UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Oceanographic factors affecting the catchability of Pacific Ocean perch, Sebastes alutus (Gilbert) Scott, Beth Emily 1990

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1990_A6_7 S35_5.pdf [ 5.7MB ]
Metadata
JSON: 831-1.0302382.json
JSON-LD: 831-1.0302382-ld.json
RDF/XML (Pretty): 831-1.0302382-rdf.xml
RDF/JSON: 831-1.0302382-rdf.json
Turtle: 831-1.0302382-turtle.txt
N-Triples: 831-1.0302382-rdf-ntriples.txt
Original Record: 831-1.0302382-source.json
Full Text
831-1.0302382-fulltext.txt
Citation
831-1.0302382.ris

Full Text

OCEANOGRAPHIC FACTORS AFFECTING THE CATCHABILITY OF PACIFIC OCEAN PERCH Sebastes alutus (Gilbert). By BETH EMILY SCOTT B.Sc, Simon Fraser University, 1983. A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Zoology)  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA July 1990 © Beth Emily Scott, 1990  In  presenting this  degree at the  thesis  in  University of  partial  fulfilment  of  of  department  this thesis for or  by  his  or  requirements  British Columbia, I agree that the  freely available for reference and study. I further copying  the  representatives.  an advanced  Library shall make it  agree that permission for extensive  scholarly purposes may be her  for  It  is  granted  by the  understood  that  head of copying  my or  publication of this thesis for financial gain shall not be allowed without my written permission.  Department The University of British Columbia Vancouver, Canada  DE-6 (2/88)  ii  Abstract  A main concern in fisheries science has been to identify an accurate index of fish abundance. An underlying paradigm in the science has been that the amount of effort (calculated in hours and standardized for boat size) spent fishing was the best variable to be used to account for the variation in catches. The use of the ratio, catch per unit of effort (cpue), assumes that variations in fish abundance are due to human-controlled processes above the ocean's surface. It does not account for variation due to oceanographic processes that affect fish behaviour and movement patterns below the ocean's surface. This study investigated the possibility that oceanographic factors such as temperature, salinity and depth could have effects on the variations observed in the apparent abundance of a demersal rockfish, Pacific Ocean Perch (Sebastes alutus. Gilbert). Simultaneous monitoring of physical variables and fish abundance estimation was achieved by attaching oceanographic equipment to the fishing gear of commercial vessels, monitoring the acoustic equipment and sampling the fish catch. It was found that Perch prefer a temperature range from 6.7 °C, down to at least 4.8 ° C and that their movement patterns are linked to the movement of these temperatures by coastal wind patterns. Perch prefer areas with steep bathymetry, characterized by frontal activity due to interactions between the local bathymetry and tidal currents. Concerns that sampling only from highly successful commercial vessels may have biased abundance estimates, prompted the analysis of historical records of fish catch and government research surveys. Analyses between different boat sizes, different areas and different seasons from the original historical data base and a corrected subset revealed that it was mainly differences between areas that was responsible for the biasing of estimates. Deeper areas predictably produced more fish for all sizes of boats, but were fished more often by the larger boats used in the study. Therefore the field abundance estimates are likely to be biased towards areas of larger perch concentrations.  iii  Table of Contents Section  Page  Abstract  ii  List of Tables  vi  List of Figures  vii  Acknowledgments  ix  Chapter 1. General Introduction  1  Chapter 2. At Sea Monitoring of Physical and Biological Factors Affecting the Abundance of Perch  6  Introduction  6  Methods  7  Study Species, Site and Trawl Vessels  7  Measurements of Fish Abundance  9  Fish Abundance Estimates via Acoustic Equipment  11  Oceanographic Studies  11  Results  12  Study Sites and Trawl Vessels  12  Measurements of Fish Abundance  13  Fish Abundance Estimates - Effects of Biological Factors  18  Sexual Maturity  18  Stomach Analysis  18  Oceanographic Studies  19  Large Spatial and Temporal Scale Pattern in the QCS  19  Small Spatial and Temporal Scale Pattern within Perch Sites  24  Temperature, Salinity and Depth within Perch Sites  26  Fish Abundance versus Temperature, Salinity and Depth  26  iv  Discussion  28  Study Sites and Trawl Vessels  28  Measurements of Fish Abundance  31  Fish Abundance Estimates - Effects of Biological Factors  32  Oceanographic Studies  33  Large Spatial and Temporal Scale Pattern in the QCS  33  Small Spatial and Temporal Scale Pattern within Perch Sites  33  Temperature, Salinity and Depth within Perch Sites  35  Fish Abundance versus Temperature, Salinity and Depth Conclusions  Chapter 3. Comparisons of DFO Commercial Data Base with Field Data  35 38  39  Introduction  39  Methods  40  Results  43  Effects of Boat, Area, Month and Year  51  Boat  51  Area  51  Month  60  Year  65  Discussion Effects of Boat, Area, Month and