Zachary S. Larson. May Final Report. Prepared for the County of Del Norte by,

USING DIDSON (DUAL FREQUENCY IDENTIFICATION SONAR) TO MONITOR CHINOOK AND STEELHEAD ESCAPEMENT IN THE SMITH RIVER, DEL NORTE COUNTY, CALIFORNIA, 2014-2015 Final Report Prepared for the County of Del Norte by, Zachary S. Larson May 2015 THIS PROJECT WAS FUNDED BY THE CALIFORNIA DEPARTMENT OF FISH AND WILDLIFE STEELHEAD CATCH REPORT RESTORATION CARD PROGRAM GRANT NUMBER P1381004 ZACK LARSON & ASSOCIATES P.O. BOX 1400 CRESCENT CITY, CA 95531

TABLE OF CONTENTS LIST OF TABLES... 2 LIST OF FIGURES... 2 ABSTRACT... 3 INTRODUCTION... 4 STUDY AREA... 4 MATERIALS AND METHODS... 8 DIDSON Fish Counts... 10 Systematic Hourly Sampling Analysis... 11 Rowdy Creek Fish Hatchery Weir Count Data... 11 Length Frequency Analysis... 11 Escapement Estimates... 12 RESULTS... 12 DIDSON Fish Counts... 13 Systematic Hourly Sampling Rate Analysis... 13 Rowdy Creek Fish Hatchery Weir Counts... 13 Length Frequency Analysis... 18 Escapement Estimates... 19 DISCUSSION... 22 CONCLUSION... 23 ACKNOWLEDGEMENTS... 24 LITERATURE CITED... 24 APPENDIX... 26 1

LIST OF TABLES Table 1. Historic salmon and steelhead run-size estimates for northern California rivers (USFWS, 1960).... 7 Table 2. Adult salmon and steelhead migration times in the Smith River.... 7 Table 3. The percent error by month associated with 20 minute per hour sampling rate..... 17 Table 4. The percent error by week associated with 20 minute per hour sampling rate..... 17 Table 5. The percent error by Run associated with 20 minute per hour sampling rate.... 17 LIST OF FIGURES Figure 1. Location of the Smith River Basin, Del Norte County, California.... 5 Figure 2. Location of the Smith River DIDSON Fish Counting Station at river mile 6 near the town of Smith River, Del Norte County, California.... 6 Figure 3. Aerial view of the Smith River DIDSON Fish Counting Station at the public boat launch facility located at river mile 6.... 8 Figure 4. Left bank DIDSON deployment, power and communication system... 9 Figure 5. Image of computer output showing split screen function used for measuring fish.... 11 Figure 6. Monthly means of estimated angler caught Chinook salmon and steelhead in the Smith River (CDFG 1997-2006)... 12 Figure 7. Daily upstream and downstream fish counts at the Smith River DIDSON Fish Counting Station vs river discharge, 2014-2015.... 14 Figure 8. Daily fish counts at the Smith River DIDSON Station compared to U.S.G.S stream gauge data.... 15 Figure 9. Counts of Chinook and steelhead from the Rowdy Creek weir compared to counts of adult fish made at the Smith River DIDSON Fish Counting Station, 2014-2015..... 16 Figure 10. Weekly average fish lengths and error measured during the 2014-2015 season..... 18 Figure 11. Weekly average fish lengths and error measured during the 2014-2015 season..... 19 Figure 12. Length frequency histograms of fish measured with DIDSON at the Smith River DIDSON Fish Counting Station 2014-2015..... 20 Figure 13. Estimates of escapement of Chinook and steelhead in the Smith River using angler data to apportion DIDSON fish counts, 2010-2015... 21 Figure 14. Estimates of escapement of Chinook and steelhead in the Smith River using a December 15 cutoff date to apportion DIDSON fish counts, 2010-2015.... 21 2

USING DIDSON (DUAL FREQUENCY IDENTIFICATION SONAR) TO MONITOR CHINOOK AND STEELHEAD ESCAPEMENT IN THE SMITH RIVER, DEL NORTE COUNTY, CALIFORNIA, 2014-2015 ABSTRACT Dual Frequency Identification Sonar (DIDSON) was used to count adult anadromous fish migrating through the lower Smith River, Del Norte County, California during the 2014-2015 spawning season. Two long-range DIDSON units operated continuously at river mile 6 for 180 days from 23 Sep 2014 to 31 Mar 2015. The total recording efficiency for the DIDSON units operating 24 hours per day was 94%. A total of 44,009 adult fish were counted passing upstream and 18,326 fish were counted passing downstream. Daily DIDSON fish counts were compared to hatchery weir counts at Rowdy Creek, a Smith River tributary located 2 miles downstream from the DIDSON station. Peak counts of Chinook at the weir coincided with peak DIDSON counts during October and November. Peak steelhead counts at the weir also coincided with peak DIDSON counts from late December through March. The ratio of Chinook to steelhead at the weir approached 1:1 in early December during a period of low fish counts at both the weir and DIDSON facilities. The weir data supports 15 December split between Chinook and steelhead runs. A systematic hourly sampling rate of 20 minutes per hour was applied to the entire data set and the percent error for the season ranged from 2.8 to 6.8 percent. The percent error was highest during the steelhead kelt migration in March, followed by Chinook migration in October and November. The percent error was lowest during the peak steelhead run time from late December through early February. DIDSON telephoto lenses were used to generate length frequency information throughout the season. Weekly means of total lengths of fish were highest during mid-november at 90 cm and lowest during the last week of March at 64 cm. Standard Deviations of weekly mean lengths were highest during October likely due to the presence of Chinook jacks. Standard Deviations of weekly mean lengths were lowest from mid-december through March suggesting a dominant steelhead presence. Two approaches were taken to apportion DIDSON counts for estimating Chinook and steelhead escapement. Estimates for the 2014-2015 season based on historical angler catch data were 14,865 Chinook and 12,875 steelhead. Using the 15 December cutoff date for dividing the Chinook and steelhead runs, escapement estimates were 15,796 Chinook and 11,944 steelhead. Based on DIDSON data and Rowdy Creek weir data from 2010-2015 annual escapement of Chinook and steelhead in the Smith River likely ranges from 15,000 to 25,000 Chinook and 10,000 to 15,000 steelhead. Coho salmon were not considered in this study as their population is suspected to be in the hundreds of individuals. Using DIDSON to monitor Chinook and steelhead escapement to the Smith River should continue in order track trends fish abundance, improve management of recreational fisheries and maintain the fisheries into the future. 3

