Lower Skeena River Angling Creel Survey 2014

Transcription

1 Lower Skeena River Angling Creel Survey 2014 FINAL REPORT Prepared for: Pacific Salmon Commission Vancouver, BC Fisheries and Oceans Canada Prince Rupert, BC Kitsumkalum First Nation Terrace, BC Prepared by: Dave Robichaud, Anita C. Blakley, and Karl K. English LGL Limited environmental research associates 9768 Second Street Sidney, BC, V8L 3Y8 18 March 2015

2

3 EXECUTIVE SUMMARY The Lower Skeena River sport fishery can yield substantial catches. It has been surveyed annually since The three objectives for this study were: 1) to provide monthly catch estimates (June to September) for all salmon species caught in the sport fishery on the waters of the Skeena River downstream of Terrace; 2) to inspect as many Chinook salmon as possible and record the incidence of adipose fin clips; and 3) to collect length and age data of Chinook salmon. This report documents the results from survey efforts conducted from June to September, The design for the 2014 lower Skeena creel survey was similar to that from 2010 to 2013, and was based on similar recreational fishery surveys conducted on the Skeena, Nass and Peace rivers. The 2014 creel survey produced catch estimates with relatively large standard errors resulting from small catches and high variability in catch rates. Sockeye salmon (Oncorhynchus nerka), which was the most abundantly caught fish (5,787 fish caught, SE = 759), was also the most retained species (4,249 fish harvested, SE = 616; 73% of fish retained). Pink salmon (O. gorbuscha), though often caught (5,045 caught, SE = 1,224) was infrequently retained (646 fish kept, SE = 126; 13% retention). Coho salmon (O. kisutch) was the third most commonly caught species (4,119 fish caught, SE = 701), and was retained 57% of the time, with a total of 2,360 (SE = 502) fish harvested. Steelhead (O. mykiss) were frequently caught (3,031 fish caught, SE = 428), but harvest was negligible (2% of total Steelhead catch). The few harvested Steelhead likely resulted from anglers who were unaware of the regulations or could not distinguish between Steelhead and other species. Overall Chinook salmon catches totaled 2,888 large fish (SE = 303) and 2,116 jacks (SE = 459). These fish were the most likely to be retained (87-93% harvest). Consistent with the fishing regulations, no Chinook salmon were harvested in Stratum One (Ferry Island to Lakelse) on or after 8 August. In 2014, overall angling effort (230,627 angler hours, SE = 32,407) approached (96%) that reported from It is the first time since 2010 that effort levels have been so high. The time-series illustrates the degree of effort fluctuation among years. Average Chinook salmon CPE in was 46-49% of the estimates from , and 44% of that from Given the relatively good fishing conditions, the lower 2013 and 2014 Chinook catch rates (relative to the surveys) may be the result of lower returns to the Skeena River. Catches rates of Steelhead and Pink and Coho salmon were at the lowest levels since the study began in Sockeye salmon catch rates recovered from the 2013 crash year. LGL Limited Page i

4 TABLE OF CONTENTS EXECUTIVE SUMMARY...i LIST OF TABLES... iii LIST OF FIGURES... iv LIST OF APPENDICES... v INTRODUCTION... 1 SCOPE... 1 OBJECTIVES... 2 METHODS... 3 August Creel Data... 4 Angler Activity Patterns... 6 Catch Per Effort Estimation Angler Effort Estimation Catch Estimation Statistical Analyses RESULTS Angler Interviews Angler Demographics and Behaviours Angler Activity Patterns Catch Per Effort Estimates Angler Effort Estimates Catch Estimates Harvest (Retention) Estimates Biosampling DISCUSSION Creel Estimates Precision of the Results Accuracy of the Results RECOMMENDATIONS ACKNOWLEDGEMENTS LITERATURE CITED APPENDICES LGL Limited Page ii

5 LIST OF TABLES Table 1. Locations where anglers were counted and interviews conducted in Table 2. Summary of scheduled sampling effort by month for AM and PM strata and weekend/holiday and weekday strata Table 3. The amount of data available to estimate angler activity patterns... 8 Table 4. Table 5. The sample size of angler CPE data for each river stratum, and for each month, day type, and access method Summary of observed sampling effort by month for AM and PM strata and weekend/holiday and weekday strata Table 6. Fishing method and gear types used by interviewed anglers, by month Table 7. Table 8. Table 9. Catch per effort estimates for six fish taxa, by month, river stratum and access method Statistical tests of the effect of month, river stratum, and access method on median catch per effort estimates for the 6 taxa surveyed Retention per effort rates for six fish taxa, by month, river stratum and access method Table 10. Effort estimates, by month, day type, river stratum and access method Table 11. Statistical tests of the effect of month, day type, river stratum and access method on median daily effort estimates during the study period Table 12. Estimated catch of six fish taxa in three geographic strata, by month Table 13. Statistical tests of the effect of month, river stratum and access method on median catch-per-day estimates for the 6 taxa surveyed Table 14. Estimated harvest of six fish taxa in three geographic strata, by month Table 15. Statistical tests of the effect of month, river stratum and access method on median harvest-per-day estimates for six fish taxa surveyed Table 16. Comparison of the catch and effort estimates for the 2003, and creel surveys for the same spatial strata and fishing periods Table 17. Table 18. Comparison of the average CPE estimates for the 2003 and creel surveys for the same spatial strata and fishing periods Estimates of the average length of a fishing trip and average number of anglers fishing each day by river stratum and month in LGL Limited Page iii

6 LIST OF FIGURES Figure 1. Map of the lower Skeena River showing the major fishing sites and boundaries for the three river strata... 2 Figure 2. Angler activity patterns, by month... 9 Figure 3. Angler activity patterns, by river stratum and day type Figure 4. Angler activity patterns, by access method Figure 5. The monthly percent of interviewed anglers that were with a professional guide and their location of residence Figure 7. The target species, as declared by interviewed anglers Figure 8. Angler activity patterns, by day type and river stratum Figure 9. Angler activity patterns, by day type and river stratum Figure 10. Figure 11. Figure 12. Figure 13. Figure 14. Figure 15. Box-whisker plots showing the distribution of shore-access CPE estimates for each month, by taxon Box-whisker plots showing the distribution of daily effort estimates for each access method Box-whisker plots showing the distribution of catch-per-day estimates for each month, by taxon Box-whisker plots showing the distribution of catch-per-day estimates by access method for each taxon Box-whisker plots showing the distribution of harvest estimates by month, for each taxon Box-whisker plots showing the distribution of harvest estimates by access method for each taxon Figure 16. Chinook salmon with adipose fin present and missing Figure 17. Total effort estimates for each month, LGL Limited Page iv

7 LIST OF APPENDICES Appendix 1. Data forms Appendix 2. Proposed survey schedule and actual shifts worked Appendix 3. Boat-based effort data, including angler counts (by access method) and survey dates and times Appendix 4. Total catch estimates (harvest + release) for each taxon, by month, LGL Limited Page v