Year Conclusions  67 69 70  V  Chapter 4. Analysis of DFO Research Survey Information Introduction D F O Research Surveys  72 72 72  Methods  73  Results  74  Discussion  76  Conclusions  77  Chapter 5. Comparison of the Upwelling Index with the Depth of Catch  78  Introduction  78  Methods  79  Results  81  Discussion and Conclusions  81  Chapter 6. General Discussion  82  Discussion  82  Conclusion  85  Literature Cited  88  Appendix A. DFO Sub-Management Areas - Names and Sizes  97  vi  List of Tables  Table  Page  2-1. Mean values and standard errors for all abundance estimates calculated on a per-tow basis 2- H. Correlation coefficients (r),  14 values and probability levels of  relationships between abundance estimates and tow characteristics from tows containing perch  17  3- 1. Correlation coefficients (r) and their respective level of significance (p) for all the combinations of relationships between abundance estimates and tow characteristics for each data set 3-II. Regression parameters for three relationships between the data sets  44 49  3-IQ. Comparisons between the actual amount of perch caught at sea per-area per-trip, to the corresponding entries in the DFO groundfish data base  50  3-IVa. Comparisons of effort (hr) categories by boat size  52  3-IVb. Comparisons of depth (m) categories by boat size  53  3-Va. Comparison of catch (mt) categories by area  55  3-Vb. Comparison of proportion (%) categories by area.  56  3-Vc. Comparison of cpue (mt/hr) categories by area  57  3-Vd. Comparison of effort (hr) categories by area  58  3-Ve. Comparison of depth (m) categories by area  59  3-Via. Comparison of cpue (mt/hr) categories by month  61  3-VIb. Comparison of depth (m) categories by month  62  vii  List of Figures Figure  Page  2.1 Bathymetry of Queen Charlotte Sound and Hectic Strait  8  2.2 Relationships between field estimates of fish abundance and tow characteristics . . . . 16 2.3 Temperature and Salinity profiles taken from areas containing perch  20  2.4 Isopleths of temperature and salinity from perch tows  22  2.5 Pictorial representation of the wind changes along the B. C. coast and corresponding effects on the deeper off-shore waters 2.6 Three consecutive tows for perch around a small seamount near Cape St. James  23 25  2.7 Average temperature and salinity values for the horizontal section of the tows targeting on perch  27  2.8 The relationships between abundance estimates and the mean temperatures at which perch were caught  29  2.9 Isopleths of abundance estimates  30  2.10 Satellite image showing areas of cool water in Queen Charlotte Sound  36  3.1 Relationships between abundance estimates and tow characteristics for the DFO historical groundfish data base  45  3.2 Relationships between abundance estimates and tow characteristics for the subset of the DFO data base containing only those trips with more than 25% perch in the total catch per-area  46  3.3 Relationships between abundance estimates and tow characteristics for the subset of the DFO data base containing only those trips with less than 20 hours of effort per-area  47  3.4a Mean per-trip catch of perch versus the mean depth of catch for each area  54  3.4b Areas fished in each year  54  viii  3.5  Mean monthly values and standard error bars for all abundance estimates and tow characteristics  3.6  Mean monthly values for all abundance estimates and tow characteristics for three groups of areas  3.7  63  64  Annual means and standard error for abundance estimates and tow characteristics.... 66  4.1 Survey information from the Goose Island Gully region  75  5.1 Mean monthly depth of commercial perch catch versus monthly upwelling i n d e x . . . . 80  ix  Acknowledgements  I thank P. A. Larkin, P. LeBlond, J. D. McPhail, T. Parsons, R. I. Perry and N. J. Wilimovsky for their informative inputs and encouragement as well as their excellent recommendations upon reviewing this manuscript. I wish to acknowledge the Science Council of British Columbia, the Deep Sea Trawlers Association, especially Doug March (president) for their financial support. I would like to express my appreciation to all the skippers and crew of the vessels whose cooperation made this study possible. Thanks to Carl Walters for the opportunity to par-take in this adventure. Thanks to fellow institute animals and oceanographic graduates, without whom I never would have made it. And very special thanks goes to X. Lambin for his constructive criticism and unfailing support.  