INTRODUCTION During the 2014-2015 Chinook salmon (Oncorhynchus tshawytscha) and steelhead (Oncorhynchus mykiss irideus) spawning migrations in the Smith River, Del Norte County, California, a dual-frequency identification sonar (DIDSON) fish counting station operated to quantify escapement. DIDSON is multibeam, high-definition imaging sonar that transmits sound pulses and converts the returning echoes into digital images. When used to monitor anadromous fish escapement DIDSON is deployed from a submerged, stationary position near the water s edge to ensonify a subsection or an entire cross-section of a submerged channel. Images of fish passing through the ensonified area are recorded and observed like video and the acoustic data is analyzed with software. Fish counts and reasonably accurate length measurements can be manually derived or semi-automated with software. The advantages of using DIDSON to monitor the status and trends of fish in anadromous waters include: the ability to operate in rivers too wide or powerful for barrier weirs or too turbid for visual observations; collecting protected species data without handling fish or affecting critical habitat; monitoring anadromous fish migrations 24 hours per day; validation of spawning ground surveys, and; observations of behavioral patterns of anadromous fish and detections of rare events (e.g. sturgeon).this study follows previous DIDSON studies in the Smith River designed to develop methods for monitoring anadromous fish migrations in relatively large rivers. The objectives of this study were to: 1. Count adult anadromous fish migrating through the lower Smith River during the 2014-2015 spawning season. 2. Apply 33% systematic hourly sampling rate to the continuous data set and analyze results. 3. Compare Rowdy Creek Fish Hatchery weir data with daily DIDSON fish counts. 4. Conduct a length frequency analysis of fish measured with DIDSON. 5. Estimate escapements of Chinook and steelhead to the Smith River and compare previous studies. 6. Establish baseline anadromous fishery information for the Smith River. STUDY AREA The Smith River is the fourth largest coastal river in California and has the unique status as the Golden State s last major free flowing river that drains directly into the Pacific Ocean (Figure 1). The Smith River is famous for water clarity, its world-class salmon and steelhead fishing and its ancient redwood forest groves (SRAC 2002). The watershed includes large areas of the Six Rivers and Rogue-Siskiyou National Forests, Redwood National Park, and California State Parks. The watershed also contains the Smith River National Recreation Area and the entire river system is designated a Wild and Scenic River, a Key Watershed under the Pacific Northwest Forest Plan, and a Salmon Stronghold Basin under the North American Salmon Stronghold Partnership. The Smith River originates in the Siskiyou Mountains of the Klamath Geologic Province and enters the Pacific Ocean 4 miles south of the Oregon Border in Del Norte County, California. The total watershed area is approximately 1,862 km² (720 mi²) with 236 km² in Oregon in the North Fork sub-basin. The elevation ranges from sea level to 1,958 m (6424 ft.). The total length of stream channel is 5,000 km of which about 483 km is available for anadromous fish (McCain et al. 1995). Annual rainfall amounts range from approximately 1,778 mm (70 in.) at the coast to over 3,800 mm (150 in.) in the interior. The precipitous terrain and elevation drop creates a fast response time of stream flow to rainfall events. The Smith River has 4

the largest mean annual runoff per area of any major watershed in California (Rantz, 1969) and is prone to periodic large scale flooding, particularly during rain-on-snow events. Mean monthly flow values in the Smith River, based on USGS gauging station data, range from about 8.5 cubic meters per second (cms) or 300 cubic feet per second (cfs) in August and September, to around 230 cms or 8,000 cfs in December and January. The status of Smith River Chinook salmon, coho salmon and steelhead trout and their associated Evolutionary Significant Units (ESU s) are included in the Appendix along with a list of all anadromous fish in the Smith River. Only coho salmon are listed under the Endangered Species Act (1973) as a Threatened Species in the Southern Oregon and Northern California Coast ESU. Otherwise fall Chinook salmon and steelhead populations in the Smith River are considered healthy and continue to support popular recreational fishing opportunities. There is no evidence of a long term decline in the fall-run Chinook population, but data are limited (Moyle 2002). In 1997 the California Department of Fish and Wildlife recognized the Smith River as being California s most robust steelhead population but overall there was widespread concern that Klamath Mountains Province (KMP) steelhead populations were in general depressed from historic levels (CDFG 1997). Smith River OREGON CALIFORNIA Figure 1. Location of the Smith River Basin, Del Norte County, California. 5

