Assessment Protocol for the Commercial Harvest of Pacific Oysters (Crassostrea gigas) in British Columbia

Transcription

1 Canadian Science Advisory Secretariat (CSAS) Researc Document 04/07 Pacific Region Assessment Protocol for te Commercial Harvest of Pacific Oysters (Crassostrea gigas) in Britis Columbia Tammy C. Norgard, Graam E. Gillespie and Micelle I. Bigg Fiseries and Oceans Canada 390 Hammond Bay Road Nanaimo, BC V9T 6N7 December 04

2 Foreword Tis series documents te scientific basis for te evaluation of aquatic resources and ecosystems in Canada. As suc, it addresses te issues of te day in te time frames required and te documents it contains are not intended as definitive statements on te subjects addressed but rater as progress reports on ongoing investigations. Researc documents are produced in te official language in wic tey are provided to te Secretariat. Fiseries and Oceans Canada Canadian Science Advisory Secretariat 00 Kent Street Ottawa ON KA 0E6 ttp:// Correct citation for tis publication: Her Majesty te Queen in Rigt of Canada, 04 ISSN Norgard, T.C., Gillespie, G.E. and Bigg, M.I. 04. Assessment Protocol for te Commercial Harvest of Pacific Oysters (Crassostrea gigas) in Britis Columbia. DFO Can. Sci. Advis. Sec. Res. Doc. 04/07. viii + 43 p.

3 TABLE OF CONTENTS ABSTRACT... VII RÉSUMÉ... VIII INTRODUCTION... PACIFIC OYSTER... BIOLOGY... HARVEST... BC MANAGEMENT AND ASSESSMENT FRAMEWORKS... 3 MANAGEMENT FRAMEWORK... 3 ASSESSMENT FRAMEWORK... 4 BC ASSESSMENT PROTOCOL... 5 DESCRIPTION... 5 EVALUATION... 5 PROPOSED DFO ASSESSMENT PROTOCOL... 6 STRATIFIED RANDOM SAMPLING SURVEY METHOD AND ANALYSIS... 6 OPTIMAL QUADRAT SIZE...0 OPTIMAL SAMPLING INTENSITY... MORTALITY ESTIMATES AND HARVEST RATES... 3 MORTALITY ESTIMATES...3 HARVEST RATES...4 DISCUSSION... 5 FUTURE WORK...7 RECOMMENDATIONS... 8 ACKNOWLEDGEMENTS... 8 REFERENCES... 8 TABLES... 3 FIGURES APPENDIX APPENDIX... 4 APPENDIX iii

4 LIST OF TABLES Table. Landings (Kt) and landed values (x06 $Cdn) of cultured and commercially arvested sellfis in Britis Columbia, Table. Area and number of strata surveyed at eac location by te BC MoA during te surveys from 005 to Table 3. Estimated abundance (oysters/quadrat), biomass (kg/quadrat) and precision for Pacific Oyster surveys using four quadrat sizes... 5 Table 4. Results of Wiegert s cost-benefit analysis using four quadrat sizes Table 5. Results of estimated precision analysis Table 6. Estimated mortality by maximum age using Hoenig s (983) metod Table 7. Summary of results for te Gulland (97), Scaefer(954) and Fox (970) arvest rate models for Pacific Oysters using maximum ages of 0, 30 and 40 years iv

5 LIST OF FIGURES Figure. Areas supporting commercial arvest of Pacific Oysters (Crassostrea gigas) in Jervis and Secelt Inlets, Lasqueti Island and Texada Island, Strait of Georgia, Britis Columbia. Source: BC MoA (0f) Figure. Areas supporting commercial arvest of Pacific Oysters (Crassostrea gigas) in Desolation Sound, Britis Columbia. Source: BC MoA (0g) Figure 3. Areas supporting commercial arvest of Pacific Oysters (Crassostrea gigas) in Barkley and Nootka Sounds, West Coast of Vancouver Island, Britis Columbia. Source: BC MoA (0) Figure 4. Areas supporting commercial arvest of Pacific Oysters (Crassostrea gigas) in Stuart Cannel and Malaspina Inlet, Strait of Georgia, Britis Columbia. Source: BC MoA (0i) Figure 5. Survey design using Stratified Random Sampling Figure 6. Nested quadrats - smallest is te 5 cm x 5 cm, 50 cm x 50 cm, 75 cm x 75 cm, up to 00 cm x 00 cm Figure 7. Results of te precision analysis for eac beac surveyed Figure 8. Results of te precision analysis for Neck Point at all 4 quadrat sizes Figure 9. Results from Sack Island stratum Figure 0. Results from Sack Island stratum v

6 LIST OF TABLES AND FIGURES: APPENDICES Appendix Table. Statistical notation for reviewed survey analyses Appendix Figure. Agreement between te Province of Britis Columbia and te Dominion of Canada (9) Appendix Figure. Oyster arvest permit application. Source: BC MoA (0e) vi

7 ABSTRACT Tis document evaluates te assessment and management frameworks utilized by te BC Ministry of Agriculture (BC MoA) for te commercial arvest of wild Pacific Oyster (Crassostrea gigas) in B.C. and proposes an assessment protocol for use by Fiseries and Oceans Canada (DFO) and Industry to guide DFOs management and regulation of te fisery wic was delegated in 0. Te primary results of te study include te recommendation to utilize a stratified random sampling survey design on ig density discrete Pacific Oyster beds using a square quadrat size of 75 cm x 75 cm or greater. Te recommended sampling intensity is 0 quadrats per ectare wit a minimum sample size of 5 quadrats per stratum. vii

8 Protocole d'évaluation visant la pêce commerciale de l'uître du Pacifique (Crassostrea gigas) en Colombie-Britannique RÉSUMÉ Ce document évalue les cadres de gestion et d'évaluation utilisés par le ministère de l'agriculture de la C.-B. pour la récolte commerciale d'uîtres du Pacifique (Crassostrea gigas) sauvages dans la province et propose un protocole d'évaluation que Pêces et Océans Canada (MPO) et l'industrie pourront utiliser afin de guider la gestion et la réglementation de la pêce par le MPO, responsabilité qui lui a été déléguée en 0. Les résultats primaires de l'étude sont notamment la recommandation d'utiliser pour les relevés des bancs d'uîtres individuels à aute densité une conception d'écantillonnage aléatoire stratifié en utilisant un quadrat carré de 75 cm sur 75 cm ou plus. L'intensité recommandée de l'écantillonnage est de 0 quadrats par ectare, avec une taille d écantillon minimale de 5 quadrats par strate. viii

9 INTRODUCTION Legislative rigts to Canada s inland and coastal fiseries were initially vested in te Federal Government by te Britis Nort America Act of 867, wic provided jurisdiction over all tidal and non-tidal fiseries except tose in Quebec (Quayle 969, 988; Parisien 97). Troug a series of petitions, legal decisions and agreements, jurisdiction over tidal fiseries came to rest wit te Federal Government (tey are also responsible for anadromous species) and jurisdiction over non-tidal fiseries wit te Provincial Government (Parisien 97). Te sole tidal fisery exception was oysters; an agreement between Britis Columbia (BC) and te Dominion of Canada in 9 delegated responsibility for oyster arvests to te Province (Appendix Figure ). In December 00, Justice C.E. Hinkson ruled on a petition before te BC Supreme Court, tereafter referred to as te Hinkson decision. Te ruling concluded tat aquaculture (wit te exception of marine plant cultivation) was, by definition, a fisery rater tan agriculture and terefore management and regulation fell under Federal rater tan Provincial jurisdiction. In addition to sellfis aquaculture, te Province also managed and regulated te arvest of wild Pacific Oysters from Crown foresore. Tis fisery may ave its roots in te recovery of stock wased outside of lease boundaries prior to successful recruitment and establisment of Pacific Oyster stocks beyond aquaculture tenures. Altoug not explicitly addressed in te Hinkson decision, tis endeavor is also clearly a fisery and tus belongs under Federal jurisdiction as per te pit and substance of te Hinkson decision. Tis document evaluates te assessment and management frameworks utilized by te BC Ministry of Agriculture (BC MoA) for te wild oyster arvest and proposes Pacific Oyster assessment protocol for use by te Fiseries and Oceans Canada (DFO) and Industry as DFO assumes responsibility for management and regulation of te fisery. PACIFIC OYSTER Te Pacific Oyster, Crassostrea gigas (Tunberg 793) is a non-indigenous species introduced to BC for aquaculture (Quayle 964, 969, 988; Gillespie et al. 0). Its native range is from Sakalin Island and coastal Russia troug Japan to Kyusu, Cina, Korea, Souteast Asia and Pakistan (Coan et al. 000). Tey ave been introduced and ave establised populations in many countries worldwide (Ruesink et al. 005, Gillespie et al. 0). Te Pacific Oyster was introduced extensively on te west coast of Nort America in te early 900s, and was first brougt into BC in 9 or 93 (Bourne 979, Gillespie et al. 0). Small scale introductions continued and large scale importation of seed oysters began in 95. Successful reproduction was reported in Ladysmit Harbour in 95, 96 and 93, followed by successful dispersal beyond te arbour in 936 (Elsey 93, 934; Elsey and Quayle 939; Quayle 964, 969, 988; Bourne 979). Widespread reproductive success was reported in 94, 958 and 96 resulting in te establisment of Pacific Oysters trougout te Strait of Georgia. Tey were transplanted to te west coast of Vancouver Island (Esperanza Inlet; Barkley, Clayoquot and Kyuquot Sounds) in 937; tey are now establised in suitable abitats Morton v. Britis Columbia (Agriculture and Lands), 009 BCSC 36, Docket S Oyster aquaculture primarily occurs on Crown foresore tenured from te Province of BC. Pacific oyster stocks on untenured foresore are considered wild (Bourne 979, IEC International 006).