1  Chapter 1  General Introduction  A widespread problem in ecology is the estimation of the abundance of a species. This problem is especially acute in fishery science where the majority of the data used to analyze all aspects of fish populations comes from catches of commercial fishermen. Commercial fisheries target on abundant and market-desired species. The fact that commercial data are biased has long been recognized and has led to the development of survey sampling by research vessels (Doubleday and Rivard 1981). However, work at sea is expensive and time consuming, so that research surveys seldom, if ever, have the same spatial or temporal degree of coverage as the commercial fleet. Regardless of the source of data for estimation of fish abundance, the information is characterized by a large degree of spatial and temporal variability (Byrne et al. 1981). Until there is a substantial increase in the accuracy of abundance estimation, policy and management decisions cannot be made with a high degree of confidence. Spatial variability in abundance of a species usually indicates that the species is clumped in space, which in turn suggests habitat preference. Temporal variability within the same location, indicates small scale (10 -100 km) movements in space, such as daily or seasonal migrations, reflecting habitat preferences and movement/migration patterns. If the habitat preferences and reasons for movements were known and qualified, much of the variability could be categorized and the level of uncertainty in fish abundance estimation greatly reduced. To date, little effort in fisheries science has been directed towards integrating quantitatively defined fish habitat and abundance estimation. There has been a great deal of work that focuses on the effects of the environment on fish physiology and behaviour (for an extensive review see Laevastu and Hayes 1981). As well, there have been many studies that focus on large scale (100's of km) environmental variables and correlate them with fluctuations in recruitment patterns (Leggett etal. 1984, Shepherd etal. 1984, Drinkwater and Myers 1987, Norton 1987,  2  Frank and Carscadden 1989). Despite the fact that the marine environment is known to be a shifting, continuously fluctuating window through which we view fish behaviour, little has been done to incorporate this knowledge into abundance estimation. Recent efforts by Perry et al. (1988) , Carscadden et al. (1989) and Lough et al. (1989), have shown that it is possible to identify preferred water mass characteristics with associated fish species as well as to correlate abundance with slight differences in water mass characteristics. Work by Rose and Leggett (1989) shows that movement patterns and apparent abundance estimates of cod (Gadus morhua) are directly related to movement patterns of water masses. What is also apparent from these studies is that the physical marine environment is only now becoming understood and measurable at a level sufficiently detailed to enable correlation of fish habitat preference with water mass characteristics. An underlying paradigm in fisheries science is that the cause of much of the variability in abundance estimation is due to the mechanical and human parameters such as the size of boat used or the knowledge level of the skipper (Ricker 1944, Gulland 1956, Beverton and Holt 1957). Until recently little consideration has been given to the effects of processes occurring below the surface, such as tidal currents or large scale wind events that influence fish behaviour and which may therefore, have drastic effects on abundance estimation (Jones and Scholes 1980, Rose and Leggett 1989). The belief that the characteristics of gear and the skill of fishermen are the major factors which influence fish catch is reflected in the standard equation used to calculate fish abundance:  Fish abundance (mt) = [catch (mt) / effort (hr)] X catchability (hr)  (1)  Effort and catchability are defined in units oftimespent fishing and standardized for vessel size. The term catch/effort is commonly referred to as catch per unit of effort or cpue (Gulland 1964). Catchability is the probability of catching an individual fish per unit of standardized time (Palaheimo and Dickie 1964). Cpue is used as an index of abundance under the assumption that  3  the catchability term remains constant (Gulland 1964) despite the fact that Palaheimo and Dickie (1964) clearly indicated that catchability does not remain constant but varies inversely with stock abundance and the area occupied by a stock. Since then, numerous studies on a variety of fisheries have concluded that catchability is not constant and is indeed inversely related to both stock abundance and stock area (MacCall 1976, Peterman and Steer 1981, Crecco and Savoy 1985, Winters and Wheeler 1985, Crecco and Overholtz 1990). The inappropriate assumptions concerning the behaviour of fish and fishermen that lead to the inverse relationship between catchability, stock abundance and size are 1) that fish are evenly distributed, 2) that effort is distributed randomly and 3) that each unit of effort is independent (Radovich 1976). The second and third assumptions relate directly to effort, while the initial assumption of even distribution of fish leads to the second violated assumption. It is the definition of effort that is at the heart of the problem and is the underlying paradigm that needs to be reassessed if the precision of abundance estimation is to be improved. An alternative method of defining effort was suggested by Treschev (1964), who was attempting to standardize effort between different fisheries. He derived formulae for the volume of water "swept" for each of the commonly used commercial gear types. Treschev argued that standardizing effort as a unit of volume would allow simple comparisons between different fisheries as well as within a fishery. Effort measured as volume is also more biologically meaningful as it identifies the actual amount of water column that is affected by the fishing gear. For a trawl fishery, effort would be defined in terms of the area of the net opening multiplied by the distance the net was towed.  