. Regressions of river kilometer of carcass recovery (Y) on day of passage at Gold Ray Dam (X) for chinook salmon, preimpoundment broods (1974-78 returns) compared with postimpoundment broods (1986-87 returns). Day-of-year calendar is in APPENDIX D.

P for difference Years N Regression SE of slope r 01-evations @Slope@s

241 Y 1974-78 = 278.1 - 0.236X 0.026 -0.53 1986-87 258 Y 0.250 0.067 = 265.4 - 0.175X 0.020 -0.49

Results from correlation analyses of factors that could have influenced the spawning distribution of wild females were inconclusive. Annual estimates of the percentage of females that spawned upstream of Elk Creek correlated significantly with (1) flow during migration, (2) flow during spawning, and (3) water temperature during spawning (Table 4). Significant correlations among most of the variables made it difficult to identify factors that influenced the spawning distribution of wild chinook salmon that spawned upstream of Gold Ray Dam.

Table 4. Correlation matrix of variables used in analyses of spawning distribution of wild female chinook salmon, 1974-87.

Spawning Flow Water temperature distributions m-i-g-r-a-t-i-o-nb-s-p-a-w-n-in-g-c -sP--a-wn-i-n-gd--Incubationu

Spawning distribution 1.00 Migration flow -0.87f 1 00 Spawning flow -0.67f 0'63f 1.00 -O'81f Spawning temperature 0.80f -0.8of 1.00 Incubation temperature -0.49 0.53 0.35 -0.44 1.00 a Percent spawned b of wild females that upstream of Elk Creek. Mean flow (cfs) near McLeod during August. c Mean flow McLeod d (cfs) near during September-October. Mean maximum (*C) McLeod e water temperature near during September-October. Mean maximum water temperature (OC) near McLeod during October-January in year i-4 (period that eggs and alevins incubated in redds). f P < 0.05 in two-tailed test.

9 Water temperature four years earlier when fish incubated in the gravel as eggs and alevins was not correlated with spawning distribution of wild females (Table 4). This finding suggested that river physical factors to which adults were exposed prior to and during spawning had a greater influence on spawning distribution compared with environmental factors to which the broods were exposed during the incubation period. Data included in these analyses are presented in Appendix Table A-5.

Spawning distribution of chinook salmon was less related to spawning time than to migration time. Amnng tagged fish, spawning location correlated sianificantlv with SDawnino time in three of the seven Years that we sampled (Tible 5). In contrast, sp-awning location correlated significantly with migration timing in six of seven years (Table 2). Comparisons of correlation coefficients indicated that spawning distribution was more highly related to migration time than spawning time for tagged fish recovered from (Z preimpoundment broods = 3.05) and postimpoundment broods (Z = 4.46).

Table 5. Correlations of river kilometer of tag recovery and day of tag recovery for chinook salmon tagged at Gold Ray Dam and recovered as spawned carcasses.

Year N r P

1974 50 0.05 0.710 1975 27 -0.25 0.209 1976 45 -0.66 <0.001 1977 9 -0.93 <0.001 1978 110 -0.24 0.011 1986 98 -0.16 0.100 1987 160 -0.14 0.077

Spawning Time

Analysis of tag recoveries indicated that the spawning time of fall chinook salmon overlapped the spawning time of spring chinook salmon for postimpoundment broods that spawned in the area between the pool behind Gold Ray Dam and Trail Creek (RK 205-240). In this area, both races of chinook salmon spawned primarily between late September and the end of October (Figure 2), and peaked during the second week of October. Tag recoveries indicated that fall chinook salmon spawned an average of 5 days later than spring chinook salmon. We did not estimate the spawning time of fall chinook salmon in the area farther upstream because few spawned upstream of Trail Creek. Also, we did not estimate the spawning time of fall chinook salmon that originated from preimpoundment broods because few spawned upstream of Gold Ray Dam (Satterthwaite 1987).

Combined tag recoveries from all areas indicated that among spring chinook salmon, adults that originated from postimpoundment broods spawned

10 40 - L23 SPRING CHINOOK SALMON IIIIIIIII FALL CHINOOK SALMON

30 -

0 LI 0 20 - z U W OL 1 0

0 SEPTEMBER OCTOBE

Figure 2. Estimated spawning time of chinook salmon that originated from postimpoundment broods. Data were from tagged carcasses recovered downstream of Trail Creek, 1986-87.

later than adults that originated from preimpoundment broods (Figure 3). Among preimpoundment broods, spawning peaked during late September. Spawning by postimpoundment broods peaked 2 weeks later. From these tag recoveries, we estimated that 70% of the preimpoundment broods spawned during September, compared with only 38% of the postimpoundment broods.

Counts of unmarked carcasses also indicated that spring chinook salmon originating from postimpoundment broods spawned later than preimpoundment broods (Figure 4). In the area closest to Lost Creek Dam (RK 245-252), spawning by preimpoundment broods peaked during the middle or later portion of September, while spawning by postimpoundment broods peaked during early October. Time of spawning in the RK 235-245 area also peaked later, but the change was I week rather than 2 weeks (Figure 4). Because tagging studies showed that few fall chinook salmon spawned in these areas, later spawning by spring chinook salmon must be responsible for the change in spawning time.