8

9 INTRODUCTION Prior to 2010, sporadic surveys have been conducted to estimate Chinook salmon (Oncorhynchus tshawytscha) catch in the Lower Skeena River sport fishery. Specifically, creel surveys that encompassed most of the Chinook salmon fishing season were performed only in 1995, 2001, and In 2010, the Pacific Salmon Commission (PSC) Northern Endowment Fund supported the development and implementation of a new set of creel surveys and angling effort counts designed to provide reliable estimates of angler effort and catches of salmon (four species) and Steelhead (Oncorhynchus mykiss). In all five years of study since 2010, the spatial strata from the 2003 creel survey program (conducted by J.O. Thomas) were used in order to facilitate comparisons between the current and results and those from or before Surveys conducted in documented a substantial reduction in fishing effort (92,373 to 103,806 angler-hours per year) from that estimated for 2003 (203,587 angler-hours; Robichaud and English 2011, 2012, Robichaud et al. 2013) which may have been the result of poorer fishing conditions and a lower abundance of Chinook salmon. The angler effort estimated for 2013 showed evidence of an approach toward 2003 levels, with total angler effort of 158,746 anglerhours (Robichaud et al. 2014). This report documents the results from a fifth year of survey efforts conducted from early June to late September The report describes the methods used to derive estimates of angler activity pattern, catch and fishing effort. SCOPE The geographic scope of the lower Skeena creel survey study in 2014 was the same as in the previous four years, i.e., from Ferry Island near Terrace BC to the Kwinitsa boat launch near the mouth of the Skeena River, a distance of 150 km (Figure 1). For the purposes of data collection and analysis, the study area was divided into three geographic strata, selected to align with previous surveys. These strata were: 1) Terrace to the Lakelse Confluence; 2) the Lakelse Confluence to the Exstew Confluence; and 3) the Exstew Confluence to the Kwinitsa boat launch. The temporal scope of the full study was from June through September Survey schedules were designed to provide sufficient data to derive catch and efforts estimates for each month and for each river zone. The study included creel analysis of all major local sport fish species, including all five salmon species and Steelhead. LGL Limited Page 1

10 Figure 1. Map of the lower Skeena River showing the major fishing sites and boundaries for the three river strata used for the 2014 creel survey design. OBJECTIVES The three objectives for this study were: 1) to provide monthly catch estimates for all salmon species caught in the sport fishery on the waters of the Skeena River downstream of Terrace; 2) to inspect as many Chinook salmon as possible and record the incidence of adipose fin clips such that awareness factors for the expansion of voluntary CWT head submissions may be developed; and 3) to collect length and age data from Chinook salmon. These objectives were addressed by conducting effort counts and angler interviews throughout the lower Skeena recreational fishery from June through September The precision goal for the study was to estimate total Chinook salmon catch within 25% of the true value 19 times out of 20. Creel survey strata included temporal separation by month, including weekdays and weekend days. Spatial stratification was similar to past surveys, including three river strata LGL Limited Page 2

11 (described above). All local fishing areas (i.e., those accessed by shore-based and by boat-based sport fishers) along the Skeena River in these strata were included in our survey design. The incidence of adipose fin clips for Chinook and Coho salmon (Oncorhynchus kisutch) were obtained by examining fish during angler interviews. Scale samples were collected from Chinook salmon, stored in scale books and used for age analysis. METHODS The study area spans a very large geographic area, making it unreasonable to obtain a complete and direct (interview-based) census of the entire catch. Therefore, our approach relied on statistical methods to estimate catch by the multiplication of angler effort estimates by catch per effort estimates, for each river stratum, month, day type (weekday vs. weekend), and species. For each river stratum during each month, fishing effort was estimated by counting anglers during boat-based river-surveys; and catch per effort was estimated from interviews (see data forms in Appendix 1). Interviews were conducted at known access points, and also from a boat (after conducting the effort surveys). During interviews, anglers were asked about their catch, effort, and fishing locations. They were also asked about their hourly fishing activity patterns on the current and previous day, and whether or not they were finished their fishing activity for the day. Data collected during interviews included: Angler effort number of anglers, total fishing effort (in angler-hours), fishing location, access location, target species, and gear/bait used; Angler activity the hours during which angling activity was conducted on the day of the interview if the fishing trip was complete and on the previous day, if fishing occurred; Fish kept number of fish caught and kept, by river stratum and by species for the five main salmonids types: Chinook (with jacks tallied separately) salmon, Coho salmon, Sockeye salmon (Oncorhynchus nerka), Pink salmon (Oncorhynchus gorbuscha), and Steelhead; Fish released number of fish caught and intentionally released, by species and by river stratum; Whether or not the catch was verified and counted; Whether or not the trip was guided by a professional; Angler demographics region or community of origin; Angler access methods (shore vs. boat); and Timestamp, including date, month, day type (i.e., weekday vs. weekend/holiday) and time of day. The methods used were identical to the previous four years of this study (Robichaud and English 2011, 2012, Robichaud et al. 2013, 2014), originally adapted from those developed and documented for the Georgia Strait Creel Survey (English et al. 2002). The methods used to LGL Limited Page 3

12 estimate the statistical precision associated with creel survey catch and effort estimates are based on those documented in English et al. (2002) and Blakley et al. (2003). This procedure provides a statistically unbiased estimate of catch per effort, provided the anglers interviewed are representative of the entire fishery. To ensure this, the interview schedule was designed to capture data from representative fishermen in each river stratum, on both day types, and over all time periods of the day. The main shore-based interviewing locations are listed in Table 1. The locations surveyed were selected from all available access points, based on their geographical distribution and the amount of fishing activity that was assumed to be conducted from that site. Within each geographic region, the busiest (i.e., most accessible) access points were selected preferentially in order to obtain the maximum number of interviews. This approach was based on two important observations: 1) the variability in CPE (catch-per-effort) among fishing parties landing at a single access point tends to be as great as the variability in CPE among different access points within a geographic area; and 2) CPE and effort can vary substantially both within and between days at a single site (English et al. 2002). Under these conditions it is better to obtain a large number of interviews covering all temporal strata for a small number of sites than to sample a larger number of sites and obtain fewer interviews and less complete temporal coverage for any specific site. Nevertheless, these access-point interviews were supplemented with boat-based interviews collected opportunistically following the boat-based effort estimation surveys. Sampling schedules were designed to ensure adequate coverage in all river strata, on both day types (Table 2). Detailed monthly survey schedules are provided in Appendix 2. Complete counts of anglers were conducted during peak fishing periods on most weekend days and usually on three of the five available weekdays each week. Angler interviews were to be conducted for a random sample of the anglers encountered during the roving effort surveys and additional survey effort was scheduled for each of the major angler access points. The roving surveys provided complete coverage of the fishing area, but the data collected was usually for incomplete fishing trips. Surveys at major access points provided more opportunities to interview anglers at the end of their daily fishing trips, especially if surveyors work PM shifts. In order to remove the known fishing effort biases associated with incomplete fishing trips, we used information of fishing activity for the previous day (yesterday line times) to derive fishing activity patterns and estimates of the average number of hours fished each day. Shore interviewing sessions were to be separated into AM and PM shifts, with AM shifts occurring between 6:00 and 14:00 and PM shifts from 14:00 to 22:00. The majority of the roving boat surveys were conducted in the afternoon from 13:00 to 21:00. The exact timing of work shifts changed over the study period in concert with the shortening daylight hours. Surveyors were allocated 30 minutes at each end of the shift to access the local survey sites and one hour to access the more remote sites. In total, 148 shifts were scheduled (Table 2), including 74 roving boat surveys and 74 access point surveys. August Creel Data Starting on 8 August, no fishing for Chinook salmon was permitted from Lakelse River mouth upstream to the Skeena River Overpass at Ferry Island. Because of this fishing regulation LGL Limited Page 4