In 1960 the US Fish and Wildlife Service provided a survey report, at the request of the Department of the Interior, to establish a baseline of fish and wildlife resources in northern California (USFWS 1960, 1965) (Table 1). However, estimates given here which are based on little or no data should be used only in outlining the major and critical factors of the resource (CDFW 1965). The basin-wide estimates of spawning abundance in the Smith River exclusive of catch were: Chinook salmon, 15,000 fish; coho salmon, 5,000 fish; and steelhead, 30,000 fish. Except for the Trinity River, recent estimates of steelhead run sizes are not available for any of the major runs in the KMP (CDFW 1997). Moyle (2002) described the Smith River as a relatively unaltered stream that never supported a large Chinook population. He reported annual estimates of Chinook salmon in the Smith River are generally 15,000-30,000 fish. The Smith River DIDSON Fish Counting Station is located 1.8 river miles upstream from the mouth of Rowdy Creek (RM 4), a lower Smith River tributary (Figure 2). In 1970, the governor signed Senate Bill #1047, which authorized the Kiwanis Club of Crescent City to build the first privately owned fish hatchery in California. The hatchery facilities included a barrier weir fish trap that provides one of the longest set of annual anadromous abundance data for the Smith River. In general peak counts of Chinook salmon at the hatchery facility occur during November and typically cease around 15 December (Andrew Van Scoyk, pers. comm.). Coho salmon, though very few are observed at the weir, have passed through the facility in October and again in late December. Although Rowdy Creek salmon and steelhead run times overlap, peak numbers of steelhead at the hatchery are observed in mid-january. These run times are similar to the general run-timings observed in the Smith River (Table 2). The weir data provides opportunities for species run timing comparisons and additions of weir counts to basin-wide escapement counts or estimates. Figure 2. Location of the Smith River DIDSON Fish Counting Station at river mile 6 near the town of Smith River, Del Norte County, California. 6

Table 1. Historic salmon and steelhead run-size estimates for northern California rivers (USFWS, 1960). River Basin Chinook salmon Coho salmon Steelhead Smith River 15,000 5,000 30,000 Klamath River 100,000 20,000 400,000 Redwood Creek 5,000 2,000 10,000 Mad River 5,000 2,000 6,000 Eel River 25,000 15,000 100,000 Mattole River 5,000 2,000 12,000 Table 2. Adult salmon and steelhead migration times in the Smith River. Species Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct Nov Dec Chinook salmon Coho salmon Winter steelhead 7

MATERIALS AND METHODS Site Selection Selection of the study site was based on findings from the pilot study and included its location upstream from tidal influence and downstream from major Smith River Mainstem and tributary spawning grounds (Larson 2012). Additional considerations for siting the station included appropriate channel geometry, availability of grid power, ownership, vehicular access and site security. The site included the State of California owned 3- acre parcel (APN# 105-02-002) currently managed by Del Norte County for boat launch and river access. Alexandre EcoDairy Farms, owners of land on the opposite side of the river, granted access for the study so the two DIDSON units could operate between the opposing river banks. Smith River DIDSON Station Operation The Smith River DIDSON Station includes a small building connected to grid power and internet, two Long Range DIDSON packages, two telephoto DIDSON lenses, 200-feet device communication cables, two laptop computers, point-to-point wireless bridge network components, a bank of 12-volt deep cycle batteries, and a motorized boat to maintain units. DIDSON units were deployed with H mounts and positioned facing each other from opposite sides of the river to ensonify a cross section of the river channel (Figure 3). The right bank sonar and components were hardwired into the computer network. The remotely operated left bank sonar was connected to the network with the point-to-point wireless bridge (figure 4). Figure 3. Aerial view of the Smith River DIDSON Fish Counting Station at the public boat launch facility located at river mile 6. DIDSON 1 is located on the right bank and DIDSON 2 is located on the left bank. (Photo by Justin Garwood, CDFW). 8

Right bank and left bank DIDSON units were operated simultaneously with one laptop computer while another laptop was used for data analyses. Data was recorded 24 hours per day, stored in 20 minute files and saved in weekly file folders. Data was analyzed onsite with Sound Metrics DIDSON software version 5.26. All data and analyses were saved to 4-terrabyte (TB) drives and backed-up to a second set of drives. The DIDSON units were deployed using an H-mount, made of schedule 40, 2-inch aluminum pipe with fittings that allow for easy assembly and adjustment (Figure 4). Units were attached to H-mounts using RAM double socket parts for easy sonar adjustment. We used the Microsoft Remote Desktop application for Apple Iphones to log on the operating computers and remotely view the adjustments in real-time. A laser range finder was used to measure the distance between the two units to ensure no overlap during low flows. New additions to the counting station included an internet connection installed in the building and a mobile solar power station for the Left Bank DIDSON. The system includes photovoltaic panels mounted on an aluminum frame, a wireless router and antenna, topside box, power monitor, sun tracking device and a battery box. A telephoto lens was added to the Right Bank DIDSON. Figure 4. Left bank DIDSON deployment, power and communication system. Left: The DIDSON mount is the simple H design (shown with DIDSON LR and Telephoto lens). Right: Solar panels connected to battery power and wireless communication devices. 9