Among spring chinook salmon spawning in Big Butte Creek, postimpoundment broods also spawned later than preimpoundment broods (Figure 4). We estimated that an average of 52% of the postimpoundment broods spawned during September, compared with an average of 34% of the preimpoundment broods. Spawning of postimpoundment broods peaked I week later than preimpoundment broods (Figure 4). Data included in analyses of data from unmarked carcasses in all areas are presented in Appendix Tables A-6 through A-8.

11 40 SPRING CHINOOK SALMON LZ3 PRaMPOUNDMENT BROODS IIIIIIIII POSnMPOUNDUENT BROODS

30 -

I--0 U- 0 ZU- Z U.1 7

10 -

@ I M j- 0 I - I - SEPTEMBER OCTOB R

Figure 3. Estimated spawning time of spring chinook salmon based on tagged carcasses recovered in all areas. Preimpoundment broods were represented by fish tagged and recovered during 1974-78. Postimpoundment broods were represented by fish tagged and recovered during 1986-87.

Annual estimates of the mean time of spawning for unmarked spring chinook salmon ranged between late September and early October in Big Butte Creek and in the two areas of the Rogue River (Figure 5). These data also showed that postimpoundment broods spawned later than preimpoundment broods. Mean time of spawning was significantly later for postimpoundment broods spawning in the RK P RK mainstem (P = 0.004 for 235-245 and < 0.001 for 245-252).

We did not detect a significant difference (P = 0.11) in the mean time of spawning between preimpoundment and postimpoundment broods that spawned in Big Butte Creek (Figure 5). Failure to detect a difference may have been related to small sample sizes. A sensitivity analysis suggested that we would have detected a change had we sampled 2 more years when postimpoundment broods spawned in Big Butte Creek, provided that time of spawning was similar to the other 3 years we sampled. Data included in these analyses are presented in Appendix Table A-9.

We also found no significant change in the median date of spawning by females that returned annually to Cole M. Rivers Hatchery (P - 0.185). The median date of spawning averaged 5 October for females that returned during 1975-80. During 1981-87, the median date of spawning by females that entered the hatchery averaged 3 October. Based on these results, we conclude that a decreased survival rate among progeny of early spawning adults was the factor that caused later spawning among wild adults produced from postimpoundment broods. Data used in this analysis are presented in Appendix Table A-10.

12 30 - M 1974-80 BIG BUTTE CREEK 1981+1986-87

20 -

10 .

not sampled 0

30 - M 1974-80 1981-87 RK 245-252

0 20 - 0 F-

LLJz

30 - E22 1974-80 1981+1986-87 RK 235-245

20 -

10-

not sampled d 0 SEPTEMBER OCTOBER

Figure 4. Estimated spawning time of spring chinook salmon based on counts of unmarked carcasses of spawned females. Fish recovered during 1974-80 composed preimpoundment broods. Fish recovered during 1981-87 composed postimpoundment broods.

13 BIG BUTTE CREEK 10/10 -

10/05 -

09/30

09/25 - not sampled

RK 245-252 10/10 - z z 10/05 - a. V)

U- C) 09/30 LLJ -

09/25 -

RK 10/10 - 235-245

10/05

09/30 -

09@/25 not sampled

1974 1976 1982 1986

Figure S. Estimated mean date of spawning for spring chinook salmon based on the count of unmarked carcasses. Open circles represent preimpoundment broods and closed circles represent postimpoundment broods. Brackets represent 95% confidence intervals associated with the means.

14 We found a significant relationship (r = 0.89, P < 0.001) between water temperature during the period eggs and alevins incubate in the gravel and the mean date of spawning by wild female spring chinook salmon in the Rogue River upstream of Elk Creek. The positive relationship indicated that adults spawned later when cohorts were exposed to increased water temperature four years earlier during incubation. Physical parameters of the river, including flow during the period of adult migration, and flow and water temperature during spawning, were not significantly correlated with spawning time. Data used in these analyses are presented in Appendix Tables A-5 and A-9.

Analysis of tag recoveries showed that spawning time correlated positively with migration time at Gold Ray Dam. Early migrants tended to spawn early, while late migrants tended to spawn later (APPENDIX C). Regressions of spawning time on migration time were significant for each of the 7 years of tagging (Table 6).

Table 6. Regressions of day of tag recovery (Y) on day of tagging (X) for chinook salmon tagged at Gold Ray Dam and recovered as spawned carcasses. Day-of-year calendar is in APPENDIX D.

Year Regression SE of slope N r P

1974 Y = 238.3 + 0.282X 0.112 60 0.31 0.014 1975 Y = 217.4 + 0.391X 0.068 32 0.72 <0.001 1976 Y = 238.4 + 0.26OX 0.039 50 0.70 <0.001 1977 Y = 259.6 + 0.178X 0.028 10 0.91 <0.001 1978 Y = 252.3 + 0.189X 0.026 126 0.55 <0.001

Y 1986 = 272.9 + 0.098X 0.020 ill 0.42 <0.001 1987 Y = 272.4 + 0.114X 0.019 171 0.42 <0.001

on Regressions of spawning time migration time did not differ (P = 0.08 for slopes and 0.12 for elevations) among preimpoundment broods. Also, on regressions of spawning time migration time did not differ (P = 0.57 for slopes and 0.17 for elevations) among postimpoundment broods. Consequently, we pooled data from 1974-78 to represent preimpoundment broods and pooled data from 1986-87 to represent postimpoundment broods.