10 on te west coast of Vancouver Island sout of Brooks Peninsula (Gillespie 007; Gillespie et al. 0). Tere is also confirmed reproductive success of Pacific Oysters in Skidegate Inlet, Haida Gwaii (Sloan et al. 00; Gillespie et al. 0) and reported occurrence of natural-set Pacific Oysters from Tasu Sound on te west coast of Haida Gwaii (Gillespie, unpublised data). BIOLOGY Pacific Oysters are protandric ermaprodites, initially spawning as males and ten may become females during te winter season (Gillespie et al. 0). Tey are broadcast spawners wit a pelagic larval period of 3-4 weeks depending on temperature (Gillespie et al. 0). Teir natural distribution in BC is limited to locations wit warmer water temperatures tat are required to stimulate gonadal development, spawning and te metamorposis of larvae. Altoug spawning can occur at temperatures between 6-34 ºC and salinities ranging from 0-4%; temperatures of 0-5 ºC and salinities of 35% are considered optimal (Gillespie et al. 0). However, te range of Pacific Oysters can be expanded by manual introduction to microabitats. Adults are sessile and te only excange between sites is troug larval transport or uman intervention. Adults grow relatively quickly in te first few years after settlement and growt slows wit maturity and senescence. Longevity and age structure of populations are not documented due to difficulties in establising aging metods and criteria. New metods for aging Pacific Oysters ave been tested on Pacific Oysters in Cina (Harding and Mann 006), but tese metods still need to be tested for te Pacific Oysters in BC. Bot te literature and local knowledge suggest tat Pacific Oysters can live for decades (Quayle 988, Pauley et al. 988). Pacific Oyster populations in BC generally occur in mid to ig intertidal zones on ard substrates (Bourne 979, Ruesink et al. 005) but can vary depending on te environmental conditions of te site. Fisermen ave noted tat Pacific Oysters are lower in te intertidal zone on te west coast of Vancouver Island (K. Vautier, Pacific Oyster fisermen, Parksville, BC, personal communication, 0). A preferred settlement substrate is oyster sell and large aggregations form if populations are not disturbed; under appropriate conditions tey can form reefs on gravel banks at te tidal mout of small streams (Gillespie et al. 0). Harvestable populations of Pacific Oysters may be present on bedrock walls and outcrops were successful larval recruitment occurs on a regular basis. In all but a few locations in BC, successful recruitment on a large scale is sporadic. Pacific Oyster populations can exibit local recruitment events tat will sustain populations for a number of years. However, populations can become epemeral if larval recruitment is irregular. HARVEST Aquaculture Following depletion of Olympia Oyster (Ostrea lurida [Carpenter, 864]) populations (Gillespie 009) and brief attempts to culture Eastern oysters (Crassostrea virginica [Gmelin, 79])(Carlton and Mann 996), te aquaculture industry in BC moved to almost complete reliance on Pacific Oysters in te 90s (Quayle 969, 97, 988; Bourne 979). Early aquaculture efforts were limited to select arbours and bays in te Strait of Georgia, but Pacific Oysters were subsequently transplanted to te west coast of Vancouver Island in te 930s, were tey flourised. Early attempts to establis culture operations in nortern BC in te 960s were not successful (Quayle 97, Bourne 979).

11 Cultured oysters accounted for annual landings of 4.9 to 8.3 tousand metric tonnes (Kt) and annual landed values between $5.0 and $8.9 million (M) dollars between 996 and 00 (Table ; BC MoA 999-0a,b). Fiseries Te commercial arvest of wild Pacific Oysters from untenured Crown foresore began after significant widespread recruitment events in 94 and 958 (Bourne 979). Participants in tis fisery are mainly aquaculturists wo use te oysters collected as seed stock for teir leases; tere also are number of First Nations communities involved in te fisery. Historically, te fisery as primarily occurred in spring monts on selected beaces on te east and west coasts of Vancouver Island (Figure to Figure 4). Recently, te fisery was regulated by te Provincial Government troug issuance of annual permits; eac permit identified an Individual Quota (IQ) and eac arvest site (wic may support multiple permits) ad associated Total Allowable Catces (TACs)(IEC International 006). Te arvested oysters were utilized as supplemental seed for furter grow out on aquaculture tenures or as product going directly to market. Between 998 and 003, tese landings accounted for. to.7% of total commercial production of Pacific Oysters in BC and te remaining 97.3 to 98.9 comes from te aquaculture industry (Table ; BC MoA 999-0a,b; IEC International 006). In 005 a arvest of 58 tonnes was valued between $6,000 to $48,000 for 97 quotas over 5 individuals (IEC International 006). Commercial arvests of wild Pacific Oysters are not explicitly documented in Provincial reports of seafood production; if tracked at all, tey are included in an Oter category (wit squid, octopus and oter unspecified sellfis)(bc MoA 999-0a,b). Tis category accounted for between <00 and 700 t of arvest and landed values of $0.M to $4.8M between 996 and 00 (Table ). Tere are also noncommercial arvests of Pacific Oysters by First Nations and in te recreational sector. Statistics on landings and values from tese fiseries are extremely limited. BC MANAGEMENT AND ASSESSMENT FRAMEWORKS MANAGEMENT FRAMEWORK Te commercial fisery for wild Pacific Oysters provided oysters bot for re-stocking of leases for eventual sale (after an appropriate grow-out or relay period) and direct sale to processors. Information in te following section comes largely from discussions wit Provincial staff; no publised management plan for Pacific Oysters in BC exists. Licensing Fisers wising to arvest Pacific Oysters would apply annually for Individual Quotas (IQs) at specific sites; applications (Appendix Figure ; BC MoA 0e) were accepted up to January 3 of te fising year (fising year begins January and ends December 3). Harvest permits were issued for a 30-day period (BC MoA 0d). Te permit was not specific to te older, oters could arvest and land product in te name of te license older. In recent years, permits were issued to corporate entities, not limited to individuals (wic as implications regarding transferability). Fisers could apply for IQs at multiple sites but not multiple IQs at a given site; te fisery was essentially controlled by te number of licensed sites and overall Total Allowable Catces (TAC). Fisers paid an application fee of $75 (non-refundable in policy, but many exceptions in practice occurred)(bc MoA 00e). Fisers also paid a royalty of $5/t post-arvest, based on self- 3

12 reported landings documented on daily arvest logs (timing of submission was not specified). Tere was little audit of te logbook program, wic was used to document landings and landed value of te fisery by te Province. Fisers were required to obtain a Fisers Registration Card from DFO (annual requirement). Movement of oysters from waters classified as contaminated to leases required dual licensing from te Provincial and Federal governments (Management of Contaminated Fiseries Regulations). Autority to arvest was also limited by DFO/CFIA (Canadian Food Inspection Agency) area closures for biotoxin issues. Individual Quotas, Total Allowable Catces and Harvest Rates Defined IQs were not less tan one ton and not more tan 0 t (tonne) per fiser per site (i.e., t IQ 0 t). Overall TAC for any arvested site was managed to a target of 0% of estimated biomass; discussions wit Provincial staff indicated tat final arvest rates were generally in te range of 0-4% of te estimated biomass. Core sites received multiple applications; if te overall TAC could not support te number of applications a lottery system was used to select successful applicants. Preferred timing of te fisery was late February to June, avoiding late-summer issues wit Vibrio paraaemoliticus and reduced condition of oysters after spawning. In practice, multiple permits per arvester and biotoxin closures often resulted in te fisery continuing troug September (and occasionally into October). Consultation, primarily wit First Nations and upland owners, was conducted annually and very time consuming; tis led to complaints from prospective arvesters regarding delays in permit issuance. Sites excluded from consideration included recreational map reserves, First Nations map reserves, many areas fronting National or Provincial Parks or Ecological Reserves and most contaminated areas 3 (BC MoA 0c). Also excluded from consideration was te entire soutern Gulf Islands region. Permit duration, notification requirements, Vibrio and product quality issues, biotoxin considerations and te vagaries of arvester activities complicated enforcement activities. Lack of validation of reported landings made defense against allegations of overarvest difficult (altoug it was unclear weter public perception of overarvest was due to regulated fisery activities or illegal arvests [poacing]). ASSESSMENT FRAMEWORK Te Provincial assessment framework was based on annual estimates of biomass of Pacific Oysters in proposed arvest areas. Te framework specified post-arvest assessment of areas wit a follow-up visual assessment in te spring to ensure winter mortality or unregulated arvest ad not drastically affected te post-arvest estimates. Some of tese areas were consistently requested (Figures -4) but te open-ended nature of te management framework (i.e., fisers could apply for wicever sites merited interest) greatly complicated assessment requirements for te fisery. In general, te framework aspired to annual assessments of arvest areas in te fall, post-arvest. However, te January deadline for applications precluded complete assessment of requested arvest sites in te fall. 3 We ave not assessed alignment of Provincial and DFO recreational closures. 4