DIDSON Fish Counts All recorded data was processed and analyzed for a complete census. Fish counts were organized into a database for future systematic sampling investigations. A copy of the data was created each week for data processing, quality control and archiving. DIDSON files and Excel spreadsheet files were stored in weekly file folders. Count data from a single data collection week from both banks was combined to form weekly count data Master file workbooks. Within each master workbook, count data was organized by the bank and day of collection. The total time per day recorded by bank was used for the calculation of recording efficiency or how many hours and minutes of data were actually recorded with two sonars per day. Fish were counted and recorded as Upstream Migrants or Downstream Migrants if they passed completely through the sonar beam array. Fish were not counted if they entered or exited the array without completely passing the upstream and downstream boundaries of the array. We used this method to avoid counting fish again with the other DIDSON or double counting fish swimming in and out of the sonar beam array. Only fish measuring greater than approximately 50 cm were counted as adults. Files containing more than 50 fish were counted several times and the average count was recorded. The Zoom and Measure tools in the software (also echogram function) were used to distinguish small fish from larger fish (>50cm), and fish from mammals and birds. Files were processed with the Convolved Samples Over Threshold (CSOT) function to remove empty file portions before performing fish counts. We used the following simple flux model described by Xie et al. (2002) to estimate total upstream movement of adult anadromous fish observed with DIDSON (in Cronkite et al. 2006). N = U D (1) where N = the net upstream flux, U = the upstream actively migrating fish and D = the downstream actively migrating fish. Milling fish or fish that are not actively migrating can also be accounted for in this model, provided these fish eventually move upstream through the ensonified area (Cronkite et al. 2006). Since steelhead are iteroparous, spawned-out fish or kelts, have to be removed from the downstream estimate since the model relates only to actively migrating fish, and kelts would have been included in the spawning population as upstream fish. Downstream fish were removed after 15 December to develop steelhead escapement estimates. Summed over a 24 hour period, this model produces daily escapement estimates that are compiled to estimate the total spawning population entering a river. The recording efficiency of the DIDSON station operation is the amount of time data were recorded during a day divided by the total possible time data could be recorded. Recording efficiency was calculated for each unit as follows. E = R+B T (2) Where E = Recording efficiency, R = total right bank sonar file time recorded, B = total left bank sonar file time recorded, and T = 48 hours. 10

Systematic Hourly Sampling Analysis Continuous sampling occurred throughout the season in 20 minute intervals, 24 hours per day. A sampling rate of 20 minutes per hour was applied to the data set in order to investigate the error associated with a systematic hourly sampling rate throughout the season. The first, second and third 20 minute intervals of the hour were included in the investigation. Error for the 20 minute sampling rate was calculated by week, month and season. Comparisons of the reduced sampling rate were also made between the run time of fallrun Chinook (1 Oct 14 Dec) and winter steelhead (15 Dec 31 Mar). Kelts (downstream steelhead) were not removed from this dataset. Rowdy Creek Fish Hatchery Weir Count Data The Rowdy Creek Fish Hatchery provided Chinook and steelhead data collected from their barrier weir trap located in lower Rowdy near the Dominie Creek confluence. Recent changes to the Rowdy Creek Fish Hatchery Management Plan required a 50 percent reduction of barrier weir operation. The weir trap is now operated every other day throughout the Chinook and steelhead run time. Trap data were organized to correspond with daily fish count data from the Smith River DIDSON Fish Counting Station. Length Frequency Analysis Daily count sheets were used to randomly select CSOT files containing fish over a 24 hour period for every day of the season. The selected CSOT files from an entire week were merged into one weekly file that could be analyzed in one DIDSON (.ddf) data file. The weekly CSOT file folders correspond to dates of Weekly Count folders. Echograms were created for each week of merged CSOT files and procedures outlined in the DIDSON manual were used to manually measure the total length (TL) of each fish (Figure 5). Every fish image measuring greater than approximately 50 cm, not including sturgeon, was included in length frequency calculations. Text files from exported echogram data were copied into excel and used to generate length frequency information. Acoustic signatures from mammals and birds were used to discriminate fish from mammals (e.g. seal, otter, beaver). Figure 5. Image of computer output showing split screen function used for measuring fish. 11

Escapement Estimates Two methods were used to apportion DIDSON fish counts for Chinook and steelhead escapement estimates. First we used the Rowdy Creek weir data and fish length frequency information from DIDSON to define a split between the Chinook and steelhead runs. Numbers of Chinook trapped at the Rowdy Creek weir decreased precipitously during the first week of December while numbers of steelhead trapped increased in mid-december. Typically Chinook cease entering the facility by 15 December (Andrew Van Scoyk pers. comm.). The average length of fish at the DIDSON fish counting station also decreased during early December and the length frequency distribution narrowed suggesting the absence of Chinook. Therefore 15 December was used to split the runs into two distinct groups, or runs, with steelhead being counted for the latter group. Coho were not considered in the estimates because CDFW spawning ground survey estimates for coho were in the hundreds of fish. Second, Chinook salmon and steelhead escapement estimates were made using angler data collected by the Department of Fish and Wildlife from 1997 through 2006 (figure 6). Attempts were also made to include coho salmon but due to the paucity of available adult data, the negative bias in the recreational catch data (no recreational coho salmon fishery) and only hundreds of returning adults we did not include coho salmon in the estimate. Ratios of species caught by month were averaged for the years surveyed and applied to DIDSON count data from 2010-2015. Downstream fish passage numbers were removed from the simple flux model after 15 December for both years to account for downstream steelhead presumably already spawned (kelts). Ratios of species caught per month were then applied to the monthly DIDSON fish counts. Figure 6. Monthly means of estimated angler caught Chinook salmon and steelhead in the Smith River (CDFG 1997-2006) RESULTS 12