We found that the relationship between spawning time and migration time changed after Lost Creek Dam became operational. Slopes of pooled preimpoundment and postimpoundment regressions differed significantly (Table 7). To make these findings easier to interpret, we used preimpoundment and postimpoundment regressions to predict the mean date of carcass recovery for spring chinook salmon that passed Gold Ray Dam on four dates representative of migration timing.

Results indicated that spawning time changed for spring chinook salmon that passed Gold Ray Dam from May through early July (Figure 6). Spawning time did not change for adults that migrated during late July and August. The

15 Table 7. Regressions of day of carcass recovery (Y) and day of passage at Gold Ray Dam (X) for chinook salmon, preimpoundment broods (1974-78 returns) compared with postimpoundment broods (1986-87 returns). Day-of-year calendar is in APPENDIX D.

P for difference Years N Regression SE of slope r Elevations Slopes

Y 1974-78 278 = 246.6 + 0.221X 0.018 0.59 < V Y VVI 1986-87 282 - 273.3 + 0.104X 0.014 0.41

10/31 0 PREIMPOUNDMENT BROODS 0 POSTIMPOUNDMENT BROODS

0 10/21 T LLJ T 0 of 0 L- I o LU 0 0 10/11 W F- 0 0 25 I W of 10/01 T - 0

1 MAY 1 JUNE 1 JULY 1 AUGUST

DATE OF MIGRATION AT GOLD RAY DAM

Figure 6. Predicted mean date of carcass recovery for spring chinook salmon passing Gold Ray Dam on four dates encompassing the period of migration. Preimpoundment broods (predicted from 1974-78 data) are compared with postimpoundment broods (predicted from 1986-87 data). Brackets represent 95% confidence intervals associated with the predicted values.

change in spawning time was most evident among early migrants. We estimated that among adults that passed Gold Ray Dam on I May, postimpoundment broods spawned an average of 12 days later compared with preimpoundment broods. In contrast, among adults that migrated on I July, the regressions predicted that postimpoundment broods spawned an average of 5 days later than preimpoundment

16 broods. These changes caused the temporal disparity in spawning time between early and late migrating adults to be less among postimpoundment broods than preimpoundment broods (Figure 6).

DISCUSSION

We found that the spawning distribution of spring chinook salmon and fall chinook salmon overlapped in the Rogue River upstream of Gold Ray Dam. Both races spawned in the area between the reservoir behind Gold Ray Dam and Trail Creek. Many spring chinook salmon, but few fall chinook salmon, spawned upstream of Trail Creek. Overlap in the spawning distribution of both races increased after Lost Creek Dam began full operation because spring chinook salmon from postimpoundment broods spawned farther downstream than preimpoundment broods.

Fall chinook salmon spawned in areas used by late-spawning spring chinook salmon. Because the difference in spawning time between the averaged 5 races less than days, we believe that few redds of spring chinook salmon were disturbed by fall chinook salmon. This belief is based on the assumption that female spring chinook salmon spent 14 days at the site of their redd and prevented other females from spawning at that site (Neilson and Banford 1983).

Fishery managers should be concerned about fall chinook salmon spawning in conjunction with spring chinook salmon. In the Rogue River basin, spring chinook salmon spawn only in the area upstream of Gold Ray Dam. In contrast, fall chinook salmon spawn in widely distributed areas throughout the Rogue River basin (Cramer et al. 1985). With increased spawning between races, racial differences in migration timing at Gold Ray Dam may become A less distinct. greater proportion of wild spring chinook salmon may migrate during July-August rather than during May-June. Such a shift in migration timing is not without precedence among chinook salmon (Slater 1963; Kwain and Thomas 1984).

Protracted interracial spawning would probably lead to a decline in the freshwater harvest of spring chinook salmon in the Rogue River. Late migrating adults contribute at lower rates to the river fisheries compared with early migrants. Early migrants pass through the fishery in the lower river (RK 5-43) at a time of optimal flow for harvest (Cramer et al. 1985). Early migrants also contribute better than late migrants to the fishery upstream of Gold Ray Dam because of a longer period of residence within the area of the fishery.

We found that adults that originated from postimpoundment broods spawned farther downstream in the Rogue River than adults that originated from preimpoundment broods. However, a similar percentage of preimpoundment and postimpoundment broods spawned in Big Butte Creek. This finding suggested that operation of Lost Creek Dam was responsible for the downstream shift in the spawning distribution of spring chinook salmon in the Rogue River.

We also found that adults that originated from postimpoundment broods spawned later than adults that originated from preimpoundment broods. The change in spawning time was greatest in the area immediately downstream of Lost Creek Dam and diminished with distance downstream. Although not as

17 pronounced as in the mainstem, we also noted a later spawning time for adults in Big Butte Creek. The simultaneous change in spawning time for adults in the mainstem and in Big Butte Creek suggested that operation of Lost Creek Dam may not have been responsible for the change in the spawning time of spring chinook salmon.