13 DESCRIPTION BC ASSESSMENT PROTOCOL Provincial assessments followed standard procedures for intertidal assessments, i.e., te expansion of density estimates of te desired caracteristics of te population over an estimated area representative of tat population. Te surveys were limited to dayligt low tides (unspecified dept) and generally started in Marc eac year. Te spring assessment was largely visual, wit expert surveyors confirming bed areas using GPS from small boats and providing a subjective expert-based confirmation of density. Table lists te beaces and number of beds/strata surveyed by te Provincial crew between 005 and 00. Provincial survey protocols were documented from discussions and a single joint survey undertaken wit Provincial and DFO staff in 0. Bed area was determined by one member of te survey crew walking te subjectively determined perimeter of te oyster bed and taking regular positions on a and-eld GPS unit. Te density of oysters (in terms of biomass) was obtained by survey crew members tossing tree m survey quadrats in a apazard fasion witin te bed area. Oysters were counted inside te quadrat if alf or more of teir eigt was witin te quadrat; on two pre-determined sides of te quadrat oysters close enoug to alf-in were included in te count and oysters on te remaining two sides were excluded. Live oysters were cleaned of attaced sell and substrate, weiged in aggregate and te total weigt of live oysters for te quadrat recorded. Tese data were later combined in a spatial analysis (undocumented) to produce estimates of total biomass for te stratum. EVALUATION A table of istoric estimates provided by te Province indicated tat a number of sites ad not been surveyed recently, weter because tey ad not been requested or due to diminising budgets (Table ). Conservative TACs were proposed for some sites based on previous survey estimates tat were - years old and anecdotal information. Original survey data were not available from te Province, only estimates of bed area, density and total biomass were provided to DFO; none included estimates of variability (confidence intervals). Estimates of biomass were not reproducible by matematical combination of area and density estimates; te exact metod by wic te estimates were derived could not be determined. Delineation of bed margins is less troublesome for oysters, wic grow on te surface, tan for infaunal bivalves. For oysters, visual determination is required, wereas digging test oles and establising density tresolds is required for infaunal bivalves. Repeatability of bed area estimation may vary somewat between different surveyors, but relatively consistent bed boundaries are usually discernible. Delineation of oyster bed areas usually results in a conservative estimate of biomass, as it excludes a portion of te population tat occurs scattered in low density outside of distinct beds. Te apazard sampling (ad oc, potentially purposive) protocol used in visually selecting quadrat locations introduces te possibility for surveyors to bias (eiter upward or downward) te biomass estimate, resulting in biased estimates of mean and variance. Tis could reduce te reliability of biomass estimates provided by tird-party or Industry surveyors. Additionally, te lack of true randomization (i.e., all potential sampling elements ave equal probability of selection) violates a major assumption of probability sampling-based metods, including simple or stratified random and systematic random designs (Kronlund et al. 998). 5

14 Te m quadrat was likely appropriate to reduce edge effect (te determination of weter or not an oyster is to be included in te sample) and te protocol furter assists in defining two of four quadrat edges tat are inclusive of oysters too close to call and two tat exclude tese oysters. Te sampling intensity utilized was very low (tree quadrats per bed/stratum) and just meets te minimum for calculation of informative estimates of variance. Weter tis intensity is appropriate could not be determined witout original survey data or estimates of variance from previous surveys, neiter of wic were available. Expert-based visual surveys were used eac spring to assess weter population levels ad canged radically from formal fall survey estimates (and may ave been used more widely as Provincial program support diminised). Because assessment responsibilities were transferred between agencies witout significant overlap for mentoring and development of expertise in DFO, considerable time and resources may be required before reliance on subjective estimates of biomass are considered reliable. PROPOSED DFO ASSESSMENT PROTOCOL Tis protocol as been developed to assist potential arvesters in conducting surveys and data collection of wild Pacific Oysters on beaces in wic discrete beds of oysters are found. Discrete beds are tose were well defined beds of oysters can be visually determined on beaces. In general, Pacific Oyster populations may be found in discrete beds of single or clustered oysters loose on te surface of te beac or individual oysters cemented to ard substrate (large rocks or bedrock), at times including vertical surfaces. Tis protocol provides key guidance on sampling and data collection metodology, optimal quadrat size and sampling intensity for discrete oyster beds. Tis study also gives te background and rationale beind te importance of determining accurate Pacific Oyster population abundance and biomass estimates. Te ultimate goal in development of te protocol is to ensure tat accurate and standardized stock information is collected so tat it can ten be utilized by DFO to develop IQs and TACs in te sort term and sustainable arvest strategies for specific beaces in te long term. STRATIFIED RANDOM SAMPLING SURVEY METHOD AND ANALYSIS Te first objective of te protocol was to establis te primary metodology for surveying Pacific Oyster beds tat will be commercially arvested. A review of seven different survey protocols was completed wen developing survey metods for Olympia Oysters (Norgard et. al. 00). Te results of te Olympia Oyster study selected te Two Stage sampling design for Olympia Oyster because teir populations are often more patcy and tis survey design spreads te surveying across te bed. Wereas te populations of Pacific Oysters to be surveyed for commercial interest will be quite dense and fairly evenly distributed; terefore we recommend using a Stratified Random Sampling (StRS) design wic is already widely in use for bivalve species in BC. Te StRS metod to survey Pacific Oysters was tested wen DFO conducted surveys during summer low tides in 0, at Sack Island ( N, W) and Neck Point (49 4. N, W) in Nanaimo, BC. Tese beaces were cosen because te Pacific Oyster populations were dense and evenly distributed across te beaces. Witin te boundaries of te discrete beds, sampling units or strata (non-overlapping groups) were defined and a simple random sample was drawn from eac group (quadrats) (Figure 5). Determining te number of strata is dependent on te pysical caracteristics of te beaces 6

15 and prior knowledge of te site. Stratification can be useful in dividing te beac into manageable survey units to account for specific population caracteristics. For example, ig density areas of a bed may be partitioned into a stratum separate from areas of lower density. Tus, one stratum may differ markedly from anoter but variability witin te stratum would be small. If no prior knowledge of te beac exists it is possible to stratify by substrate type or tidal elevation. In te case of Pacific Oysters, epifaunal beds are relatively easily delineated and strata represent distinct aggregations. A key feature of te StRS metod is tat a sample (quadrat) is selected from eac stratum independently of oter strata and quadrats can be randomly placed trougout strata using a random number generator. Randomization provides a fair and repeatable means of avoiding bias in te selection of sampling sites (Kronlund et al. 998). For te 0 surveys in Nanaimo, te primary surveyor walked te beac using a Trimble GPS (Trimble Pro XT) logging waypoints to delineate te boundaries of discrete oyster beds. At Neck Point, tree surveyors undertook te same procedure to determine boundaries of te oyster bed wit te Trimble. Te result was tat eac surveyor mapped almost identical beds to eac oter, confirming tat te boundaries of te discrete bed were obvious and discernible (Tammy Norgard, Fiseries and Oceans, Nanaimo, BC, unpublised data). A single stratum was setup to cover as muc of tis discrete bed as possible. At Sack Island, two strata were establised to account for two iger density areas tat were separated by a low density area and because of te curvature of te beac. By setting up two separate strata we were able to reduce variability wit eac stratum. At te Nanaimo sites, te strata were divided into m by m quadrats. At Neck Point 80 random quadrats were selected for sampling witin one stratum of size 5 m. At Sack Island 5 and 5 quadrats in stratum (95 m ) and stratum (50 m ) respectively were selected. At eac quadrat location te total weigt of oysters was recorded for four different nested quadrat sizes, details on tis examination is found in te Optimal Quadrat Size section (p. 0). To obtain weigts, surveyors started from te smallest sized quadrat (5 cm x 5 cm) and broke excess sell and rock from oysters using a small ammer and ten weiged all te oysters togeter. Te number of oysters was recorded and oysters were ten placed in a metal basket to obtain an aggregate weigt using a and eld digital fis weiging scale. Te weigt of te basket was subtracted to obtain te true oyster weigt in kilograms. Tis was repeated for eac quadrat size at eac sampling site witin te stratum. From tese weigts an estimate of abundance and biomass was calculated for eac stratum (Table 3). Tis metodology does not include sampling oysters outside te stratum (Figure 5); terefore te biomass and abundance estimates are not extrapolated to determine biomass or abundance of oysters of te entire bed. Tis results in a conservative estimate of biomass tat can be used to set a sustainable arvest rate. Te survey data collected from bot beac surveys in 0 were used to determine optimum quadrat size and sampling intensity for Pacific Oysters. Estimating te Population Mean One objective of te survey is to estimate te mean density (number and weigt) of oysters in te survey area. Statistical notation for te equation listed in tis section is provided in Appendix Table and are list below; H stratum index, maximum strata number, 7

16 i N y-value index, total number of sampling units (quadrats) in te population, N total number of sampling units in stratum, n number of units (quadrats) in te sample, or sample size, n number of units in te sample from stratum, y estimated population mean density in stratum, y i y-value i in stratum (number of oysters), µ population mean, τ y population total estimated population mean, Vˆ ( y) estimated variance of te population mean, τˆ ( τˆ ) estimated population total, V ˆ estimated variance of te population total, s sample variance in stratum, z α / t-value may be replaced wit tis estimator for large sample sizes a variable witin Sattertwaite s approximation Te mean density over te surveyed area (weigted mean of strata densities) for a given beac or arvest area is estimated as: y = N Te variance of te mean is estimated as: were Vˆ N H ( y) = = H = N N y N n N s n () () 8

