Sockeye salmon smolt abundance and inriver distribution in Bristol Bay, Alaska: results from the Kvichak, Ugashik, and Egegik rivers in 2013

Prepared for

Bristol Bay Science and Research Institute Box 1464, Dillingham, AK 99576

September 2014

Sockeye salmon smolt abundance and inriver distribution in Bristol Bay, Alaska: results from the Kvichak, Ugashik, and Egegik rivers in 2013

by

Matthew J. Nemetha, Guy D. Wadea, Don J. Deganb, and Michael R. Linka

aLGL Alaska Research Associates, Inc. 2000 W. International Airport Road, Suite C1 Anchorage, Alaska 99502

bAquacoustics, Inc. P.O. Box 1473 Sterling, Alaska 99672-1473

for

Bristol Bay Science and Research Institute Box 1464, Dillingham, Alaska 99576

September 2014

Suggested format for citation:

Nemeth, M. J., G. D. Wade, D. J. Degan, and M. R. Link. 2014. Sockeye salmon smolt abundance and inriver distribution in Bristol Bay, Alaska: results from the Kvichak, Ugashik, and Egegik rivers in 2013. Report prepared by LGL Alaska Research Associates, Inc., Anchorage, AK, and Aquacoustics, Inc. Sterling, AK, for the Bristol Bay Science and Research Institute, Dillingham, AK, 55 pp. + appendices.

Smolt monitoring on three Bristol Bay rivers in 2013

EXECUTIVE SUMMARY

Information on the abundance of sockeye salmon (Oncorhynchus nerka) smolts can help characterize freshwater and marine productivity, set biological escapement goals, and forecast adult returns. In 2013, the Bristol Bay Science and Research Institute (BBSRI) used sonar arrays to estimate the hourly, daily, and seasonal abundance of sockeye salmon smolts migrating from three rivers in the Bristol Bay region of Alaska. This was the sixth study year on the Kvichak River (all consecutive), fourth on the Ugashik River, and second on the Egegik River. On each river, a series of upward-looking sonar transducers were placed across the river bottom, perpendicular to the water flow, in an array that ranged from 5 to 12 transducers. Each array was operated separately to generate independent estimates of smolt abundance. Smolt distribution and run timing were also described to help assess factors that may affect abundance estimates. Finally, smolts were also captured to characterize the age, weight, and length of the migrating population. The following are the major results of the study in 2013.

Kvichak River • Two independent sonar arrays were operated from 18 May through 13 June in 2013. Sonar arrays operated reliably, with minimal down time or sound interference. Smolt abundance was estimated to be 41.9 million (95% CI = 36.9 – 46.9 million) at Site 1 and 34.0 million (95% CI = 31.0 – 37.0) at Site 2. The study period appeared to encompass the entire run. Smolt run timing was slightly earlier than in prior years, and ranged 22 May to 9 June, with a midpoint of 27 May. Water temperatures were the highest recorded at the run midpoint (6.5 °C) since project inception in 2008 (previous range 4.5 to 6.3 °C).

• Smolt distribution and behavior were similar to prior years. Laterally, most smolts migrated down the left side of the channel (downstream perspective) where the water was relatively deep, although smolts were also detected across the entire river. Vertically, most smolts migrated within the upper 1.0 m of the water column, with a higher percentage towards the surface at night than during the day. Approximately half the smolt run migrated during daylight hours, with passage per hour therefore being highest during the relatively few hours of darkness. Water velocity ranged from 1.3 to 1.8 m/sec in areas with highest smolt passage.

• All smolts were either age -1 or age-2, with the proportion of age-1 (45%) being somewhat lower and the proportion of age-2 (55%) somewhat higher than in prior years. Within each age class, smolt body length and weight were similar to prior years. Tissues samples were collected for genetic analysis, to be reported separately by the Alaska Department of Fish and Game (ADF&G).

• A third array of new sonar pods using different transducers was successfully tested. Abundance and distribution data were analyzed and compared to data from the established transducers at sites 1 and 2 for validation purposes, but not used to characterize the smolt run. This array will be deployed on another river in 2014.

iv Smolt monitoring on three Bristol Bay rivers in 2013

Ugashik River • Two sonar arrays were operated from 18 May through 13 June in 2013, at the same sites used since 2010 and near the site of past studies by ADF&G. Sonar arrays operated reliably, although 33% of the events at Site 1 needed to be interpolated due to down time and sound interference. Smolt abundance was estimated to be 3.1 million (CI = 2.8 – 3.4 million) at Site 1 and 3.2 million (CI = 2.6 – 3.8 million) at Site 2. The study period appeared to encompass the entire run, and was similar in overall timing to comparable study years 2010 and 2012. Smolt run timing ranged from 22 May to 7 June, with a midpoint of 2 June. Water temperatures were higher at the run midpoint (6.8 °C) than in 2010 or 2012.

• Smolt distribution and behavior were similar to prior years and to the Kvichak River in 2013. Laterally, most Ugashik River smolts migrated down the center to center-left of the channel (downstream perspective), where the water was relatively deep. Low numbers of smolts were also detected across the entire river. Vertically, most smolts migrated within the upper 1.0 m of the water column, with a higher percentage towards the surface at night than during the day. Approximately 40-45% of the smolts migrated during daylight hours, with passage per hour therefore highest during the relatively few hours of darkness. Water velocity ranged from 1.4 to 2.5 m/sec in areas with highest smolt passage.

• All smolts were either age -1 or age-2, with the proportion of age-1 (68%) being somewhat higher and the proportion of age-2 (32%) somewhat lower than in prior years. Within each age class, smolt body length and weight were relatively high compared to prior years.

Egegik River • One sonar array was operated from 22 May through 12 June in 2013, at the same site attempted in 2011 and near the site of past studies by ADF&G. The array operated reliably, although 26% of the events needed to be interpolated due to down time and sound interference. Ice precluded full deployment from 18 May through 22 May. The study period appeared to encompass the entire run, which had a midpoint of 31 May. The overall run timing was more compressed than on the Kvichak and Ugashik rivers in 2013.

A smolt abundance estimate was not generated due to lack of reliable water velocity data. Smolt density in time and space was able to be calculated, however, which allowed us to describe the relative run timing and distribution of smolts.

• Laterally, most Egegik River smolts migrated down the right half of the river; the relatively uniform river depth prevented obvious associations between depth and smolt passage. Several aspects of smolt distribution and timing were different from the Kvichak and Egegik rivers in 2013. Vertically, only about 60% of the smolts migrated within the upper 1.0 m of the water column, a smaller fraction than on the Kvichak or Egegik rivers. Also different was that Egegik River acoustic data showed smolts nearer the surface in the day than at night, in contrast

v Smolt monitoring on three Bristol Bay rivers in 2013

to the other two rivers. Finally, only 37% of the smolts migrated during daylight hours, a lower percentage than on the other two rivers.

• Too few smolts were captured to conduct age, weight, or length analysis.

vi Smolt monitoring on three Bristol Bay rivers in 2013

TABLE OF CONTENTS

EXECUTIVE SUMMARY ...... iv Kvichak River ...... iv Ugashik River ...... v Egegik River ...... v TABLE OF CONTENTS ...... vii LIST OF TABLES ...... viii List of Figures ...... ix List of photos ...... x List of appendices ...... x INTRODUCTION ...... 1 OBJECTIVES ...... 2 STUDY AREA ...... 2 Bristol Bay ...... 2 Kvichak River ...... 3 Ugashik River ...... 3 Egegik River ...... 4 METHODS ...... 4 Sonar System Design Summary...... 4 Components and General Design ...... 4 Kvichak River Deployment and Operation ...... 5 Ugashik River Deployment and Operation ...... 6 Egegik River Deployment and Operation ...... 6 Data Collection and Analysis...... 6 Pre-processing and Echo Integration of Sonar Data ...... 7 Estimating Smolt Abundance ...... 7 Model Assumptions ...... 8 Smolt Distribution ...... 9 Environmental Conditions ...... 9 Smolt Age, Weight, and Length ...... 9 RESULTS ...... 11 Kvichak River ...... 11 Data Pre-processing ...... 11 Smolt Abundance and Run Timing ...... 11 Smolt Distribution ...... 12 Environmental Conditions ...... 12 Sampling Smolt for Age, Weight, and Length ...... 12 Ugashik River ...... 13 Data Pre-processing ...... 13 Smolt Abundance and Run Timing ...... 13 Smolt Distribution ...... 13 Environmental Conditions ...... 13 Smolt Age, Weight, and Length ...... 13 Egegik River ...... 14 Data Pre-processing ...... 14

vii Smolt monitoring on three Bristol Bay rivers in 2013

Smolt Abundance and Run Timing ...... 14 Smolt Distribution ...... 14 Environmental Conditions ...... 15 Smolt Age, Weight, and Length ...... 15 DISCUSSION ...... 15 Kvichak River ...... 15 Smolt Abundance ...... 15 Assumptions and Uncertainty in Abundance Estimates ...... 16 Distribution and Run Timing ...... 16 Smolt Age, Weight, and Length ...... 16 Ugashik River ...... 16 Smolt Abundance ...... 16 Distribution and Run Timing ...... 17 Egegik River ...... 17 Smolt Abundance ...... 17 Distribution and Run Timing ...... 18 CONCLUSION AND RECOMMENDATIONS ...... 18 ACKNOWLEDGEMENTS ...... 19 LITERATURE CITED ...... 21 TABLES ...... 23 FIGURES ...... 35 PHOTOS ...... 52 APPENDICES ...... 56

LIST OF TABLES

Table 1. Historical ice cover dates on Lake Iliamna, in the Kvichak River drainage. .. 24

Table 2. Historical ice cover dates on Lower Ugashik Lake (Ugashik River drainage) and Becharof Lake (Egegik River drainage)...... 25

Table 3. Annual sonar pod placements on the Kvichak, Ugashik, and Egegik rivers since project inception in 2008...... 26

Table 4. Mean water velocity (m/s) by date at sonar pods on the Kvichak, Ugashik, and Egegik rivers, 2013...... 28

Table 5. Annual abundance estimates of sockeye salmon smolts, by site, on the Kvichak, Ugashik, and Egegik rivers since project inception in 2008...... 29

Table 6. Estimated daily abundance and annual proportion of sockeye salmon smolts at Site 1 on the Kvichak River, 2013...... 30

Table 7. Estimated daily abundance and annual proportion of sockeye salmon smolts at Site 2 on the Kvichak River, 2013...... 31

viii Smolt monitoring on three Bristol Bay rivers in 2013

Table 8. Estimated daily abundance and annual proportion of sockeye salmon smolts at Site 1 on the Ugashik River, 2013...... 32

Table 9. Estimated daily abundance and annual proportion of sockeye salmon smolts at Site 2 on the Ugashik River, 2013...... 33

Table 10. Median migration date of sockeye smolts from the Kvichak, Ugashik, and Egegik rivers since project inception in 2008, along with corresponding water temperature… ...... 34

LIST OF FIGURES

Figure 1. The Bristol Bay region, showing the Kvichak, Ugashik, and Egegik and six other main rivers that produce sockeye salmon targeted in commercial, sport, and subsistence fisheries...... 36

Figure 2. Map of Kvichak River and Lake Clark drainages in Southwestern Alaska, showing locations of sonar sites 1 and 2 operated near the village of Igiugig, 2013...... 37

Figure 3. Map of Ugashik River watershed, showing locations of sonar systems and fyke net operated near the outlet of Lower Ugashik Lake, 2013...... 38

Figure 4. Map of the Egegik River watershed, showing locations of sonar system and fyke net operated near the outlet of Becharof Lake, 2013...... 39

Figure 5. Conceptual drawing of the smolt sonar system used in 2013...... 40

Figure 6. Smolt estimates at Site 1 and 2 on the Kvichak River showing periods when abundance was estimated via linear interpolation, 2013...... 41

Figure 7. Proportion of total smolts by hour of the day at sonar sites on the Kvichak, Ugashik, and Egegik rivers in 2013...... 42

Figure 8. Vertical distribution of sockeye salmon smolts migrating in darkness (2300 – 0500 hrs) and light at sites 1 and 2 on the Kvichak River in 2013. Difference in vertical axes between sites reflects shallower depths at Site 2...... 43

Figure 9. Annual run timing of sockeye salmon smolts on the Kvichak River since project inception in 2008...... 44

Figure 10. Water depth and percent distribution of sockeye salmon smolts at Site 1 and Site 2 sonar pods on the Kvichak River in 2013, showing pod distances (m) from right bank………… ...... 45

Figure 11. Smolt estimates at Site 1 and 2 on the Ugashik River showing periods when abundance was estimated via linear interpolation, 2013...... 46

ix Smolt monitoring on three Bristol Bay rivers in 2013

Figure 12. Annual run timing of sockeye salmon smolts on the Ugashik River since project inception in 2010...... 47

Figure 13. Vertical distribution of sockeye salmon smolts migrating in darkness (2300 – 0500 hrs) and light at sites 1 and 2 on the Ugashik River in 2013...... 48

Figure 14. Water depth and percent distribution of sockeye salmon smolts at Site 1 and Site 2 sonar pods on the Ugashik River in 2013 ………… ...... …….49

Figure 15. Annual run timing of sockeye salmon smolts on the Egegik River since project inception in 2010...... 50

Figure 16. Vertical distribution of sockeye salmon smolts migrating in darkness (2300 – 0500 hrs) and light on the Egegik River in 2013...... 50

Figure 17. Water depth and percent distribution of sockeye salmon smolts at Site 2 sonar pods on the Egegik River in 2013...... 51

LIST OF PHOTOS

Photo 1. Black line showing the approximate location of Site 1 sonar array used in 2008 – 2013 on the Kvichak River (note Weather Port on bank)...... 53

Photo 2. Black line showing the approximate location of Site 2 sonar array used in 2008 and 2010 – 2013 on the Kvichak River...... 53

Photo 3. Black lines showing the approximate location of the two sonar arrays used on the Ugashik River in 2010 – 2013...... 54

Photo 4. Sonar pod on shore during deployment with wire rope anchor harness running through eyebolts on the upstream side to prevent tipping...... 54

Photo 5. Sonar pods mounted (large orange pod is the split beam) on a welded aluminum sled with attached power, network cable, and power/control housings...... 55

Photo 6. Inclined Plane Trap (IPT) used to capture sockeye salmon smolts on the Kvichak River in 2013...... 55

LIST OF APPENDICES

Appendix A1. Daily catch and estimated age structure of sockeye salmon smolt on the Kvichak River, 2013...... 57

Appendix A2. Daily age, weight, and length of sockeye salmon smolts subsampled from Kvichak River catches in 2013...... 58

x Smolt monitoring on three Bristol Bay rivers in 2013

Appendix A3. Estimated age and weight of total daily catch of sockeye salmon smolts migrating from the Kvichak River in 2013...... 60

Appendix A4. Annual estimated age composition of sockeye salmon smolts emigrating from the Kvichak River...... 61

Appendix A5. Daily age, weight, and length of sockeye salmon smolts subsampled from Ugashik River catches in 2013...... 64

Appendix A6. Estimated age and weight of total daily catch of sockeye salmon smolts migrating from the Ugashik River in 2013...... 65

Appendix A7. Annual estimated age composition of sockeye salmon smolts emigrating from the Ugashik River...... 66

Appendix A8. Daily age, weight, and length of sockeye salmon smolts subsampled from Ugashik River catches in 2012...... 69

Appendix A9. Estimated age and weight of total daily catch of sockeye salmon smolts migrating from the Ugashik River in 2012...... 70

Appendix B1. Daily climate and hydrological observations made at 0800 and 2000 hours near the Kvichak River sonar site, 2013...... 71

Appendix B2. Daily climate and hydrological observations made at 0800 and 2000 hours near the Ugashik River sonar site, 2013...... 73

Appendix B3. Daily climate and hydrological observations made at 0800 and 2000 hours near the Egegik River sonar site, 2013...... 75

xi Smolt monitoring on three Bristol Bay rivers in 2013

INTRODUCTION

The value of monitoring salmon smolt abundance and populations characteristics in Bristol Bay rivers has been long recognized, even as the scope of the monitoring program has changed. Historically, the impetus for monitoring sockeye salmon (Oncorhynchus nerka) smolts in Bristol Bay has been to help salmon management by improving preseason forecasts of returning adult salmon abundance, and to help understand the relationship between parent escapement and smolt production and how this changes over time. Understanding productivity as a function of escapement is useful for refining escapement goals used to manage the fishery. Although less of an original impetus, there is also value to monitoring smolts simply because they are an early sentinel of changes in population characteristics such as body size and weight, age structure, and abundance. These characteristics can reflect changes in the forage base or in environmental conditions that can subsequently affect the returns of adult salmon.