The possibility that Lost Creek Dam did not change the spawning time of wild spring chinook salmon assumes that few, if any, of the fry produced in the mainstem strayed as adults and spawned in Big Butte Creek. Because some straying may have occurred, we also examined the spawning time of female spring chinook salmon that returned to Cole M. Rivers Hatchery. Because the spawning time of wild adults changed, without a simultaneous change in the spawning time of adults that returned to the hatchery, we concluded that environmental conditions during the last few months prior to spawning were not responsible for the change in spawning time of wild adults.

During earlier work, we found that the operation of the dam did not affect the spawning time or spawning distribution of broods produced during preimpoundment years that returned to spawn in postimpoundment years (Cramer et al. 1985). The change in spawning time and spawning distribution of broods produced during postimpoundment years led us to conclude there was a change in the genetic composition of the stock. Changes in survival rate of juveniles or adults affected the genetic composition of the stock, and altered life history characteristics. Selective breeding practices among hatchery stocks is a good example of such a mechanism (Larkin 1981).

We believe that decreased survival among the progeny of spring chinook salmon that migrated early resulted in a downstream shift in spawning distribution among cohorts that survived to spawn. This conclusion was based on the finding that early migrants spawned farther upstream than late migrants. The decrease in the relative abundance of early migrants resulted in an increase in the relative abundance of late migrants that spawn in areas farther downstream.

We also believe that decreased survival among the progeny of early migrants resulted in a later time of spawning among cohorts that survived to spawn. This conclusion was based on the finding that early migrants spawned earlier than late migrants. The decrease in the relative abundance of early migrants resulted in an increase in the relative abundance of late migrants that spawn later. The relative abundance of early migrants could have decreased because of increased rate of adult mortality. Although the rate of natural mortality in the river did not increase after Lost Creek Dam became operational (Cramer et al. 1985), the rate of fishing mortality in the river probably increased during 1974-87. Angler effort increased greatly during that period (telephone conversation on 22 November 1989 with Michael Jennings, ODFW, Roseburg, Oregon). In addition, the closing date for the fishery in the area upstream of Gold Ray Dam changed from 15 July during 1973-77 to 31 July during 1978-87. Because early migrants resided in the area of the fishery for a longer period of time than late migrants, increased fishing effort may have caused a disproportionate increase in the harvest of early migrants compared with late migrants.

18 Cramer et a]. (1985) estimated that anglers harvested 33% of the spring chinook salmon that passed Gold Ray Dam during 1981. Actually, harvest rates could have been 50% and 10% for adults that passed Gold Ray Dam during May and July, respectively. Unfortunately, we do not have the data needed to assess the potential for differential harvest rates. To determine the relationship between migration timing and harvest rate, an intensive study including tagging, prespawning mortality surveys, and angler surveys throughout the fishery upstream of Gold Ray Dam would have to be conducted.

Differential production of juveniles from different groups of spawners is another factor that might be responsible for the change in the spawning time and spawning distribution of spring chinook salmon. A localized decrease in juvenile production could decrease the localized abundance of cohorts that returned to spawn because anadromous salmonids tend to spawn in natal areas (Ricker 1972; Horrall 1981). Factors postulated that might have decreased juvenile production in the localized area downstream from Lost Creek Dam included (1) changes in spawning habitat, (2) redd dewatering, and (3) increased water temperature during the autumn and early winter.

Quality of spawning habitat for spring chinook salmon might have decreased because of increased rates of armoring or sedimentation of gravel. The quantity of spawning habitat may have decreased with cessation of gravel recruitment from areas upstream of the dam. We believe such changes, if they occurred, would develop gradually. In contrast, observed changes in spawning time and spawning distribution of spring chinook salmon were abrupt, and exhibited minimal variation among postimpoundment broods.

We are more certain that redd dewatering was not responsible for changes in spawning time and spawning distribution of spring chinook salmon. Flow augmentation during spawning did not begin until 1981, yet changes in spawning parameters were evident among the 1977-80 broods. Consequently, we believe that increased water temperature is the more likely factor responsible for the relative decrease of the early spawning portion of the run.

Satterthwaite (1987) estimated that the operation of Lost Creek Dam decreased production of spring chinook salmon fry by an average of 33% annually during 1977-85. Increased water temperature during the time that eggs and alevins incubated in the gravel was identified as a possible causal factor. Increased water temperature, which resulted in an accelerated rate of development of eggs and alevins, caused spring chinook salmon fry to emerge earlier than they would have under natural conditions (Cramer et al. 1985). Studies of coho salmon in Oregon coastal streams also implicated early emergence as a cause of decreased production (Nickelson et al. 1986). If early emergence decreased the production of spring chinook salmon fry in the Rogue River, the effect would have been greatest near Lost Creek Dam. Simulations of water temperature by the USACE indicated that the operation of Lost Creek Dam increased water temperature during the incubation period (November-January) by an average of 2.0'C at RK 248 and by an average of 1.30C RK at 202. Resultant differences in juvenile production may have caused subsequent spatial and temporal differences in the number of cohorts that survived to spawn.