DIDSON Fish Counts The DIDSON station operated from 23 September 2014 through 31 March 2015. The station was shut down due to high flows on 21 December and 7 February. High flows also resulted in partial days of operation on 6 February and 11 December. The recording efficiency of the DIDSON station operating continuously 24 hours per day was 94%. A total of 44,009 fish were counted passing upstream and 18,326 fish were counted passing downstream (Figure 7). The highest daily net count of fish migrating upstream was 825 fish and occurred on 01 November 2014. This was followed by other peak counts on 30 October 2014 (786 fish) and 2 November 2014 (743 fish). Counts of more than 400 fish per day occurred ten times during the season between 29 September 2014 and 6 November 2014, one of which days occurred during the steelhead run time on 16 January 2015. The total number of fish counted during the Chinook run-time period from 23 September 2014 to 15 December 2014 was 15,796. After 15 December 2014 fish were counted passing downstream but were assumed to be moribund Chinook or steelhead kelts and were not subtracted from the upstream counts. Furthermore downstream passage of fish declined rapidly after 15 December as noted in previous Smith River studies. A total of 11,944 fish were counted passing upstream during the steelhead run-time from 15 December 2014 to 31 March 2015 (Figure 8). Spikes of downstream fish passage occurred in mid-january, early February and late March. These downstream fish passage spikes were attributed to fall or early winter steelhead kelts. Systematic Hourly Sampling Rate Analysis A sampling rate of 33 percent, or 20 minutes per hour, was applied to the entire 2014-2015 data set (Tables 3,4,5). Results from the entire season, without subtracting kelts after 15 December 2014, averaged 4.8 percent error. The percent error for the first, second and third 20 minute sampling interval from 01 October 2014 to 14 December 2014 averaged 6.5 percent. The percent error during the steelhead run-time from 15 December 2014 to 31 March 2015 averaged 4.2 percent. Within the steelhead run-time period, percent error was lowest during January and averaged 2.9 percent (peak steelhead migration). The percent error was highest during March averaged 202 percent for the first, second and third 20 minute intervals (steelhead kelt migration). Rowdy Creek Fish Hatchery Weir Counts Comparisons of DIDSON fish counts to Rowdy Creek weir counts in 2014/2015 were similar to the results from the 2010-2013 seasons. Peak fish passage rates of Chinook and steelhead during the fall and winter at the Rowdy Creek weir coincided with peak passage rates at the DIDSON fish counting station (Figure 9). During three complete seasons of DIDSON operation and one winter steelhead DIDSON operation (2012-2013), Rowdy Creek weir data indicates overlap of Chinook and steelhead runs from the last week in November to mid-january. The first steelhead at the Rowdy Creek weir was caught 26 November 2014 and the last Chinook was caught 17 January 2015. However the ratios of steelhead to Chinook in the last half of November, and Chinook to steelhead in the first half of January were low at 0.01 and 0.02, respectively. The Chinook to steelhead ratio approached 1:1 during the first half of December. Overall counts of fish at the Rowdy Creek and DIDSON facilities were lowest during mid-december suggesting a separation of the Chinook and steelhead run times. 13

Figure 7. Daily upstream and downstream fish counts at the Smith River DIDSON Fish Counting Station vs river discharge, 2014-2015. 14

Figure 8. Daily fish counts at the Smith River DIDSON Station compared to U.S.G.S stream gauge data. The first Y axis corresponds to the net upstream numbers of fish and the second Y axis corresponds to river discharge in cubic feet per second (c.f.s.) 15

Figure 9. Counts of Chinook and steelhead from the Rowdy Creek weir compared to counts of adult fish made at the Smith River DIDSON Fish Counting Station, 2014-2015. The first Y axis corresponds to the DIDSON fish counts while the second Y axis corresponds to Rowdy Creek weir counts. The shaded area represents DIDSON fish counts and the red and blue bars represent Chinook and Steelhead, respectfully. 16

Table 3. The percent error by month associated with 20 minute per hour sampling rate. Downstream fish or kelts were not removed. Data from 23 September to 30 September were not included. Estimated Monthly Counts and Error Actual Total Count Total Using First 20 min. of Each Hour Total Using Second 20 min. of Each Hour Total Using Third 20 min. of Each Hour % Error % Error % Error Oct. 1-31 8401 7710 9.0% 9222 8.9% 8271 1.6% Nov. 1-30 5671 5379 5.4% 6234 9.0% 5403 5.0% Dec. 1-31 2129 2190 2.8% 2265 6.0% 1938 9.9% Jan. 1-31 5277 5115 3.2% 5508 4.2% 5208 1.3% Feb. 1-28 3084 3249 5.1% 3111 0.9% 2892 6.6% Mar. 1-31 213 456 53.3% 234 9.0% -48 543.8% TOTALS 24775 24099 26574 23664 Table 4. The percent error by week associated with 20 minute per hour sampling rate. Downstream fish or kelts were not removed. Data from 23 September to 30 September were not included. Estimated Weekly Counts and Error Actual Total Count Total Using First 20 min. of Each Hour Total Using Second 20 min. of Each Hour Total Using Third 20 min. of Each Hour % Error % Error % Error 10/1/2014-10/7/2014 931 891 4.5% 984 5.4% 918 1.4% 10/8/2014-10/14/2014 968 528 83.3% 921 5.1% 1455 33.5% 10/15/2014-10/21/2014 2782 2715 2.5% 3183 12.6% 2448 13.6% 10/22/2014-10/28/2014 2306 2364 2.5% 2520 8.5% 2034 13.4% 10/29/2014-11/4/2014 3939 3693 6.7% 4524 12.9% 3600 9.4% 11/5/2014-11/11/2014 1374 1095 25.5% 1620 15.2% 1407 2.3% 11/12/2014-11/18/2014 990 978 1.2% 888 11.5% 1104 10.3% 11/19/2014-11/25/2014 411 369 11.4% 471 12.7% 396 3.8% 11/26/2014-12/2/2014 606 663 8.6% 639 5.2% 516 17.4% 12/3/2014-12/9/2014 485 516 6.0% 516 6.0% 429 13.1% 12/10/2014-12/16/2014 155 174 10.9% 159 2.5% 132 17.4% 12/17/2014-12/23/2014 186 171 8.8% 207 10.1% 180 3.3% 12/24/2014-12/30/2014 895 933 4.1% 873 2.5% 879 1.8% 12/31/2014-1/6/2015 1542 1524 1.2% 1713 10.0% 1389 11.0% 1/7/2015-1/13/2015 1316 1224 7.5% 1404 6.3% 1320 0.3% 1/14/2015-1/20/2015 1113 1164 4.4% 1137 2.1% 1038 7.2% 1/21/2015-1/27/2015 879 909 3.3% 876 0.3% 852 3.2% 1/28/2015-2/3/2015 1298 1194 8.7% 1266 2.5% 1434 9.5% 2/4/2015-2/10/2015 678 696 2.6% 648 4.6% 690 1.7% 2/11/2015-2/17/2015 939 1020 7.9% 984 4.6% 813 15.5% 2/18/2015-2/24/2015 544 522 4.2% 627 13.2% 483 12.6% 2/25/2015-3/3/2015 404 513 21.2% 345 17.1% 354 14.1% 3/4/2015-3/10/2015 352 396 11.1% 366 3.8% 297 18.5% 3/11/2015-3/17/2015-149 -150 0.7% -156 4.5% -141 5.7% 3/18/2015-3/24/2015-144 -15 860.0% -123 17.1% -294 51.0% 3/25/2015-3/31/2015-25 12 308.3% -18 38.9% -69 63.8% TOTALS 24775 24099 26574 23664 Table 5. The percent error by Run (based on a 15 December end date for Chinook salmon) associated with 20 minute per hour sampling rate. Downstream fish or kelts were not removed. Data from 23 September to 30 September were not included. Estimated Species Counts and Error Actual Total Count Total Using First 20 min. of Each Hour Total Using Second 20 min. of Each Hour Total Using Third 20 min. of Each Hour % Error % Error % Error 10/1/2014-12/15/2014 14911 13938 7.0% 16395 9.1% 14409 3.5% 12/16/2014-3/31/2015 9864 10161 2.9% 10179 3.1% 9255 6.6% TOTALS 24775 24099 2.8% 26574 6.8% 23664 4.7% 17