For all these reasons, the University of Washington began estimating age, length, and abundance of sockeye salmon smolts on the Wood and Kvichak rivers in the early 1950s (Burgner 1968). The Alaska Department of Fish and Game (ADF&G) expanded smolt research in Bristol Bay in the 1960s, then began experimenting with hydroacoustics for estimating abundance in the early 1970s (ADF&G; Wade et al. 2010). Smolt programs continued through the 1990s but interest had waned due to the budget cuts, statistical uncertainty, and apparent usefulness of abundance estimates. By 2002, ADF&G had discontinued sockeye salmon smolt projects in Bristol Bay (Crawford and Fair 2003; Wade et al. 2012b).

Interest in smolt data renewed in the mid-2000s due to increased discussion of escapement goal changes after the Alaska Board of Fisheries adopted the Policy for the Statewide Escapement Goals (5 AAC 39.223); this caused Bristol Bay goals to be evaluated more frequently and more extensively in the next few years, during which time there was much discussion about raising the goals (Baker et al. 2009). The renewed interest prompted BBSRI to begin designing a new sonar system that drew on extensive evaluations of the prior approaches by several researchers (e.g., Crawford and Fair 2003; Ruggerone and Link 2006; Maxwell et al. 2009; Wade et al. 2010). The new design was a series of custom-designed, up-looking sonar pods (each with a transducer and echosounder) that could be deployed in ways to sample the entire river width. This digitally-based sonar system was first tested on the Kvichak River in 2008 (Wade et al. 2010), then expanded with pilot years on the Ugashik River in 2010 and the Egegik River in 2011.

In 2013, BBSRI continued the project to enumerate sockeye salmon smolts on the Kvichak (sixth consecutive year), Ugashik (fourth year, not all consecutive), and Egegik rivers (second year, with the first being in 2011). Sockeye salmon smolt abundance was estimated using sonar systems that were the same bottom-founded, up-looking design as described in Wade et al. (2010). Smolts (hereafter, assumed to be sockeye salmon unless otherwise noted) were also collected with an incline plane trap and fyke nets to estimate age, body size, stock composition and to verify species composition (Kvichak and Ugashik rivers only). Physical site data and smolt behavior and distribution information

1 Smolt monitoring on three Bristol Bay rivers in 2013

were again collected to help interpret the sonar estimates and identify problems or anomalies that could influence the abundance estimates.

This report provides hourly, daily, and seasonal abundance of smolts in 2013 and characterizes horizontal, vertical, and diel distribution of smolt schools as they migrate. The report also discusses possible sources of uncertainty in the abundance estimates (e.g., smolt swimming speed, distribution) and how these could theoretically bias the final results. Descriptions of smolt behavior in these systems also benefits other studies by helping guide various sampling efforts, identifying differences among river systems, and documenting how factors such as ice and water discharge may affect smolt migrations.

Smolt age, weight, and length (AWL) data are important elements of the freshwater production of salmon, and were also collected as part of the field portion of this study. The AWL data complement the abundance estimates, helping to understand the overall health of the population leaving freshwater and to make inferences about smolt survival at sea. Over time, these datasets can be used to make preseason forecasts of adult returns and to refine system-specific escapement goals for Bristol Bay sockeye salmon (Crawford and West 2001; Baker et al. 2006, 2009). The AWL analyses contained here were conducted separately by ADF&G.

In 2012, ADF&G began collecting genetic samples from smolts as the first of three years of sampling designed to estimate the stock composition of smolts from the Lake Clark and Lake Iliamna drainages. Samples were again collected during the 2013 field season. Analyses of those samples are currently being conducted and results will be made available in a separate report by ADF&G in 2015.

OBJECTIVES

The objectives of the 2013 study were to:

1. Estimate the hourly, daily, and seasonal abundance of sockeye salmon smolts migrating from the Kvichak, Ugashik, and Egegik rivers. 2. Characterize the vertical, horizontal, and diel distribution of smolts emigrating from all three rivers. 3. Estimate the age, weight, and length of sockeye salmon smolts from all three rivers. 4. Estimate stock composition of sockeye salmon smolts from the Kvichak River. 5. Test additional sonar units on the Kvichak River for future deployment.

STUDY AREA

Bristol Bay The Kvichak, Ugashik, and Egegik rivers are three of the nine main rivers that produce sockeye salmon targeted in the Bristol Bay commercial, subsistence, and sport fisheries (Figure 1). The commercial sockeye salmon fishery in Bristol Bay is the largest in the world; over the 20-year period from 1991 through 2010, total sockeye salmon returns

2 Smolt monitoring on three Bristol Bay rivers in 2013

averaged 37.3 million fish (Salomone et al. 2011). The Alaska Department of Fish and Game manages Bristol Bay stocks for river-specific escapement goal ranges using pre- season forecasts and several in-season indicators of run strength.

Subsistence fishing for sockeye salmon in Bristol Bay has occurred since inhabitance and continues to be an important source of protein for local residents (Morstad et al. 2010). In 2009, subsistence harvest of sockeye salmon for the Kvichak River/ Iliamna Lake subdistrict totaled 46,772 fish from 187 permits, and in the Igiugig region totaled 1,071 from 5 permits (Salomone et al. 2011). Ugashik district harvest totaled 1,061 sockeye from a total of 15 harvest permits. In addition to the subsistence fishery, sockeye salmon have been an essential segment of the sport fishing industry for both the Kvichak and Ugashik drainages. From 2003 through 2007, the annual sport fish harvest of sockeye salmon in the Kvichak and Ugashik rivers averaged 1,461 and 2,789 fish, respectively (Dye and Schwanke 2009).

Kvichak River The Iliamna watershed is located in southwest Alaska and drains an area of 16,830 km2 (Figure 2). This watershed includes Lake Clark and Iliamna Lake. Iliamna is the largest lake in Alaska, with an area of 2,622 km2 and a volume of 115.3 km3 (Quinn 2005). Lake Clark (267 km2) is located north of Iliamna Lake and flows into Iliamna Lake via the Newhalen River. Lake Clark is glacially fed, causing turbidity at the head of the lake; this turbidity diminishes as it reaches the Newhalen River. The Kvichak River connects Iliamna Lake to the ocean and flows southwest for approximately 106 km where it enters Kvichak Bay, in the northeastern corner of Bristol Bay. The Kvichak River is a clear- water stream exiting the western end of Iliamna Lake, near the village of Igiugig, which is approximately 14 m above sea level.

Mean annual discharge for the Kvichak River collected near Igiugig from 1968 to 1986 ranged from 361 m3/s to 729 m3/s and averaged 503 m3/s (USGS 2008). Peak discharge occurs during August, September, and October; the lowest discharge typically occurs during March, April, and May. From 1970 through 2001, total duration of ice coverage for Lake Iliamna varied from 39 d to 161 d and had an average breakup date of 13 May (Table 1; Crawford and Fair 2003).

The initial 1.2 km of the Kvichak River below Iliamna Lake is a single channel; downstream, the river is mostly braided with a few exceptions (Figure 2). The river forms a single channel 3.5 and 7.0 km downstream from the lake; these two sites have been the locations of smolt studies from 1976 to present, with the exception of Site 2 having been ~ 3 km upstream in 2009 (Photo 1; Photo 2; Maxwell et al. 2009; Wade et al. 2012b). This study used the upper site as Site 1 (N 59.2924, W 155.9550) and the lower site as Site 2 (N 59.3042, W 155.9715).

Ugashik River The Ugashik River drainage is located on the northern portion of the Alaska Peninsula and flows westerly into Ugashik Bay, in the southernmost region of Bristol Bay (Figure 3). The Ugashik River watershed consists of the Upper and Lower Ugashik lakes, the

3 Smolt monitoring on three Bristol Bay rivers in 2013

Ugashik River connecting the lower lake to Ugashik Bay, and the King Salmon and Dog Salmon rivers. The Ugashik lakes are relatively large, with surface areas of 177 km2 for the upper lake and 208 km2 for the lower lake; the entire watershed drains 4,205 km2.

The Ugashik River is approximately 60 km long and is an alluvial river with a meandering channel pattern that is highly braided in some sections. Just below Lower Ugashik Lake, the river is confined to a single channel for a short distance (~150 m) and then spreads out into a highly braided region for the next 2 km before reaching a shallow lagoon. The river is tidally influenced downstream of the braids near the lake outlet. There are no USGS stream flow data available for this area. Ice cover data for the lakes were collected during the 1980s and 1990s, during which time total duration of ice coverage varied from 51 d to 135 d (Table 2; Crawford and West 2001).

The sonar systems were located approximately 80 m from the Lower Ugashik Lake outlet (N 58.0600, W 156.8860), at or near the same locations used for smolts studies by BBSRI in 2010 and 2012 and by ADF&G in the 1980s and 1990s Photo 3; Crawford and West 2001).

Egegik River The Egegik River watershed is on the Alaska Peninsula, approximately 67 km south of King Salmon, AK (Figure 4). The head of the watershed is Becharof Lake, the second largest lake in Alaska. Becharof Lake is drained by the Egegik River, which flows westerly for ~45 km and empties into the Egegik Bay, in the southeastern portion of Bristol Bay. The King Salmon River is another major river in the drainage that empties into Egegik River near Egegik Bay. The watershed drains an area of 8,442 km2; Becharof Lake has a surface area of 1,132 km2.

The Egegik River has a single channel for the first 5 km, then widens out into a lagoon before splitting into braided channels. The stretch of river from the lake outlet to the lagoon is generally clear water; downstream of this point, the water is more turbid due to tidal influences. Water levels and velocities are tidally influenced for the entirety of the river. The USGS has not collected stream flow data for this area. Ice cover data for Lake Becharof were collected during the 1980s and 1990s, during which time total duration of ice coverage varied from 39 d to 128 d (Table 2; Crawford and West 2001).

In 2013, the sonar system was 4.5 rkm downstream of the outlet (N 58.0605, W 156.8893) in the same approximate location as smolt studies conducted by BBSRI in 2011 and by ADF&G in the 1980s and 1990s (Crawford and Fair 2003).

METHODS

Sonar System Design Summary

Components and General Design Each sonar system consisted of 5 to 12 up-looking sonar pods joined in line to form an array (Figure 5), along with supporting hardware, controllers, data communication and

4 Smolt monitoring on three Bristol Bay rivers in 2013

storage, and a power supply system. Each pod was mounted on an aluminum sled (Photo 4) designed to remain upright on the river bottom, and all sleds were tethered together by wire rope to form the array. Each sonar array was then connected to a shoreside control box by a power and data cable, allowing data communication, storage, and system monitoring. Individual data files were collected continuously and then stored in 1 hour blocks. Power to the array was supplied by a bank of 12 V batteries and supplemented with gas or thermoelectric generators; more detail about components and specifications are provided by Wade et al. (2012a).

Optimal sonar sites were where the river was confined to a single channel, with a gradual bottom cross section suitable for towing the array across the river. In 2013, all sonar arrays were deployed in the same manner, following the methods described by Wade et al. (2010). A towline was attached to the chain on the first sled, and the array was pulled across the river by boat or with a chainsaw winch. Once in a suitable location, the ends of the sonar array were anchored to each bank.

Once operational, the sonar systems collected data 24 hours per day for the entire season. Each system was checked twice daily, generally at 0800 hours and 2300 hours, to ensure adequate power supply and operation. Data stored streamside for each of the arrays were downloaded onto a portable computer each day and examined visually using specialized software (EchoView® 5.1 by Myriax Software Pty. Ltd., Tasmania, Australia).

Each sonar pod was a 24 V, low-power acoustic sounder and transducer contained within a machined aluminum housing (22 cm diameter x 19 cm high) and designed to send a data stream back to shore via an Ethernet cable (Photo 4; Wade et al. 2010). Each pod was outfitted with a 7.5 o (at -3 dB) single beam transducer, manufactured by BioSonics or Airmar. In addition to the single beam pods, one split-beam pod (Simrad model ES120-7C) was integrated into one sonar array on each river to estimate the acoustic target strength of individual smolts (Photo 5). A pool calibration of each sonar pod was performed prior to field deployment (Wade et al. 2012a). Tests in prior years show each transducer can detect targets from within 0.5 cm of the water surface, down to 71 cm from the face of the transducer (Wade et al 2012b). Within 71 cm of the transducer face, detections are unreliable due to “near field” effect. Transducer specifications are described in more detail in Wade et al. (2012b).

Kvichak River Deployment and Operation Site 1 was set in a section of river 102 m wide. This array had a total of eight pods, one of which was the split-beam transducer. The first pod (T1) was set 17 m from the right bank in 2.5 m of water. Pods were then spaced 10 m apart, with the last one (T8) 77 m from the right bank and 10 m from the left bank (Table 3). The bottom profile of Site 1 began with a gently sloping right bank to a maximum depth of 3.9 m, in the region 57-77 m from shore (Table 3). The array was operated 19 May – 13 June.

Site 2 was set in a section of river 130 m wide. Site 2 had 12 single-beam transducers that formed two arrays placed end-to-end across the river, one with eight pods and the other with four pods. The first pod (T1) was set 4 m from the right bank in 1.5 m of water. Most pods were again spaced 10 m apart, with the last one (T12) 108 m from the right

5 Smolt monitoring on three Bristol Bay rivers in 2013

bank and 22 m from the left bank, in 3.3 m of water (Table 3). The bottom profile of Site 2 was relatively even across its width. Site 2 was operated from 18 May through 12 June.

The new sonar units to be field tested were arranged in a third array of six pods, 20 m downstream from Site 1.

Ugashik River Deployment and Operation Site 1 was set in a section of river 43 m wide. This array had five pods (all single-beam) evenly spaced every 5 m, from 20 to 40 m from the right bank. The river bed was deepest in the middle, with T3 being 30 m from the right bank in 3.3 m of water (Table 3). The array was operated from 20 May through 13 June.

Site 2 was set in a section of river 61 m wide. This array also had five pods (four single- beam, one split-beam), evenly spaced every 5 m, from 32 to 53 m from the right bank. The river bed was slightly sloped towards the left bank, with the deepest pod (T4) 48 m from the right bank, in 2.8 m of water (Table 3). The array was operated from 18 May through 12 June.

In 2011, water velocities on the Ugashik River were high enough to move the sonar arrays downstream, preventing abundance estimates for that year (Wade et al. 2012a). To prevent this from happening in 2013, each sled from Site 1 was equipped with two eyebolts on the upstream side (Photo 4). Prior to deployment, a 7.9 mm (5/16”) diameter wire rope was anchored on the river bottom upstream of the site; as the pods were deployed they were “threaded” onto the cable to help prevent tipping.

Egegik River Deployment and Operation The Egegik River array was placed in a section of river 80 m wide. This array had seven sonar pods (all single-beam), evenly spaced every 10 m, from 4 to 64 m from the left bank. The river depth dropped off quickly near each shore and had a U-shaped channel; most pods were between 3.6 and 4.1 m deep (Table 3). The array was operated from 22 May through 12 June.

Data Collection and Analysis Smolt abundance was estimated by using the echo integration of the sonar data to eventually estimate mean smolt density per hour (smolts per m2) at each transducer each hour, within each 0.2 m vertical cell of the water column. Smolt densities for all 0.2 m strata were summed to provide total water column densities. Smolt density was multiplied by water current speed to estimate smolt passage per hour at each cell (i.e., flux). Linear interpolation was then used to estimate smolt passage among transducers and between transducers and the riverbank, in 1 m increments (and also in 0.2 m vertical cells). Hourly abundance was the total passage of smolts across the entire river, each hour.

The resulting estimates of smolt abundance per hour, across the stream (i.e., each transducer), and at each vertical cell (0.2 m) in the water column allowed estimates of the vertical, lateral, and diel distribution of smolts each day.

6 Smolt monitoring on three Bristol Bay rivers in 2013

Additionally, smolts were captured throughout the sampling period to collect age, weight, and length data (AWL). These data were used to summarize the size and age structure of the smolt run on each river.

Pre-processing and Echo Integration of Sonar Data Data files were pre-processed (using EchoView® 5.4 software) by removing acoustic noise events generated by ice, boat passage through the sample area, wind/rain events, and any interference among transducers. The distinction between noise events and smolts was obvious the majority of the time. If the technician could not distinguish between smolts and noise, that region of data was excluded from the analysis. For hours where data were removed, estimates were linearly interpolated based on the values surrounding these events.

Next, sonar data were echo integrated. Sockeye salmon smolts aggregate in schools too dense for the sonar to detect single targets accurately; therefore abundance estimates could not be calculated by counting individual fish. Instead, echo integration summed all backscatter cross-sections from multiple targets (i.e., smolts) in a given sample volume, producing a backscatter coefficient. Once average target strength of a single fish was known, the number of fish was estimated from the backscatter coefficient.