Effort (m^) = area of net opening (m^) X time towed (hr) X speed (m/hr)  (2)  Fish abundance (mt) = [catch (mt) / effort (m^)] X catchability (m^)  (3)  Thus:  4  When effort is measured in units of volume, catchability also has to be in units of volume. The catch/effort term in equation (3) represents catch per unit of volume and I will refer to it as cpuv. Treschev's ideas have not taken hold, perhaps because the use of equation (3) requires a definition of catchability in volumetric terms. The biological interpretation of volumetric catchability is the 3-dimensional space a fish species can potentially occupy (ie. fish habitat). Fish abundance estimates using cpuv are simply the ratios of the volume of water swept during fishing operations to the volume of water potentially occupied by a species. Catch per unit of volume, cpuv, is a direct measurement of density. Effort and catchability defined as volumes, demand that the dimensions of potential fish habitat be known. Seasonal, as well as year to year changes in the geographical area and depth range that any particular fish species inhabits must be investigated before the use of this formula is possible. Using this alternative methodology, catchability will not vary inversely with the area or density of stock size because catchability will be defined as the probability of catching a fish within the geographical range and depth that the stock presently occupies. Abundance estimation, calculated in this manner, can be assessed from a site-specific viewpoint up to a global scale, depending on the amount of detailed knowledge of the habitat preferences of the species in question. The main benefit of thinking of effort in terms of volume swept and using cpuv as an indicator of abundance is that it forces attention to fish behaviour and processes below the surface instead of focusing solely on fishing gear and fishermen's behaviour. The purpose of this study is to investigate the possibility of quantifying fish habitat using some measurable physical characteristics of the water masses in which they are found. To identify physical oceanographic characteristics or processes that influence the catchability of a particular fish species differences in apparent abundance estimates (cpue and cpuv) are compared to the temperature, salinity and depth in the exact locations of fish catch. The hypothesis being investigated is that those physical factors which affect the catchability, (the probability of a fish being caught), will give insight into which oceanographic characteristics can be used to define habitat as well as indicate which processes influence movement patterns.  5  A demersal rockfish, the Pacific Ocean Perch Sebastes alums (Gilbert), (in this study referred to as perch), was chosen as the species of study because it is abundant in the commercial catch and there is a relatively large amount of relevant biological and historical information available (Alverson 1960, Westrheim 1970, Gunderson 1971,72, 74, Westrheim 1975). To identify the effects of physical factors on the catchability of perch, simultaneous measurements of both the water mass characteristics (temperature, salinity, depth) and abundance estimation of perch were made on board commercial trawl vessels over a ten month period in the Queen Charlotte Sound region of the northeast Pacific coast. The information was analyzed for patterns between catch rates and physical characteristics of the ocean. It was found that a combination of temperature, depth and bathymetric preferences could be used to predict changes in apparent abundance. The restriction of the field data collection from a few vessels and sites did not allow separation of the effects of vessels and sites on abundance estimation. Large historical data sets available for perch were examined in an attempt to determine how these variables may have influenced the field results. The remainder of this thesis is divided into 5 chapters. Chapter 2 deals with the information collected on commercial vessels, Chapter 3 covers the investigations of the historical perch catch records. Chapters 4 and 5 explore other biological and physical data sets concerning perch and the Queen Charlotte Sound region. The final chapter contains an encompassing discussion and conclusion.  