Length Frequency Analysis Length frequency measurements have vastly improved with the addition of telephoto lenses fitted to the long range DIDSON units. Total lengths of 4,459 fish were measured from the Right and Left Banks from 01 October 2014 through 31 March 2015 (Figures 10, 11). The average total length of fish from the entire 2014-2015 season was 76 cm with a standard deviation of 16.25. The average length was highest during the week of 12 November 2014 and lowest during the week of 25 March 2015. The average lengths of fish during the Chinook run time was highest during the week of 12 November 2014. During the steelhead run the highest average length occurred the week of 24 December 2014. Standard Deviations of fish measured with DIDSON were largest during the first half of October 2014 likely due to a greater presence of jacks and lowest during mid-january 2015 likely due to the dominant presence of steelhead. Length frequency histograms also indicate differences in length distributions throughout the study period (Figure 12). Fish lengths during October are distributed widely indicating the presence of multiple age classes of Chinook and a strong presence of jacks and possibly large half-pounder steelhead. November fish length distributions may indicate the presence of older age classes of Chinook and a limited number of jacks. During December the length frequency distribution narrows and shifts to the left representing the end of the Chinook run and the presence of steelhead. January through March distributions shift further to the left indicating the presence of steelhead. Figure 10. Weekly average fish lengths and error measured during the 2014-2015 season. A total of 4,459 fish were measured using a long range DIDSON fitted with a telephoto lens at the Smith River DIDSON Fish Counting Station. The Y axis is length (cm). 18

Escapement Estimates We estimated the escapement of Chinook and steelhead in the Smith River upstream from river mile 6 (e.g. Rowdy Creek) using (1) available information from historic Smith River angler survey data and (2) a cutoff date supported by length frequency information, fish behavior, and Rowdy Creek weir data (Figures13,14). Based on monthly Chinook to steelhead catch ratios from a compilation of angler surveys the 2014-2015 escapement estimates of Chinook and steelhead were 14,400 and 12,800 fish, respectively. Using the cutoff date of 15 December Chinook and steelhead escapement estimates were 15,796 and 11,906 fish, respectively. However these estimates do not include escapement to the Rowdy Creek sub-basin and other small tributaries downstream from the DIDSON site. Figure 11. Weekly average fish lengths and error measured during the 2014-2015 season. A total of 4,459 fish were measured using a long range DIDSON fitted with a telephoto lens at the Smith River DIDSON Fish Counting Station. The First Y axis is the total number of upstream fish while the second Y axis corresponds to average fish length (cm). 19

Figure 12. Length frequency histograms of fish measured with DIDSON at the Smith River DIDSON Fish Counting Station 2014-2015. The Y axis corresponds to the number of fish measured. 20

Numbers of Fish Numbers of Fish 25000 20000 18681 19354 15000 15771 14674 14865 12875 10000 9600 Chinook Steelhead 5000 0 2010-2011 2011-2012 2012-2013 2013-2014 2014-2015 DIDSON Study Year Figure 13. Estimates of escapement of Chinook and steelhead in the Smith River using angler data to apportion DIDSON fish counts, 2010-2015. Estimates do not include Rowdy Creek or tributaries downstream from the DIDSON site. No data was collected for Chinook in 2012 or for Chinook and steelhead in 2013-2014 25000 21806 20681 20000 15796 15000 10000 12646 13345 9600 11944 Chinook Steelhead 5000 0 2010-2011 2011-2012 2012-2013 2013-2014 2014-2015 DIDSON Study Year Figure 14. Estimates of escapement of Chinook and steelhead in the Smith River using a December 15 cutoff date to apportion DIDSON fish counts, 2010-2015. Estimates do not include Rowdy Creek or tributaries downstream from the DIDSON site. 21