17 s = n n n y i ( ) i= i= yi y i= n = An estimate of te total number or weigt (were y is eiter number or weigt depending on te estimate) of oysters in te survey area can be obtained by expanding te mean estimate over te total surveyed area: τˆ = H Ny = = n N y y i Te variance associated wit tis estimate can be calculated as: were s Vˆ = Confidence Intervals N H ( y) = = N N n N n n n y i ( ) i= i= yi y i= n = n y i n s n Confidence intervals for population parameters can be computed in a variety of ways for stratified random sampling. Te coice of te metod may depend on te sample size witin eac stratum, or on weter normality is assumed. Wen te sample size witin eac stratum is greater tan 30 units, ten te normal approximation may be used. For te population mean: y ( y ) ± t d Vˆ α /, (7) n (3) (4) (5) (6) For te population total: ( τ ) τ τ ˆ ˆ α ˆ /, d V ± (8) 9

18 were t is te upper α/ point of Student s t distribution wit d degrees of freedom computed using Sattertwaite s approximation. If sample sizes are large, ten te t-value may be replaced z α / wit. Wen sample sizes are small (as a rule of tumb, less tan 30) an adjustment to te degrees of freedom for te t-statistic is appropriate. Te adjustment is called Sattertwaite s approximation (Sattertwaite 946): were H as = d = (9) H ( as ) = n a N ( N n ) =. (0) n If all stratum sizes are equal and all sample sizes are equal, ten te degrees of freedom are n- H n = H, were n =. An alternative to assuming te normal distribution is to use resampling (bootstrap) tecniques to compute a non-parametric estimate of te confidence interval. Tis metod is described by Rao and Wu (988), Sitter (99), and Kronlund et al. (998). OPTIMAL QUADRAT SIZE A second objective of te study was to determine optimal quadrat size for sampling Pacific Oyster. Population estimates are a function of te caracteristic of quadrats in relation to te distribution of te species on te beac (Kronlund et al. 998). In tis study, te criterion for optimal quadrat size was based on te quadrat size tat ad te least edge effect, lowest variance, best tradeoff for cost (time) and practicality. Te experiment used nested quadrats at Sack Island and Neck Point in Nanaimo, BC, to determine te optimal quadrat size (Krebs 998, Wiegert 96, Kronlund et al. 998). Four sizes of square quadrat were used: 5 cm x 5 cm (0.065m ), 50 cm x 50 cm (0.5m ), 75 cm x 75 cm (0.565m ) and 00 cm x 00 cm (m ) (Figure 6). Edge effect can occur wen a sampler decides weter to count an oyster tat lay on te quadrat edge as inside or outside te quadrat. Upwardly biased estimates may result from keen samplers counting oysters tat lay on quadrat edges as being inside te quadrat. Training of samplers can reduce te edge effect. However, it is better to coose a larger quadrat sizes (75 cm x 75 cm or 00 cm x 00 cm) because tey ave smaller edge-to-area ratios and reduced edge effects (Wiegert 96). If no edge effect was present, it would increase te likeliood tat all mean and biomass estimates from te four quadrat sizes (standardized to m ) would be te same. Results from surveys at Neck Point and Sack Island sow tat edge effect was present in tese surveys 0

19 since te estimates were not te same for eac of te quadrat sizes.. For example, in Table 3, te igest mean biomass (7.8 kg/m and 0.64 kg/m ) and igest mean abundance (58.00/m and 68.48/m ) respectively were seen in te smallest quadrat at almost all sites (5 cm x 5 cm). Wiegert s (96) metod was used to analyze te data collected in nested quadrat experiments for Sack Island and Neck Point to determine optimal quadrat size. Tis metod proposes tat te two factors of primary importance in determining sample size are te relative variability in te oyster population and relative cost (time or effort) required to assess abundance or biomass of eac quadrat size. Te time required to count te number of Pacific Oysters in eac of te four quadrat sizes in bot strata and at Sack Island was calculated and utilized in Wiegert s analysis. No time data were collected for te Neck Point survey so te time data from te Sack Island survey was used for te analysis of tis survey. Tis analysis was completed using timed results from te stratum and stratum at Sack Island (Table 4). Te optimal cost x variance is calculated using te Weigert analysis. Tis analysis multiplies a standardized relative variance by a standardized cost (Time). Standardized relative variance = (Standard deviation) x Standardized (Minimum standard deviation) Te lowest standardized cost x standardized variance is te optimal quadrat size. Te smallest quadrat (5 cm x 5 cm) was te optimal quadrat in almost all cases using tis analysis. However, te smallest quadrat sowed relatively ig variance in biomass and abundance estimates and te largest edge effect. Wen only considering te amount of time to complete surveying of eac quadrat relative to te quadrat area (Cost/Quadrat Area column from Table 4) te 75 cm x 75 cm quadrat was te optimal quadrat size (Table 4). OPTIMAL SAMPLING INTENSITY Te tird objective of te study was to determine optimal sampling intensity for any given potential Pacific Oyster arvest site. Two metods were used to determine optimal sample size for Pacific Oyster surveys. In Metod, precision estimates were calculated using te index of dispersion (Elliot 977; Metod ). Te second metod calculates estimates of sample size utilizing a formula from Quinn and Keoug (00; Metod ). Metod Cost Kingzett and Bourne (998) completed te analysis described below to obtain estimates of precision based on istoric butter clam survey data from Seal Island, BC. Te number of sampling units required to acieve a given precision in a study may be predicted wit knowledge of te variation witin a population (typically from an initial sample or previous surveys) for randomly distributed populations. For populations were te negative binomial distribution is a suitable model (populations wit aggregated distributions), te index of dispersion statistic common (k) may be used. To calculate te required number of samples for a given precision, te standard error of aritmetic mean to ratio index of precision (d) was used. Te value of d represents te standard error as a percentage of te mean µ. Percentage confidence limits of d about te mean were calculated by incorporating te Student s t- distribution statistic in te equations (t=.96 for 95% confidence interval). For a negative binomial distribution te number of required samples (n) was solved for various levels of desired accuracy (d) using te following formula (Elliot 977):

20 t N = d + µ k Te index of dispersion ( k ) was approximated using te following formula Results are presented in Table 5, Figure 7 to Figure 0. Metod () µ k = σ () µ We compared te metods of Elliot (977) and Quinn and Keoug (00) and produced identical results. Quinn and Keoug (00) used te equation: α σ z n d were n is sample size, zα is te z value from a standard normal distribution for te cosen α (we used.96 wic is 0.05 for te 95% confidence interval), σ is te variance of te population, and d is te maximum allowable absolute difference between te true population mean te estimated population mean (tested for a range of 0% to 00% in Table 5). Te calculated optimal number of quadrats to obtain a specified precision about te mean are sown in Table 5. All estimates give te approximate number of samples tat would be needed to obtain precision of te mean wit 95% confidence. Estimated sample numbers for eac survey give a general indication of te precision tat sould be obtainable in future surveys wen te mean and standard deviations are unknown (Table 5, Figure 7 to Figure 0). Te optimal sample size was determined on te result of utilizing Metods and and setting te target for precision of 30% (or better). Te precision of 30% as been found to be reasonable measurement of precision for bivalve surveys. Based on tese criteria, future surveys of discrete beds require sampling of 6 to 4 quadrats using a 75 cm x 75 cm quadrat size and 4 to quadrats using a 00 cm x 00 cm quadrat size. Biological Sampling Pacific Oysters ave been studied in detail in relation to aquaculture for te past 00 years but studies of wild oyster populations in BC are very limited (Gillespie et al. 0). Tere is limited understanding of life istory parameters suc as growt and recruitment in BC. Pacific Oyster growt is relatively rapid in te first two years and is influenced by wave action, temperature and location on te beac and all tese affect te in sell sape (Quayle 969). Aging studies of Pacific Oysters using te cross section of valves as been successful for oysters collected in Cina and could elp determine growt and maximum age for wild Pacific Oysters in BC, but te metod as yet to be validated (Harding and Mann 006). Provincial government assessment work as largely focused on weigt estimates wic did not provide data on populations and stock types (IEC International 006). A 006 survey of arvesters sowed tat wile oysters may ave been present on te beac, tey may not ave been te required types or sizes for market needs. Collection of survey information to include size classes and cluster sample (weigt clusters) as been suggested as being potentially useful for te industry. Tis data could be used to partition estimates of biomass or abundance (3)