6  Chapter 2 At Sea Monitoring of Physical and Biological Factors Affecting the Abundance of Perch  Introduction  Identifying effects that physical oceanographic variables such as temperature, salinity and depth, have on the catchability of a fish species requires simultaneous monitoring of water mass characteristics and fish abundance estimates (Perry et al. 1988). Measuring the physical environment on a site-specific (small scale - 10 km basis), and comparing differences in fish catch may provide a method of identifying preferred fish habitat. Temperature has long been used for fish habitat definition, especially in freshwater (Brandt et al. 1980, Christie and Regier 1988), but the marine environment is more complex and other variables such as the density structure of the water column also need to be examined (Longhurst 1984). In the ocean, both temperature (T) and salinity (S) are used to characterize water masses. For example TS "signatures" are used to track the California undercurrent off the British Columbian coast (Freeland and Denman 1982). For the present study, TS signatures of the water masses in which perch were found were compared between different areas and seasons to identify any repeating patterns of occurrence. Small scale phenomena (10 km) were of particular interest in identifying site-specific reasons for fish aggregation. Collecting information on board commercial vessels allows direct observations of both fishermen behaviour and actual amount of fish catch providing the opportunity to appreciate and correct for any biases in the apparent abundance estimates. Per-tow estimates must be examined because fish behaviour may change on a seasonal or even daily scale due to such things as sexual maturity and feeding behaviour. Abundance estimates can be calculated in a variety of ways. To determine which estimate or combination of estimates best represent changes in apparent abundance, a variety of estimates are examined along with corresponding tow characteristics.  7  Methods  Study Species, Site and Trawl Vessels Pacific Ocean Perch (Sebastes alutus). perch, a demersal rockfish, has been commercially harvested in the northeast Pacific for more than 30 years. Records of locations and amounts of catch have been stored by the Department of Fisheries and Oceans (DFO) since the 1950's and DFO research surveys have been conducted on the species since the early 1960's to the present. Adult perch are thought to migrate to minimum depths of 140 m in the summer and to depths greater than 400 m, perhaps even as deep as 1000 m, in the winter months (Alverson 1960, Lyubimova 1968, Westrheim 1970, Gunderson 1971). Independent of depth, perch prefer rocky, steep bathymetry (Carlson and Haight 1976, Carlson and Straty 1981). This association with bottom type, slope and lack of large scale annual movement (less than 100 km) suggests that they have a strong preference for a particular habitat. The majority of the perch landed in British Columbia waters (64 % from 1980 to 1985, Tyler et al, 1986) are caught in the Queen Charlotte Sound region (QCS) off the north coast of British Columbia, Canada (Figure 2.1). The region extends from the southern end of the Queen Charlotte Islands (52.20 °N, 131.00 °W), to the northerntipof Vancouver Island (50.00 °N, 128.00 °W). The continental shelf runs parallel to the coastline, defining the oceanic boundaries of the sound. The QCS has an abundance of steep canyon walls and rocky pinnacles. The major topographical features of the QCS are three large gullies; Goose, Mitchell and Moresby (Figure 1), that run perpendicular to the coastline. They are shallow near the coast (100 m) and gradually deepen (300 m) as they spread out over the shelf edge. At the shelf break, the depth quickly drops to 2000 m. There are numerous smaller, steep canyons along the continental shelf. The trawl vessels used in the study were chosen from a list of vessels that brought in perch catches over all seasons from the QCS and did so consistently for the preceding three years (1985 - 1987). The use of various vessels and skippers provided more knowledge of the fishing grounds than the use of any one fishermen would have because skippers tend to make use of only  8  9  those sites with which they are most familiar and they do not all use the same sites. No information is given as to the precise location of tows, in order to provide the confidentiality requested by the fishermen. However, the codes for the DFO sub-management areas that the sites are found within are stated and the names and sizes of the DFO areas are given in Appendix A. Vessels ranged in length from 25 m to 45 m and horsepower from 800 to 1000 hp. All vessels used a "Box net" when they were targeting for perch. Box nets are small, square-shaped trawl nets used for their maneuverability over rocky bottom. The box nets used during the study ranged in spread between the wings from 18 to 26 m and in opening height from 4 to 6 m. To examine seasonal changes, sampling trips occurred at intervals less than eight weeks apart from April 1 to December 5 1989 and were generally twelve days in duration. The length of a trip was generally 10 days of fishing time and two days of travel time, but both could be longer in bad weather. The five trips were taken during the first weeks of April and May, mid to late July and September and early December. Perch were usually the first species to be targeted and 3-6 tows were directed at perch during a trip. Rockfish regulations in 1988 limited the amount of fish caught on a species-specific and area-specific basis each trip. The limit for perch ranged from 9 to 36 metric tons per trip.  