13 Table 1. Locations where anglers were counted and interviews conducted in The main shore-based interview locations are marked with an X. Zone # Zone Name Site # Site Name Shore Survey 1 Ferry Island to 1 Upper Ferry Is. X Lakelse Confluence 2 Lower Ferry Is. X 3 Power Line Bar 4 Cottonwoods 5 Hells Gate Bar 6 Kalum Boat Launch X 7 Alberta Bar 8 Kraut Bar 9 New Remo Bar 10 Old Remo Bar 11 Chicken Bar 12 Turd Island 13 Lakelse Confluence 2 Lakelse Confluence 14 Delta Bar to Exstew Confluence Mile Bar (Shames R. top) Mile Bar (Shames R. bottom) 17 Esker Bar 18 Shames Bar (Konaham) 19 Exstew Bar 3 Exstew Confluence to 20 Camp Wanahoot China Bar 21 Gitnadoix Bar Mile Bar (bottom of Andesite) 23 Andesite Bar (river right) X 24 Exchamsiks Mouth (river left) 25 Salvus Bar (river right) 26 Kasiks River (Snowbound) X 27 China Bar X LGL Limited Page 5

14 Table 2. Summary of scheduled sampling effort (number of interviewer shifts by shore-based survey site and for boat surveys) by month for AM and PM strata and weekend/holiday (WE) and weekday (WD) strata. Boat Survey Shore Survey Kwinista Kalum Terrace China Bar Grand Month Daytype AM PM AM PM Total AM PM AM PM Total Total * June WD WE July WD WE August WD WE September WD WE * Does not include two training days per surveyor that occurred 2-3 June, change, it made sense to treat the first week of August as part of July for the purposes of analysis. In this report, estimates for catch and effort in August will be based on the period 8-31 August. Angler Activity Patterns Two weighting factors were used together with the interview-derived angling activity data to estimate the daily fishing activity pattern (English et al. 2002). The first weighting factor, W1, expanded the numbers of days spent interviewing in each river stratum, to account for the total number of days available for sampling. That is, it was assumed that the daily activity pattern recorded during the interview shifts in river stratum s, were consistent for river stratum s, even during the days when no interviews occurred. A specific W1 was calculated for each river stratum during each month and day type: Nmd W1mds = (Eqn. 1) Kmds where N md was the total number of type d days in month m; and K mds was the number of days during which interviews occurred in river stratum s, on type d days during month m. The second weighting factor, W2, expanded the numbers of interviews conducted, to account for the anglers that were not interviewed. That is, it was assumed that the activity pattern recorded during the interview shifts also held for those anglers that were not interviewed. A specific W2 was calculated for each surveyor (u) operating on each surveying date (k) in each river stratum during each month and day type: LGL Limited Page 6

15 W L =, (Eqn. 2) mdsku 2mdsku Amdsku where L mdsku was the number of anglers observed and A mdsku was the number of anglers interviewed by surveyor u, during surveying date k, in river stratum s, during day type d, and month m. We used the term A mdsfkuqt to denote the number of anglers reporting activity during time-block t, that were part of the fishing party (q) that was interviewed by surveyor u on survey date k, in river stratum s, with access method f, during month m, and on day type d (n mdsfkq was used to denote the total number of anglers that were part of that fishing party). The two correction factors were applied, and the data were summed over surveyors, survey dates and fishing parties (within month, day type, stratum, access method and time-block): A' mdsft = W1mds ååå ( W2mdsku Amdsfkuqt ). (Eqn. 3) k u q Summing the adjusted number of anglers over the 16 time-blocks gave: T' mdsf =å A' mdsft. (Eqn. 4) The proportion of anglers (P mdsft ) that were active during in each of 16 hourly time-blocks (t) was calculated for each month, day type, river stratum and access method: A' mdsft Pmdsft =. æ ö (Eqn. 5) ç W1mds åå( W2mdsk ) nmdsfkq è k q ø For this calculation, current day activity was included only if the anglers said their trip was finished for the day. Regardless, prior day activity was included in the analyses, being careful to assign the data to the correct temporal categories. For example, if an interview was conducted on a Monday, the prior day activity data would be counted under day type = weekend. It should be noted that the ratio of interviewed-to-not-interviewed anglers was not known for the day prior to the interview, thus W2 weights were assigned a value of 1 when processing prior day activity data. Using this method, 48 unique angler activity patterns were to be estimated (i.e., 4 months 2 day types 3 river strata 2 access methods, see Figure 2, Figure 3 and Figure 4). To reliably describe angler activity, a relatively large number of anglers (~ 60) needed to be interviewed in each of the 48 blocks. In the end, some blocks contained too few interviews (Table 3), so it was decided to pool activity data over access method (too little boat angler data in River Stratum Two and Three), and to group the monthly data into two episodes, e (Episode One: 1 June to 7 Aug; and Episode Two: 8 Aug to 30 Sept). The equation for angler activity was thus ååa' mdsft m f Pedst =, æ ö (Eqn. 6) ç W1mds W2mdsk nmdsfkq åå t åå( ) m f è k q ø For which the data were only summed over the months that were part of the episode in question. The associated variance was: LGL Limited Page 7

16 2 ( Pdst)(1 -Pdst ) SP =. edst æ ö ååç W1mds åå( W2mdsk nmdsfkq ) m f è k q ø (Eqn. 7) The average number of hours fished per angler (G eds ) was calculated for each episode / day type / river stratum combination using weighted observations: åå( T' mdsf ) m f Geds =, æ ö (Eqn. 8) ç W1mds W2mdsk nmdsfkq åå åå( ) m f è k q ø but the variance was calculated from the raw interview data (rather than from the weighted values) using the standard formula. The same two adjustments (W1 and W2) were applied to the angler demographic data. Table 3. The amount of data (number of anglers) available to estimate angler activity patterns, for all levels of each factor (data from June to September 2014). Access River Stratum Month Day Type Method One Two Three June Weekday Boat Shore Weekend Boat Shore July Weekday Boat Shore Weekend Boat Shore August Weekday Boat Shore Weekend Boat Shore September Weekday Boat Shore Weekend Boat Shore LGL Limited Page 8

17 Proportion of anglers active 100% 80% 60% 40% 20% 0% June (n=594) 100% August (n=816) 80% 60% 40% 20% 0% Proportion of anglers active 100% 80% 60% 40% 20% 0% July (n=1560) < 7 AM 7-8 AM 8-9 AM 9-10 AM AM 11-noon noon-1 PM 1-2 PM 2-3 PM 3-4 PM 4-5 PM 5-6 PM 6-7 PM 7-8 PM 8-9 PM > 9 PM 100% 80% 60% 40% 20% 0% September (n=314) < 7 AM 7-8 AM 8-9 AM 9-10 AM AM 11-noon noon-1 PM 1-2 PM 2-3 PM 3-4 PM 4-5 PM 5-6 PM 6-7 PM 7-8 PM 8-9 PM > 9 PM Time block Time block Figure 2. Angler activity patterns, by month, from interview data collected from June to September LGL Limited Page 9

18 Proportion of anglers active 100% 80% 60% 40% 20% 0% S1 (n=1961) 100% S2 (n=442) 100% Weekday (n=1697) Proportion of anglers active 80% 60% 40% 20% 0% Proportion of anglers active 80% 60% 40% 20% 0% 100% S3 (n=881) 100% Weekend/Holiday (n=1587) Proportion of anglers active 80% 60% 40% 20% 0% < 7 AM 7-8 AM 8-9 AM 9-10 AM AM 11-noon noon-1 PM 1-2 PM 2-3 PM 3-4 PM 4-5 PM 5-6 PM 6-7 PM 7-8 PM 8-9 PM > 9 PM Proportion of anglers active 80% 60% 40% 20% 0% < 7 AM 7-8 AM 8-9 AM 9-10 AM AM 11-noon noon-1 PM 1-2 PM 2-3 PM 3-4 PM 4-5 PM 5-6 PM 6-7 PM 7-8 PM 8-9 PM > 9 PM Time block Time block Figure 3. Angler activity patterns, by river stratum (left column) and day type (right column) from interview data collected from June to September LGL Limited Page 10