19 Also, if early emergence decreased fry production, the effect would be greatest on eggs deposited by early spawners. A lower rate of survival among progeny of early spawning parents would be reflected in a later time of spawning among adults that survived to spawn, as we observed. Taylor (1980) found strong heritability in the spawning time of pink salmon. Progeny of early spawning parents spawned earlier than the progeny of later spawning parents. The time of maturation is dependent primarily on a genetically based response to the amount of daylight hours. Alteration of the photoperiod changed the spawning time of spring chinook salmon held in a hatchery (Zaugg et al. 1986).

Many researchers have hypothesized that the varied spawning time among stocks of Pacific salmon is a genetically based adaptation to localized regimes of water temperature during the period that eggs and alevins incubate in the gravel (Bams 1969; Ricker 1972; Godin 1981; Miller and Brannon 1981). Assuming that this hypothesis is true, then later spawning by spring chinook salmon in the Rogue River would compensate for the early emergence of fry caused by increased water temperature during the time that eggs and alevins incubate in the gravel. Because the increase in water temperature diminished with distance downstream, one would expect that any change in spawning time would also diminish with distance downstream, which is what we observed in this study.

The later time of spawning may not have been accompanied by a later time of migration among wild adults. Unfortunately, we could not evaluate the migration timing of postimpoundment broods that passed Gold Ray Dam. The high percentage of unmarked fish among hatchery adults prevented us from confidently estimating migration timing of wild adults separately from hatchery adults.

Recent changes in outflow temperature from Lost Creek Dam may shift the spawning of spring chinook salmon to a spatial and temporal distribution more like the distribution of preimpoundment broods. Since the fall of 1984, release temperature during incubation of eggs and alevins decreased compared with release temperature during postimpoundment years. If an increase in fry production and subsequent adult returns occurs because eggs and alevins develop at a slower rate, then an increase in the relative abundance of early spawners should occur in the vicinity of Lost Creek Dam.

We are now evaluating the effect of modified water temperature on the production of spring chinook salmon. As part of this project, we are estimating the abundance of wild spring chinook salmon and wild fall chinook salmon that migrate upstream of Gold Ray Dam. If we find no increase in the relative abundance of the spring race compared with the fall race, then a factor(s) other than incubation temperature is likely responsible for the downstream shift in spawning distribution and spawning time noted for the first postimpoundment broods.

ACKNOWLEDGEMENTS

We thank the numerous seasonal personnel who assisted with the field work. We also thank Michael Flesher and William Noll for leading sampling

20 crews during most of the study. Mary Buckman, Wayne Burck, Lyle Calvin, Alan McGie, and several anonymous individuals reviewed and improved the report with their comments.

REFERENCES

Bams, R.A. 1969. Adaptations of sockeye salmon associated with incubation in stream gravels. Pages 71-87 in T.G. Northcote, editor. Symposium on salmon and trout in streams. University of British Columbia, H.R. MacMillan Lectures on Fisheries, Vancouver, Canada.

Cramer, S.P., T.D. Satterthwaite, R.R. Boyce, and B.P. McPherson. 1985. Lost Creek Dam fisheries evaluation phase I completion report. Volume 1. Impacts of Lost Creek Dam on the biology of anadromous salmonids in the Rogue River. Oregon Department of Fish and Wildlife, Fish Research Project DACW57-77-C-0027, Portland.

Godin, J-G.J. 1981. Migrations of salmonid fishes during early life history phases: daily and annual timing. Pages 22-50 in E.L. Brannon and E.O. Salo, editors. Proceedings of a symposium on salmon and trout migratory behavior. University of Washington, Seattle. Hankin, D.G. 1982. Estimating escapement of Pacific salmon: marking practices to discriminate wild and hatchery fish. Transactions of the American Fisheries Society 111:286-298.

Horrall, R.M. 1981. Behavioral stock-isolating mechanisms in Great Lakes fishes with special reference to homing and site imprinting. Canadian Journal of Fisheries and Aquatic Sciences 38:1481-1496.

Jones, K.K. 1988. Stock assessment of anadromous salmonids. Oregon Department of Fish and Wildlife, Fish Research Project AFC-130, Annual Progress Report, Portland.

Kwain, W. and E. Thomas. 1984. The first evidence of spring spawning by chinook salmon in Lake Superior. North American Journal of Fisheries Management 4:227-228.

Larkin, P.A. 1981. A perspective on population genetics and salmon management. Canadian Journal of Fisheries and Aquatic Sciences 38:1469-1475.

McNeil, W.J. 1964. Redd superimposition and egg capacity of pink salmon spawning beds. Journal of the Fisheries Research Board of Canada 21:1385-1396.

Miller, R.J., and E.L. Brannon 1981. The origin and development of life history patterns in Pacific salmonids. Pages 219-227 in E.L. Brannon and E.O. Salo, editors. Proceedings of a symposium on salmon and trout migratory behavior. University of Washington, Seattle.

21 Neilson, J.D., and C.E. Banford. 1983. Chinook salmon (Oncorhynchus tshawytscha) spawner characteristics in relation to redd physical features. Canadian Journal of Zoology 61:1524-1531.

Nickelson, T.E., M.F. Solazzi, and S.L. Johnson. 1986. Use of hatchery coho salmon (Orcorhynchus kisutch) presmolts to rebuild wild populations in Oregon coastal streams. Canadian Journal of Fisheries and Aquatic Sciences 43:2443-2449.