Estimating Smolt Abundance Backscatter coefficient was calculated over a given range from each transducer to produce the area backscatter coefficient/m2 (ABC). The ABC was calculated in 1-h x 0.2 m depth intervals, then divided by the mean sigma (target strength in linear domain) to obtain the smolt density for each cell. Smolt density for each cell was a measure of mean smolt count/cross sectional area sampled, normalized to smolt density/m2 for each strata. Fish density/m2 was then multiplied by the water velocity to obtain the smolt flux, which gave the number of smolts/hour/meter of river cross section sampled at each pod.

Several models were considered for expanding the pod-specific estimates to the entire cross section of river. We chose to linearly interpolate between pods to estimate smolt passage for areas not sampled. Likewise, we interpolated over the distance between the end pods and the river banks, which were assigned values of zero passage. This method yielded a river-wide estimate of smolt passage at each site.

Smolt passage was not subsampled through time because counts were continuous from beginning to end of the enumeration project. When portions of the total season were missed due to shutdowns and environmental noise, the missing hours for each transducer were filled using linear interpolation between adjacent hours. The season total abundance and variance of the mean for each site were estimated by the following:

K SA = ∑ HAj (1) j=1

7 Smolt monitoring on three Bristol Bay rivers in 2013

n ∑Tij HA = i=1 × ES (2) j n j

n d m ES j = ∑∑Tij + (Ti+1, j −Tij ) (3) i=1 m=0 d

K Var(SA) = ∑[Var(HAj )] (4) j=1

n 2 ∑(Tij −T. j ) fpc 2 Var(HA )= i=1 ⋅ × ES (5) j n −1 n j

A − a fpc = (6) A −1

th where, SA = smolt abundance, HAj = smolt abundance for the j hour, ESj = scalar that expands each hourly average across transducers to the entire stream, m = number of meters after the ith transducer for which the interpolation was being generated, d = number of meters between transducer i and i+1 (i and i+1 could also represent either bank for which smolt passage was assigned a value of zero), K = number of hours for which counts were estimated over the entire season, n = number of transducers across the th th river, Tij = count for the i transducer in the j hour, Var(SA )= variance of SA, T. j = average count across all i transducers for the jth hour, fpc = finite population correction, a = cross sectional area ensonified by all transducers, and A = total cross sectional area for which the estimate was expanded. Normal 95% confidence intervals were produced for SA estimated at each site. Estimates of variance include uncertainty due to subsampling the water column, but not uncertainty from estimating the scaling factor during echo integration

Smolt abundance was estimated hourly, and expanded to calculate daily and total season abundance. Abundance was compared between sites within each river, and among hours within each site. Diel timing of downstream movement was described by comparing hourly abundances during dark and daylight hours. For the purpose of this study, daylight was defined as the hours from 0500 to 2259 hours and darkness from 2300 to 0459 hours.

Model Assumptions Four main assumptions about smolt behavior were needed to produce reliable abundance estimates comparable across years:

1. Smolts travel at or near the same speed as the river water velocity. 2. Within the water column, smolts do not travel so low as to be undetected by the sonar transducers, or so high as to be indistinguishable from surface noise.

8 Smolt monitoring on three Bristol Bay rivers in 2013

3. Among years, the proportion of any smolts traveling in the undetectable areas of the water column does not vary. 4. Mean target strength may be used to scale echo integration.

Violations of these assumptions could bias the final estimate (Wade et al. 2012a).

Smolt Distribution Smolt abundance estimates were used to describe smolt distribution each hour in lateral and vertical dimensions, as well as during hours of relative light and dark. For lateral distribution (across the river), river width was divided into relatively large strata with bounds on either side of the pods and toward the shore. The hourly abundance estimate in each stratum was summed for all depths to give the total amount of smolts in that section of river. For vertical distribution, abundance was calculated in 0.2 m depth strata from the surface down to the transducer’s near field (~ 71 cm; Wade et al. 2012b). Abundance was then summed for all hours of daylight and darkness to compare diel patterns.

Environmental Conditions Water velocity was measured at a depth of 1 m at each pod, from a boat anchored 2 to 3 m downstream of each pod. Measurements were taken for one minute, three times at each transducer to give an arithmetic mean. Water velocity on the Kvichak River was measured three times at each site, roughly at the beginning, middle, and end of the sonar system operating dates (Table 4). At the Kvichak River, a model 622 Gurley Price meter (Gurley Precision Instruments, Troy, NY) was used to take measurements. Velocities were calculated based on the GPI conversion table. Water velocities were measured four times on the Ugashik River and one time on the Egegik River, using a FP111 digital flow meter made by Global Water Instrumentation, Inc. (Sacramento, CA).

Weather and other hydrologic data were recorded at the Kvichak and Ugashik rivers using a Watch Dog 2000® weather station (Spectrum Technologies, Plainfield, IL). The weather station was configured for hourly measurements of air temperature (°C), relative humidity (%), rainfall (mm), wind direction (degrees), wind speed (km/h), and wind gusts (km/h). One weather station was operated near the primary sonar site on the Kvichak and Ugashik rivers for the duration of each project (Appendix B). Hourly water temperature was measured on the Kvichak and Ugashik rivers using a Tidbit® v2 TempLogger.

Water velocity on the Egegik River was also measured with an acoustic Doppler current profiler (ADCP; Argonaut SL 0.5 MHz, SonTek, San Diego CA).

Smolt Age, Weight, and Length A subsampling design was used to estimate the daily age structure, mean body length, and mean body weight of smolts migrating from each river. For all river systems, the sampling sites were in the approximate location as the sites ADF&G had used since 1956 (Crawford and West 2001). On the Kvichak River, an inclined plane trap (IPT) was used to capture smolts (Photo 6). This trap was modeled after a similar trap that operated on the Kasilof River (Todd 1994). On the Ugashik and Egegik rivers, smolts were captured using a standard fyke net with a rigid 4’ x 4’ opening.

9 Smolt monitoring on three Bristol Bay rivers in 2013

Each calendar day, up to 100 smolts were sampled for age, weight, and length (AWL). Up to another 500 smolts were sampled for length only, thereby developing a regression relationship between length and age and between length and weight for each sampling day. These two groups (i.e., up to 600 smolts) were combined to generate a mean length, weight, and age structure for the daily sample. Sample sizes were based on binomial proportions for the two major age groups (ages 1 and 2); a sample size of 400 smolts would simultaneously estimate the percentage of each age class within 5% of the true percentage 95% of the time (Cochran 1977).

For age, the relationship between length and age of the 100 aged fish was used to identify a discriminant length that minimized classification error (or age-length key; Bue and Eggers 1989), chosen such that the number of age-1 smolt classified as age-2 smolt was equal to the number of age-2 smolt classified as age-1 smolt. This discriminant length was calculated postseason, from all fish aged throughout the season, then applied retroactively to the rest of the fish sample (i.e., measured for length, but age not measured directly) to generate the daily and final age estimates. Age-3 smolts were not included in this analysis because too few samples were collected.

For weight, the relationship between length and weight of the 100 aged fish was estimated using a least squares linear regression. Based on paired weight-length data obtained from smolt sampled for age, weight, and length, we estimated weights (Wj) of age j smolt measured only for length as explained by (Ricker 1975):

β = α L , W j j (1)

where

Lj = fork length of an age j smolt, and α and β = parameters which determine the y-axis intercept and the slope of the line.

Due to the variability of age and size composition estimates among subsamples (e.g., incline plane trap catches) taken the same day, daily mean weight (W ) and age  proportions (Pj ) were estimated as the mean of subsampled values:

m  ∑w  ∑  k  k=1  nk  W = , m (2)

10 Smolt monitoring on three Bristol Bay rivers in 2013

where

m = number of subsamples collected during a sampling period, wk = observed weights from subsample k, and nk = number of observations in subsample k; and

m  n j,k    ∑   k=1 nk P j = , m (3)

where

nj,k = number of observations of age j in subsample k.

To keep data together from each nightly sampling session, all fishing times, fish catches, and age-length-weight sampling data were recorded by smolt day. A smolt day was a 24- h sampling period starting at 1200 hours and ending at 1159 hours the next calendar day.

All AWL analysis and reporting was performed by the Alaska Department of Fish and Game, Division of Commercial Fisheries, Region II (Appendices A1–A9).

RESULTS

Kvichak River

Data Pre-processing At Site 1, sonar data were collected for 602 hours, from 19 May (1400 hours) through 13 June (1500 hours). At Site 2, sonar data were collected for 599 hours, from 18 May (2200 hours) to 12 June (2000 hours).

Environmental noise from wind, ice, or vessels precluded analysis of 39 hourly blocks of data from Site 1 (6% of total) and 7 hourly blocks from Site 2 (1% of total; Figure 6). These blocks were removed from the dataset and replaced using interpolations from surrounding hourly blocks.

Smolt Abundance and Run Timing At Site 1, sockeye salmon smolt abundance was estimated at 41,860,976 (95% CIs = 36,858,100 – 46,863,852;Table 5; Table 6). At Site 2, estimated abundance was 34,000,549 (95% CIs = 31,048,648 – 36,952,450;Table 5; Table 7). Hourly smolt passage peaked at each site during darkness, defined as 2300 to 0500 (Figure 7). Many smolts also migrated during daylight, however, accounting for 50% of the run at Site 1 and 56% of the run at Site 2 (Figure 8).

11 Smolt monitoring on three Bristol Bay rivers in 2013

Overall run timing was slightly early for the first 75% of the smolt run as compared to prior years (Figure 9), although the main peaks in catch per unit effort (25 May to 28 May) were similar to prior years. Both sonars appeared to have been deployed before any major smolt movements, based on patterns between the dates of deployment on 18 and 19 May and the first smolt peaks on 22 May.

Smolt Distribution Vertical distribution was split into 0.2 m depth strata down to a maximum depth of 3.0 m at Site 1 and a maximum depth of 2.6 m at Site 2. At both sites, smolt distribution was highly skewed towards the surface, especially during dark hours (Figure 8). For both sites more than 97% of smolts detected at night were in the upper 1.0 m, with most of these in the upper 0.4 to 0.6 m. By contrast, smolts traveling during daylight hours tended to have a deeper vertical distribution, with only 67% to 78% of the smolts in the upper 1.0 m (Figure 8).

Smolts were detected by all sonar pods at each site (i.e., across the entire river), but were disproportionately distributed. At both sites, between 80% and 85% of the smolts were concentrated in the left side of the river channel; this corresponded with the deepest part of the river at Site 1, but not at Site 2 (Figure 10).

Environmental Conditions Water temperatures were approximately 3 °C when sonar pods were deployed on 19 May, then peaked at 11.6 °C on 6 June. Precipitation was recorded on five dates, but water color remained clear on all but two days (Appendix B1).

Sampling Smolt for Age, Weight, and Length On the Kvichak River, smolt sampling for age, weight, and length began on 19 May and ended 10 June (Appendix A1). Sampling was conducted daily, except for 22 through 24 May when ice flow prevented sampling. A total of 1,075 smolts were aged, weighed, and measured for length to use in developing the length-age and the length-weight relationships (Appendix A2). These relationships were then applied to another 6,752 smolts measured for length during the season (Appendix A3).

Of the 7,827 smolts sampled in 2013, 45% were estimated at age-1 and 55% at age-2 (Appendix A4). Age-1 smolts had an estimated mean length of 86 mm and mean weight of 5.9 g; Age-2 smolts had an estimated mean length of 106 mm and mean weight of 10.8 g (Appendix A4). The discriminant body length used to separate the two age classes was 95 mm.

The age class percentage shifted over the course of the season. Age-2 smolts dominated the samples from 19 May through 27 May; thereafter, the majority of smolts were age-1 on most days (Appendix A1). Previously, age-2 smolts emigrated earlier than age-1 smolts (Crawford and West 2001).

12 Smolt monitoring on three Bristol Bay rivers in 2013

Ugashik River

Data Pre-processing At Site 1, sonar data were collected for 586 hours, from 20 May (0000 hrs) through 13 June (1500 hours). At Site 2, sonar data were collected for 593 hours, from 18 May (1800 hours) to 12 June (0900 hours).

Environmental noise from wind, ice, or vessels precluded analysis of 192 hourly blocks of data from Site 1 (33% of total) and 78 hourly blocks from Site 2 (13% of total; Figure 11). These blocks were removed from the dataset and replaced using interpolations from surrounding hourly blocks.

Smolt Abundance and Run Timing At Site 1, sockeye salmon smolt abundance was estimated at 3,060,357, with a 95% CI of +/- 335,010 (Table 8Table 6). At Site 2, estimated abundance was 3,200,639, with a 95% CI of +/- 550,402 smolts (Table 6; Table 9). Hourly smolt passage peaked at each site at the onset of darkness (Figure 7). Many smolts also migrated during daylight hours, however, accounting for 45% of the run at Site 1 and 39% of the run at Site 2.

Overall smolt run timing followed the same general trend as previous years (Figure 12), with a series of highest peaks between 1 and 7 June (Figure 11). Both sonar arrays appeared to have been deployed before any major smolt movements, based on patterns between the dates of deployment on 18 and 20 May and the first notable smolt detections on 22 May.

Smolt Distribution Vertical distribution was split into 0.2 m depth strata down to a maximum depth of 2.0 m at Site 1 and a maximum depth of 1.6 m at Site 2. Smolts were surface-oriented at both sites, though more so at Site 1 than Site 2 (Figure 13). Smolts had a deeper vertical distribution during daylight hours than at night. At night, over 95% of the smolts at each site travelled in the upper 1.0 meter of the water column (Figure 13).

Smolts were detected by all sonar pods at each site (i.e., across the entire river), but were disproportionately distributed. At both sites, between 80% and 85% of the smolts were in the center to center left of the river channel, corresponding roughly to the deepest part (Figure 14).

Environmental Conditions Water temperature was 2.1 °C at the initial measurement the morning of 18 May, then rose thereafter before peaking at 7.3 °C on 5 June. Precipitation was recorded on 11 dates. Water color was either murky or glacial (six days), or clear (all other days; Appendix B2).

Smolt Age, Weight, and Length On the Ugashik River, smolts were sampled for age, weight, and length on every second day between 21 May and 10 June (Appendix A5). A total of 590 smolts were aged,

13 Smolt monitoring on three Bristol Bay rivers in 2013

weighed, and measured for length to use in developing the length-age and the length- weight relationships (Appendix A5). These relationships were then applied to another 55 smolts measured for length during the season (Appendix A6).

Of the 645 smolts sampled in 2013, 68% were estimated at age-1 and 32% at age-2 (Appendix A7). Age-1 smolts had an estimated mean length of 97 mm and mean weight of 8.0 g; Age-2 smolts had an estimated mean length of 122 mm and mean weight of 16.5 g (Appendix A7). The discriminant body length of 108 mm was used to separate the two age classes. The majority of fish were age-1 on all days sampled except 27 and 29 May.

Egegik River

Data Pre-processing Sonar data were collected at the only array (Site 1) for 492 hours, from 22 May (1900 hrs) through 12 June (0600 hours). Equipment malfunctions or environmental noise (e.g., wind, ice, or vessels) precluded analysis of 129 hourly blocks of data (26% of total).

Smolt Abundance and Run Timing Sonar data were processed to the point of estimating relative density of smolts per m2, thereby allowing estimates of relative run timing, cross-channel distribution, and vertical distribution. Water velocity data was unreliable due to differing tidal influences across the river at the same time, however, preventing estimates of the actual number of smolts migrating per unit time. Abundance was therefore not estimated.

Most of the run emigrated between 26 May and 2 June, a more compressed run timing than on the Ugashik or Kvichak rivers in 2013 (Figure 15). The sonar system appeared to have been deployed before any major smolt movements, based on patterns between the dates of deployment on 22 May and the first main smolt detections 26 May.

Hourly smolt passage peaked at each site at the onset of darkness (Figure 7), with the hours between 2300 and 0500 accounting for 73% of the estimated smolt passage.

Smolt Distribution Vertical distribution was split into 0.2 m depth strata down to a maximum depth of 2.0 m at Site 1 and a maximum depth of 1.6 m at Site 2. Smolts were surface-oriented at both sites, though much less so than on the Kvichak and Ugashik rivers; on the Egegik River, only 60% of the smolts were in the top 1.0 m of the water column (Figure 16). The diel timing of this vertical distribution also differed from the Kvichak and Ugashik rivers; on the Egegik River, smolt abundance in the uppermost 0.4 m of the water column was greater during the day than at night Figure 16).

Smolts were detected by all sonar pods at each site (i.e., across the entire river), with approximately 70% of the run estimated to have been on the right half of the river (looking downstream). Relatively uniform transducer depths prevented assessing whether smolt density varied with depth (Figure 17).