19 Proportion of anglers active 100% 80% 60% 40% 20% 0% Boat (n=164) < 7 AM 7-8 AM 8-9 AM 9-10 AM AM 11-noon noon-1 PM 1-2 PM 2-3 PM 3-4 PM 4-5 PM 5-6 PM 6-7 PM 7-8 PM 8-9 PM > 9 PM 100% 80% 60% 40% 20% 0% Shore (n=3120) < 7 AM 7-8 AM 8-9 AM 9-10 AM AM 11-noon noon-1 PM 1-2 PM 2-3 PM 3-4 PM 4-5 PM 5-6 PM 6-7 PM 7-8 PM 8-9 PM > 9 PM Time block Time block Figure 4. Angler activity patterns, by access method, from interview data collected from June to September Catch Per Effort Estimation Catch per effort (and, similarly, harvest per effort) was estimated for each species of fish from interviews of anglers. For each interview (i), the month (m), day type (d) and access method (f) was recorded, along with the catch (C) of each species (r), the number of anglers (A), and the number of hours spent fishing (H) in each river stratum (s). Using these data, catch per effort was calculated as: Cmdsfri CPEmdsfri =. (Eqn. 9) A H Ideally, mean CPE would have been calculated for each month, river stratum, day type, access method and species. However, too few anglers were available to be interviewed to provide adequate sample size (n ~3) to reliably estimate CPE and its variance for each of the 48 blocks (Table 4). As CPE was expected to change with month, river stratum and access method, it was decided to pool interview data by day type. In most cases, mean CPE was calculated by summing the catch for all n mdsf interviews, pooling over day type, and dividing by the total number of angler-hours of fishing effort recorded for these interviews: mdsfi mdsfi LGL Limited Page 11

20 CPE ˆ msfr = n d i= 1 mdsf åå( Amdsfi Hmdsfi ) d i= 1 nmdsf åå C mdsfri The variance for the estimate of mean catch per effort was calculated as: S 2 CPE ˆ msfr = nmdsf å i= 1 ( CPE ˆ )- nmdsf 2 i= 1 msfri ( n -1) mdsf å ( CPE ˆ ) n mdsf msfri. (Eqn. 10) 2 (Eqn. 11). In several instances, the month/access method/river stratum-specific sample size was too low, even after the data were pooled over day type (Table 4). Two modifications were made. First, in both June and July (includes the first week of August), sample sizes were limited for boat anglers in Strata 2 and 3, thus the month-specific boat-based CPE from Stratum 1 was applied to all areas. Second, CPE data were pooled over access method in late August and September due to low interview counts obtained from boats after 7 Aug. The CPE of fish that were harvested was calculated by repeating the creel analyses with released fish excluded from the interview database. Angler Effort Estimation To obtain statistically valid estimates of angler effort, anglers were counted during surveys conducted from a boat traveling through the study area. The study area was split into two subsections, the first comprising River Strata One and Two, and the second being equivalent to River Stratum Three. On any given survey day, anglers were enumerated in one of the two study area subsections. Table 2 shows the number of boat surveys scheduled for each month and day type. Each survey was supposed to cover the entire subarea with the start and end times recorded for each of the River Strata surveyed (Appendix 1). The schedule included surveys on every weekend day and usually three of the five weekdays each week. During survey o (conducted during month m and on day type d), observers tallied the total number of anglers (boating and shore-based counted separately, f) that were actively fishing at time t in sub-stratum u (within river stratum s), V mdsfout. These tallies were pooled by substratum. Since angling occurs over the course of the entire day, the number of anglers that were observed at the moment of the survey was divided by the proportion of average daily number of shore and boat-based anglers active (P edst ) during the time block when the observations were recorded, and multiplied by the average number of hours fished per angler (G esd ). These adjusted tallies were summed over the duration of the survey, to calculate the total number of anglerhours of fishing on the day of the survey, by river stratum and access method, B mdsfo : B mdsfo å æ Vmdsfout ö u = ç G å esd. t ç P (Eqn. 12) esdt è ø LGL Limited Page 12

21 Table 4. The sample size of angler CPE data (i.e., the number of interviewed parties reporting catch and effort) for each river stratum, and for each month, day type, and access method (data from June to September 2014). River Stratum Month Day Type Access June Weekday Boat Shore Weekend Boat Shore July - 7 Aug Weekday Boat Shore Weekend Boat Shore Aug Weekday Boat Shore Weekend Boat Shore September Weekday Boat Shore Weekend Boat Shore These estimates were then averaged over the number of surveys conducted, n mds, as: Bˆ mdsf = n mds å o= 1 B mdsfo. (Eqn. 13) nmds Total monthly fishing effort, was calculated for each day type, river stratum and access method by multiplying the average daily effort by the number days of day type d that occurred in month m: E = Bˆ N. (Eqn. 14) mdsf mdsf md The variance of B mdsfo was calculated using the standard formulas for combining the variance of products and quotients of two independent random variables (Goodman 1960): if z = x / y, Var( z) = ( y ) Var( x) + ( x y ) Var( y). (Eqn. 15) 2 2 if z = xy, Var( z) = ( y ) Var( x) + ( x ) Var( y) LGL Limited Page 13

22 Thus, S / = ( P ) S + ( V P ) S and V P esdt V esdt P esdt æ V ö (Eqn. 16) 2 SB = ( Gesd) SV / P + ç SG P esd, è esdt ø where the variance of the observed angler counts S 2 V was calculated from the raw data as: nmds æ ( ) 2 ö nmds ç å Vmdsfo 2 o= 1 åvmdsfo -ç o= 1 ç nmds (Eqn. 17) ç 2 é Nmd - nmds ù SV = è ø mdsf ê ú. nmds -1 ë Nmd -1 û The variance of the estimate of the total monthly fishing effort was: SE = S ˆ Nmd (Eqn. 18) mdsf The standard error of the estimate of the total monthly fishing effort, after pooling over day types, was: S Emsf d Bmdsf 2 SE mdsf = å (Eqn. 19) n mds Catch Estimation Total catch was calculated for each month, river stratum and species by multiplying total angling effort by catch per effort, and then summing over day type and access method: C ( ˆ msr = åå Emdsf CPEmsfr ). (Eqn. 20) f d The standard errors for these catch estimates were derived using the Goodman (1960) equation: æ S 2 CPE S msfr 2 E S mdsf CPE S ö msfr Emdsf SC = msr åå Emdsf + CPEmsfr +. ç f d nmsf nmdsf nmsf n (Eqn. 21) è mdsf ø Because the Chinook salmon fishery above Lakelse was closed on 8 August, it made the most sense to include the first week of August in with July, and to treat the last three weeks of August separately. As a result, it was necessary that comparisons of catch among months, river strata and access method be made using catch per day in lieu of catch per month. Catch per day was calculated by diving the monthly catch estimates by the number of days included in each period (30, 38, 24, and 30 days for the four months, respectively). To estimate the monthly number of fish that were harvested, the creel analyses were repeated with released fish excluded from the interview database. LGL Limited Page 14