Ricker, W.E. 1972. Heredity and environmental factors affecting certain salmonid Mulations. Pages 27-160 in T.G. Northcote, editor. The I . stock concept in Pacific salmon.- University of British Columbia, H.R. MacMillan Lectures on Fisheries, Vancouver, Canada.

Satterthwaite, T.D. 1987. Rogue basin fisheries evaluation program, effects of Lost Creek Dam on spring chinook salmon in the Rogue River, Oregon: An update. Oregon Department of Fish and Wildlife, Fish Research Project DACW57-77-C-0027, Portland.

Slater, D.W. 1963. Winter-run chinook salmon in the Sacramento River, California, with notes on water temperature requirements on spawning. U.S. Fish and Wildlife Service, Special Scientific Report (Fisheries) 461, Washington, D.C.

Snedecor, A.W., and W.G. Cochran. 1967. Statistical methods, 6th edition. Iowa State University Press, Ames. Taylor, S.G. 1980. Marine survival of pink salmon fry from early and late spawners. Transactions of the American Fisheries Society 109:79-82. USACE (United States Army Corps of Engineers). 1983. Water quality investigations - 1981, Lost Creek Lake Project, Rogue River, Oregon. Portland District, Portland. van den Berghe, E.P., and M.R. Gross. 1986. Length of breeding life of coho salmon (Oncorhynchus kisutch). Canadian Journal of Zoology 64:1482-1486. Zar, J.H. 1984. Biostatistical analysis, 2nd edition. Prentice-Hall, Inc., Englewood Cliffs, New Jersey. Zaugg, W.S., J.E. Bodle, J.E. Manning, and E. Wold. 1986. Smolt transformation and seaward migration in 0-age progeny of adult spring chinook (Oncorhynchus tshawytscha) matured early with photoperiod control. Canadian Journal of Fisheries and Aquatic Sciences 43:885-888.

22 I

APPENDIX A

Tables of Data Relating to Studies of Chinook Salmon

23 Appendix Table A-1. Estimated composition of spawned carcasses of female chinook salmon found in the Rogue River between Gold Ray Dam and Shady Cove, 1974-87. The area was not surveyed during 1982-85.

RK 205-212 RK 212-223 RK 223-235 Year Wild Hatchery Wild Hatche r-y --r-Wi d Hatchery

1974 109 a 192 2 401 3 1975 159 1 160 0 302 2 1976 112 0 267 1 310 1 1977 44 1 64 3 103 2

1978 232 0 325 0 837 3 1979 85 0 213 0 503 0 1980 I40 0 229 0 592 2 1981 141 0 168 6 187 3

1986 302 10 533 71 815 27 1987 633 42 671 84 886 156

Appendix Table A-2. Estimated composition of spawned carcasses of female chinook salmon found in the Rogue River between Shady Cove and Cole M. Rivers Hatchery, and in Big Butte Creek, 1974-87.

RK 235-245 RK 245-252 Big Butte Creek Year Wild Hatchery Wild Hatchery Wild Hatchery

1974 438 2 826 5 428 5 1975 306 2 586 11 424 11 1976 313 1 526 21 352 3 1977 118 1 272 5 239 1 1978 1,570 24 1,924 39 11119 22

1979 925 2 1,229 29 532 3 1980 960 2 951 32 474 15 1981 251 11 397 26 288 6 1982 577 32 ------1983 267 20

1984 364 10 1985 -- -- 898 171 -- -- 1986 680 40 723 184 363 108 1987 965 36 1,206 95 571 177

24 Appendix Table A-3. Distribution of tagged spring chinook salmon recovered as spawned carcasses in the Rogue River basin, 1974-78 and 1986-87. Data were grouped so that areas of the mainstem were of similar distance.

River kilometer of the Rog ue 241!rRiv Big Butte Year 205-11@6217@-228 2@29240 252 Creek

1974 8 2 17 24 9 1975 3 2 3 18 6 1976 4 7 9 21 5 1977 1 0 0 4 1 1978 11 14 16 59 16

1986 10 11 19 22 12 1987 32 17 36 43 11

Appendix Table A-4. Distribution of tagged fall chinook salmon recovered as spawned carcasses in the Rogue River basin, 1974-78 and 1986-87. Data were grouped so that areas of the mainstem were of similar distance.

River kilometer of 12h; R 09 ue Ripr Big Butte 24 Year 205-216 217-228 _ 24 252 Creek

1974 0 0 0 0 0 1975 1 0 0 0 0 1976 4 0 0 0 0 1977 2 1 2 0 0 1978 6 2 1 0 0

1986 21 4 8 3 1 1987 17 4 5 6 0

25 Appendix Table A-5. Data used to assess factors that affected the distribution and spawning time of wild female spring chinook salmon that spawned in the Rogue River, 1974-87. Data on spawning time are presented in Appendix Table A-9.