21 into specific size classes as is done wit size limit tresolds in oter commercial bivalve surveys. Oyster eigt is te largest measurement and is te distance between te umbo and te ventral valve margin (Galtsoff 964). Heigt frequency distributions can be used to provide information on te recent recruitment (<30 mm) events but eigt data for te larger oysters becomes less useful as teir sape depends on te abitat in wic tey grow. If eigt frequency data was to be collected for oysters, typical requirements for random sample selection and consistent measurement of sell size (fide Gillespie and Kronlund [999] for intertidal clams) are vital. Biological samples sould be randomly selected at te quadrat level, and te final quadrat completed once te sample size tresold is acieved. Furter work is required to determine meaningful measurements (sell lengt, eigt, tickness, total weigt, recovered weigt, or some combination of tese metrics) of oysters required depending on te purpose. Te sample size for biological measurement sould be appropriate to acieve a reasonable level of precision in measurement and sould accurately represent te oyster population of a specific beac. One study as suggested tat a minimum sample size of 0 times te number of eigt classes in te sample would be a reasonable compromise between effort and precision (Gerritsen and McGrat 007). Using tis guideline for Pacific Oysters (approximate maximum eigt of 300 mm in BC), 0 classes of 30-mm bins would give a sample size of 00 oysters. But, if time and funding permits, more samples sould be measured. Biological Sampling by Industry Industry will be utilizing tis protocol to conduct surveys; terefore te expectation for te level of biological sampling must be practical in terms of time and cost. If possible we recommend obtaining a sample of 00 random oyster eigts recorded from various quadrats witin a stratum to obtain potential recruitment data Te IEC International (006) report suggested tat size class sample and cluster sampling may be of interest to industry and sould also be considered. Formal consultation wit industry will be required to finalize potential biological sampling requirements for te assessment, depending on te needs of industry and for fisery management. MORTALITY ESTIMATES AND HARVEST RATES Introduction of a discussion on mortality estimates and arvest rates is important in providing additional context to tis assessment protocol. No formal arvest rates models were used by te Provincial Government in te management of te Pacific Oyster fisery. Furter, te lack of fisery data does not facilitate te use of complex fisery models to set arvest rates. As more data become available, models sould be reviewed to set new arvest rates. MORTALITY ESTIMATES Te simplest metods to set arvest rates require mortality estimates wic ave not been calculated for Pacific Oysters in BC. Hoenig s (983) model to estimate mortality rates in datapoor situations as been used in studies for Manila clams (Gillespie et al.998b) and scallops (Surry et. al. 0) but requires knowledge of te maximum age. Altoug maximum age of Pacific Oysters as not been directly estimated in BC, two documented anecdotal estimates of maximum age were 0 years (Quayle 969) and 40 years (Pauley et al. 988). Hoenig (983) described relationsips between mortality and te maximum observed age for fis, cetaceans, and molluscs and provided coefficients specifically for molluscs based on a data set tat included clams, cockles, gastropods, oysters, and scallops. Te relationsip was: 3

22 ln( M ) = a + b ln( t were M is te instantaneous natural mortality, and t max is te maximum observed age. Te values for a and b for molluscs were.3 and -0.83, respectively (Hoenig 983). Estimated mortality rates ranged from 6% to 36% using maximum age estimates of 40 to 5 years, respectively (Table 6). Recent studies of winter mortality in Pacific Oysters from Denmark, Sweden and Norway sowed te lowest annual rates of 5% (75% annual survival) at te lowest latitude in Denmark wit an increase to 87% and 55% respectively (Strand et al. 0). If we assume tat our winters are not as ars as tose experienced in Denmark, ten we can assume tat te BC annual mortality rates are somewere below 5%. A 5% annual mortality corresponds to a constant, instantaneous mortality rate of approximately 0.8 (year - ). HARVEST RATES Mortality estimates calculated above were used in Gulland s (97) arvest rate model: max ) (4) MSY = XMB 0 (5) were MSY is te maximum sustainable yield, X is a constant, M is te natural mortality, and B 0 is te unexploited or virgin biomass. Lauzier et al. (005), Boutillier et al. (998) and oters ave used values of X=0.. In oter unpublised studies of butter clams a value of X=0.5 as been used (Gillespie, unpublised data). Gulland s (97) model is often used to provide preliminary estimates of MSY in new and developing fiseries, but may not be te best coice of model and sould not be used for fiseries in wic tere is already significant exploitation (Garcia et al. 989). Gulland s equation as been criticized by Francis (974), Deriso (98) and Beddington and Cooke (983), among oters, and it is now generally recognized tat fising mortality is often lower tan M in equation 5 and tat using a value of X=0.5 overestimates MSY. Terefore values of X 0.5 migt be preferred. One approac to estimating MSY in data limited situations as been proposed by Garcia et al. (989). Teir (gross) approximation of MSY is based on te Fox (970) or Scaefer (954) surplus production models and requires only one year of estimates of catc and biomass. Te assumptions of tis metod include: Biological processes are deterministic; Catcability is not density-dependent; Te fisery acts on a single stock wit stable age/size distributions (equilibrium), and tat fisery caracteristics are canging slowly; Natural mortality rate (M) is known; Te relationsip between M and fising mortality rate (F msy ) is of te form F msy =XM were X is a constant tat depends on stock parameters; and Observations of current biomass (B c ) and current yield (Y c ) are available for one year only. Note tat F msy and M correspond to annual rates rater tan instantaneous rates following te notation of Garcia et al. (989): 4

23 F msy = XM (6) were X is a constant tat depends on stock parameters (Gillespie et al. 998a). For te Scaefer model, given one year of biomass estimates (B C ) and catc (Y C ), MSY is given by te following: MSY ( FMSY BC ) = F MSY B C Y Similarly, for te Fox model, MSY is given by te following: Y C C F BC (7) msy MSY = F B e (8) msy C Te biomass estimate B C is te biomass estimate for one year, and bot te catc and biomass referred to sould ave te same age or size structure (Garcia et al. 989). Te Scaefer model becomes unstable as F approaces F msy and Y C and B C approac zero. Tus, te Scaefer model sould not be used wen tere is a ig level of effort or te stock as been badly overfised. Similarly, te Fox model becomes unstable as B C approaces zero, but tis occurs at extremely ig values of F (Gillespie et al. 998). Table 7 contains an evaluation of MSY for te Gulland, Scaefer and Fox models for tree values of F: 0.38, 0.8, 0.0 and 0.6, based on maximum ages 5, 0, 30 and 40 years, respectively. A range of values for X (0.-0.5) in te relationsip F msy =XM (Garcia et al. 983) were tested in te models at tree potential maximum ages. Te data from te 00 Pacific Oyster fisery arvest and biomass estimates were used in tis analysis (Table 7). Harvest rates ranged between 3 and 7% for te tree models wen te constant of X=0. was selected. Wereas te rates ranged between 5 and 8% wen a constant of X=0.5 was used. Terefore, from tis preliminary review, arvest rates between 3% and 8% may be appropriate to apply to te wild Pacific Oyster fisery, but a more toroug review of tis is required before any recommendations can be made. DISCUSSION Te wild Pacific Oyster fisery is small in terms of tonnage arvested, and financial value to Industry wen compared to Pacific Oyster aquaculture arvest. However, arvest biomass of wild Pacific Oyster doubled between 996 and 005 and growt of tis fisery could continue if Industry is able to increase demand for its product and create new markets. Data derived from future assessment of wild Pacific Oysters and fisery information will be important in informing DFO weter te trend in growt of te industry is continuing. Responsibility for data collection, management and analysis will need to be defined in order to ensure te utility and reliability of data for use in setting arvest rates. We ave recommended implementation of te Stratified Random Sampling (StRS) design as te best survey metod to attain accurate and conservative biomass estimates during assessment of discrete beds wit uniform density. Tis survey design requires tat strata be defined to cover individual beds on a beac, tat quadrats are randomly placed witin te stratum and tat oysters witin tose quadrats are weiged. Tis design recommends surveying eac bed/stratum separately ten combining all te results for eac stratum into beac biomass estimate. DFO recommendation for implementation of tis protocol for arvest of wild Pacific 5

24 Oysters can be rationalized in tat it is a standardized scientifically defensible metodology tat eliminates surveyor bias, te precedent and requirement for a similar protocol is used by depuration clam fisery, and because te apazard approac (wic relies eavily on visual assessment) is subjective and as inerent bias tereby reducing reliability of data used setting future arvest rates. Implementation of tis sampling design as important implications for Industry given tat tis design is more systematic, requires more time (and likely increased cost) to sample tan te apazard approac and also recommend some form of biological sampling. DFO s review of Provincial Government surveys from 005 to 00 indicate a very large number of strata were completed eac year and on some beaces up to 34 strata were assessed (Table ). Te StRS survey is more time consuming wic will result in a reduction in te number of strata assessed on eac beac. In order to utilize StRS, expertise witin Industry may ave to be gained or assessments be contracted by Industry to qualified surveyors. However, te advantage of te StRS metod to Industry is tat no GPS mapping of te beds need to be conducted during set up te survey. In eiter case a sampling manual detailing assessment metodology and elements of tis protocol sould be developed to ensure consistency of data. DFO will be responsible in ensuring Industry conforms wit te assessment protocol role, wic can be acieved troug monitoring, enforcement and troug discussions wit Industry. Monitoring not only provides DFO important wit information about oyster assessments but also a monitoring protocol can be developed to provide information on ecosystem impacts. Tis monitoring can provide information about recruitment and can be designed to evaluate factors influencing conservation and sustainability of te fisery. Poto documentation of site condition prior to and post-arvest could be utilized as a tool in monitoring. Te 75 cm x 75 cm quadrat is te optimal recommended quadrat size for sampling Pacific Oysters on beaces in BC because it sowed a low level of variance, took te least time for te amount of area surveyed and was less affected by edge effect. Te 00 cm x 00 cm quadrat as similar results to te 75 cm x 75 cm quadrat and may be used for te ease of selecting te random quadrats in te survey set up. Te additional time it takes to complete te beac survey using te larger quadrat, in addition to te reduced practicality of tis size of quadrat sould be considered wen setting up te survey protocol. DFO is aware tat Industry may currently utilize round ( m diameter) equipment for assessment and tis will ave to be switced to square quadrats to be consistent wit tis protocol and reduce potential variations in data. Te recommended sampling intensity of tis protocol is iger in order to acieve a reasonable level of precision. Te requirement for increased sampling intensity and precision can be justified wen comparing survey of oter bivalves suc as Manila clams. For example, te survey precision of 30% for Manila clams as been produced fairly reliable estimates of biomass (Norgard, unpublised data). We found for Pacific Oysters survey to obtain approximately 30% survey precision of discrete beds a sampling range of between 6 and 4 quadrats would be required wen using a 75 cm x 75 cm quadrat size and a range between 4 and quadrats would be required wen using a 00 cm x 00 cm quadrats. Wen comparing to te sampling intensity under Provincial management of 3 quadrats/strata, tis new protocol requires an increase to a minimum of 5 quadrats/strata and optimally 0 quadrats/ectare. So we ave recommended a sampling intensity of 0 quadrats per ectare wit a minimum sample size of 5 quadrats per stratum. DFO will evaluate data collected from assessment of surveys using te metodology recommended witin tis protocol and will consult wit Industry to determine weter te recommended sampling intensity requires adjustment or survey design requires modifications. At tis time, DFO may also consider weter to expand and broaden te assessment protocol to include oter beac types (oter tan oysters distributed in discrete beds). 6