14 Smolt monitoring on three Bristol Bay rivers in 2013

Environmental Conditions Water temperature was 4.0 °C at the initial measurement the morning of 2 May, but did not consistently rise for another week. Water temperatures peaked at 9.0 °C the evening of 8 June. Precipitation was recorded on seven dates. Water color was clear on all but one day (Appendix B3).

Smolt Age, Weight, and Length Insufficient numbers of smolt were caught to assess age, weight, or length.

DISCUSSION

In 2013, sampling on the Kvichak and Ugashik rivers was able to address the main objectives of smolt abundance, spatial distribution, run timing, and AWL structure. Abundance estimates were consistent between two arrays on each of the two rivers, and the full run appeared to have been captured based on early season patterns. On the Egegik River, sonar data was reliable for estimating relative run timing and spatial distribution, but not for abundance. Although Egegik River sonar was able to generate relative smolt density vertically, horizontally, and by time period, we did not have the detailed water flow data needed to estimate the actual numbers of smolts migrating per hour across the river. Better instrumentation in future years will be able to capture the varying tidal influences across the river. Low smolt catches on the Egegik River also precluded AWL analysis.

Results from 2013 support the idea that this project must be prepared to sample smolts when ice is flowing in the river, but that it may not be necessary in years when ice is out early. In 2013, ice was gone from the rivers by mid-May, and smolts did not begin migrating until a few days later. In 2011, ice remained unusually late in the Kvichak and Egegik rivers, and substantial numbers of smolts migrated with the ice during a smolt run that was later and more protracted than in 2013 (Wade et al. 2012b). The overlap between ice flows and smolt run timing will need to be tracked in future years to better assess the situations in which under-ice sampling is needed.

Kvichak River

Smolt Abundance The 2013 abundance estimate at each site was within the range reported on the Kvichak River over the prior five seasons (2008-2012; Table 5). In the past we have assumed the Site 1 estimate to be the more accurate of the two sites because it is located on a narrower portion of the river with a higher proportion of smolts migrating through the ensonified area, and/or Site 1 had better operational conditions for a given season. Both arrays functioned well in 2013; with no apparent reason for the 21% difference in estimated smolt abundance between sites, we recommend using Site 1 for the annual point estimate in 2013 (41.9 million smolts).

15 Smolt monitoring on three Bristol Bay rivers in 2013

This was the second year that Site 2 used a design of two sub-arrays, intended to provide more complete cross-river coverage than in past years. The design posed no complications and should be retained in future years.

Assumptions and Uncertainty in Abundance Estimates Swimming speed is used to calculate smolt flux (smolt/hour), which is used to estimate hourly, daily, and yearly abundance (Mueller et al. 2006). Smolt swimming speed is assumed to be at or near water velocity, based on work conducted over a three-day span in the early portion of the smolt migration by Maxwell et al. (2009). Although the best available data, the relatively short study period would not have been able to detect seasonal differences in swimming speed.

Error in swimming speed could bias smolt abundance estimates either high or low (Wade et al. 2010), and would have the greatest impact on indexing the smolt run if it swimming speed varies as a function of the overall abundance of the smolt run. Also, if swimming speed increases with smolt size, we would underestimate the contribution of the larger age-2 fish to the overall run. In the current study no attempt has been made to verify that smolts traveled at or near river velocity, or if this changes throughout the season.

Distribution and Run Timing Although run timing was earlier, smolt distribution was similar to prior years. Across the channel, smolts were most prevalent in the center to center-left of the river, in areas at or approaching the maximum river depth. Vertically, smolts were predominantly in the upper 1 meter of the water column, with a tendency to be distributed deeper during the day than night. The diel pattern of increased density during nighttime hours was also consistent with prior years. All of this information increases our understanding of smolt migration behavior, and provides a baseline useful for evaluating changes in future annual abundance estimates.

The early run timing in 2013 was consistent with Quinn’s (2005) suggested correlation of spring water temperatures and smolt run timing on the Kvichak River. The first 75% of the run migrated earlier than in any of the other five years of the project, and the water temperatures oat the run midpoint (27 May) were the highest midpoint temperatures (6.7 °C).

Smolt Age, Weight, and Length Age-2 smolts typically migrate earlier than age-1 smolts in the Kvichak River (Quinn 2005), as was seen again in 2013. The estimated age structure (45% age-1 and 55% age- 2) provides consistent comparisons with prior years, but may not represent the entire run in 2013 because it was not weighted by estimated daily abundance.

Ugashik River

Smolt Abundance Smolt abundance estimates in 2013 were the lowest of three years recorded on this project. Each array provided consistent estimates (3.0 vs 3.2 million smolts), approximately one-fourth the number estimated from identically-operated arrays in 2012.

16 Smolt monitoring on three Bristol Bay rivers in 2013

The 2013 estimates were even further reduced from 2010, but operational changes from 2010 may account for some of this difference. In 2010, spacing between the pods had been wider, and both arrays combined into one to provide sufficient coverage (Wade et al. 2013).

In 2010 and 2012, many Ugashik River smolts were located high enough in the water column that estimates may have been biased low (because some smolts at the water surface may not have been detectable). This was not as likely a factor in 2013, when smolts were not as concentrated at the water surface, further emphasizing the reduced abundance estimate from 2012 to 2013.

The entirety of the run appeared to have been captured in 2013, based on proportional estimates in the early and late season. This differed from 2012, when the project may have ended before the entire run was complete (Wade et al. 2012b).

The same caveats about swimming speed and abundance estimates discussed for the Kvichak River, above, apply to the Ugashik River estimates.

Distribution and Run Timing Across the river, smolts were disproportionately present at the thalweg, in the middle of the river, similar to prior years, and to the pattern seen on the Kvichak River. As on the Kvichak River, smolts were predominantly in the upper 1.0 m of the water column, and the vertical distributions tended to be deeper in daylight hours than at night. Smolts were not as concentrated in the upper 0.5 m of the water column as they were in 2012 and 2010, however (Wade et al. 2012b, 2013). Finally, the increased movement during nighttime hours was consistent with the Kvichak River and with prior years on the Ugashik River.

As on the Kvichak River, water temperature (6.8 °C) was higher at the run midpoint than in prior years (Table 10). Unlike the Kvichak River, however, this did not translate into a noticeably early run timing (Figure 12).

Smolt Age, Weight, and Length Ugashik River smolts were large (both length and weight) in 2013 compared to prior years. In 2013, age-1 Ugashik River smolts were similar in size to prior years on the Ugashik River, but age-2 smolts were some of the largest average sizes recorded and were closer to the long-term size of age-3 smolts than of age-2 (Appendix A7).

As in most years, Ugashik River smolts were larger than Kvichak River smolts and had a higher proportion of age-1 fish (Appendix A4; Appendix A7).

Egegik River

Smolt Abundance Although the sonar portion of the Egegik River work functioned reliably in 2013, site conditions and unreliable water velocity data meant that we could not confidently generate abundance estimates. The Egegik River differs from the other two study rivers

17 Smolt monitoring on three Bristol Bay rivers in 2013

in that water velocities at the sonar site are affected by tidal stage and thus highly variable in time and across the river. We were not confident that this variability was effectively measured in 2013, and using basic assumptions resulted (e.g., uniform vs. variable flow through the season) yielded large differences in estimated smolt abundance. Instead, we focused on sonar performance and on relative smolt abundance (run timing and inriver distribution) in advance of future monitoring efforts. The same caveats about swimming speed and abundance estimates discussed for the Kvichak and Ugashik rivers, above, would apply to any future Egegik River estimates.

Distribution and Run Timing The entirety of the run appeared to have been captured in 2013, based on proportional estimates in the early and late season.

Several aspects of smolt distribution and timing were different from the Kvichak and Egegik rivers in 2013. In contrast to the other two rivers, Egegik River smolts were less likely to migrate during the day, were more likely to be near the surface in the day than at night, and were distributed deeper in the water column overall. The run midpoint was similar to the Ugashik River, but water temperatures at that time were lower (Table 10).

CONCLUSION AND RECOMMENDATIONS

Upward-looking sonar arrays appear to be an effective way to estimate smolt abundance on select rivers systems in Bristol Bay. The arrays are clearly able to detect large changes in smolt abundance in both lateral and vertical segments of the river; these detections can be integrated across vertical, lateral, and time strata and used to generate abundance estimates at hourly, daily, and seasonal time intervals. The accuracy of the estimates as absolute numbers of fish cannot be validated without known abundances. However, the relatively low measurement error of two independent, sonar-based estimates from significantly different sites in the same river supports the belief that the estimates are a relatively precise index of daily abundance within each year; how well we can index the smolt abundance among years will be more difficult to determine. The usefulness in forecasting adult salmon returns will emerge as more smolt and adult return data are gathered. Large and unpredictable variability in marine survival may limit usefulness in this regard, but not affect the value in helping to understand variation in freshwater productivity and to refine escapement goals.

Sonar array configuration and duration of sampling can affect abundance estimates at each site, so it is necessary to have adequate spatial and temporal coverage. Several sonar changes implemented in 2012 (Wade et al. 2012b) were retained in 2013 and functioned well.

Swimming speed is an important metric for calculating absolute smolt abundance. Speed is used to calculate abundance and is currently based on data collected by ADF&G over a relatively short period of time (3 days). If there are diel or seasonal changes in smolt swimming speed, or if it varies by body size of smolt these are not currently taken into account when calculating abundance. Future work should estimate smolt swimming speeds over time and over different water velocities.

18 Smolt monitoring on three Bristol Bay rivers in 2013

Refinements to the BBSRI smolt sonar over the years have brought improvements to the operation of the program and increased our confidence in the abundance estimates and in the sources of uncertainty. Sonar hardware and software modifications have made operation of the sonar easier to understand and operate, reducing mistakes in the field. Physical changes to the sonar sleds and configuration of arrays have improved spatial coverage and ultimately the utility of smolt abundance estimates. Moving forward, BBSRI will continue to look for ways to improve the efficiency and accuracy of the project with the goal of a sustainable program that will be useful for future generations of stakeholders in Bristol Bay. Key field changes recommended for 2014 are as follows:

1. Investigate smolt swimming speeds on at least one river.

2. On the Egegik River, add a second sonar site, obtain accurate water velocity through space and time, and increase smolt catches for AWL analysis.

ACKNOWLEDGEMENTS

Kvichak River – We thank Corey Graham (BBSRI) and Chris Sewright and Dirk Middleton (ADF&G) for sonar operation and smolt sampling. Christina Salmon- Wassille, AlexAnna Salmon and Sandy Alvarez (Igiugig Village Council) provided logistical assistance for the field sampling. ADF&G supplied housing and equipment (boats, generators, etc.) essential to the project.

Ugashik River – We thank Jace Bures and Drew Stinnett (BBSRI) for sonar operation and fish sampling. Bob and Carol Dreeszen provided room and board, plus valuable assistance with logistics and operations.

Egegik River – We thank Kris Shippen, Colby Johnson, and Deven Lisac for sonar operation and fish sampling.

Anna-Maria Mueller and Jacqui Degan (Aquacoustics) assisted with the processing of all acoustic data. Dr. Scott Raborn (LGL) developed the statistical methods for estimating smolt abundance and Sean Burril (LGL) assisted with data assembly and analysis. Fred West (ADF&G) managed the scale aging operation, analyzed the AWL data, and supplied descriptions of the smolt sampling and AWL methods. Vicki Priebe (Happy Computer Services) assisted with report formatting and publication.

Funding for the entire project was provided by BBSRI, the Alaska Sustainable Salmon Fund (AKSSF Project 44630), the Pebble Fund, and BBSRI’s benefactors: Alaskan Leader, Bristol Leader, Arctic Fjord, and Arctic Storm.

This report was prepared by the Bristol Bay Science and Research Institute by Matthew Nemeth, Guy Wade, Don Degan, and Michael Link, with partial support from AKSSF 44630, National Oceanic and Atmospheric Administration, U.S. Department of Commerce, administered by the Alaska Department of Fish and Game. The statements, findings, conclusions, and recommendations are those of the author(s) and do not

19 Smolt monitoring on three Bristol Bay rivers in 2013 necessarily reflect the views of the National Oceanic and Atmospheric Administration, the U.S. Department of Commerce, or the Alaska Department of Fish and Game.

20 Smolt monitoring on three Bristol Bay rivers in 2013

LITERATURE CITED

Baker, T.T., L.W. Fair, R.A. Clark, and J.J. Hasbrouck. 2006. Review of salmon escapement goals in Bristol Bay, Alaska, 2006. Alaska Department of Fish and Game, Fishery Manuscript Series No. 06-05, Anchorage.

Baker, T.T., L.W. Fair, F.W. West, G.B. Buck, X. Zhang, S. Fleischman, and J. Erickson. 2009. Review of salmon escapement goals in Bristol Bay, Alaska 2009. Alaska Department of Fish and Game, Fishery Manuscript Series No. 09-05, Anchorage.

Bue, B.G., and D.M. Eggers. 1989. An age-length key for sockeye salmon. Alaska Department of Fish and Game, Division of Commercial Fisheries, Regional Information Report 2D89-5, Anchorage.

Burgner, R.L. 1968. Further studies of Alaska sockeye salmon. University of Washington, Seattle.

Cochran, W.G. 1977. Sampling Techniques. John Wiley and Sons, New York, New York.

Crawford, D.L., and F.W. West. 2001. Bristol Bay sockeye salmon smolt studies for 2000. Alaska Department of Fish and Game, Regional Information Report 2A01- 12, Anchorage.

Crawford, D.L., and L.F. Fair. 2003. Bristol Bay salmon smolt studies using upward- looking sonar, 2002. Alaska Department of Fish and Game, Commercial Fisheries Division. Regional Information Report No. 2A03-17, Anchorage.

Dye, J.E., and C.J. Schwanke. 2009. Report to the Alaska Board of Fisheries for the recreational fisheries of Bristol Bay, 2002, 2008, and 2009. Alaska Department of Fish and Game, Special Publication No. 09-14, Anchorage.

Maxwell, S., A. Mueller, D. Degan, D. Crawford, L. McKinley, and N. Hughes. 2009. An evaluation of the Bendix smolt counter used to estimate outmigrating sockeye salmon smolt in the Kvichak River, Alaska, and the development of a replacement sonar, 2000-2001. Alaska Department of Fish and Game, Fishery Manuscript No. 09-02, Anchorage.

Morstad, S., M. Jones, T. Sands, P. Salomone, G. Buck, and F. West. 2010. 2009 Bristol Bay area annual management report. Alaska Department of Fish and Game, Fishery Management Report No. 10-25, Anchorage.

Mueller, A.M., D.J. Degan, R. Kieser, and T. Mulligan. 2006. Estimating sockeye salmon smolt flux and abundance with side-looking sonar. North American Journal of Fisheries Management 26: 523-534.

Quinn, T.P. 2005. The behavior and ecology of Pacific Salmon and Trout. University of Washington Press, Seattle, Washington.

21 Smolt monitoring on three Bristol Bay rivers in 2013

Ricker, W.E. 1975. Computation and interpretation of biological statistics of fish populations. Bulletin of the Fisheries Research Board of Canada No. 191, Ottawa, Canada.

Ruggerone, G.T., and M.R. Link. 2006. Collapse of Kvichak sockeye salmon production brood years 1991-1999: Population characteristics, possible factors, and management implications. Unpublished report prepared by Natural Resources Consultants, Inc. and LGL Alaska Research Associates, Inc. for the North Pacific Research Board, Anchorage, AK. xiv + 103 p.

Salomone, P., S. Morstad, T. Sands, M. Jones, T. Baker, G. Buck, F. West, and T. Kreig. 2011. 2010 Bristol Bay area annual management report. Alaska Department of Fish and Game, Fishery Management Report No. 11-23, Anchorage.

Todd, G.T. 1994. A lightweight, inclined-plane trap for sampling salmon smolts in rivers. Alaska Fishery Research Bulletin 1 (2):168-175.

USGS. 2008. National Water Information System: Web Interface, 15300500. Retrieved on 20 December 2008 from the United States Geological Survey webpage: http://waterdata.usgs.gov/ak/nwis/nwismap/?site_no=15300500&agency_cd=US GS.

Wade, G.D., D.J. Degan, M.R. Link, and S.W. Raborn. 2010. Evaluation of an up- looking sonar system designed to enumerate sockeye salmon smolts on the Kvichak River, 2008. Report prepared by LGL Alaska Research Associates, Inc., Anchorage, AK, and Aquacoustics, Inc. Sterling, AK, for the Bristol Bay Science and Research Institute, Dillingham, AK, 45 p + Appendix.