Spawning Qow Inc!ba!Zriotnuperature Year distribution Migration' Spawning' Spawnlng@

1974 55.1 1,446 1,305 5.1 9.7 1975 56.0 1,520 1,433 4.3 10.5 1976 51.6 1,617 1,277 5.0 10.3 1977 64.3 1,047 1,174 6.1 12.9 1978 53.1 1,866 1,294 5.4 10.8

1979 53.0 1,899 1,258 5.6 10.6 1980 44.4 1,988 1,276 5.8 10.0 1981 49.1 2,026 1,413 8.1 9.0 1982 2,148 1,813 6.7 9.5 1983 2,521 2,049 7.0 9.8

1984 2,920 1,740 7.2 8.7 1985 2,255 1,580 7.0 8.0 1986 32.7 2,142 1,446 6.9 9.1 1987 38.9 2,214 1,447 6.9 8.5 a Percentage that spawned upstream of RK 245 among wild females that spawned of RK 205. b upstream Mean flow (cfs) near McLeod during August. c Mean flow McLeod during September-October. d (cfs) near Mean maximum temperature (QC) near McLeod during October-January in year i-4. e Mean maximum temperature (QC) near McLeod during September-October.

26 Appendix Table A-6. Number of unmarked female chinook salmon counted weekly as spawned carcasses in the Rogue River between Shady Cove and Elk Creek (RK 235-245), 1974-87. The area was not surveyed during 1982-85. Week-of- year calendar is in APPENDIX D.

WeTk-of-ye:r Year 38 39 40 4 2 43 44 45

1974 8 52 78 131 87 52 32 1975 -- -- 42 59 75 73 55 4 1976 3 45 101 82 52 31 5 1977 -- 18 35 27 25 9 1978 21 176 381 508 290 167 44

1979 201 218 345 104 41 18 -- -- 1980 64 ill 241 318 155 62 11 1981 19 40 75 95 84 -- 35 -- 1986 14 48 110 158 221 127 68 30 1987 18 47 109 211 195 237 137 23

Appendix Table A-7. Number of unmarked female chinook salmon counted weekly as spawned carcasses in the Rogue River between Elk Creek and Cole M. Rivers Hatchery (RK 245-252), 1974-87. Week-of-year calendar is in APPENDIX D.

Week-of-year Year 38 39 40 41 42 43 44 45 46

1974 89 84 160 162 130 95 76 30 -- 1975 26 92 165 109 85 72 17 31 1976 35 88 148 142 95 32 7 1977 9 -- -- 45 82 73 46 16 4 1978 81 255 441 482 441 205 42 -- 1979 101 -- 272 380 255 136 102 12 1980 56 107 206 202 252 114 38 1981 6 -- -- 59 98 94 97 51 18 1982 7 34 126 138 186 87 42 21 1983 13 25 52 65 52 26 27 19

1984 1 7 21 52 121 112 47 11 -- 1985 9 36 112 160 239 170 178 71 35 1986 16 35 91 164 211 96 238 48 -- 1987 14 43 104 250 343 282 272 149 53

27 9'0 T 6'ZSZ 9,0 T V,V8z S'o T zIt8z L861 6'0 T SIM L'O T SIM S'o T VIM 9861 810 T 1118Z S86T 6'0 T V'I8Z VW

SIT T 9'9LZ C861 8'0 T S'ILZ Z861 t'l T SIM 6'0 T 9'183 O'l T 8,09Z 1861 O'l T 9'LLZ L'O T SIM 9'0 T UtLZ 0861 O'l T 9'9LZ 9'0 T S'ZLZ 9'0 T O'OLZ 6L61

9'0 T L'ILZ S'O T SIM G'O T UELZ BL61 ZIT T T'083 III T SIM III T I'M LL61 Z'I T 9'693 8'0 T L'OLZ O'l T S'ELZ 9L61 III T I'M 6'0 T L'69Z III T Z'OLZ SL61 III T CIM S'O T L'VLZ O'l T SIM VIM

jaaao allng 6[8 M-SOZ M M-SEZ AN JROA LeAJajUj OMONPOO '/,S6 T JW-10-NRP ueOW

aePuOLeO JR@.(-40-,(eo 'uOwLes @OOUL43 'a MUM u! s[ I LLPJ aAam saeumeds awos asneoaq MIAR jou sem BuLumeds 4o J9440 JOI P04MISOI I OWL4 @qi 'LB-tL61 'Maoij a;jnq 6[9 u[ pUe JOAL8 an6ob 041 UL UOWLeS @OOUL43 P'ajeWL4s3 -6-v' XLpua@dV BuLads apwaj pa@jpwun 4o 6uLumeds jo awl aLqP.L

Et Ell C61 SIT zli 9S oz t L861 11 OL VS W 06 Ss 11 C 9861 z 12 tz SL zt BE E -- 1861 I 99 68 U Ell E6 80 31 0861 BE IL LEI 98 Vol V9 Ss -- U61 VI 9S III LRZ SSz 89Z 9ZI tz MI -- C 61 80 IL 19 SC -- LL61 8 ST LZ VI 89 CS III 81 S 9L61 9 lz Sz 90 COT M 96 BE 9 SL61 -- LI 98 LS C9 901 Ss St -- VL61

9v St tt Et zt it Ov 6C BE- JeaA Jeak-JO-MOOM

XIaN3ddV UL SL JPPuOLP3 leak-40-108M 'SS-ZS61 6uunp PB.(8AAns 'a . I I 4OU sem v9JR a4i 'L8-tt6l '(Z-0 )lb) jaaj3 aling BLS UL S8SSeOJe3 paumeds se 41aam pajunoo UOWLes AOOuL4D OLL'tuaJ paj.Aewun jo AaqwnN -8-V 8Lqe.L XLpueddV Appendix Table A-10. Median date of spawning of female spring chinook salmon spawned annually at Cole M. Rivers Hatchery, 1975-87. Data received from Michael Evenson, ODFW, Cole M. Rivers Hatchery, Trail, Oregon.