Measurements of Fish Abundance To compare fish abundance estimates on a per-tow basis, all tows which contained perch were sampled in order to calculate species composition. To overcome the non-random sorting of fish in the net by size, shape andtimingof catch, the samples were collected systematically throughout the sorting of the tow as the fish passed along a conveyor belt. For the vessels without conveyor belts the samples were taken near the four comers and center area of the deck or bin in which the fish were dumped. An exploratory trip was taken in August 1987 to test the feasibility of the use of an STD-12 (described below) on board commercial vessels, as well as to assess a cumulative basket sampling technique for species composition estimates. Sampling baskets (40 X 30 X 55 cm) of fish were sorted one at a time. Data for successive baskets were  10  added cumulatively, until there was less than a 5% change in relative abundance of the species in question. The results indicated that if a species made up 10 - 30% of the catch by weight, at least six baskets were needed to determine relative abundance with + 5% precision. If the species made up between 31 - 60% of the catch, at least four baskets were needed, and for percent values between 61% - 100 %, three baskets were needed for the same level of precision. No fewer than six baskets per tow were used for the collection of species composition information. The definition of various measurements concerned with catch were as follows: Catch = weight (mt) of perch caught in a tow.  Proportion of perch in the catch by weight (proportion) = catch (mt) of perch per-tow / total catch (mt) of all species per-tow.  Catch per unit of effort (cpue mt/hr) = catch (mt) of perch per-tow / hours fished (hr).  Catch per unit of volume towed (cpuv g/m^) = catch (g) of perch per-tow / (area of the net opening (m^) X speed of the vessel while towing (m/hr) X the hours towed (hr).  Hours-towed = the elapsed time, from the net being lowered off the back deck, to the moment it returned to the stern ramp. Thetimeit takes for the net to sink and be hauled back is variable between vessels but the use of similarly powered vessels in this study showed the average time of set and haul procedures for perch tows to be 21 min, with a range of 13 to 30 min and an average towingtimeof 2.6 hours. Sexual maturity was defined using the scale described in Westrheim (1975). Due to the extreme pressure change when the fish are brought up through the water column, many stomachs were everted, making it difficult to make meaningful comparisons of fullness. Investigation of 20 intact stomachs per-tow was considered sufficient for a general assessment of the species diet.  11  Fish Abundance Estimates via Acoustic Equipment Records of fish abundance, depth and time were taken manually by viewing the fish sounder at regular time intervals. A fish sounder is an acoustic device that sends out a signal which is reflected by gaseous substances in the fishes swim bladders, enabling the fishermen to "see" the amount of fish the boat is passing over. Due to the variety of sounders (Furuno, Epsco Cromascope, Sitex and Sea-tex), the use of a digitizer to analyze the signals was not practical. To assess the abundance of fish being targeted, readings were taken directly off the sounder. Sounder screens were set to laterally scroll at a range of rates of one screen replacement every 2 to 5 min, so the length oftimebetween readings was adjusted according to the screen replacement rate. Perch form less dense aggregations than species such as pacific cod (Gadus macrocephalus) and therefore appear as a hazy, raindrop like signal on the sounder (pers. observation). For species such as cod, schools show up as compact blobs which appear darker in colour when abundance is greater. Aggregations of perch, do not tend to change colour with increasing abundance, instead they cover relatively larger vertical depth ranges which makes the height of the signal a good indicator of abundance. Therefore the relative abundance of perch was assessed with an arbitrary scale of signal height that ranged from 1 to 10, representing the height and density of the fish signal.  Oceanographic Studies Physical oceanographic information was collected with a salinity temperature, depth recorder called an STD-12 (manufactured by Applied Microsystems LTD). It is a self-contained recording device that can be lowered over the side of a boat without connecting cables and can be programmed to record the salinity, temperature and depth of the water column. Recording can be done continuously (up to eight scans per second), or at chosen pressure (depth) increments or at chosen time increments. To maximize information and record the entire tow, time intervals were set to 25 or 30 sec for shorter tows of 2 to 3 hr. For those tows that the skipper indicated may be dragged for longer than 3 hours, changes in pressure equal to  12  differences of 3 m of depth were used as the recording interval. The STD-12 was hooked to the headrope of the trawl for each tow that the skipper indicated would be targeting on perch. It was also hooked on some randomly selected tows within areas that perch are known to inhabit but were tows for which perch was not being targeted. Placing the STD-12 on the headrope allowed continuous monitoring of the water through which the net was towed, as well as providing temperature and salinity profiles of the water column when the net was set and hauled back. Due to the length of time required to transfer the recorded data to a portable personal computer (20 to 35 min), I could not monitor all consecutive tows for perch. As well, the STD-12 model could not withstand all of the harsh field conditions and 45% of perch tow data were lost due to equipment failure. The calibration of the STD-12 was tested by staff from both the Institute of Ocean Sciences and Applied Microsystems (Sidney, Vancouver Island, Canada) at the beginning of the study, in the middle (July) and again after the last trip in December.  