Wade, G.D., D.J. Degan, M.R. Link, and S.W. Raborn. 2012a. Estimates of hourly, daily, and seasonal sockeye salmon smolt abundance on the Kvichak and Ugashik Rivers in 2010. Report prepared by LGL Alaska Research Associates, Inc., Anchorage, AK, and Aquacoustics, Inc. Sterling, AK, for the Bristol Bay Science and Research Institute, Dillingham, AK, 61 p + Appendix A.

Wade, G.D., D.J. Degan, M.R. Link, and M.J. Nemeth. 2012b. Monitoring of sockeye salmon smolt abundance and inriver distribution using sonar on the Kvichak, Egegik, and Ugashik rivers in 2011. Report prepared by LGL Alaska Research Associates, Inc., Anchorage, AK, and Aquacoustics, Inc. Sterling, AK, for the Bristol Bay Science and Research Institute, Dillingham, AK, 64 p.

Wade, G.D., D.J. Degan, M.R. Link, and M.J. Nemeth. 2013. Monitoring sockeye salmon smolt abundance and inriver distribution using sonar on the Kvichak and Ugashik rivers in 2012. Report prepared by LGL Alaska Research Associates, Inc., Anchorage, AK, and Aquacoustics, Inc. Sterling, AK, for the Bristol Bay Science and Research Institute, Dillingham, AK, 65 p.

22 Smolt monitoring on three Bristol Bay rivers in 2013

TABLES

23 Smolt monitoring on three Bristol Bay rivers in 2013

Table 1. Historical ice cover dates on Lake Iliamna, in the Kvichak River drainage.

Winter of Freeze-up datea Break-up date Ice cover days 1976 - 1977 4-Feb 2-May 88 1977 - 1978 11-May 1978 - 1979 3-May 1979 - 1980 3-May 1980 - 1981 1981 - 1982 9-Jan 25-May 137 1982 - 1983 1983 - 1984 1984 - 1985 11-Feb 5-Jun 115 1985 - 1986 18-Jan 12-May 115 1986 - 1987 13-Feb 23-May 39 1987 - 1988 26-Jan 1988 - 1989 13-Jan 1989 - 1990 9-Jan 22-May 134 1990 - 1991 7-Jan 1991 - 1992 27-Jan 4-May 98 1992 - 1993 22-Jan 3-May 102 1993 - 1994 16-Feb 5-May 79 1994 - 1995 11-Jan 22-May 132 1995 - 1996 12-Jan 5-May 114 1996 - 1997 23-Dec 8-May 137 1997 - 1998 5-Jan 26-Apr 112 1998 - 1999 30-Dec 28-May 150 1999 - 2000 30-Dec 6-May 128 2000 - 2001 2001 - 2002 20-May 2002 - 2003 11-Apr 2003 - 2004 2004 - 2005 12-May 2005 - 2006 19-May 2006 - 2007 17-May 2007 - 2008 15-May 2008 - 2009 20-May 2009 - 2010 22-May 2010 - 2011 13-May 2011 - 2012 22-May 2012 - 2013 17-May aData provided by ADF&G; most information is from local air charter companies and considered anecdotal. 2013 data is from the University of Washington, Fishery Research Institute.

24

monitoringSmolt on three BayBristol rivers 2013in

Table 2. Historical ice cover dates on Lower Ugashik Lake (Ugashik River drainage) and Becharof Lake (Egegik River drainage). Lower Ugashik Lake Becharof Lake Winter of Freeze-up datea Break-up date Ice cover days Freeze-up datea Break-up date Ice cover days 1981 - 1982 12-May 1982 - 1983 18-Jan 1983 - 1984 16-Jan 16-Jan 16-May 121 1984 - 1985 11-Feb 14-May 92 11-Feb 3-May 82 1985 - 1986 26-Feb 9-May 74 26-Feb 27-Apr 61 1986 - 1987 12-Mar 12-Mar 1987 - 1988 9-Dec 24-Mar 106 24-Mar 1988 - 1989 17-Jan 10-May 113 17-Jan 27-Apr 101 1989 - 1990 21-Feb 25-Apr 63 21-Feb 25-Apr 64 1990 - 1991 8-Jan 4-Feb 1-Apr 57

25 1991 - 1992 27-Jan 4-May 97 27-Jan 10-May 104

1992 - 1993 20-Jan 31-Mar 70 23-Jan 31-Mar 68 1993 - 1994 16-Feb 8-Apr 51 25-Feb 4-Apr 39 1994 - 1995 24-Jan 28-Apr 94 24-Jan 28-Apr 95 1995 - 1996 8-Jan 15-Apr 97 8-Jan 28-Mar 80 1996 - 1997 13-Dec 26-Apr 135 13-Dec 19-Apr 128 1997 - 1998 5-Jan 4-Apr 89 6-Jan 4-Apr 89 1998 - 1999 22-Jan 19-May 117 5-Feb 28-May 113 1999 - 2000 25-Dec 7-Apr 104 2-Jan 12-Apr 101 2000 - 2009b 2009 - 2010 10-May 2010 - 2011 10-Dec 8-Feb Lake ice broke 4 times. 10-May 2011 - 2012 15-Nov 16-May Ice flowed until 22-May. 2012 - 2013 aData provided by ADF&G; most information is from local air charter companies and considered anecdotal. bADF&G smolt program discontinued in 2001. c Smolt crews arrived on 10-May and noted 20 - 30 % of lake ice remaining.

monitoringSmolt on three BayBristol rivers 2013in

Table 3. Annual sonar pod placements on the Kvichak, Ugashik, and Egegik rivers since project inception in 2008. Depth is water depth in meters, range (R) is distance from shore in meters, and RR and RL are right and left sides of river when facing downstream. Blank cells indicate no pods used; NA indicates pods used but no data available. 2013 2012 2011 2010 2009 2008 River Site Transducer Depth (m) R (m) Depth (m) R (m) Depth (m) R (m) Depth (m) R (m) Depth (m) R (m) Depth (m) R (m) Kvichak River 1 RR bank 0 0 0 0 0 0 0 0 0 0 0 0 1 1 2.5 17 1.7 19 1.8 13 1.9 15 2.4 29 2.4 19 1 2 2.8 27 2.3 29 2.1 22 2.2 25 2.9 39 2.5 29 1 3 3.1 37 2.7 39 2.6 35 2.7 34 3.5 49 3.0 39 1 4 3.7 47 3.1 49 3.1 45 3.3 44 3.9 59 3.6 49 1 5 3.9 57 3.6 59 3.5 53 3.6 54 4.4 69 4.0 59 1 6 3.9 67 3.7 69 3.7 65 3.8 65 3.5 79 4.3 69 1 7 3.9 77 3.8 79 3.9 74 4.1 75 2.2 89 3.0 79

1 8 NA 87 3.4 89 3.2 83 3.5 86 NA NA 1.9 89 1 RL bank 0 102 0 114 0 101 0 103 0 103 0 103 26 2 RR bank 0 0 0 0 0 0 0 0 0 0 0 0 2 1 1.5 4 2.2 15 2.8 35 2.8 41 3.7 37 2.9 42 2 2 2.7 14 2.9 26 2.6 45 2.6 51 3.8 47 2.7 52 2 3 3.1 24 3.0 36 2.5 56 2.5 62 4.0 57 2.7 62 2 4 3.1 34 2.8 46 2.5 66 2.5 72 4.3 67 2.7 72 2 5 3.3 38 2.7 57 2.4 75 2.6 81 4.3 77 2.6 82 2 6 3.1 48 2.7 68 2.4 85 2.6 91 2.6 87 2.5 92 2 7 3.1 58 2.6 78 2.3 95 2.8 101 1.4 97 2.6 102 2 8 2.9 68 2.5 88 2.7 105 3.4 111 2 9 2.9 78 2.5 99 2 10 2.7 88 2.6 110 2 11 2.7 98 3.3 120 2 12 3.3 108 3.4 131 2 RL bank 0 130 0 146 0 128 0 134 0 119 0 130 Ugashik River 1 RR bank 0 0 0 0 0 0 0 0 1 1 2.3 20 1.7 22 1.6 35 0.9 16 1 2 3.1 25 2.6 27 2.3 40 1.3 18

monitoringSmolt on three BayBristol rivers 2013in

Table 3. Continued. 2013 2012 2011 2010 2009 2008 River Site Transducer Depth (m) R (m) Depth (m) R (m) Depth (m) R (m) Depth (m) R (m) Depth (m) R (m) Depth (m) R (m) 1 3 3.3 30 3.2 34 2.4 45 2.5 22 1 4 2.7 35 3.0 38 2.5 50 2.6 27 1 5 1.9 40 2.0 42 2.2 55 3.3 32 1 6 2.8 37 1 7 1.9 42 1 RL bank 0 43 0 44 0 60 0 45 2 RR bank 0 0 0 0 0 0 2 1 1.8 32 1.8 41 1.3 16 2 2 2.1 37 2.2 46 1.6 51 2 3 2.3 43 2.4 50 2.2 56 2 4 2.8 48 2.0 54 2.5 60 2 5 2.5 53 2.0 64

2 RL bank 0 61 0 57 0 66

27 Egegik River

1 RL bank 0 0 0 0 1 1 NA 4 0.8 11 1 2 3.6 14 1.9 15 1 3 3.9 24 2.8 28 1 4 4.1 34 3.0 39 1 5 4.1 44 3.3 50 1 6 4.0 54 3.2 60 1 7 3.8 64 3.2 71 1 RR bank 0 80 0 90 2 RL bank 0 0 2 1 1.1 15 2 2 2.0 26 2 3 2.5 37 2 4 2.8 47 2 5 3.3 57 2 6 2.9 68 2 7 2.3 79 2 RR bank 0 90

monitoringSmolt on three BayBristol rivers 2013in

Table 4. Mean water velocity (m/s) by date at sonar pods on the Kvichak, Ugashik, and Egegik rivers, 2013. Blank cells indicate no pods used; NA indicates pods used but no data available Average water velocity (m/s) at sonar transducer (T) River Array Date T1 T2 T3 T4 T5 T6 T7 T8 T9 T10 T11 T12 Kvichak River 1 25-May 0.7 1.1 1.4 1.3 1.6 1.6 1.6 1.5 6-Jun 0.7 0.9 1.1 1.6 1.7 1.6 1.8 1.6 11-Jun 0.7 1.1 1.4 1.7 1.7 1.8 1.8 1.6 2 24-May 0.5 0.5 0.7 0.9 0.9 1.0 1.1 1.0 1.0 1.0 0.8 0.8 5-Jun 0.6 0.8 1.0 1.1 1.3 1.3 1.5 1.3 1.4 1.3 1.1 1.0 11-Jun 0.7 0.9 1.0 1.2 1.2 1.4 1.4 1.5 1.5 1.4 1.1 1.0 Ugashik River

28 1 22-May 1.1 1.8 2.0 2.2 1.3

30-May 1.2 1.7 2.1 2.4 2.0 5-Jun 1.2 1.7 2.1 2.5 2.0 11-Jun 1.3 1.8 2.4 2.4 1.5 2 22-May 1.6 1.6 1.9 1.4 NA 30-May 1.8 2.1 2.3 1.8 NA 5-Jun 1.7 2.1 2.4 1.7 NA 11-Jun 1.2 2.1 2.3 1.4 NA Egegik River 1 21-May 0.3 1.2 1.8 2.0 1.9 1.5 1.5

Smolt monitoring on three Bristol Bay rivers in 2013

Table 5. Annual abundance estimates of sockeye salmon smolts, by site, on the Kvichak, Ugashik, and Egegik rivers since project inception in 2008. Difference River Year Site 1 Site 2 Smolt Percent Kvichak 2008 30,786,980 26,965,627 3,821,353 13.2 2009 35,247,209 38,755,938 (3,508,729) -9.5 2010a 57,320,620 Not operated 2010b 15,805,698 15,891,807 (86,109) -0.5 2011 48,806,237 41,730,658 7,075,579 15.6 2012 49,198,830 47,011,636 2,187,194 4.5 2013 41,861,000 34,001,000 7,860,000 20.7 Ugashik 2010 20,400,000 Not operated 2012 11,193,920 11,064,475 129,445 1.2 2013 3,060,357 3,200,639 (140,282) -4.5 Egegik 2011 9,907,344 8,860,449 1,046,895 11.2 2013 Not reported Not operated a Estimate for entire run, Site 1 only, 24 May - 13 June. b Estimate for partial run when both sites operated concurrently, 31 May - 13 June.

29 Smolt monitoring on three Bristol Bay rivers in 2013

Table 6. Estimated daily abundance and annual proportion of sockeye salmon smolts at Site 1 on the Kvichak River, 2013. Site 1, Kvichak River Abundance Percentage of total Date Daily 95% CI Cumulative Daily Cumulative 19-May 8,869 6,472 8,869 0 0 20-May 22,716 17,116 31,585 0 0 21-May 170,528 152,879 202,112 0 0 22-May 3,124,776 2,904,284 3,326,888 7 8 23-May 3,143,071 2,938,348 6,469,959 8 15 24-May 2,039,193 1,972,866 8,509,152 5 20 25-May 4,810,694 4,452,501 13,319,847 11 32 26-May 7,377,852 6,103,867 20,697,699 18 49 27-May 3,455,908 3,339,308 24,153,606 8 58 28-May 7,241,432 7,422,435 31,395,038 17 75 29-May 525,357 672,248 31,920,395 1 76 30-May 92,877 99,577 32,013,272 0 76 31-May 815,685 871,512 32,828,957 2 78 1-Jun 588,711 583,812 33,417,668 1 80 2-Jun 723,239 678,725 34,140,907 2 82 3-Jun 631,433 412,465 34,772,340 2 83 4-Jun 2,637,467 2,009,507 37,409,807 6 89 5-Jun 1,035,556 1,033,382 38,445,363 2 92 6-Jun 977,172 574,889 39,422,535 2 94 7-Jun 628,248 676,794 40,050,783 2 96 8-Jun 388,233 346,347 40,439,016 1 97 9-Jun 613,131 557,866 41,052,147 1 98 10-Jun 458,837 383,535 41,510,984 1 99 11-Jun 140,809 108,805 41,651,793 0 100 12-Jun 136,850 187,921 41,788,643 0 100 13-Jun 72,333 66,572 41,860,976 0 100 Total 41,860,976 5,002,876 100

30 Smolt monitoring on three Bristol Bay rivers in 2013

Table 7. Estimated daily abundance and annual proportion of sockeye salmon smolts at Site 2 on the Kvichak River, 2013. Site 2, Kvichak River Abundance Percentage of total Date Daily 95% CI Cumulative Daily Cumulative 18-May 598 283 598 0 0 19-May 18,649 11,979 19,247 0 0 20-May 20,274 13,162 39,521 0 0 21-May 156,317 119,902 195,838 0 1 22-May 2,286,175 1,312,716 2,482,014 7 7 23-May 2,755,346 2,005,205 5,237,360 8 15 24-May 2,103,433 1,722,042 7,340,793 6 22 25-May 5,094,586 4,300,634 12,435,378 15 37 26-May 4,209,072 3,045,628 16,644,450 12 49 27-May 2,017,342 1,221,342 18,661,792 6 55 28-May 3,901,128 2,588,178 22,562,920 11 66 29-May 314,750 273,112 22,877,670 1 67 30-May 137,745 121,033 23,015,415 0 68 31-May 956,921 787,013 23,972,336 3 71 1-Jun 414,618 304,778 24,386,954 1 72 2-Jun 609,077 421,152 24,996,031 2 74 3-Jun 834,450 515,169 25,830,481 2 76 4-Jun 3,605,894 2,271,145 29,436,375 11 87 5-Jun 795,862 584,441 30,232,238 2 89 6-Jun 1,285,163 951,940 31,517,400 4 93 7-Jun 633,295 496,961 32,150,696 2 95 8-Jun 454,262 270,009 32,604,958 1 96 9-Jun 769,602 528,769 33,374,560 2 98 10-Jun 518,204 308,636 33,892,764 2 100 11-Jun 63,462 54,367 33,956,226 0 100 12-Jun 44,323 35,517 34,000,549 0 100 Total 34,000,549 2,951,901 100