25 Te development of tis protocol as resulted in consideration of a recommended arvest rate. It is DFOs policy to apply te precautionary approac and adaptive management approac to management of fiseries wic are important considerations in setting future sustainable arvest rates. It sould be noted tat te arvest rates recommended witin tis protocol were developed in te absence of key information about Pacific Oyster in BC waters. Gaps in understanding currently exist in te following areas: Te size of te Pacific Oyster stock in BC; Te potential effects of climate cange (sea level rise and variability in ocean temperature) on te population and fisery; Lack of age data from wic to reliably estimate mortality; Te variables affecting recruitment on beaces in BC; and Te ecological impacts of tis wild fisery. Since mortality estimates are currently unavailable for wild Pacific Oyster, te only metod we used to set arvest rates was based on te Hoenig s (983) model. Tis model as been used in studies to estimate mortality rates in Manila clams (Gillespie et al.998b) and scallops (Surry et. al. 0) but requires knowledge of te maximum age of te species. Longevity estimates of Pacific Oysters range from 0 years (Quayle 969) to 40 years and te model yielded mortality rates ranging from 6% to 36%. One approac to estimating MSY in data limited situations as been proposed by Garcia et al. (989). Teir (gross) approximation to MSY is based on te Fox (970) or Scaefer (954) surplus production models and requires only one year of estimates for catc and biomass. We applied tese models to te catc (0 tonnes) and biomass (5090 tonnes) data from 00 Pacific Oyster fisery provided to DFO by te Province and found tat arvest rates between 3% and 8% would be appropriate to apply to te wild Pacific Oyster fisery. DFOs arvest range recommends a larger range tan te Provincial government arvest rate of 0 to 4% arvest rates but wic fall wit te suggested range of te recommended arvest rates in tis study. As more fisery, assessment and mortality rate data become available furter analysis and possible modification of DFOs recommended arvest rates sould be undertaken. DFO will also ave to consider if or wen to undertake furter study to improve understanding information gaps given te relatively small size of tis commercial fisery, te DFO resources available and importance relative to oter Departmental priorities. In summary, we ave conducted an evaluation of te assessment and management frameworks utilized by te BC MoA for wild oyster arvest and ave proposed a Pacific Oyster assessment protocol for assessing biomass on beaces. Also, we ave recommended a range of arvest rates to be applied by managers to surveyed beaces. All of tis information was derived from extremely limited data. Collection of iger volumes of more diverse data could support more sopisticated assessment advice in te future. FUTURE WORK To facilitate transfer of standards and acceptable protocols to Industry or tird-party surveyors, a formal survey manual for Pacific Oyster surveys sould be developed (following ratification of te metods and standards presented erein). Te current advice is for relatively ig density populations in discrete beds (te preferred arvest sites). Future work can be done to develop survey protocols for oter population types (e.g., oysters attaced to bedrock or rock walls). 7

26 Te use of oyster sells or ligaments to determine age of oysters would provide a better understanding of age composition, recruitment and mortality rates (and tus a more sopisticated approac to determining acceptable arvest rates). Over te longer term, tis work could allow development of age-based assessments of oyster populations. Biological sampling protocols sould be developed in consultation wit Industry. Provision of minimal assessment advice (i.e., biomass estimates and preliminary arvest rates for fisery management) does not require biological samples. However, any advance in te assessment framework, and te quality of advice provided to managers and Industry, will require increased sampling of biological caracteristics of arvested populations. RECOMMENDATIONS. Stratified Random Sampling survey metods sould be used on relatively ig-density discrete beds. Formal aderence to randomization for locating quadrats prevents bias, allows establised probability statistics to be used and improves defensibility of tirdparty or Industry assessments.. We recommend a quadrat size of no less tan 75 cm x 75 cm. Smaller quadrat sizes exibited iger variance, more edge effect and appeared to be more affected by smallscale patciness. Larger quadrat sizes did not exibit tese problems to te same degree, and te 75 cm x 75 cm quadrat size outperformed te 00 cm x 00 cm quadrat in cost effectiveness (and to some extent in practicality). 3. We recommend a sampling intensity of 0 quadrats per ectare wit a minimum sample size of 5 quadrats per stratum. Tis sampling intensity will be reviewed as more survey results become available. ACKNOWLEDGEMENTS We wis to tank Paige Borrett for ours of tireless field and data work on Pacific Oysters; Sean MacConnacie for is GIS support, collaboration and advice formulating objectives; our survey crew of Leslie Barton, Sara Davies, Amy Ganton, Selee Hamilton, Georg Jorgensen and Rezia Kan. We would also like to tank Brian Kingzett (Vancouver Island University), Steven Scut (DFO) and participants in te CSAS Regional Review Process for teir reviews of tis Working Paper REFERENCES BC MoA Britis Columbia seafood industry. Year in review 998. BC Ministry of Agriculture, Policy and Competitiveness Branc, Victoria. 8 p. BC MoA Britis Columbia seafood industry. Year in review 999. BC Ministry of Agriculture, Policy and Competitiveness Branc, Victoria. 8 p. BC MoA. 00. Britis Columbia seafood industry. Year in review 000. BC Ministry of Agriculture, Policy and Competitiveness Branc, Victoria. 6 p. BC MoA. 00. Britis Columbia seafood industry. Year in review 00. BC Ministry of Agriculture, Policy and Competitiveness Branc, Victoria. 8 p. BC MoA Britis Columbia seafood industry. Year in review 00. BC Ministry of Agriculture, Policy and Competitiveness Branc, Victoria. 8 p. 8

27 BC MoA Britis Columbia seafood industry. Year in review 003. BC Ministry of Agriculture, Policy and Competitiveness Branc, Victoria. 8 p. BC MoA Britis Columbia seafood industry. Year in review 004. BC Ministry of Agriculture, Policy and Competitiveness Branc, Victoria. 8 p. BC MoA Britis Columbia seafood industry. Year in review 005. BC Ministry of Agriculture, Policy and Competitiveness Branc, Victoria. 6 p. BC MoA Britis Columbia seafood industry. Year in review 006. BC Ministry of Agriculture, Policy and Competitiveness Branc, Victoria. 6 p. BC MoA Britis Columbia seafood industry. Year in review 007. BC Ministry of Agriculture, Policy and Competitiveness Branc, Victoria. 6 p. BC MoA. 00. Britis Columbia seafood industry. Year in review 008. BC Ministry of Agriculture, Policy and Competitiveness Branc, Victoria. 6 p. BC MoA. 0a. Britis Columbia seafood industry. Year in review 009. BC Ministry of Agriculture, Policy and Competitiveness Branc, Victoria. 6 p. BC MoA. 0b. Britis Columbia seafood industry. Year in review 00. BC Ministry of Agriculture, Policy and Competitiveness Branc, Victoria. 3 p. BC MoA. 0c. Commercial Fiseries. Harvesting of wild oysters in Britis Columbia. BC MoA. 0d. General terms for wild oyster arvesting: Regulations Concerning te Harvesting of Oysters from Vacant Crown Foresore. Accessed September 04. BC MoA. 0e. Wild oyster arvesting application form: Application for a Permit to Harvest Oysters from Vacant Crown Foresore. Accessed September 04. BC MoA. 0f. Jervis Inlet/Strait of Georgia areas. Accessed September 04. BC MoA. 0g. Desolation Sound areas. Accessed September 04. BC MoA. 0. West Coast Vancouver Island areas. Accessed September 04. BC MoA. 0i. Contaminated areas. Accessed September 04. Beddington, J. and Cooke, J.G., 983. Te potential yield of fis stocks. FAO Fis. Tec. Pap., 4: 47 pp. Bourne, N Pacific Oysters, (Crassostrea gigas Tunberg), in Britis Columbia and te Sout Pacific Islands. In Exotic species in mariculture. Edited by R. Mann. MIT Press, Cambridge, MA. pp Boutillier, J.A., Lauzier, R.B., Pillips, A.C. and Barton, L Framework for a Tanner crab (Cionoecetes tanneri and C. angulatus) fisery in waters off te West Coast of Canada. Canadian Stock Assessment Secretariat Researc Document 98/5. Carlton, J.T. and Mann, R Transfers and World-wide introductions. p In: V.S. Kennedy, R.I.E. Newell and A.F. Eble [eds.]. Te Eastern oyster Crassostrea virginica. Maryland Sea Grant College Publ. UM-SG-TS Coan, E.V., Valentic Scott, P. and Bernard, F.R Bivalve Seasells of Western Nort America. Santa Barbara Museum of Natural History Monograps. Santa Barbara, CA. Cocran, W.G Sampling tecniques. 3rd Ed. Jon Wiley and Sons, New York. Deriso, R.P. 98. Relationsip of fising mortality and growt and te level of maximum sustainable yield. Can. J. Fis. Aquat. Sci., 39:

28 Elliot, J.M Some metods for te statistical analysis of samples of bentic invertebrates. Sci. Publ. Freswater Biol. Assoc p. Elsey, C.R. 93. Japanese oysters breed in Ladysmit Harbour, B.C. Fis. Res. Board Can. Pac. Prog. Rep. 5: 3-5. Elsey, C.R Distribution of oyster larvae. Fis. Res. Board Can. Pac. Prog. Rep. : 9-0. Elsey, C.R. and Quayle, D.B Distribution of oyster larvae from te 936 spawning in Ladysmit Harbour. Fis. Res. Board Can. Pac. Prog. Rep. 40: 8-0. Fox, W.W., Jr An exponential surplus-yield model for optimizing exploited fis populations. Trans. Am. Fis. Soc. 99(): Francis, R.C., 974. Relationsip of fising mortality to natural mortality at te level of maximum yield under te logistic stock production model. J. Fis. Res. Board Can. 3: Garcia, S., P. Sparre, and J. Csirke Estimating surplus production and maximum sustainable yield from biomass data wen catc and effort time series are not available. Fiseries Researc 8: 3-3. Galtsoff, P. B. S Te American oyster, Crassostrea virginica Gmelin. U.S. Fis. Wildlife Service. Fis. Bull. 64:iii p. Gerritsen, H.D. and McGrat, D Precision estimates and suggested sample sizes for lengt-frequency data. Fis. Bull. 06: 6-0. Gillespie, G.E Preliminary review of experimental arvest rates in te depuration fisery for intertidal clams. DFO Can. Stock Assess. Sec. Res. Doc. 000/. 57 p. Gillespie, G.E Distribution of non-indigenous intertidal species on te Pacific Coast of Canada. Nipp. Suis. Gakk. 73(6): Gillespie, G.E Status of te Olympia Oyster, Ostrea lurida Carpenter 864, in Britis Columbia, Canada. J. Sellfis Res. 8(): Gillespie, G.E. and Kronlund, A.R A manual for intertidal clam surveys. Can. Tec. Rep. Fis. Aquat. Sci p. Gillespie, G.E., Kronlund, A.R. and Heritage, G.D. 998a. Intertidal clam stock estimates for selected depuration arvest beaces 994. p In: B.J. Waddell, G.E. Gillespie and L.C. Walters [eds.]. Invertebrate working papers reviewed by te Pacific Stock Assessment review Committee (PSARC) in 995. Pt.. Bivalves. Can. Tec. Rep. Fis. Aquat. Sci. 4. Gillespie, G.E. Kronlund, A.R. and Heritage, G.D. 998b. Assessment of Manila clam stocks at Savary Island, Britis Columbia 995. p In: B.J. Waddell, G.E. Gillespie and L.C. Walters [eds.]. Invertebrate working papers reviewed by te Pacific Stock Assessment review Committee (PSARC) in 995. Pt.. Bivalves. Can. Tec. Rep. Fis. Aquat. Sci. 4. Gillespie, G.E., Norgard, T.C. and Scurra, F.E. 00. Status of Manila clam (Venerupis pilippinarum) stocks in Area 7, wit a proposal for active management of a data-limited fisery. DFO Can. Stock Assess. Sec. Res. Doc. 00/ p. 0

29 Gillespie, G.E., Bower, S.M., Marcus, K.L. and Keiser, D. 0. Biological synopsis for tree exotic mollusks, Manila clam (Venerupis pilippinarum), Pacific Oyster (Crassostrea gigas) and Japanese scallop (Mizuopecten yessoensis) licenced for aquaculture in Britis Columbia. DFO Can. Sci. Advis. Sec. Res. Doc. 0/03. v + 97 p. Gulland, J.A., 97. Te Fis Resources of te Ocean. Fising News (Books), West Byfleet, 55 pp. Harding J.M. and R. Mann, 006. Age and Growt of Wild Suminoe (Crassostrea Ariakensis, Fajita 93) and Pacific (C. Gigas, Tunberg 973) Oysters from Lanzou Bay, Cina. Journal of Sellfis Researc, Vol. 5, No., Hendricks, A.A Te Matematical Teory of Sampling. Te Scarecrow Press, New Brunswick. N.J. Hoenig, J.M Empirical use of longevity data to estimate mortality rates. Fisery Bull. (U.S.) 8: IEC International Wild oyster fisery review report. Report submitted to BC Ministry of Agriculture and Lands, Courtenay, BC. 33 p. + app. Kingzett, B.C. and Bourne, N.F Assessment of Intertidal Clam Populations Surveys at Seal Island, Britis Columbia, pp In: B.J.Waddell, G.E. Gillespie, and L.C. Walters [eds.]. Invertebrate Working Papers reviewed by te Pacific Stock Assessment Review Committee (PSARC) in 995. Part. Bivalves. Can. Tec. Rep. Fis. Aquat. Sci. 4. Krebs, C.J Ecological Metodology. Addison Wesley Longman, New York. Kronlund, A.R., Gillespie, G.E., and Heritage, G.D Survey metodology for intertidal bivalves. pp In: B.J.Waddell, G.E. Gillespie, and L.C. Walters [eds.]. Invertebrate Working Papers reviewed by te Pacific Stock Assessment Review Committee (PSARC) in 995. Part. Bivalves. Can. Tec. Rep. Fis. Aquat. Sci. 4. Lauzier, R.B., Walters, L.C., Hajas, W., Lessard, J. and Wylie, E Progressing to a scientifically-based assessment and management system for renewed commercial pink and spiny scallop fiseries off te Britis Columbia coast. DFO. Can. Sci. Advis. Sec. Res. Doc. 005/ p. + i. Norgard, T., Davies, S., Stanton, L. and Gillespie, G.E Evaluation of survey metodologies for monitoring Olympia Oyster (Ostrea lurida Carpenter, 864) populations in Britis Columbia. DFO Can. Sci. Advis. Sec. Res. Doc. 00/006. iv + 56 p. Parisien, R.W. 97. Te Fiseries Act: origins of Federal delegation of administrative jurisdiction to te Provinces. Report submitted to Policy Planning and Researc Service, Environment Canada. 36 p. + app. Pauley, G.B., Van Der Ray, B. and Troutt, D Species profiles: life istories and environmental requirements of coastal fises and invertebrates (Pacific Nortwest) Pacific Oyster. U.S. Fis Wildl. Serv. Biol. Rep. 8(.85). U.S. Army Corps of Engineers, TR EL p. Quayle, D.B Distribution of introduced marine molluscs in Britis Columbia waters. J. Fis. Res. Board Can. (5): Quayle, D.B Pacific Oyster culture in Britis Columbia. Fis. Res. Board Can. Bull p.

30 Quayle, D.B. 97. Pacific Oyster raft culture in Britis Columbia. Fis. Res. Board Can. Bull p. Quayle, D.B Pacific Oyster culture in Britis Columbia. Can. Bull. Fis. Aquat. Sci p. Quinn, G.P and Keoug, M.J. 00. Experimental Design and Analysis for Biologists. Cambridge University Press, Cambridge, UK. Rao, J.N.K. and Wu, C.F.J Resampling inference wit complex survey data. J. Amer. Stat. Soc. 83(40): 3-4. Ruesink, J.L., Lenian, H.S., Trimble, A.C., Heiman, K.W., Miceli, F., Byers, J.E. and Kay, M.C Introduction of non-native oysters: ecosystem effects and restoration implications. Annu. Rev. Ecol. Evol. Syst. 36: Sattertwaite, F.E An approximate distribution of estimates of variance components. Biometrics Bull. : 0-4. Scaefer, M.B Some aspects of te dynamics of populations, important for te management of te commercial fiseries. Bull. Inter-Amer. Trop. Tuna Comm. (): 56 p. Sitter, R.R. 99. A resampling procedure for complex survey data. J. Amer. Stat. Assoc. 87: Sloan, N.A., Bartier, P.M. and Austin, W.C. 00. Living marine legacy of Gwaii Haanas. II: Marine invertebrate baseline to 000 and invertebrate-related management issues. Parks Canada Tec. Rep. Ecosystem Sci p. Strand, A., Blanda, E., Bodvin, T., Davids, J. C., Jensen, L. F., Holm-Hansen, T. H., Jelmert, A., Lindegart, S., Mortensen, S., Moy, F. E., Nielsen, P., Norling, P., Nyberg, C., Cristensen, H. T., Vismann, B., Holm, M. W., Hansen B. W., and Dolmer, P. 0. Impact of an icy winter on te Pacific Oyster (Crassostrea gigas Tunberg, 793) populations in Scandinavia. Aquatic Invasions 7(3): Surry, A.M., Fong, K.H., Ruterford, D.T. and Nguyen, H. 0. Update to te assessment framework for te pink and spiny scallop (Clamys rubida and C. astata) dive fisery in waters off te west coast of Canada. DFO Can. Sci. Advis. Sec. Res. Doc. 0/3. x+ 65 p. Wiegert, R.G. 96. Te selection of an optimum quadrat size for sampling te standing crop of grasses and forbs. Ecology 43: 5-9.