Results  Study Sites and Trawl Vessels  The use of commercial vessels restricted sampling to only those areas within the Queen Charlotte Sound region regularly fished by the individual skippers. Locations within seven (5, 8, 11, 18, 21, 22, 29) of the DFO sub-management areas were sampled but few of these areas were repeatedly sampled because fishermen could catch the trip limit of perch in as few as three tows and would do so in only one sub-management area (usually at one location). Possible biases and confounding effects on abundance estimation by sampling within only a few sub-management areas on different, larger than average vessels will be addressed in chapter 3.  13  Measurements of Fish Abundance The amount of perch caught, proportion, cpue, cpuv, the amount of hours-towed and depth of capture were all calculated on a per tow basis. The measurements for only those tows containing perch, were grouped by sub-management area, per-trip (Table 2.1). Comparisons were made for all the abundance measurements (catch, proportion, cpue, cpuv) and tow characteristics (hours-towed, depth), between seasons and areas, where possible. Relationships between all tow variables were examined for correlation. All statistical analyses were performed using SAS procedures (SAS Institute, Carey, NC) and unless otherwise stated, level of significance is p<0.05. Abundance estimates (catch, proportion, cpue, cpuv) and tow characteristics (hourstowed, depth) were grouped by month and compared using the Kruskal-Wallis non-parametric anova. All variables showed significant differences between months (p>0.04, Table 2.1) except for effort values (hours-towed, p>0.26, Table 2.1). Tests for seasonal differences between sub-management areas could only be performed on two of the seven areas visited because only two were repeatedly sampled. Area 22 (South of Cape St. James) was fished in April and July by the same skipper, however two different vessels were used for the fishing. Area 5 (Triangle) was fished in May and December by two different skippers on two different vessels. In tests (Kruskal-Wallis non-parametric anova) between estimates within areas 22 and 5, there were significant differences between the amount of effort (hours-towed) for both areas (p=0.05 and p=0.03, respectively, df=l). The only other test to reveal interesting differences was within area 22 for the variable cpuv. The cpuv values for area 22 in April and July are marginally significantly different (p=0.08, df=l). In April, female perch were releasing larvae. The correlations between perch abundance estimates and tow characteristics compared against each other, showed that except for relationships containing effort (hours-towed) all correlations were positive and statistically significant (r values range from 0.55 to 0.95, p values  Table 2.1 Mean values and standard errors for all abundance estimates calculated on a per-tow basis. The information has been grouped per-area per-trip. The p values given at the bottom of each column is the level of significants of the differences between monthly means. Bold-face type and italic numbers represent those effort values that are significantly different and cpuv estimates that are marginally significant within area 22 between April and July. The numbers which are just bold-face represent significant differences between effort values within area 5 between May and Dec. All tests for significance were performed using a KruskalWallis non-parametric ANOVA. Month  Area  #of Tows  April  Mean  SE  Catch (mt)  Mean  SE  Effort (hr)  Mean  SE  Proportion  Mean  SE  Cpue (mt/hr)  Mean  Cpuv (g/m^)  21  1  4.3  -.-  0.8  -.-  0.96  -.-  5.7  -.-  11.9  22  5  12.5  2.5  2.7  0.2  0.89  0.1  4.6  0.9  5  4  1.7  0.4  4.8  1.0  0.26  0.1  0.4  8  3  1.3  0.5  2.4  0.6  0.56  0.1  11  2  2.3  0.9  2.1  0.1  0.68  18*  1  1.8  -.-  3.0  -.-  22  4  8.6  2.8  1.8  Sept  29  4  3.9  0.8  Dec  5  3  0.5  0.0  May  July  Comparisons between months p>0.00  SE  Mean  SE  Depth (in)  -.-  293  -  9.6  1.8  298  7  0.1  0.7  0.2  220  18  0.6  0.2  1.0  0.3  264  33  0.2  1.1  0.5  2.1  0.9  284  10  0.29  -.-  0.6  -.-  0.7  -.-  126  -  0.4  0.64  0.17  4.5  0.7  5.5  0.9  276  11  2.5  0.0  0.71  0.04  1.6  0.3  1.6  0.3  211  7  1.8  0.5  0.26  0.10  0.3  0.1  0.9  0.4  208  2  p=0.26  p=0.01  (*) tow was only one in which perch were not targeted and yet were caught.  