23 Statistical Analyses Comparisons of CPE, effort, catch or harvest by month, stratum, or access method were performed within each fish taxa using Kruskal-Wallis tests (Zar 1984). For these tests, the Bonferroni adjustment was used to control experiment-wise alpha levels within each species (e.g., to look for effects of month, river stratum and access method on Coho CPE, three Kruskal- Wallis tests would be performed, thus the adjusted alpha would be 0.05/3 = 0.017). When statistically significant effects were observed, post-hoc Kruskal-Wallis tests were used to compare between each possible pair, where the false discovery rate was controlled using the Benjamini Hochberg procedure (Benjamini and Hochberg 1995). RESULTS Angler Interviews Over the four month study period, 4,263 anglers were questioned during 1,842 interviews. Of the 4,263 anglers interviewed, 1,120 (26.3%) reported on completed fishing trips, and 2,164 (50.8%) reported their previous-day s fishing activity. Interviews were conducted during 144 survey shifts (Table 5, Appendix 2). Completed survey shifts represented 97% of the scheduled survey effort (99% of the shore-based effort, and 96% of the boat-based effort). The actual Table 5. Summary of observed sampling effort (number of interviewer shifts by shore-based survey site and for boat surveys) by month for AM and PM strata and weekend/holiday (WE) and weekday (WD) strata. Boat Survey Shore Survey Kwinista Kalum Terrace China Bar Grand Month Daytype AM PM AM PM Total AM PM AM PM Total Total 1,2 June WD WE July WD WE August WD WE September WD WE Total % of scheduled 99% 96% 97% 1 In September, one boat-survey day was moved to an end of season for a clean-up day. 2 In September, one shore-survey day was moved to an end of season for a clean-up day and two shore days were moved to assist boat crew that were short staffed. LGL Limited Page 15

24 survey schedule included four fewer days than the proposed schedule (144 vs 148). In September, two survey days were moved to the end of the season for project clean-up, as well as, two shore-survey days were cancelled so that shore surveyors could assist the short-staffed boat crew. Angler Demographics and Behaviours A minority (2%) of anglers were with a professional guide, although the percentage increased over the course of the summer from 0.9% in June to 8.7% in September (Figure 5). Anglers were overwhelmingly (89%) from British Columbia, including 52% from the Skeena area and 37% from the rest of the province. The portion of anglers from outside the province grew over time, ranging from 6% in June to 25% in September (Figure 5). Overall, 90% of non-guided anglers were BC residents, whereas only 42% of guided anglers were from BC. Throughout the summer, the majority (82%) of anglers were still fishing (Table 6), but the method was most popular in June and July (93% of anglers). Fly casting increased in popularity over the summer, ranging from 3.1% in June, and increasing to 27% by September (Table 6). Bottom bouncing was never very common, but was most popular in August (14% of anglers). Based on responses given during interviews, only about half of one percent of anglers were backtrolling, all during the month of June. Because of concerns about the fishing method, surveyors made note of all back-trolling activity that was observed during their shifts. In all, 12 boats were observed to be back-trolling (4 in June, 5 in July, and 3 in August). Note that these counts are likely under- estimates of the actual 100% Percent of Anglers Interviewed 80% 60% 40% 20% Overseas US Rest of Canada Rest of BC Skeena Percent Guided 0% June July August September Month Figure 5. The monthly percent of interviewed anglers that were with a professional guide (line) and their location of residence (bars), categorized as either being from the local area (Skeena), from anywhere in the rest of British Columbia, the rest of Canada, the United States, or locations overseas. LGL Limited Page 16

25 amount of back-trolling that occurred, as no effort has been made to account for days when no surveyors were working, etc. Clearly, the amount of back-trolling was under-reported during the creel interviews, according to which no back-trolling occurred at all after the month of June. Most anglers (83%) used spin and glow lures, either with or without bait The popularity of flies increased over the summer along with the increase in the fly-fishing method (Table 6). The majority of anglers (Figure 6) were targeting Chinook salmon (62%), followed by Coho salmon (13%), Sockeye salmon (6.8%) and Pink salmon (0.5%). Only 5.0% of anglers said they were targeting Steelhead, and of these, about half were simultaneously targeting one or more salmon species. 9.5% of anglers said they were targeting multiple species of salmon, but not Steelhead. No anglers said they were targeting trout. Table 6. Fishing method and gear types used by interviewed anglers, by month. Month # of Anglers Interviewed Fly Cast Bottom Bounce Fishing Method Spin Cast Still Other June % 0.5% 3.2% 93% 0.3% July % 1.3% 0.6% 93% 0.0% August % 14.0% 5.2% 66% 0.0% September % 5.8% 9.9% 57% 0.0% Gear Type Month Fly Spin+Glow Spin+Glow+Bait Spinners Spoon Wool Other June 3.1% 57% 38% 0.0% 1.1% 0.2% 0.7% July 5.2% 62% 32% 0.0% 0.7% 0.2% 0.6% August 15% 49% 18% 0.3% 4.7% 8.4% 5.0% September 27% 49% 10% 0.0% 8.6% 2.5% 2.0% No target Steelhead Steelhead + salmon Coho salmon Chinook salmon Pink salmon Sockeye salmon Several salmon Figure 6. The target species, as declared by interviewed anglers in June to September Proportions of anglers targeting steelhead are shown in green, blue and purple, offset from the remainder of the pie. LGL Limited Page 17

26 Angler Activity Patterns As described in the Methods Section (above), low sample sizes required interview data to be pooled across access methods and for months to grouped into two episodes (1 June to 7 Aug; and 8 Aug to 30 Sept), resulting in the 12 angler activity patterns estimates (2 day types 3 river strata 2 episodes) shown in Figure 7 and Figure 8. Angler activity patterns from 1 June to 7 Aug differed markedly from those between 8 Aug and 30 Sept. As day length declined over the study period, activity patterns in the second Episode declined in the earliest and latest hours of the day and there was greater midday activity. Catch Per Effort Estimates In order to obtain adequate sample sizes for CPE estimation, interview data were pooled over day type. In most cases, the pooled number of interviews was >= 3 (Table 4). Two exceptions were made. First, in both June and July (including the first week of August), sample sizes precluded reliable estimation of CPE for boat anglers in Strata 2 and 3, thus the month-specific boat-based CPE from Stratum 1 was applied to all areas. Second, CPE data were pooled over access method in late August and September due to low interview counts obtained from boats after 7 Aug. After pooling (as described above), CPE estimates were calculated for each species by month, river stratum and access method (Table 7). Species included in the other fish category included Chum salmon (Oncorhynchus keta), Rainbow trout (O. mykiss), Cutthroat trout (O. clarkii), Dolly Varden (Salvelinus malma), Bull trout (S. confluentus), and sculpins (Cottus sp.) that were not identified to species, where several species are referred to collectively and colloquially as bullhead. Month appeared to have had a significant effect on CPE of Coho Salmon (Table 7). Indeed, median Coho salmon CPE was higher in September (0.087) than it was in August (0.024) or in June or July (0 to fish per angler hour; Figure 9), but none of the post-hoc pairwise comparisons were statistically significant after controlling the false discovery rate (i.e., using the Benjamini Hochberg procedure). For Chinook salmon highest CPEs were early in the summer, whereas for Pink salmon and Steelhead, CPE peaked later in the summer. However, sample size limitations (n =12) meant that statistical tests were conservative, and among-month differences were not statistically significant after the Bonferroni adjustment (Table 8). There were no significant effects of river stratum or access method on CPE for any species (Table 8). The retention per effort rates (i.e., the rate of fish harvest, with units of fish per unit effort, where released fish were excluded) are shown for each species by month, river stratum and access method in Table 9. Steelhead, Pink salmon, and other fish were infrequently kept. Given the no-harvest fishing regulation, Chinook salmon caught in the Ferry Island to Lakelse stratum after 6 August were never retained. LGL Limited Page 18