DISCUSSION This study follows a two-year DIDSON fish counting station pilot study designed to monitor the total escapement of Chinook, coho and steelhead to the Smith River, excluding Rowdy Creek and tributaries below river mile 6, from October through March 2010-2012 Figure 13 (Larson 2013a). Another study was conducted at the same location and focused on the Steelhead migration from 15 December 2012 through March 2013 (Larson 2013b). These studies along with recent work from 2014-2015 had the primary objective of testing the effectiveness of multi-beam sonar as a tool for monitoring anadromous salmonids in northern California pursuant to the Coastal Salmonid Monitoring Plan (Adams et al 2011). Results from our studies suggest that multi-beam sonar is a valuable component for monitoring the long term trends of anadromous salmonid abundance and assessing species status, throughout the Pacific Northwest and California. However the technology by itself does not provide a means of apportioning fish species, particularly for those species having depressed populations, in rivers with overlapping run times. We believe this drawback can be overcome with concomitant survey methods and should not detract from the value of annual DIDSON data in anadromous streams. Obtaining this additional information may be accomplished through physical means (seine), video during periods of adequate visibility, and/or a complete on the ground angler survey. These additional pieces of information may also provide hatchery influence information not available during the preparation of this report. The Threatened Southern Oregon Northern California coho salmon population in the Smith River is estimated to be several hundred fish by the CDFW while Klamath Mountains Province Steelhead and SONCC Chinook populations in the Smith River are each greater than ten thousand fish. Therefore we do not consider coho in our DIDSON counts or population estimates. Additionally little is known about the timing of entry or details of coastal cutthroat migrations in the Smith River or their population size. Weir trap studies in Smith River tributaries indicate the majority of adult coastal cutthroat trout are under 50 cm in length falling below the length threshold for adult fish (Chinook and steelhead) in our study. Therefore coastal cutthroat trout were not considered in our study. Smith River angler survey data and Rowdy Creek weir data illustrate Chinook salmon and steelhead run time (time of river entry) overlap. Chinook salmon typically stage in the Smith River estuary and mainstem in September and the run begins to build after the first significant rise in river stage. Chinook typically cease entering the river around mid-december and the vast majority of adult salmonids migrating into the Smith River past mid-december are steelhead (Andrew Van Scoyk pers.comm.). Coho salmon periodically occur in the recreational fishery during November and December though are not usually observed at spawning grounds until late December and January. Therefore coho timing of entry likely spans from the latter part of the Chinook run in late November into the early part of the steelhead run in mid-december. As in previous Smith River DIDSON study periods Chinook salmon exhibited upstream and downstream passage behavior (milling) while steelhead almost exclusively travelled upstream. The dates at which milling behavior decreased or ceased are likely when the run transitioned from the overlap of Chinook and steelhead runs to a discrete steelhead run time. Sampling Rate Analysis Continuous fisheries sampling data sets are extremely rare (Xie and Martens 2014). We analyzed the count data from the continuous data collected from the 2014-2015 DIDSON study and estimated the error 22

associated with sampling at the reduced rate of 33% or 20 minutes per hour. Greater percent error values were associated with the reduced sampling rate during the Chinook migration in October and the steelhead kelt migration in March. The greater error in October is likely due Chinook salmon tendencies to stage and mill during low flows, typically in October on the Smith River. Steelhead kelts migrating downstream increase in abundance during March and frequently travel in larger schools. The steelhead kelt migrations are likely responsible for the greater error associated with the reduced sampling rate in March. Reduced sampling rate error values were lowest during January likely due to the consistent upstream movements of steelhead during the afternoon and early evening hours. Length Frequency Analysis The addition of the second telephoto lens improved sonar image resolution and reduced positive bias experienced while measuring fish with standard lenses. Length frequency information appears to support our Chinook and Steelhead run time separation date of mid-december. However, this is likely to vary from season to season based on environmental factors affecting marine growth of salmon and steelhead and age class compositions. In addition more information is needed to compare DIDSON length measurements fish to actual fish lengths. In 2014 three Chinook salmon caught by anglers in view of the sonars were documented in sonar files, measured, and compared to the actual lengths. This limited investigation was not included in this report however actual lengths of fish were within centimeters of measurements made with DIDSON telephoto lenses. Escapement Estimates Estimates of basin-wide Chinook and steelhead escapement are important for the long term management of populations. Approaches taken here to estimate escapement for these species are the result of the lack of available and current information for more in-depth analyses. However we feel somewhat confident these are the most accurate estimates of escapement available for the Smith River. Estimates are provided in Figures 13 and 14. The 2014-2015 estimates of escapement for Chinook and steelhead, not including Rowdy Creek escapement, were from 14,900 to 15,800 Chinook and 12,000 to 12,900 steelhead. Based on DIDSON data and Rowdy Creek weir data from 2010 to 2015 we estimate the annual Chinook escapement to be from 15,000 to 25,000 fish and steelhead escapement to be from 10,000 to 15,000 fish. CONCLUSION When used in conjunction with conventional fisheries sampling techniques, DIDSON should reduce the uncertainty associated with current status assessments of imperiled species, help tease out ocean conditions and other climatic factors affecting populations, and improve the management of existing recreational, tribal and commercial fisheries. Long term funding is needed for monitoring Smith River anadromous fish populations with DIDSON. 23