31 Smolt monitoring on three Bristol Bay rivers in 2013

Table 8. Estimated daily abundance and annual proportion of sockeye salmon smolts at Site 1 on the Ugashik River, 2013. Site 1, Ugashik River Abundance Percentage of total Date Daily 95% CI Cumulative Daily Cumulative 20-May 29,975 43,603 29,975 1 1 21-May 29,706 22,292 59,681 1 2 22-May 46,945 64,253 106,626 2 3 23-May 17,540 25,100 124,165 1 4 24-May 15,024 20,937 139,189 0 5 25-May 29,849 52,780 169,038 1 6 26-May 86,384 154,824 255,422 3 8 27-May 142,919 256,881 398,341 5 13 28-May 199,454 358,940 597,795 7 20 29-May 51,270 70,039 649,065 2 21 30-May 104,552 107,221 753,617 3 25 31-May 326,110 317,537 1,079,727 11 35 1-Jun 280,653 291,514 1,360,380 9 44 2-Jun 180,552 147,186 1,540,932 6 50 3-Jun 188,308 222,602 1,729,241 6 57 4-Jun 192,270 241,171 1,921,511 6 63 5-Jun 180,803 149,955 2,102,313 6 69 6-Jun 300,624 251,426 2,402,938 10 79 7-Jun 241,573 326,261 2,644,510 8 86 8-Jun 80,479 99,158 2,724,990 3 89 9-Jun 57,881 51,509 2,782,870 2 91 10-Jun 82,354 124,335 2,865,224 3 94 11-Jun 101,086 158,166 2,966,311 3 97 12-Jun 36,296 29,406 3,002,607 1 98 13-Jun 57,750 73,777 3,060,357 2 100 Total 3,060,357 335,010 100 100

32 Smolt monitoring on three Bristol Bay rivers in 2013

Table 9. Estimated daily abundance and annual proportion of sockeye salmon smolts at Site 2 on the Ugashik River, 2013. Site 2, Ugashik River Abundance Percentage of total Date Daily 95% CI Cumulative Daily Cumulative 18-May 2,413 2,388 2,413 0 0 19-May 1,876 1,834 4,289 0 0 20-May 2,822 3,546 7,110 0 0 21-May 69,696 66,335 76,807 2 2 22-May 100,427 146,189 177,234 3 6 23-May 6,212 5,173 183,445 0 6 24-May 11,186 9,996 194,631 0 6 25-May 45,361 51,273 239,992 1 7 26-May 44,834 44,073 284,827 1 9 27-May 65,186 66,245 350,012 2 11 28-May 107,575 106,171 457,587 3 14 29-May 49,493 68,297 507,080 2 16 30-May 162,550 189,525 669,631 5 21 31-May 228,382 271,186 898,013 7 28 1-Jun 353,369 470,213 1,251,382 11 39 2-Jun 206,604 160,914 1,457,986 6 46 3-Jun 301,035 342,537 1,759,021 9 55 4-Jun 236,128 335,514 1,995,149 7 62 5-Jun 333,873 399,951 2,329,022 10 73 6-Jun 239,053 294,591 2,568,075 7 80 7-Jun 361,164 578,491 2,929,238 11 92 8-Jun 57,287 69,233 2,986,526 2 93 9-Jun 72,082 80,761 3,058,608 2 96 10-Jun 89,776 128,029 3,148,384 3 98 11-Jun 32,436 30,986 3,180,820 1 99 12-Jun 19,818 22,432 3,200,639 1 100 Total 3,200,639 550,402 100 100

33 Smolt monitoring on three Bristol Bay rivers in 2013

Table 10. Median migration date of sockeye smolts from the Kvichak, Ugashik, and Egegik rivers since project inception in 2008, along with corresponding water temperature.

Median migration Water temp. River Year date (°C) Kvichak 2008 1-Jun 6.3 2009 28-May 6.0 2010 27-May 4.7 2011 28-May 4.5 2012 27-May 5.1 2013 27-May 6.5 Ugashik 2010 2-Jun Pending 2012 1-Jun 5.5 2013 2-Jun 6.8 Egegik 2013 31-May 3.8

34 Smolt monitoring on three Bristol Bay rivers in 2013

FIGURES

35 Smolt monitoring on three Bristol Bay rivers in 2013

Figure 1. The Bristol Bay region, showing the Kvichak, Ugashik, and Egegik and six other main rivers that produce sockeye salmon targeted in commercial, sport, and subsistence fisheries.

36 Smolt monitoring on three Bristol Bay rivers in 2013

Figure 2. Map of Kvichak River and Lake Clark drainages in Southwestern Alaska, showing locations of sonar sites 1 and 2 operated near the village of Igiugig, 2013.

37 Smolt monitoring on three Bristol Bay rivers in 2013

Figure 3. Map of Ugashik River watershed, showing locations of sonar systems and fyke net operated near the outlet of Lower Ugashik Lake, 2013.

38 Smolt monitoring on three Bristol Bay rivers in 2013

Figure 4. Map of the Egegik River watershed, showing locations of the sonar system and fyke net operated near the outlet of Becharof Lake, 2013.

39 Smolt monitoring on three Bristol Bay rivers in 2013

Figure 5. Conceptual drawing of the smolt sonar system used in 2013.

40 Smolt monitoring on three Bristol Bay rivers in 2013

2.5 Kvichak River, Site 1

2.0 Sonar data Interpolated data 1.5

1.0 Interpolations 0.5 Smolt abundance Smolt abundance in millions

0.0 19-May 22-May 25-May 28-May 1-Jun 4-Jun 7-Jun 10-Jun 13-Jun Date 2.5 Kvichak River, Site 2

2.0 Sonar data Interpolated data

1.5

1.0

Interpolations 0.5 Smolt abundance Smolt abundance in millions

0.0 18-May 22-May 25-May 28-May 31-May 3-Jun 6-Jun 9-Jun Date

Figure 6. Daily smolt estimates at Site 1 and 2 on the Kvichak River showing periods when abundance was estimated via linear interpolation, 2013.

41 Smolt monitoring on three Bristol Bay rivers in 2013

20 Kvichak River 18 16 Site 1 14 Site 2 12 10 smolts 8 6 4 Percentageof total sockeye 2 0 20 18 Ugashik River 16 14 12 10 smolts 8 6 4

Percentageof total sockeye 2 0 20 18 Egegik River 16 14 12 10 smolt 8 6 4 2 Percentageof total sockeye 0 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 0 1 2 3 Hour of Day

Figure 7. Proportion of total smolts by hour of the day at sonar sites on the Kvichak, Ugashik, and Egegik rivers in 2013. Shading shows hours considered nighttime (2300 – 0500) during the study period.

42 Smolt monitoring on three Bristol Bay rivers in 2013

Proportion of total run

0% 10% 20% 30% 40% 50% 0.0 Day 0.4 Dark 0.8

1.2

1.6 Kvichak River, Site 1 Depth (m) Depth 2.0

2.4

2.8

0.0

0.4

0.8

1.2

Depth (m) Depth Kvichak River, Site 2 1.6

2.0

2.4

2.8

Figure 8. Vertical distribution of sockeye salmon smolts migrating in darkness (2300 – 0500 hrs) and light at sites 1 and 2 on the Kvichak River in 2013. Water depth was 2.8 m at Site 1 and 2.0 m at Site 2.

43 Smolt monitoring on three Bristol Bay rivers in 2013

Smolt run timing 100%

75%

2013 Mean 2008-2012 Min 2008-2012 50% Max 2008-2012

Cumulative proportion 25%

0%

Date Figure 9. Annual run timing of sockeye salmon smolts at Site 1 on the Kvichak River since project inception in 2008.

44 Smolt monitoring on three Bristol Bay rivers in 2013

35 Kvichak Site 1 0 Smolt 30 Transducers Right Bank 1 25 Left Bank Depth (m) 20 2 15 Depth (m) Depth

10 3

Percentage of total sockeye smolt 5

0 4

River Strata (m)

35 0 Kvichak Site 2 30 1 25

20 2 15 Depth (m) Depth

10 3

Percentage of total sockeye smolt 5

0 4

River Strata (m)

Figure 10. Water depth and percent distribution of sockeye salmon smolts at Site 1 and Site 2 sonar pods on the Kvichak River in 2013, showing pod distances (m) from right bank. Water depth unavailable for 83-95 m at Site 1.

45 Smolt monitoring on three Bristol Bay rivers in 2013

3.0 Ugashik River, Site 1

Interpolated data 2.0 Sonar data

1.0 Smolt abndance Smolt abndance x 100,000

0.0 20-May 23-May 26-May 29-May 1-Jun 4-Jun 7-Jun 10-Jun Date 3.0 Ugashik River, Site 2

Interpolated data 2.0 Sonar data

1.0 Smolt abundance Smolt abundance x 100,000 0.0 18-May 21-May 24-May 28-May 31-May 3-Jun 6-Jun 9-Jun Date

Figure 11. Smolt estimates at Site 1 and 2 on the Ugashik River showing periods when abundance was estimated via linear interpolation, 2013.

46 Smolt monitoring on three Bristol Bay rivers in 2013

100% Smolt run timing

Mean 1991 - 2000 75% 2010 Both sites 2012 Both sites 2013 50%

25% Cumulative proportion Cumulative

0%

Date Figure 12. Annual run timing of sockeye salmon smolts on the Ugashik River since project inception in 2010. 2012 and 2013 data are for both sonar sites combined.

47 Smolt monitoring on three Bristol Bay rivers in 2013

Proportion of total run 0% 10% 20% 30% 40% 50% 0.0 Day 0.4 Dark

0.8

1.2 Site 1 Depth (m)

1.6

2.0

0.0

0.4

0.8

Depth (m) Depth 1.2 Site 2

1.6

2.0

Figure 13. Vertical distribution of sockeye salmon smolts migrating in darkness (2300 – 0500 hrs) and light at sites 1 and 2 on the Ugashik River in 2013. Difference in vertical axes between sites reflects shallower depths at Site 2.

48 Smolt monitoring on three Bristol Bay rivers in 2013

60 Ugashik Site 1 0 Smolt 50 Transducer Right Bank 1 Left Bank 40 Depth

30 2 Depth(m) 20 3

Percentageof total sockeye smolt 10

0 4 0-11 12-21 22-26 27-33 34-37 38-42 43 River Strata (m)

60 Ugashik Site 2 0

50 1 40

30 2 Depth(m) 20 3

Percentageof total sockeye smolt 10

0 4 0-25 26-34 35-41 42-46 47-51 52-58 59-61 River Strata (m)

Figure 14.Water depth and percent distribution of sockeye salmon smolts at Site 1 and Site 2 sonar pods on the Ugashik River in 2013, showing pod distances (m) from right bank.

49 Smolt monitoring on three Bristol Bay rivers in 2013

Smolt run timing, 2013 100% 90% 80% Kvichak River 70% Egegik River 60% Ugashik River 50% 40% 30%

Cumulative proportion 20% 10% 0%

Date Figure 15. Annual run timing of sockeye salmon smolts on the Kvichak (both sites combined), Ugashik (both sites combined), and Egegik (one site) rivers in 2013.

Proportion of total run 0% 10% 20% 30% 40% 50% 0.0

0.4 Dark Day 0.8

1.2

Depth (m) Depth 1.6 Site 1 2.0

2.4

2.8

Figure 16. Vertical distribution of sockeye salmon smolts migrating in darkness (2300 – 0500 hrs) and light on the Egegik River in 2013.

50 Smolt monitoring on three Bristol Bay rivers in 2013

35 Egegik River 0 Smolt 30 Transducer 1 Left Bank 25 Right Bank 20 Depth 2

15

3 Depth(m) 10 4 5 Percentageof total sockeye smolts 0 5 0-7 8-17 18-27 28-37 38-47 48-57 58-67 68-80 River Strata (m)

Figure 17. Water depth and percent distribution of sockeye salmon smolts at Site 2 sonar pods on the Egegik River in 2013, showing pod distances (m) from right bank.

51 Smolt monitoring on three Bristol Bay rivers in 2013

PHOTOS

52 Smolt monitoring on three Bristol Bay rivers in 2013

Flow

Photo 1. Black line showing the approximate location of Site 1 sonar array used in 2008 – 2013 on the Kvichak River.

Flow

Photo 2. Black line showing the approximate location of Site 2 sonar array used in 2008 and 2010 – 2013 on the Kvichak River.

53 Smolt monitoring on three Bristol Bay rivers in 2013

Lower Ugashik Flow Lake outlet

Photo 3. Black lines showing the approximate location of the two sonar arrays used on the Ugashik River in 2010 – 2013.

Photo 4. Sonar pod on shore during deployment with wire rope anchor harness running through eyebolts on the upstream side to prevent tipping.

54 Smolt monitoring on three Bristol Bay rivers in 2013

Photo 5. Sonar pods mounted (large orange pod is the split beam) on a welded aluminum sled with attached power, network cable, and power/control housings.

Photo 6. Inclined Plane Trap (IPT) used to capture sockeye salmon smolts on the Kvichak River in 2013.

55 Smolt monitoring on three Bristol Bay rivers in 2013

APPENDICES

56 Smolt monitoring on three Bristol Bay rivers in 2013

Appendix A1. Daily total catch, subsample size (n), and estimated age structure of sockeye salmon smolt on the Kvichak River, 2013. Data tables and analysis provided by the Alaska Department of Fish and Game. Incline Plane Trap (Kvichak River) Elapsed Smolt Time Total 8 am H20 Date (min) Catch n CPUE b age 1 Temp (°C) 05/19 110 236 236 2.15 0.0% 3.0 05/20 106 453 453 4.27 0.0% 3.0 05/21 55 775 613 14.09 0.7% 4.0 05/22 c - - - 3.0 05/23 c - - - 4.0 05/24 c - - - 3.0 05/25 74 775 602 10.47 8.4% 3.0 05/26 28 1,225 604 43.75 4.6% 3.0 05/27 10 1,365 618 136.50 25.6% 5.0 05/28 10 1,055 618 105.50 63.4% 8.0 05/29 115 733 539 6.37 35.9% 8.0 05/30 6 1,140 634 190.00 50.9% 9.0 05/31 23 960 600 41.74 58.6% 9.0 06/01 40 985 600 24.63 64.3% 8.0 06/02 112 365 301 3.26 27.9% 8.0 06/03 15 1,150 602 76.67 82.0% 6.0 06/04 28 1,210 600 43.21 87.6% 6.0 06/05 130 305 200 2.35 74.4% 8.0 06/06 24 1040 603 43.33 81.6% 8.0 06/07 127 188 188 1.48 71.3% 8.0 06/08 120 119 112 0.99 82.1% 7.0 06/09 121 5 5 0.04 80.0% 7.0 06/10 120 8 8 0.07 62.5% 8.0 Total 1,374 14,092 8,736 Minimum 6 5 5 0.04 0.0% 3.0 Average 69 705 437 37.54 48.1% 6.0 Maximum 130 1,365 634 190.00 87.6% 9.0 * Data and analysis provided by ADF&G. a Arrived late afternoon 5/15 with approximately 20% of Illiamna Lake still covered in ice. Setup sonar equipment 5/16 - 5/17. Deployed site 2 on 5/18 and site 1 on 5/19. b CPUE=Catch per unit effort. Total catch / elapsed fishing time. c No fishing. Ice flows in river entire day.

57

monitoringSmolt on three BayBristol rivers 2013in

Appendix A2. Daily age, weight, and length of sockeye salmon smolts subsampled from Kvichak River catches in 2013. Weight and length were measured directly; age was estimated from scales sampled from each fish. Data tables and analysis provided by the Alaska Department of Fish and Game. Age 1. Age 2. Mean Mean Mean Mean Smolt Length Std. Weight Std. Sample Length Std. Weight Std. Sample Day (mm) Dev. (g) Dev. Size (mm) Dev. (g) Dev. Size 05/19 112 4.5 12.7 1.2 100 05/20 110 4.6 12.9 1.4 100 05/21 110 4.3 12.1 1.4 100 a

05/22 05/23 a 58 a 05/24 05/25 94 3.2 7.9 1.1 6 107 5.1 11.2 1.5 94 05/26 88 5.5 5.6 0.7 3 107 5.0 11.0 1.4 97 05/27 86 5.5 5.8 0.9 33 107 4.7 10.6 1.4 67 05/28 86 4.4 5.8 0.9 61 101 5.4 9.3 1.4 39 05/29 86 6.3 6.8 1.4 31 103 4.3 11.4 1.4 69 05/30 88 4.5 6.2 0.8 49 104 4.4 10.3 1.3 51 05/31 87 4.2 6.2 0.8 47 103 6.3 10.1 1.8 53 06/01 86 4.0 5.9 0.8 50 105 5.7 10.6 1.8 50 06/02 89 5.1 6.8 1.1 22 103 6.0 10.0 1.4 78 06/03 85 4.0 5.8 0.8 79 99 8.5 8.9 2.1 21

monitoringSmolt on three BayBristol rivers 2013in

Appendix A2. Continued. Age 1. Age 2. Mean Mean Mean Mean Smolt Length Std. Weight Std. Sample Length Std. Weight Std. Sample Day (mm) Dev. (g) Dev. Size (mm) Dev. (g) Dev. Size 06/04 82 4.8 5.5 0.8 88 99 5.3 9.4 1.7 12 06/05 84 4.9 5.5 0.8 72 101 8.7 9.1 2.0 28 06/06 82 4.1 5.4 0.8 84 102 6.8 9.7 2.2 16 06/07 85 3.8 5.6 0.7 75 103 6.1 10.0 1.9 25 06/08 85 4.6 5.4 0.8 56 108 7.7 10.5 2.5 6 06/09 84 3.0 5.4 0.5 4 99 8.9 1 06/10 84 3.4 4.9 0.5 5 104 2.3 9.1 0.7 3

Total 765 1,010

59 Mean 85 5.8 106 11.1

* Data and analysis provided by ADF&G. a No fishing. Ice flows in river entire day.