Year Date Year Date

1975 10/01 1981 09/29 1976 10/05 1982 10/05 1977 10/05 1983 10/04 1978 10/03 1984 10/08 1979 10/10 1985 10102 1980 10/07 1986 10/01 1987 09/30

29 30 I

APPENDIX B

Figures Showing Annual Relationship Between Location of Carcass Recovery and Date of Tagging

31 260 - 1974

240 - v -0.05

220 -

200

260 -

1 975 % uj

240 -

Lu -0.58

0 220 -

LLJ

200 -

260 - 1976

240 -

-0.66

m 220 -

200 MAY JULY SEPTEMBER

DATE OF TAGGING

32 260

1977

240 -

LLJ 220

Ld r X -0.86 C.9 0 @<- 200

LL 0 260

1978

00 0 0 0 r % 240 -

220

r -0.53

200 - MAY JULY SEPTEM13ER

DATE OF TAGGING

Appendix Figure B-1. Annual relationship between location where tagged carcasses were recovered and the date chinook salmon were tagged at Gold Ray Dam, 1974-78.

33 260 -

1986

240 -

LLJ

U 220 LLJ 0: W

r -0.53 0 0 0 U. 0 200

LLJ

260 -

0 0 1987

0 of 0:

LIJ 0 0 0 240 > so

220

r -0.47 0

200 MAY JULY SEPTEMBER

DATE OF TAGGING

Appendix Figure B-2. Annual relationship between location where tagged carcasses were recovered and the date chinook salmon were tagged at Gold Ray Dam, 1986-87.

34 APPENDIX C

Figures Showing Annual Relationship Between Date of Carcass Recovery and Date of Tagging

35 11/15 - 1974

11/01 -

10/15 -

10/01

0.31 09/15

11/15 1 975

11/01

C.)0 LLJ 10/15

0.72

1 976

11/01 -

10/15 -

10/01

0.70 09/1 5

MAY JULY SEPTEMBER

DATE OF TAGGING

36 11/15 1977

11/01 -

10/15 -

Uj 09/15

U- 11/15 1978

11/01

10/15 r I r.

10/01

0.55 09/15

I MAY JULY SEM@MBERI

DATE OF TAGGING

Appendix Figure C-1. Annual relationship between the date tagged carcasses were recovered and the date chinook salmon were tagged at Gold Ray Dam, 1974-78.

37 11/15 - 1986

11/01 -

10/15 -

0 r 0.42 0LLJ (19/15

LL 11/15 0 - LLJ 1987

11/01 -

r 10/15 -

a 0 0 10/01 - 0 0.42 09/15 r

MAY JULY SEPTEMBER

DATE OF TAGGiNG

Appendix Figure C-2. Annual relationship between the date tagged carcasses were recovered and the date chinook salmon were tagged at Gold Ray Dam, 1986-87.

38 APPENDIX D

Relation between Gregorian day, day-of-year and week-of-year.

Gregorian day Day-of-year Week-of-year

1-7 January 1-7 1 8-14 January 8-14 2 15-21 January 15-21 3 22-28 January 22-28 4

29 January-4 February 29-35 5 5-11 February 36-42 6 12-18 February 43-49 7 19-25 February 50-56 8

26 February-4 March 57-64 9a 5-11 March 65-71 10 12-18 March 72-78 11 19-25 March 79-85 12

26 March-1 April 86-92 13 2-8 April 93-99 14 9-15 April 100-106 15 16-22 April 107-113 16 23-29 April 114-120 17

30 April-6 May 121-127 18 7-13 May 128-134 19 14-20 May 135-141 20 21-27 May 142-148 21

28 May-3 June 149-155 22 4-10 June 156-162 23 11-17 June 163-169 24 18-24 June 170-176 25

25 June-1 July 177-193 26 2-8 July 184-190 27 9-15 July 191-197 28 16-22 July 198-204 29 23-29 July 205-211 30

a Eight-day week during leap years.

39 Gregorian day Day-of-year Week-of-year

30 July-5 August 212-218 31 6-12 August 219-225 32 13-19 August 226-232 33 20-26 August 233-239 34

27 August-2 September 240-246 35 3-9 September 247-253 36 10-16 September 254-260 37 17-23 September 261-267 38 24-30 September 268-274 39

1-7 October 275-281 40 8-14 October 282-288 41 15-21 October 289-295 42 22-28 October 296-302 43

29 October-4 November 303-309 44 5-11 November 310-316 45 12-18 November 317-323 46 19-25 November 324-330 47

26 November-2 December 331-337 48 3-9 December 338-344 49 10-16 December 345-351 50 17-23 December 352-358 51 24-31 December 359-366 52b b Eight-day week.