27 Proportion of anglers active 100% 80% 60% 40% 20% 0% Stratum 1, Weekdays (n = 792) 100% Stratum 2, Weekdays (n = 121) 100% Stratum 3, Weekdays (n = 357) June 1 - Aug 7 June 1 - Aug 7 June 1 - Aug 7 80% 80% 60% 60% 40% 40% 20% 20% 0% 0% Proportion of anglers active 100% 80% 60% 40% 20% 0% Stratum 1, Weekends (n = 737) 100% Stratum 2, Weekends (n = 148) 100% Stratum 3, Weekends (n = 249) June 1 - Aug 7 June 1 - Aug 7 June 1 - Aug 7 80% 80% < 7 AM 7-8 AM 8-9 AM 9-10 AM AM 11-noon noon-1 PM 1-2 PM 2-3 PM 3-4 PM 4-5 PM 5-6 PM 6-7 PM 7-8 PM 8-9 PM > 9 PM 60% 40% 20% 0% < 7 AM 7-8 AM 8-9 AM 9-10 AM AM 11-noon noon-1 PM 1-2 PM 2-3 PM 3-4 PM 4-5 PM 5-6 PM 6-7 PM 7-8 PM 8-9 PM > 9 PM 60% 40% 20% 0% < 7 AM 7-8 AM 8-9 AM 9-10 AM AM 11-noon noon-1 PM 1-2 PM 2-3 PM 3-4 PM 4-5 PM 5-6 PM 6-7 PM 7-8 PM 8-9 PM > 9 PM Time block Time block Time block Figure 7. Angler activity patterns, by day type and river stratum, from interview data collected from 1 June to 7 August LGL Limited Page 19

28 Proportion of anglers active Proportion of anglers active 100% 80% 60% 40% 20% 0% 100% 80% 60% 40% 20% Stratum 1, Weekdays (n = 196) 100% Stratum 2, Weekdays (n = 103) 100% Stratum 3, Weekdays (n = 88) Aug 8 - Sep 30 Aug 8 - Sep 30 Aug 8 - Sep 30 80% 80% 60% 40% 20% 0% 0% Stratum 1, Weekends (n = 173) 100% Stratum 2, Weekends (n = 44) 100% Stratum 3, Weekends (n = 124) Aug 8 - Sep 30 Aug 8 - Sep 30 Aug 8 - Sep 30 80% 80% 60% 40% 20% 60% 40% 20% 60% 40% 20% 0% < 7 AM 7-8 AM 8-9 AM 9-10 AM AM 11-noon noon-1 PM 1-2 PM 2-3 PM 3-4 PM 4-5 PM 5-6 PM 6-7 PM 7-8 PM 8-9 PM > 9 PM 0% < 7 AM 7-8 AM 8-9 AM 9-10 AM AM 11-noon noon-1 PM 1-2 PM 2-3 PM 3-4 PM 4-5 PM 5-6 PM 6-7 PM 7-8 PM 8-9 PM > 9 PM 0% < 7 AM 7-8 AM 8-9 AM 9-10 AM AM 11-noon noon-1 PM 1-2 PM 2-3 PM 3-4 PM 4-5 PM 5-6 PM 6-7 PM 7-8 PM 8-9 PM > 9 PM Time block Time block Time block Figure 8. Angler activity patterns, by day type and river stratum, from interview data collected from 8 August to 30 Sept LGL Limited Page 20

29 Table 7. Catch per effort (CPE) estimates (fish per angler-hour) for six fish taxa (with large and jack Chinook reported separately), by month, river stratum and access method. Variance in parentheses. Access Chinook salmon Coho Sockeye Pink Month Method River Stratum large jacks salmon salmon salmon Steelhead Other June Shore Ferry Island to Lakelse (0.018) (0.004) (0.000) (0.000) (0.002) Lakelse to Exstew (0.004) (0.034) (0.002) Exstew to China Bar (0.001) (0.001) (0.001) Boat Ferry Island to Lakelse (0.004) (0.004) Lakelse to Exstew (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) Exstew to China Bar (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) 1 July - 7 Aug Shore Ferry Island to Lakelse (0.002) (0.003) (0.003) (0.050) (0.041) (0.000) (0.000) Lakelse to Exstew (0.006) (0.170) (0.007) (0.003) (0.000) (0.001) (0.031) Exstew to China Bar (0.001) (0.003) (0.001) (0.002) (0.043) (0.003) 0 Boat Ferry Island to Lakelse (0.003) (0.001) (0.000) (0.000) (0.000) 0 0 Lakelse to Exstew (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) Exstew to China Bar (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) 8-31 Aug Shore Ferry Island to Lakelse (0.000) (0.000) (0.019) (0.121) (1.393) (0.008) (0.001) Lakelse to Exstew (0.002) (0.009) (0.023) (0.002) (0.312) (0.143) (0.001) Exstew to China Bar (0.000) (0.001) (0.010) (0.008) (0.011) (0.004) (0.000) Boat Ferry Island to Lakelse (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) Lakelse to Exstew (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) Exstew to China Bar (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) September Shore Ferry Island to Lakelse (0.015) (0.023) (0.022) (0.027) (0.006) Lakelse to Exstew (0.036) (0.000) (0.009) (0.002) (0.000) Exstew to China Bar (0.019) (0.002) (0.002) (0.001) (0.000) Boat Ferry Island to Lakelse (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) Lakelse to Exstew (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) Exstew to China Bar (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) LGL Limited Page 21

30 Figure 9. Box-whisker plots showing the distribution of shore-access CPE estimates for each month, by taxon (with large and jack Chinook reported separately). Letters indicate statistically significant post-hoc comparisons among months (i.e., months that share a letter in common are not significantly different). On the X-axis, months are shown using a numeric shorthand as follows: 6 = June; 7 = 1 July - 7 August; 8 = 8-31 August; and 9 = September. Monthly samples sizes were 3. LGL Limited Page 22

31 Table 8. Statistical tests of the effect of month, river stratum, and access method on median catch per effort (CPE) estimates for the 6 taxa surveyed (with large and jack Chinook reported separately). P-values that are underlined are less than 0.05, but only those in bold are statistically significant after the Bonferroni adjustment. Month (Shore anglers) River Stratum (Shore anglers) Access Method (Stratum 1, Jun-Jul) Taxon χ3 2 P χ2 2 P χ1 2 df P Chinook, large Chinook, jacks Coho Pink Sockeye Steelhead Other LGL Limited Page 23