Smolt monitoring on three Bristol Bay rivers in 2013

Appendix A3. Estimated age and weight of sockeye salmon smolts measured for body length (only) on the Kvichak River in 2013. Age and weight were estimated from body length, using relationships determined from fish shown in Appendix A2. Data tables and analysis provided by the Alaska Department of Fish and Game. Age 1 a Age 2 a Mean Estimated Mean Estimated Smolt Length Std. Weight Sample Length Std. Weight Sample Date (mm) Dev. (g) Size (mm) Dev. (g) Size 5/19 113 5.1 12.8 135 5/20 113 4.3 12.8 152 5/21 90 6.6 6.5 4 112 5.1 12.7 506 5/22 b 5/23 b 5/24 b 5/25 88 5.7 6.2 45 106 5.3 10.9 462 5/26 89 3.5 6.5 25 108 4.7 11.4 481 5/27 88 4.0 6.2 126 106 5.0 10.8 394 5/28 86 4.3 5.9 333 102 4.6 10.0 188 5/29 85 5.0 5.8 161 104 5.0 10.4 274 5/30 86 4.3 5.8 272 105 5.1 10.7 259 5/31 87 3.7 6.0 304 104 5.5 10.5 195 6/01 86 3.6 6.0 335 104 5.6 10.5 164 6/02 88 4.2 6.2 61 104 5.3 10.5 136 6/03 85 4.1 5.7 414 101 4.8 9.7 87 6/04 84 4.1 5.5 437 106 7.2 11.0 62 6/05 83 4.5 5.5 76 105 6.5 10.6 23 6/06 83 4.8 5.4 408 103 4.8 10.1 95 6/07 85 4.6 5.8 59 105 6.6 10.7 29 6/08 85 4.4 5.7 36 106 5.3 10.9 14 Total 3,096 3,656 Mean 85 5.8 107 11.1 * Data and analysis provided by ADF&G. a Length-weight parameters by age group and discriminating length used to separate ages from May 19 to June 10 were: Age 1. a = 1.74260E-04 b = 2.34111 r2 = 0.727 n = 765 Age 2. a = 6.83656E-05 b = 2.56798 r2 = 0.797 n = 1010 Discriminating Length = 95.0 mm b No fishing. Ice flows in river entire day.

60

monitoringSmolt on three BayBristol rivers 2013in

Appendix A4. Annual estimated age composition of sockeye salmon smolts emigrating from the Kvichak River. Age, length, and weight estimates include smolts measured directly (e.g., Appendix A2) and estimated (e.g., Appendix A3). Data tables and analysis provided by the Alaska Department of Fish and Game. Age 1. Age 2. Age 3. Percent Mean Mean Percent Mean Mean Percent Mean Mean Year of Brood of Total Length Weight Brood of Total Length Weight Brood of Total Length Weight Total Migration Year Estimate (mm) (g) Year Estimate (mm) (g) Year Estimate (mm) (g) Estimate 1955 1953 7 89 _ 1952 93 _ _ 1951 0 _ _ 260,068 1956 1954 39 92 _ 1953 61 116 _ 1952 0 _ _ 77,660 1957 1955 72 96 7.3 1954 28 120 14.4 1953 0 _ _ 30,907 1958 1956 98 84 4.6 1955 2 114 _ 1954 0 _ _ 3,333,953 1959 1957 3 80 _ 1956 97 99 7.6 1955 0 _ _ 2,863,876

1960 1958 10 91 6.3 1957 90 108 10.3 1956 0 _ _ 614,003 61 1961 1959 72 92 6.8 1958 28 117 13.1 1957 0 _ _ 36,164

1962 1960 94 82 4.3 1959 6 110 9.9 1958 0 _ _ 1,203,000 1963 1961 3 83 4.8 1960 97 98 7.5 1959 0 _ _ 4,229,431 1964 1962 22 87 5.2 1961 78 108 9.8 1960 0 _ _ 2,061,586 1965 1963 4 90 6.8 1962 96 109 11.3 1961 0 _ _ 1,812,555 1966 1964 92 94 7.4 1963 8 114 12.6 1962 0 _ _ 275,761 1967 1965 93 86 5.9 1964 7 118 14.2 1963 0 _ _ 3,088,742 1968 1966 11 88 5.5 1965 89 104 9.2 1964 0 _ _ 6,123,683 1969 1967 52 92 5.7 1966 48 109 10.6 1965 0 _ _ 1,135,344 1970 1968 38 91 6.0 1967 62 110 11.0 1966 0 _ _ 483,638 1971 1969 93 90 5.8 1968 7 111 11.1 1967 0 _ _ 91,682,813 1972 1970 1 80 4.2 1969 99 106 10.0 1968 0 _ _ 54,623,559 1973 1971 3 86 5.1 1970 97 97 8.3 1969 0 _ _ 196,966,331 1974 1972 9 96 8.3 1971 79 111 13.1 1970 12 124 17.5 27,082,626 1975 1973 63 98 8.4 1972 37 122 16.4 1971 0 _ _ 15,632,531

monitoringSmolt on three BayBristol rivers 2013in

Appendix A4. Continued. Age 1. Age 2. Age 3. Percent Mean Mean Percent Mean Mean Percent Mean Mean Year of Brood of Total Length Weight Brood of Total Length Weight Brood of Total Length Weight Total Migration Year Estimate (mm) (g) Year Estimate (mm) (g) Year Estimate (mm) (g) Estimate 1976 1974 97 88 5.8 1973 3 121 14.2 1972 0 _ _ 111,388,180 1977 1975 38 86 5.5 1974 62 106 10.1 1973 0 _ _ 192,578,099 1978 1976 12 88 6.0 1975 88 97 7.8 1974 0 _ _ 245,591,014 1979 1977 51 90 6.0 1976 49 109 10.3 1975 0 _ _ 55,181,540 1980 1978 94 88 5.9 1977 6 110 10.7 1976 0 _ _ 192,853,007 1981 1979 89 85 5.4 1978 11 108 10.2 1977 0 _ _ 252,222,769 1982 1980 58 84 5.1 1979 39 103 9.1 1978 0 _ _ 239,721,729 1983 1981 8 80 4.9 1980 92 98 8.5 1979 0 _ _ 82,793,899

62 1984 1982 58 90 6.8 1981 42 104 10.0 1980 0 _ _ 89,489,975

1985 1983 92 85 5.3 1982 8 102 9.2 1981 0 _ _ 25,527,851 1986 1984 61 84 5.5 1983 39 107 10.4 1982 <1 102 9.1 136,733,218 1987 1985 3 82 4.5 1984 97 96 7.0 1983 <1 97 8.5 342,686,918 1988 1986 13 86 5.6 1985 87 99 8.3 1984 <1 107 9.8 100,173,692 1989 1987 95 85 5.5 1986 5 108 10.8 1985 <1 105 9.5 153,464,216 1990 1988 53 87 6.1 1987 47 105 10.5 1986 0 _ _ 88,004,103 1991 1989 72 85 5.5 1988 28 105 9.9 1987 0 _ _ 121,454,182 1992 1990 23 84 5.6 1989 77 100 9.3 1988 0 _ _ 79,490,008 1993 1991 10 86 6.0 1990 90 97 8.2 1989 0 _ _ 226,407,888 1994 1992 64 84 5.7 1991 36 102 9.5 1990 0 _ _ 83,845,472 1995 1993 95 87 6.2 1992 5 103 9.8 1991 0 _ _ 220,892,127 1996 1994 74 89 6.5 1993 26 110 11.3 1992 0 _ _ 373,166,532 1997 1995 74 88 6.8 1994 26 105 10.6 1993 0 _ _ 363,397,663 1998 1996 65 90 6.7 1995 35 112 11.9 1994 0 _ _ 295,470,850

monitoringSmolt on three BayBristol rivers 2013in

Appendix A4. Continued. Age 1. Age 2. Age 3. Percent Mean Mean Percent Mean Mean Percent Mean Mean Year of Brood of Total Length Weight Brood of Total Length Weight Brood of Total Length Weight Total Migration Year Estimate (mm) (g) Year Estimate (mm) (g) Year Estimate (mm) (g) Estimate 1999 1997 92 86 5.8 1996 8 108 10.3 1995 0 _ _ 143,543,215 2000 1998 82 86 5.8 1997 18 103 9.5 1996 0 _ _ 130,038,649 2001 1999 71 78 4.2 1998 29 102 8.5 1997 0 _ _ 325,914,951 2002 b 2000 65 80 4.5 1999 35 94 7.2 1998 0 _ _ N/A 2003 2001 64 83 5.2 2000 36 109 9.0 1999 0 _ _ N/A 2004 2002 69 90 6.2 2001 31 106 10.1 2000 0 _ _ N/A 2005 2003 100 88 5.9 2002 <1 112 11 2001 0 _ _ N/A 2006 2004 100 81 4.3 2003 <1 110 8.8 2002 0 _ _ N/A

63 2007 2005 75 81 4.5 2004 25 99 7.9 2003 0 _ _ N/A

2008 2006 74 82 4.7 2005 26 99 7.7 2004 0 _ _ N/A 2009 2007 79 84 5.5 2006 21 107 10.4 2005 0 _ _ N/A 2010 2008 61 84 4.9 2007 39 104 9.3 2006 0 _ _ N/A 2011 2009 69 86 5.6 2008 31 108 11.0 2007 0 _ _ N/A 2012 2010 76 84 5.2 2009 24 108 11.0 2008 0 _ _ N/A 2013 2011 45 86 5.9 2010 55 106 10.8 2009 0 _ _ N/A Mean 1955-2012 56 87 5.7 45 107 10.2 107 10.9 Mean 2003-2012 77 84 5.2 29 106 9.6 * Data and analysis provided by ADF&G. a Estimates of smolt numbers from 1955-1970 and 2005-2007 based on fyke net catches; estimates of smolt numbers from 1971 to 2004 based on hydroacoustic techniques. b Ice flows in 2002 prevented us from deploying the sonar until June 5, about 2 weeks later than normal. It is unknown what component of the run we counted/sampled from June 5 - 14.

monitoringSmolt on three BayBristol rivers 2013in

Appendix A5. Daily age, weight, and length of sockeye salmon smolts subsampled from Ugashik River catches in 2013. Weight and length were measured directly; age was estimated from scales sampled from each fish. Data tables and analysis provided by the Alaska Department of Fish and Game. Age 1. Age 2. Mean Mean Mean Mean Smolt Length Std. Weight Std. Sample Length Std. Weight Std. Sample Day (mm) Dev. (g) Dev. Size (mm) Dev. (g) Dev. Size 05/21 99 6.7 8.2 1.3 9 122 1.0 15.1 1.5 3 05/23 99 2.6 8.4 0.5 10 121 6.2 15.0 2.6 9 05/25 99 4.5 8.4 1.0 75 114 5.3 12.8 1.8 25 05/27 102 4.2 9.6 1.0 8 123 5.1 15.6 1.9 36 05/29 98 5.0 8.4 1.0 25 121 5.3 15.6 1.9 59

64 05/31 94 8.6 7.3 1.9 37 117 8.5 14.0 3.6 15 06/02 95 6.4 7.6 1.4 90 123 16.9 17.9 8.8 10 06/04 96 6.1 7.9 1.5 44 119 12.0 15.3 4.8 9 06/06 89 11.9 6.5 1.4 37 138 16.1 27.0 11.2 9 06/09 99 6.1 8.6 1.5 34 121 12.4 16.7 5.6 16 06/10 97 3.8 8.1 1.1 19 124 10.4 17.1 5.1 11 Total 388 202 Mean 96 7.3 121 15.9 * Data and analysis provided by ADF&G.

Smolt monitoring on three Bristol Bay rivers in 2013

Appendix A6. Estimated age and weight of sockeye salmon smolts measured for body length (only) on the Ugashik River in 2013. Age and weight were estimated from body length, using relationships determined from fish shown in Appendix A2. Data tables and analysis provided by the Alaska Department of Fish and Game. Age 1 a Age 2 a Mean Estimated Mean Estimated Smolt Length Std. Weight Sample Length Std. Weight Sample Date (mm) Dev. (g) Size (mm) Dev. (g) Size 5/25 95 5.4 7.7 13 6/02 86 13.0 6.5 22 6/09 92 7.7 7.2 16 113 6.6 12.5 4 Total 51 4 Mean 90 7.0 113 12.5 * Data and analysis provided by ADF&G. a Length-weight parameters by age group and discriminating length used to separate ages from May 25 to June 9 were: Age 1. a = 3.32E-03 b = 1.69963 r2 = 0.55589 n = 388 Age 2. a = 3.82E-06 b = 3.17156 r2 = 0.99266 n = 202 Discriminating Length = 108.0 mm

65

Smolt monitoringSmolt on three BayBristol rivers 2013in

Appendix A7. Annual estimated age composition of sockeye salmon smolts emigrating from the Ugashik River. Age, length, and weight estimates include smolts measured directly (e.g., Appendix A5) and estimated (e.g., Appendix A6). Data tables and analysis provided by the Alaska Department of Fish and Game. Age 1. Age 2. Age 3. Percent Mean Mean Percent Mean Mean Percent Mean Mean Year of Brood of Total Length Weight Brood of Total Length Weight Brood of Total Length Weight Total Migration Year Estimate (mm) (g) Year Estimate (mm) (g) Year Estimate (mm) (g) Estimate a 1958 1956 - 93 6.4 1955 - 112 11.7 1954 - - - - 1959 1957 - 90 6.1 1956 - 120 13.5 1955 - - - - 1960 1958 - 90 6.6 1957 - 104 11.0 1956 - - - - 1961 1959 - 90 6.7 1958 - 112 12.2 1957 - - - - 1962 1960 - 88 6.1 1959 - 112 12.3 1958 - - - -

1963 1961 - 90 6.1 1960 - 104 9.6 1959 - - - - 1964 1962 - 92 6.9 1961 - 118 12.7 1960 - - - - 66

1965 1963 - 94 6.9 1962 - 114 12.5 1961 - - - - 1967 1965 - 88 6.0 1964 - 113 12.2 1963 - - - - 1968 1966 - 93 6.5 1965 - 113 10.7 1964 - - - - 1969 1967 - 97 7.5 1966 - 121 14.5 1965 - - - - 1970 1968 - 97 7.7 1967 - 125 15.9 1966 - - - - 1972 1970 - 81 5.0 1969 - 112 11.2 1968 - 129 14.3 - 1973 1971 - 93 7.2 1970 - 113 11.9 1969 - 132 20.1 - 1974 1972 - 94 7.4 1971 - 119 13.6 1970 - - - - 1975 1973 - 96 7.2 1972 - 116 13.0 1971 - 125 16.7 - 1982 1980 - 88 6.3 1979 - 113 13.0 1978 - 138 22.5 - 1983 1981 71 89 7.6 1980 29 111 13.2 1979 - - - 44,033,811 1984 1982 48 87 6.8 1981 52 102 10.3 1980 0 103 11.7 158,174,626 1985 1983 37 94 8.3 1982 63 107 11.8 1981 - - - 34,101,390 1986 1984 71 87 5.8 1983 29 114 10.9 1982 - - - 53,076,253

Smolt monitoringSmolt on three BayBristol rivers 2013in

Appendix A7. Continued. Age 1. Age 2. Age 3. Percent Mean Mean Percent Mean Mean Percent Mean Mean Year of Brood of Total Length Weight Brood of Total Length Weight Brood of Total Length Weight Total Migration Year Estimate (mm) (g) Year Estimate (mm) (g) Year Estimate (mm) (g) Estimate a 1987 1985 20 94 7.9 1984 80 107 11.1 1983 0 138 24.1 26,947,225 1988 1986 85 87 5.7 1985 15 109 10.8 1984 0 128 15.6 215,968,015 1989 1987 74 90 6.5 1986 26 108 10.7 1985 - - - 126,298,122 1990 1988 28 90 6.7 1987 72 108 11.8 1986 - - - 53,627,347 1991 1989 35 92 7.7 1988 65 107 11.6 1987 - - - 73,769,877 1992 b 1990 - - - 1989 - - - 1988 - - - - 1993 1991 83 92 8.0 1990 17 109 12.5 1989 - - - 70,747,074 1994 1992 81 89 6.7 1991 19 109 11.2 1990 - - - 30,030,624

67 1995 1993 31 93 7.8 1992 69 106 11.1 1991 - - - 22,234,137

1996 1994 44 101 9.9 1993 56 114 13.5 1992 - - - 2,576,812 1997 1995 92 92 7.9 1994 8 109 12.1 1993 - - - 15,519,783 1998 1996 82 91 6.4 1995 18 110 11.1 1994 - - - 12,624,441 1999 1997 99 91 6.8 1996 1 125 17.5 1995 - - - 10,631,631 2000 1998 18 95 8.4 1997 82 112 12.5 1996 - - - 10,880,559 2001 1999 82 92 7.3 1998 18 108 11.5 1997 - - - 35,123,888 2002 2000 81 91 7.7 1999 19 110 12.7 1998 - - - 47,627,642 2003 b 2001 2000 1999 - - - - 2004 b 2002 2001 2000 - - - - 2005 b 2003 2002 2001 - - - - 2006 b 2004 2003 2002 - - - - b 2007 2005 2004 2003 - - - -

Smolt monitoringSmolt on three BayBristol rivers 2013in

Appendix A7. Continued. Age 1. Age 2. Age 3. Percent Mean Mean Percent Mean Mean Percent Mean Mean Year of Brood of Total Length Weight Brood of Total Length Weight Brood of Total Length Weight Total Migration Year Estimate (mm) (g) Year Estimate (mm) (g) Year Estimate (mm) (g) Estimate a 2008 b 2006 2005 2004 - - - - 2009 b 2007 2006 2005 - - - - 2010 2008 59 97 7.8 2007 41 117 13.2 2006 - - - - 2011 c 2009 2008 2007 - - - - 2012 2010 70 90 6.4 2009 30 108 10.8 2008 - - - - 2013 2011 68 97 8.0 2010 32 122 16.5 2009 - - - - Mean 62 92 7.0 38 112 12.3 128 17.9 * Data and analysis provided by ADF&G. 68 a No estimates of smolt numbers from 1958-1982 fyke net catches; estimates of smolt numbers from 1983-1991 and 1993-2002 based on hydroacoustic techniques. b Project not operated in 1992 and 2003-2009. No smolt data collected. c Project only operated from May 19 through May 24 before high water washed out the project. No data collected.