31 TABLES Table. Landings (Kt) and landed values (x06 $Cdn) of cultured and commercially arvested sellfis in Britis Columbia, a) Landings Category Cultured Sellfis Clams Oysters Scallops/Oter 0. <0. <0. <0. < Wild Sellfis Clams Oter Wild Oysters b) Landed value Category Cultured Sellfis Clams Oysters Scallops/Oter Wild Sellfis Clams Oter Notes: Data from BC MoA (999-0a,b) except Wild Oyster landings from IEC International (006). Cultured Oter includes mussels and scallops, depending on te year. 3 Wild Sellfis includes commercial landings of crab, srimp, prawn, scallops, sea cucumbers, geoducks & sea urcins. 4 Wild Oter includes octopus, squid and oter unspecified sellfis. 5 Not available. 3

32 Table. Area and number of strata surveyed at eac location by te BC MoA during te surveys from 005 to 00. Total Area (m ) # of Strata Total Area (m ) # of Strata Total Area (m ) # of Strata Total Area (m ) # of Strata Total Area (m ) # of Strata Total Area (m ) # of Strata Location Atrevida Reef ,843 Bird Cove , Blind Bay ,54 4 8, Boot Bay ,603 9 Carrington Bay , Comox Harbour , Davie Bay ,78 Dog Bay - - 6, East Hernando , ,857 8 False Bay ,6 5 Galiano Bay 4, Goliat Bay , Harwood Island ,8 5 Hernando Reef , Hisnit Inlet , Jane Bay - - 7, Killam Bay ,83 8 Kuper Island , Lloyd Point , Marvinas Bay , Mooya/Crescent , Mouat Bay , ,97 6 Myrtle Rocks ,060 4 Narrows Inlet , Perketts Creek , Scottie Bay , Seaford ,736 0 Seal Islets 0, Secelt Inlet , Sark Spit ,833 Single Spit - -, Smelt Bay , ,05 6 St Vincent Bay ,476 3 Stag Bay , Storm Bay , Tasis Cannel , Teakerne Arm , Teodosia , ,93 3 Toquart Bay , Toquart River , Union Bay , Vanguard Bay , West Hernando ,67 9 Westview ,643 Total 4,76 5 6, , , , ,6 30 4

33 Table 3. Estimated abundance (oysters/quadrat), biomass (kg/quadrat) and precision for Pacific Oyster surveys using four quadrat sizes. Data from eac quadrat size were standardized to m. A) Abundance Abundance Strata Abundance Estimates Survey Abundance Estimates Location Quadrat Size (cm) Stratum Stratum Area (m ) quadrat # Mean (#/m ) SD SE Strata Stock Estimate 95% Confidence Interval of te Strata Stock Estimate Strata Precision (stock est/95%) Total Stock Estimate 95% Confidence Interval of te Total Stock Estimate Neck Point 5x ,50 0,9 5.8% Neck Point 50x ,669 7,095.% Neck Point 75x ,00 5,54 0.8% Neck Point 00x ,04 5, % Precision Sack 5x ,867 37,37 5.0% 9,099 54,73 4.8% Sack 5x ,3 37, % Sack 50x ,37 6, % 87,63 36,90 9.7% Sack 50x ,304 4,04 0.5% Sack 75x ,653, % 74,039 30, % Sack 75x ,386 0, % Sack 00x ,33, % 0,56 3,89 5.5% Sack 00x ,94,8 6.7% B) Biomass Biomass Strata Biomass Estimates Survey Biomass Estimates Location Quadrat Size (cm) Stratum Stratum Area (m ) quadrat # mean (kg/m ) SD SE Strata Stock Estimate 95% Confidence Interval of te Strata Stock Estimate Strata Precision (stock est/95%) Total Stock Estimate 95% Confidence Interval of te Total Stock Estimate Neck Point 5x ,08,43 7.6% Neck Point 50x , % Neck Point 75x , % Neck Point 00x , % Precision Sack 5x ,685 9, % 6,06 3,878.7% Sack 5x ,376 0,08.8% Sack 50x ,97 6,67 37.% 5,860 9,058 7.% Sack 50x ,888 5, % Sack 75x ,559 4, % 49,49 7, % Sack 75x ,870 4, % Sack 00x ,706 5, % 56,56 7,9.6% Sack 00x ,856 4,766.9%

34 Table 4. Results of Wiegert s cost-benefit analysis using four quadrat sizes. Quadrats were not timed at Neck Point; te table below uses te mean time from Sack Island survey. Mean times for eac quadrat type differ in eac stratum at Sack Island; tis table sows times for eac quadrat size and calculates a standardized cost. Te upper part of te table uses te mean time to complete eac quadrat size at Sack Island in stratum and te lower part of te table uses te mean time for stratum. Te lowest standardized cost x standardized variance (igligted) is te optimal quadrat size for eac scenario and is calculate by multiplying te standardized cost by te standardized relative variance. Cost/Quadrat Area (sec/m) is te mean time to survey a quadrat divided by te quadrat area. Quadrat size (cm) Quadrat area (m ) Sack Island Stratum Standard Deviation Sack Island Stratum Neck Point Cost Mean Time (Mean Time from Quadrats in Stratum at Sack Island) (sec) Standardized Cost Calculated from Cost Mean Time Standardized cost x Standardized Variance Sack Island Stratum Sack Island Stratum Neck Point Cost/Quadrat Area 5x ,96,758 5, x ,77 6,75 3, x ,93 5,758, x00 5,3 5,574, Quadrat size (cm) Quadrat area (m ) Sack Island Stratum Standard Deviation Sack Island Stratum Neck Point Cost Mean Time (Mean Time from Quadrats in Stratum at Sack Island) (sec) Standardized Cost Calculated from Cost Mean Time Standardized cost x Standardized Variance Sack Island Stratum Sack Island Stratum Neck Point Cost/Quadrat Area 5x ,96,758 5, x ,77 6,75 3, x ,93 5,758, x00 5,3 5,574,

35 Table 5. Results of estimated precision analysis. Mean and standard deviation are calculated from aggregate weigt of Pacific Oysters by quadrat. Te estimated precisions are te number of quadrats required to obtain eac precision level. # of Quadrats Required to Reac te Precision Level Below Location Stratum Quadrat Size (cm) Mean (kg/m ) Standard Deviation Stratum Area (m ) Actual Sampl e # Actual Survey Precision % 0% 5% 0% 30% 40% 50% 60% 70% 80% 90% 00% Neck Point 5x Neck Point 50x Neck Point 75x Neck Point 00x Sack Island 5x Sack Island 50x Sack Island 75x Sack Island 00x Sack Island 5x Sack Island 50x Sack Island 75x Sack Island 00x

36 Table 6. Estimated mortality by maximum age using Hoenig s (983) metod. Maximum Age (yr) Mortality estimate (Hoenig 983) 5 36% 0 8% 5 4% 30 0% 40 6% 8

37 Table 7. Summary of results for te Gulland (97), Scaefer(954) and Fox (970) arvest rate models for Pacific Oysters using maximum ages of 0, 30 and 40 years. HR = Harvest Rate, and MSY= Maximum Sustainable Yield. Harvest and biomass are from te 00 fisery data provided by te Province. Hoenig (983) Mortality Rate Estimate 00 Harvest (tonnes) 00 Biomass Estimate (tonnes) Model Constant X Fising Mortatily Rate (Fmsy) Gulland Model (97) Scaefer model (954) Fox model (970) MSY HR MSY HR MSY HR Maximum Age = 5 Years % 53 5% 34 5% % 337 7% 9 6% % 45 8% 355 7% % 55 0% 40 8% Maximum Age = 0 Years % 4% 3 4% % 80 5% 5 5% % 346 7% 99 6% % 45 8% 348 7% Maximum Age = 30 Years % 0 4% 0 4% % 8 4% 8 4% % 7 5% 46 5% % 38 6% 78 5% Maximum Age = 40 Years % 4 4% 07 4% % 08 4% 05 4% % 36 5% 3 4% % 7 5% 46 5% 9

38 FIGURES Figure. Areas supporting commercial arvest of Pacific Oysters (Crassostrea gigas) in Jervis and Secelt Inlets, Lasqueti Island and Texada Island, Strait of Georgia, Britis Columbia. Source: BC MoA (0f). 30

39 Figure. Areas supporting commercial arvest of Pacific Oysters (Crassostrea gigas) in Desolation Sound, Britis Columbia. Source: BC MoA (0g). 3

40 Figure 3. Areas supporting commercial arvest of Pacific Oysters (Crassostrea gigas) in Barkley and Nootka Sounds, West Coast of Vancouver Island, Britis Columbia. Source: BC MoA (0). 3

41 Figure 4. Areas supporting commercial arvest of Pacific Oysters (Crassostrea gigas) in Stuart Cannel and Malaspina Inlet, Strait of Georgia, Britis Columbia. Source: BC MoA (0i). 33

42 Figure 5. Survey design using Stratified Random Sampling. Te baseline is establised along one side of te bed (large green amorpous sape), setting up an x and y axis grid (stratum), ten random quadrats (red squares) are selected trougout te stratum. Te grid sow all te possible sampling locations and te two areas sow tat tis design can be setup on multiple strata on te same beac. 34

43 Figure 6. Nested quadrats - smallest is te 5 cm x 5 cm, 50 cm x 50 cm, 75 cm x 75 cm, up to 00 cm x 00 cm. 35