ACKNOWLEDGEMENTS Funding for this project was awarded by the California Department of Fish and Wildlife Steelhead Catch Report Restoration Card Program. Previous financial support has been provided by the Fisheries Restoration Grants Program, Resources Legacy Fund Foundation and the Wild Salmon Center. The support from the Del Norte County Board of Supervisors allowed the grant to be administered through the County of Del Norte. The support and project outreach efforts by the Smith River Alliance, California Trout and Trout Unlimited has provided opportunities for future partnerships for monitoring Smith River fisheries. Thank you Phillip Bairrington and Justin Garwood of CDFW for the use of equipment essential to the project and Beatrijs DeWaard for contract support. This project would not have succeeded without the hard work, innovative spirit and dedication from Jesse Nolan and Ryan Nelson. Thank you Jim Waldvogel, Tom Weseloh, Ben Taylor and the Smith River Advisory Council for your assistance, knowledge and long term commitment to improving fisheries in the Smith River. LITERATURE CITED Adams, P.B., Boydstun, L.B., Gallagher, S.P. Lacy, M.K., McDonald, T., and K.E. Shaffer. 2011. California Coastal Salmonid Population Monitoring: Strategy, Design and Methods. State of California Natural Resources Agency Department of Fish and Game, Fish Bulletin 180. 82pp. CDFW, 1996. Recreational Angler Use and Catch in the Mainstem Smith River, California, August December, 1984. State of California, Department of Fish and Wildlife. Inland Fisheries, Northern Coast Region. CDFW, 1999. Recreational Angler Use, Catch and Harvest in the Smith River, Del Norte County, California, September 1997- March 1998. State of California, Department of Fish and Wildlife. Inland Fisheries, Northern Coast Region. CDFW, 2000. Recreational Angler Use, Catch and Harvest in the Smith River, Del Norte County, California, November 1998- March 1999. State of California, Department of Fish and Wildlife. Inland Fisheries, Northern Coast Region. CDFW, 2000. Recreational Angler Use, Catch and Harvest in the Smith River, Del Norte County, California, October 1999- March 2000. State of California, Department of Fish and Wildlife. Inland Fisheries, Northern Coast Region. CDFW, 2002. Recreational Angler Use, Catch and Harvest in the Smith River, Del Norte County, California, October 2000- March 2001. State of California, Department of Fish and Wildlife. Inland Fisheries, Northern Coast Region. CDFW, 2005. Recreational Angler Use, Catch and Harvest in the Smith River, Del Norte County, California, 2004-2005. State of California, Department of Fish and Wildlife. Inland Fisheries, Northern Coast Region.Project 1gl CDFW, 2006. Recreational Angler Use, Catch and Harvest in the Smith River, Del Norte County, California, 2005-2006. State of California, Department of Fish and Wildlife, Inland Fisheries, Northern Coast Region. 24

Cronkite, G.M.W., Enzenhofer, H.J., Ridley, T., Holmes, J., Lilja, J., and Benner, K. 2006. Use of highfrequency imaging sonar to estimate adult sockeye salmon escapement in the Horsefly River, British Columbia. Can. Tech. Rep. Fish. Aquat. Sci. 2647: vi + 47 p. Larson, Z.S. 2013a. Use of dual frequency identification sonar to monitor steelhead escapement in the Smith River, California. County of Del Norte, California Dept. of Fish and Wildlife Fisheries Restoration Grants Program: Grant Number P0910315. 27pp.. Larson, Z.S. 2013b. Operation of dual frequency identification sonar (DIDSON) to monitor adult anadromous fish migrations in the Smith River, California: 2-year pilot study (2010-2012). County of Del Norte, California Dept. of Fish and Wildlife Fisheries Restoration Grants Program: Grant Number P0910315. 43pp. Waldvogel, J. 2006. Spawning Escapement Estimate and Age Composition for a Tributary of the Smith River, California 23-Year Analysis. University of California Cooperative Extension. California Sea Grant College Program, Univ. of Cal., San Diego, California, 92093-0232. Publication No. T-060 2006 ISBN 1-888691-16-6. Xie, Y., A. P. Gray, F. J. Martens, J. L. Boffey and J. D. Cave. 2005. Use of Dual-Frequency Identification Sonar to Verify Salmon Flux and to Examine Fish Behaviour in the Fraser River. Pacific Salmon Comm. Tech. Rep. No. 16: 58 p. Xie, Y., and F.J. Martens. 2014. An empirical approach for estimating the precision of hydroacoustic fish counts by systematic hourly sampling. North American Journal of Fisheries Management 34:535, 2014. Xie, Y., Mulligan, T.J., Cronkite, G.M.W., and Gray, A.P. 2002. Assessment of potential bias in hydroacoustic estimation of Fraser River sockeye and pink salmon at Mission, B.C. Pacific Salmon Comm. Tech. Rep. No. 11: 42p. Personal Communication Andrew Van Scoyk, Rowdy Creek Fish Hatchery Manager. Smith River, California. 25

APPENDIX Status of fish species occurring or potentially occurring in the Smith River watershed. SPECIES ESU STATUS¹ Federal State Coho salmon (Oncorhynchus kisutch) Southern Oregon/Northern California Coasts T T Chinook salmon Oncorhynchus tshawytscha Southern Oregon/Northern California Coasts NW CSC Steelhead Oncorhynchus mykiss irideus Klamath Mountains Province NW None Coastal cutthroat trout Oncorhynchus clarki clarki Southern Oregon/California Coastal NW CSC Chum salmon Pacific Coast NW None Oncorhynchus keta Green sturgeon N/A C2 None Acipenser medirostris River lamprey N/A C2 CSC Lampetra ayresi Pacific Lamprey N/A C2 None Lampetra tridentata T Threatened NW Not warranted for listing CSC CDFG and/or California Board of Forestry Species of Concern C2 Species formerly classified as Category 2 candidates by the USFWS; these species no longer have a legal federal status CE Candidate to be listed as endangered pursuant to the California Endangered Species Act. 26