Smolt monitoringSmolt on three BayBristol rivers 2013in

Appendix A8. Daily age, weight, and length of sockeye salmon smolts subsampled from Ugashik River catches in 2012. Weight and length were measured directly; age was estimated from scales sampled from each fish. Data tables and analysis provided by the Alaska Department of Fish and Game. Age 1. Age 2. Mean Mean Mean Mean Smolt Length Std. Weight Std. Sample Length Std. Weight Std. Sample Day (mm) Dev. (g) Dev. Size (mm) Dev. (g) Dev. Size 05/27 75 7.6 3.8 1.1 19 93 5.3 7.2 1.1 22 05/28 84 9.8 5.4 1.8 71 108 11.6 11.3 3.3 29 05/30 97 5.4 8.2 1.0 48 109 10.4 11.2 3.0 52 06/01 96 5.5 7.8 1.3 48 118 8.4 13.6 2.5 52

06/03 94 6.1 7.7 1.3 25 115 6.8 12.3 1.9 9 06/05 92 5.2 6.5 1.0 96 115 8.3 12.1 2.8 4 69

06/08 93 5.3 6.9 1.2 89 117 6.9 12.6 2.4 9 06/09 92 5.4 6.5 1.2 96 125 22.0 18.5 10.5 4 06/11 91 5.7 6.5 1.2 96 110 3.1 10.2 1.7 4 Total 588 185 Mean 91 6.7 110 11.7 * Data and analysis provided by ADF&G.

Smolt monitoring on three Bristol Bay rivers in 2013

Appendix A9. Estimated age and weight of sockeye salmon smolts measured for body length (only) on the Ugashik River in 2012. Age and weight were estimated from body length, using relationships determined from fish shown in Appendix A8. Data tables and analysis provided by the Alaska Department of Fish and Game. Age 1 a Age 2 a Mean Estimated Mean Estimated Smolt Length Std. Weight Sample Length Std. Weight Sample Date (mm) Dev. (g) Size (mm) Dev. (g) Size 5/28 89 7.1 6.2 187 110 9.0 11.4 99 5/30 93 5.4 6.9 243 106 7.4 10.4 254 6/01 90 5.2 6.2 113 112 9.1 11.9 94 6/05 90 4.4 6.4 354 104 6.4 9.8 47 6/09 89 5.6 6.1 103 103 3.1 9.4 16 6/11 91 4.1 6.4 85 104 6.4 9.8 24 Total 1,085 534 Mean 91 6.4 108 8.3 10.7 * Data and analysis provided by ADF&G. a Length-weight parameters by age group and discriminating length used to separate ages from May 27 to June 11 were: Age 1. a = 2.70E-05 b = 2.74557 r2 = 0.78932 n = 525 Age 2. a = 4.34E-05 b = 2.65145 r2 = 0.88282 n = 157 Discriminating Length = 98.0 mm

70

monitoringSmolt on three BayBristol rivers 2013in

Appendix B1. Daily climate and hydrological observations made at 0800 and 2000 hours near the Kvichak River sonar site, 2013. Kvichak River environmental data 2013 Cloud Wind dir./vel. Air temp. Water temp. Water covera Precipitationb (km/h) (°C) (°C) colorc Date 8:00 20:00 8:00 20:00 8:00 20:00 8:00 20:00 8:00 20:00 8:00 20:00 19-May 4 n 0 n n NE 5 n 1.2 n 3.3 1 n 20-May 4 n 0 n N 3 S 5 2.7 6.6 2.9 3.3 1 n 21-May 4 n A n NW 0 SE 0 4.6 5.3 3.1 3.5 1 n 22-May 1 n 0 n S 2 SE 3 8.8 5.4 3.5 3.7 1 n 23-May 1 n 0 n S 5 SE 0 9.8 5.7 3.4 4.0 1 n

24-May 1 n 0 n SE 0 S 2 8.7 8.2 4.0 5.0 1 n 25-May 1 1 0 n N 0 S 2 10.4 12.6 4.8 6.9 1 n 71 26-May 1 2 0 0 SE 5 S 16 10.7 8.1 5.7 6.2 5 5

27-May 3 2 Tr 0 S 13 SE 10 11.4 10.1 5.8 7.5 2 2 28-May 1 1 0 0 S 13 S 8 10.9 9.3 7.5 9.3 1 1 29-May 1 n 0 n SE 8 N 0 11.2 23.3 8.0 8.7 1 n 30-May 1 n 0 n N 6 NW 3 15.0 15.2 8.0 9.6 1 1 31-May 3 n 0 n SE 2 E 3 13.8 15.3 11.0 10.5 1 1 1-Jun 4 4 A B W 0 SE 2 11.8 10.9 10.0 9.5 1 1 2-Jun 3 3 0 n N 0 SE 0 14.0 8.8 9.6 10.1 1 1 3-Jun 3 n 0 n S 6 SE 19 11.9 8.8 9.0 9.3 1 n 4-Jun 4 n 0 n S 13 S 11 13.4 10.2 9.4 9.9 1 n 5-Jun 3 n B n SE 2 S 3 12.3 9.7 9.9 11.0 1 n 6-Jun 1 n 0 n S 3 S 16 12.5 10.2 9.0 11.6 1 n 7-Jun 1 n 0 n S 11 S 0 12.6 13.6 10.3 10.9 1 n 8-Jun 1 n 0 n N 6 N 2 13.6 14.8 10.4 11.6 1 n

monitoringSmolt on three BayBristol rivers 2013in

Appendix B1. Continued. Kvichak River environmental data 2013 Cloud Wind dir./vel. Air temp. Water temp. Water covera Precipitationb (km/h) (°C) (°C) colorc Date 8:00 20:00 8:00 20:00 8:00 20:00 8:00 20:00 8:00 20:00 8:00 20:00 9-Jun 3 n 0 n N 5 N 10 16.3 18.4 9.3 9.6 1 n 10-Jun 4 n 0 n N 14 N 21 12.3 11.1 8.5 9.9 1 n 11-Jun 5 n Tr n W 11 N 16 10.9 13.8 8.8 9.8 1 n 12-Jun n n n n N 13 N 10 10.6 19.0 9.2 10.4 n n 13-Jun n n n n NE 0 S 8 14.9 18.7 10.1 n n n 14-Jun n n n n NW 0 n 13.0 n n n n n a b c

Sky codes Precipitation codes Water color codes n - No observation n - No observation n - No observation 72 1 - Clear sky, < 1/10 cover 0 - No precipitation 1 - Clear

2 - Cloud cover , 1/2 sky Tr - Trace 2 - Light brown 3 - Cloud cover > 1/2 sky A - Intermittent rain 3 - Brown 4 - Completely overcast B - Continuous rain 4 - Dark brown 5 - Fog or thick haze C - Snow 5 - Murky or glacial D - Snow and rain E - Hail F - Thunderstorm

monitoringSmolt on three BayBristol rivers 2013in

Appendix B2. Daily climate and hydrological observations made at 0800 and 2000 hours near the Ugashik River sonar site, 2013. Ugashik River environmental data 2013 Cloud Wind dir./vel. Air temp. Water temp. covera Precipitationb (km/h) (°C) (°C) Water colorc Date 8:00 20:00 8:00 20:00 8:00 20:00 8:00 20:00 8:00 20:00 8:00 20:00 18-May n n n n N 10 NW 16 -2.8 0.8 2.1 2.2 1 1 19-May 3 3 0 0 W 3 E 2 -2.0 11.2 2.0 3.0 1 1 20-May 3 2 0 0 E 0 S 8 1.6 8.6 2.5 3.1 1 1 21-May 3 3 Tr Tr SE 8 S 10 4.7 5.3 3.1 4.2 1 1 22-May 3 3 0 Tr SE 8 SE 6 5.7 6.3 4.1 3.2 1 1

23-May 3 2 Tr 0 SE 0 N 8 3.9 8.6 3.5 3.5 1 1 24-May 2 2 0 0 NW 2 SE 5 5.6 10.4 3.8 4.6 1 1 73 25-May 3 3 Tr 0 SE 23 E 23 5.6 6.5 4.9 5.8 5 5

26-May 3 3 0 Tr E 8 E 5 6.1 10.9 4.8 3.9 1 1 27-May 3 3 A Tr E 5 E 3 7.0 8.7 4.6 4.3 5 5 28-May 1 1 0 0 E 5 NW 10 4.1 16.2 3.6 4.3 1 1 29-May 1 1 0 Tr W 5 NW 8 7.6 14.7 4.7 5.1 1 1 30-May 2 2 0 Tr E 8 E 8 5.5 11.6 5.1 5.9 5 5 31-May 2 2 0 0 E 0 SE 0 7.1 11.2 6.1 5.0 1 1 1-Jun 3 2 0 0 S 0 SE 13 7.3 9.8 4.5 5.8 1 1 2-Jun 3 3 0 0 SE 11 E 13 7.1 7.9 6.8 6.0 1 1 3-Jun 3 3 0 0 E 5 SE 6 8.2 9.4 5.3 5.6 1 1 4-Jun 3 2 0 0 E 2 S 5 7.8 11.2 4.9 5.3 1 1 5-Jun 2 3 0 0 SE 11 E 11 7.1 13.1 5.4 7.3 1 1 6-Jun 2 2 A 0 SE 0 SE 6 7.8 11.4 5.6 6.0 1 1 7-Jun 2 2 0 0 W 11 W 5 6.6 11.9 6.3 6.2 1 1 8-Jun 3 3 Tr Tr W 10 W 16 6.8 9.0 5.5 6.4 5 5

monitoringSmolt on three BayBristol rivers 2013in

Appendix B2. Continued. Ugashik River environmental data 2013 Cloud Wind dir./vel. Air temp. Water temp. covera Precipitationb (km/h) (°C) (°C) Water colorc Date 8:00 20:00 8:00 20:00 8:00 20:00 8:00 20:00 8:00 20:00 8:00 20:00 9-Jun 3 3 0 Tr SW 13 W 11 4.3 6.3 4.9 4.7 5 5 10-Jun 2 2 Tr Tr W 10 W 10 4.1 5.3 4.6 4.7 5 5 11-Jun 2 3 Tr 0 SW 6 SE 2 2.7 15.2 4.5 5.0 1 1 12-Jun 2 1 0 0 E 3 SE 6 7.4 12.6 5.1 n 1 1 a Sky codes b Precipitation codes c Water color codes n - No observation n - No observation n - No observation 1 - Clear sky, < 1/10 cover 0 - No precipitation 1 - Clear

2 - Cloud cover , 1/2 sky Tr - Trace 2 - Light brown

74 3 - Cloud cover > 1/2 sky A - Intermittent rain 3 - Brown 4 - Completely overcast B - Continuous rain 4 - Dark brown 5 - Fog or thick haze C - Snow 5 - Murky or glacial D - Snow and rain E - Hail F - Thunderstorm

monitoringSmolt on three BayBristol rivers 2013in

Appendix B3. Daily climate and hydrological observations made at 0800 and 2000 hours near the Egegik River sonar site, 2013. Egegik River environmental data 2013 Cloud Wind dir./vel. Air temp. Water temp. Water covera Precipitationb (km/h) (°C) (°C) colorc Date 8:00 20:00 8:00 20:00 8:00 20:00 8:00 20:00 8:00 20:00 8:00 20:00 22-May 4 n Tr n SE 24 n 3.3 4.1 4.0 n 1 n 23-May 4 4 0 0 SE 0.9 NE 1 4.5 9.0 4.0 4.0 1 1 24-May 3 1 0 0 0 SE 12 9.5 11.0 0.0 0.0 1 1 25-May 2 1 0 0 E 17 E 28 5.5 9.0 0.0 4.0 1 1 26-May 1 n 0 n E 8 n 6.0 n 2.0 n 2 n

27-May 3 3 0 A 0 E 9 9.5 8.0 3.0 3.5 1 1 28-May 1 1 0 0 E 4 W 9 9.0 13.0 1.0 5.5 1 1 75 29-May 1 1 0 0 W 8 W 4 12.5 15.0 3.5 6.0 1 1

30-May 1 1 0 0 SE 22 S 11 5.0 8.0 1.0 1.0 1 1 31-May 3 3 Tr 0 W 1 0 8.5 11.0 2.0 4.0 1 1 1-Jun 4 3 0 0 E 4 S 11 8.0 9.0 4.0 6.0 1 1 2-Jun 4 4 A A SE 11 S 8 4.5 6.5 2.0 4.5 1 1 3-Jun 4 4 A Tr 0 W 4 10.0 10.0 3.5 4.5 1 1 4-Jun 3 2 Tr 0 S 1 W 4 7.0 13.0 5.0 7.0 1 1 5-Jun 1 1 0 0 E 6 S 11 12.0 8.0 5.5 8.0 1 1 6-Jun 1 1 0 0 N 6 S 3 10.0 13.0 5.0 8.0 1 1 7-Jun 1 4 0 B NW 4 W 4 15.0 13.0 7.0 6.0 1 1 8-Jun 1 1 0 0 NE 4 W 14 16.0 11.0 8.0 9.0 1 1

monitoringSmolt on three BayBristol rivers 2013in

Appendix B3. Continued. Egegik River environmental data 2013 Cloud Wind dir./vel. Air temp. Water temp. Water covera Precipitationb (km/h) (°C) (°C) colorc Date 8:00 20:00 8:00 20:00 8:00 20:00 8:00 20:00 8:00 20:00 8:00 20:00 9-Jun 4 4 0 0 W 14 W 11 7.0 8.0 7.0 7.0 1 1 10-Jun 4 4 B 0 W 11 W 8 6.0 7.0 6.0 7.0 1 1 11-Jun 4 1 Tr 0 W 8 W 3 6.0 13.0 5.0 9.0 1 1 12-Jun 1 1 0 0 E 20 E 4 9.0 13.0 7.0 9.0 1 1 13-Jun 1 1 0 0 S 9 W 3 15.0 17.0 9.5 13.0 1 1 a Sky codes b Precipitation codes c Water color codes

n - No observation n - No observation n - No observation 1 - Clear sky, < 1/10 cover 0 - No precipitation 1 - Clear 76 2 - Cloud cover , 1/2 sky Tr - Trace 2 - Light brown

3 - Cloud cover > 1/2 sky A - Intermittent rain 3 - Brown 4 - Completely overcast B - Continuous rain 4 - Dark brown 5 - Fog or thick haze C - Snow 5 - Murky or glacial D - Snow and rain E - Hail F - Thunderstorm