32 Table 9. Retention per effort (CPE of harvested fish) rates for six fish taxa (with large and jack Chinook reported separately), by month, river stratum and access method. Variance in parentheses. Access Chinook salmon Coho Sockeye Pink Month Method River Stratum large jacks salmon salmon salmon Steelhead Other June Shore Ferry Island to Lakelse (0.018) (0.004) (0.000) (0.000) 0 Lakelse to Exstew (0.004) (0.016) Exstew to China Bar (0.001) (0.001) (0.001) Boat Ferry Island to Lakelse (0.001) (0.002) Lakelse to Exstew (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) Exstew to China Bar (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) 1 July - 7 Aug Shore Ferry Island to Lakelse (0.002) (0.003) (0.003) (0.048) (0.001) (0.000) (0.000) Lakelse to Exstew (0.006) (0.134) (0.006) (0.003) (0.000) 0 0 Exstew to China Bar (0.001) (0.002) (0.000) (0.002) (0.002) 0 0 Boat Ferry Island to Lakelse (0.002) (0.001) Lakelse to Exstew (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) Exstew to China Bar (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) (see Stratum 1) 8-31 Aug Shore Ferry Island to Lakelse (0.008) (0.066) (0.002) (0.000) (0.000) Lakelse to Exstew (0.002) (0.009) (0.015) (0.000) (0.001) 0 0 Exstew to China Bar (0.000) (0.000) (0.002) (0.008) (0.001) (0.000) (0.000) Boat Ferry Island to Lakelse (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) Lakelse to Exstew (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) Exstew to China Bar (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) September Shore Ferry Island to Lakelse (0.007) (0.001) (0.004) 0 0 Lakelse to Exstew (0.003) Exstew to China Bar (0.013) (0.001) (0.000) Boat Ferry Island to Lakelse (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) Lakelse to Exstew (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) Exstew to China Bar (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) (see Shore) LGL Limited Page 24

33 Angler Effort Estimates Over the 4 month study period, a total of 73 effort surveys were conducted, covering River Stratum One, Two and Three a total of 37, 37 and 36 times, respectively. Of the 73 surveys, 42 were conducted on weekdays, and 31 on weekend/holidays (Appendix 3). The total angling effort was estimated for each month, day type, river stratum and access method (Table 10). There was a very strong statistically significant effect of access method (Table 11), whereby shore-based angling effort (median angler hours per day) far outstripped that of boatbased anglers (median 29.5 angler hours per day; Figure 10). There were no significant effects of month, day type or river stratum on fishing effort (Table 11). Table 10. Effort estimates (angler-hours per month), by month, day type, river stratum and access method. Standard errors in parentheses. River Stratum Month Day Type Access Method Ferry Island to Lakelse Lakelse to Exstew Exstew to China Bar Total June Weekday Boat 2,786 (1,377) (83) 2,906 (1,380) Shore 17,226 (9,482) 3,659 (2,080) 4,486 (2,662) 25,371 (10,066) Weekend Boat 1,394 (1,082) 97 (82) 0 1,491 (1,085) Shore 7,689 (8,492) 2,698 (2,558) 4,689 (3,117) 15,076 (9,401) 1 July - 7 Aug Weekday Boat 5,558 (2,863) 309 (187) 329 (167) 6,196 (2,873) Shore 32,096 (15,963) 8,163 (4,763) 24,379 (9,712) 64,639 (19,283) Weekend Boat 3,597 (1,908) 193 (118) 52 (29) 3,843 (1,912) Shore 15,899 (7,960) 5,892 (3,346) 13,681 (5,150) 35,472 (10,054) 8-31 Aug Weekday Boat 88 (55) (55) Shore 8,499 (4,358) 3,685 (1,957) 12,783 (10,071) 24,967 (11,147) Weekend Boat 44 (29) 212 (119) 126 (99) 382 (157) Shore 4,426 (2,428) 2,997 (1,635) 7,584 (5,466) 15,007 (6,200) September Weekday Boat 144 (103) 837 (548) 78 (58) 1,059 (560) Shore 5,288 (3,399) 3,347 (2,087) 10,553 (7,448) 19,189 (8,449) Weekend Boat (0) Shore 3,100 (2,514) 1,478 (1,145) 10,366 (11,427) 14,943 (11,756) Overall Total 107,834 (23,19233,567 (7,569) 89,226 (21,332) 230,627 (32,407) Table 11. Statistical tests of the effect of month, day type, river stratum and access method on median daily effort estimates during the study period. Bold P-values that are statistically significant after the Bonferroni adjustment. Effect Test χ 2 df P Month Day Type River Stratum Access Method < LGL Limited Page 25

34 Figure 10. Box-whisker plots showing the distribution of daily effort estimates for each access method. Letters indicate statistically significant comparisons between access methods (i.e., access methods that share a letter in common are not significantly different). Samples size were 24 for each access method. Catch Estimates Estimates of total monthly catch (Table 12) were generated by calculating E CPE, and then summing over day types and access methods. Sockeye and pink salmon were the species caught in the highest numbers (sockeye: 5,787, SE = 759; pink: 5,045, SE = 1,224). Coho salmon was the next most commonly caught species (4,119, SE = 701), followed by Steelhead (3,031, SE = 428). Chinook salmon catches totaled 2,888 large fish (SE = 303) and 2,116 jacks (SE = 459). Catch of other fish species summed to 450 individuals (SE = 195). The strongest effects of month on catch per day was for Chinook (both for large fish and for jacks), Coho and Pink salmon (Table 13). Large Chinook salmon were caught more frequently from June to July (median 4.3 to 7.3 fish per day), when 98% of the overall adult Chinook salmon catch was caught. Few were caught after the first week of August (median 0.2 fish per day), and none were caught in September (Table 12). Catches of Chinook jacks followed a similar temporal pattern to those of the large Chinook salmon, with 96% of the total catch made in June, July and the first week of August (Table 12, Figure 11). By comparison, catch rates of Coho salmon increased over time, ranging from 0 fish in June to 14 fish per day in September (Table 12, Figure 11), with 88% of the overall Coho salmon catch made after the first week of August. Pink salmon catch peaked in the last three weeks of August, when 50% of the speciesspecific catch was made (median catch of 13.2 fish per day; Table 12, Figure 11). Total daily catches resulting from shore-based angling were greater than those resulting from boat-based angling (Figure 12). The effect was statistically significant for Sockeye salmon, Steelhead and other fish (Table 13). The low overall catch resulting from boat-based angling is likely due to the fact that shore-based angling effort far outstripped that of boat-based angling (Figure 10, Table 10), especially since there was not a significant difference in CPE between access methods (Table 8). LGL Limited Page 26

35 For no species did catch per day vary significantly with river stratum (Table 13). Species-specific monthly catches from are in Appendix 4. Table 12. Estimated catch (harvest + release) of six fish taxa (with large and jack Chinook reported separately) in three geographic strata, by month. Catches are rounded to the closest whole number. Standard errors in parentheses. Chinook salmon Coho Sockeye Pink Month River Stratum large jacks salmon salmon salmon Steelhead Other June Ferry Island to Lakelse 580 (204) 412 (113) 0 37 (23) 0 25 (26) 62 (59) Lakelse to Exstew 191 (69) 103 (137) (34) Exstew to China Bar 229 (55) 14 (22) (26) 1 July - 7 Aug Ferry Island to Lakelse 942 (158) 785 (141) 173 (89) 1627 (428) 841 (339) 86 (33) 47 (27) Lakelse to Exstew 379 (93) 329 (377) 158 (81) 22 (46) 50 (13) 34 (23) 27 (159) Exstew to China Bar 518 (89) 396 (106) 150 (49) 185 (83) 839 (391) 546 (115) Aug Ferry Island to Lakelse 10 (3) 10 (11) 274 (123) 2488 (521) 1146 (983) 578 (126) 59 (25) Lakelse to Exstew 28 (30) 35 (62) 168 (101) 28 (29) 770 (381) 336 (249) 7 (19) Exstew to China Bar 11 (6) 33 (37) 494 (213) 230 (151) 614 (248) 384 (156) 44 (17) Sept Ferry Island to Lakelse (217) 797 (240) 587 (193) 744 (233) 63 (60) Lakelse to Exstew (235) 15 (4) 31 (71) 107 (38) 15 (9) Exstew to China Bar (549) 357 (173) 167 (119) 190 (99) 71 (41) Overall Total 2888 (303) 2116 (459) 4119 (701) 5787 (759) 5045 (1224) 3031 (428) 450 (195) Table 13. Statistical tests of the effect of month, river stratum and access method on median catch-per-day (harvest + release) estimates for the 6 taxa surveyed (with large and jack Chinook reported separately). P-values that are underlined are less than 0.05, but only those in bold are statistically significant after the Bonferroni adjustment. Month River Stratum Access Method Species χ3 2 P χ2 2 P χ1 2 P Chinook, large Chinook, jacks Coho Pink Sockeye Steelhead Other LGL Limited Page 27

36 Figure 11. Box-whisker plots showing the distribution of catch-per-day estimates (harvest + release) for each month, by taxon (with large and jack Chinook reported separately). Letters indicate statistically significant post-hoc comparisons among months (i.e., months that share a letter in common are not significantly different). Months are shown using a numeric shorthand as follows: 6 = June; 7 = 1 July - 7 August; 8 = 8-31 August; and 9 = September. Samples sizes were 6 per taxon per month. LGL Limited Page 28

37 Figure 12. Box-whisker plots showing the distribution of catch-per-day estimates (harvest + release) by access method for each taxon (with large and jack Chinook reported separately). Letters indicate statistically significant comparisons between access methods months (i.e., access methods that share a letter in common are not significantly different). Samples sizes were 12 per taxon per access method. LGL Limited Page 29

38 Harvest (Retention) Estimates Estimates of total monthly harvest are shown in Table 14 and Figure 13. Sockeye salmon which was the most abundantly caught fish, was also the most retained species (4,249 fish harvested, SE = 616; 73% of fish retained). Pink salmon, though often caught, was infrequently (646 fish kept, SE = 126) retained (13% retention). Coho salmon were retained 57% of the time, with a total of 2,360 (SE = 502) fish harvested. Chinook salmon were the most likely species to be harvested, taken 87-93% of the time. Consistent with the fishing regulations, no Chinook salmon were harvested in Stratum One (Ferry Island to Lakelse) after 7 August. Steelhead was the least frequently harvested species, with negligible numbers of fish harvested (2% of total catch). These results are not surprising given the no-retention regulations for Steelhead. There were statistically significant effects of month on Chinook and Coho salmon harvest per day (Table 15, Figure 13). The species-specific harvest patterns followed the catch patterns described above. There were no other statistically significant effects of month or river stratum (Table 15). Similar to the effect observed for catch (kept + released) rates, the total daily harvest (kept only) attributed to shore-based angling was greater than that from boat-based angling (Figure 14). The effect was observed for all species with non-negligible harvest (Figure 12), and was statistically significant for Coho and Pink salmon (Table 15). Table 14. Estimated harvest of six fish taxa (with large and jack Chinook reported separately) in three geographic strata, by month. Numbers are rounded to the closest whole number. Standard errors in parentheses. Chinook salmon Coho Sockeye Pink Month River Stratum large jacks salmon salmon salmon Steelhead Other June Ferry Island to Lakelse 557 (200) 364 (103) 0 37 (23) 0 12 (23) 0 Lakelse to Exstew 190 (68) 91 (95) Exstew to China Bar 208 (49) 13 (22) (22) 1 July - 7 Aug Ferry Island to Lakelse 832 (144) 699 (132) 94 (83) 1431 (405) 118 (51) 8 (5) 8 (22) Lakelse to Exstew 370 (92) 279 (335) 144 (75) 21 (46) 14 (8) 0 0 Exstew to China Bar 495 (83) 337 (84) 92 (36) 170 (83) 163 (76) Aug Ferry Island to Lakelse (80) 1959 (403) 167 (50) 10 (5) 10 (6) Lakelse to Exstew 28 (30) 35 (62) 147 (82) 21 (7) 35 (18) 0 0 Exstew to China Bar 11 (6) 22 (16) 439 (157) 230 (151) 66 (43) 44 (27) 11 (9) Sept Ferry Island to Lakelse (99) 94 (34) 84 (52) 0 0 Lakelse to Exstew (53) Exstew to China Bar (433) 286 (138) (15) Overall Total 2691 (291) 1840 (401) 2360 (502) 4249 (616) 646 (126) 74 (36) 63 (36) LGL Limited Page 30

39 Table 15. Statistical tests of the effect of month, river stratum and access method on median harvest-perday estimates for six fish taxa surveyed (with large and jack Chinook reported separately). P- values that are underlined are less than 0.05, but only those in bold are statistically significant after the Bonferroni adjustment. Month River Stratum Access Method Species χ3 2 P χ2 2 P χ1 2 P Chinook, large Chinook, jacks Coho Pink Sockeye Steelhead Other Figure 13. Box-whisker plots showing the distribution of harvest estimates by month, for each taxon (with large and jack Chinook reported separately) with notable harvest. Details as in other figures. Samples sizes were 6 per taxon per month. LGL Limited Page 31

40 Figure 14. Box-whisker plots showing the distribution of harvest estimates by access method for each taxon (with large and jack Chinook reported separately) with notable harvest. Samples sizes were 12 per taxon per access method. Biosampling Scale samples were obtained from 196 (76%) of the 257 Chinook salmon observed during angler interviews. These samples were delivered to DFO for analysis. Eight biosampled Chinook salmon were marked with an adipose fin clip. Photographs were taken from most fish to confirm if the adipose fin was missing (e.g., Figure 15). In 2014, none of the 131 observed Coho salmon were biosampled. LGL Limited Page 32

41 Figure 15. Chinook salmon with adipose fin present (upper photo) and missing (lower photo). DISCUSSION Creel Estimates Effort Comparison of the catch and fishing effort estimates derived from the 2003 creel survey data (Tallman 2004) with those derived for a similar period in (Table 16, Table 17; data from Robichaud and English 2011, 2012, Robichaud et al. 2013, 2014) revealed some substantial differences. The total angler effort estimates for the June-August period in 2010 and 2011 were remarkably similar, both only 42% of the comparable effort estimate for Effort and declined further in 2012 (36% that observed in 2003). The effort for the same time period in 2013 was 131,999 angler hours, or 65% that observed in 2003 ( % that in ). Since then, angling effort has been on the rise. The June to August effort in 2013 was 65% of that in 2003, and that in 2014 was 96% of the 2003 value. It is not within the scope of this study to determine the factors that resulted in the observed fluctuations in effort levels. The average number of anglers observed fishing in 2014 over all strata was 254 anglers per day (Table 18). This is higher than the averages of 103, 101, 98 and 152 anglers/day observed in 2010, 2011, 2012 and 2013 respectively, and were on parr with 2003 reports that daily counts of shore-based frequently exceeded 250 anglers per day (Tallman 2004). Unfortunately, a precise estimate of the average number of anglers observed each day in 2003 cannot be readily derived from the Tallman (2004) report because activity patterns were not provided. Using Tallman s (2004) reported average trip length (4.5 h; likely biased low), and his reported effort estimate of 203,587 angler hours, we calculated that, on average, 595 anglers would have been fishing each day over the 76 day study period in Even if the average trip length in 2003 was similar to that in 2011, over 300 anglers per day would be required to produce the effort estimate reported LGL Limited Page 33