p>0.00  p>0.00  p=0.04  15  are consistently >0.00 and the sample number is constant, n=27, Figure 2.2, Table 2.2). Effort (hours-towed) was not related to any other variable (Figure 2.2, Table 2.2). The relationship between cpue and cpuv has the highest correlation coefficient (r=0.95). The relationship between cpue and cpuv was further explored by splitting the information into two groups, one being the month of April when the females are releasing larvae and the other containing the remaining sampled months. A comparison between the slopes of the two groups revealed significant heterogeneity of the slopes (ANCOVA F=100.15, p=0.00, df=l, n=27). For April, the equation of the regression line was: cpuv = 2.09 * cpue + 0.00, during the rest of the year: cpuv = 1.17 * cpue + 0.26. Cpuv valuesfromarea 22, were twice as high in April as they were in July and many times higher than in all other areas during the remainder of the year. The changes in cpuv values indicate that the fish were more densely aggregated during the time of larvae release. In contrast, cpue values from area 22 did not show any difference between April and July. The cpue and cpuv per-tow estimates have high correlations to catch values (r=0.89 and r=0.86, respectively, for both p=0.00, Table 2.2). It is thus not surprising that all three abundance estimates, cpue, cpuv and catch were similarly, exponentially correlated with the proportion of perch in the catch. When the natural log of each variable was compared to proportion value the resulting correlation coefficient was r=0.78, p>0.00 for all three relationships (Table 2.2). There are three obvious outliers in the relationships cpue, cpuv and catch versus proportion-catch (Figure 2.2, 3rd row). They all show greater abundance estimates than the proportion values would indicate. The outliers all come from tows in which Sebastes reedi (Westrheim and Tsuyuki 1967) made up the majority of the remaining percentage composition of the catch, indicating that perch are found in denser aggregations in the presence of S_. reedi than they are with other fish species.  16  0 2 4 6 8  0 2 4 6 8  0 5 10  20 oo co  •••• *i  CM  o  0 2 4 6 8  0 5 10  20  *  o  0.0  1  •  •• ••  —  0 5 15 CATCH  Figure 2.2  0 2 4 6 8 CPUE  • • •  «•  •  • •  0 5 10 CPUV  * •*  •  •  o  LO  20  1  •  o  IT) CM  •  in  • •  M *  o m  o  o in  IT)  o in CM  0.4 0.8  • •  o to  • •  '0.0 0.4 0.8 PROPORTION  CM 1  o in cn  '0  •• • •• •  • • •*  • •  2 4 6 8 HOURS  Relationships between field estimates of fish abundance and tow characteristics.  The numbers plotted represent the month in which the estimates were made.  17 Table 2.2 Correlation coefficients (r), R values and probability levels of relationships between abundance estimates and tow characteristics from tows containing perch. z  Relationship catch vs cpue catch vs cpuv  Correlation co. (r) 0.89 0.86 lncatch vs proportion 0.78 catch vs effort -0.00 catch vs depth 0.55 cpue vs cpuv 0.95 lncpue vs proportion 0.78 cpue vs effort -0.36 cpue vs depth 0.59 lncpuv vs proportion 0.78 cpuv vs effort -0.22 cpuv vs depth 0.60 proportion vs effort -0.19 proportion vs depth 0.65 effort vs depth -0.07 n = 27  R  z  Significance level (p)  0.95 0.93 0.85 0.22  0.00* 0.00* 0.00* 0.98  0.49 0.84 0.89  0.00* 0.00* 0.00*  0.12 0.43 0.76 0.17 0.56 0.17 0.33 0.34  0.06 0.00* 0.00* 0.26 0.00* 0.33 0.00* 0.71  18  All four abundance estimates, cpue, cpuv, catch and proportion were significantly correlated with the depth of capture (p=0.00 for all relationships, Table 2.2), indicating that perch were either more abundant and/or in more condensed schools at the deeper end of their seasonally preferred depth range.  Fish Abundance Estimation - Effect of Biological Factors Sexual Maturity Analysis of gonadaltissueof perch during each trip indicated that spring is the time of live larval release by the females. In the first week of April 48% of the females were in a spent condition and 14% were still ripe, while 34% were resting. By early May 0% were ripe and 21% were spent, with 35% in the resting stage. For the remainder of the study, the female population was either fully recovered from reproductive activity or immature. The mature male population showed signs of gonadal development starting in July and lasting through September. The timing of gonadal maturity for each sex over the seasons is consistent with previous research by Skalkin (1967), Lyubimova (1968) and Westrheim (1975).  Stomach Analysis Stomach analyses indicated that the perch had an almost exclusive diet of euphausiids and shrimp, which were found in full (stretched to full size) stomachs from April through September. Only euphausiids were found in shrunken(l/4 full size) stomachs in December. Euphausiids in this region do not grow larger than 3.0 cm in length (Ponomareva 1966), therefore those animals greater than 3.0 cm were assumed to be shrimp. During the entire study no more than 2% of the total sampled stomachs contained fish and the only other type of food found in perch stomachs was squid, which occurred in 10% of the stomachs in April and 5% in July. It is interesting to note that in early April, 10% of the fish sampled had substances in their stomachs which had the same color and consistency of the larvae found in the female ovaries. The diet described in this study is consistent with earlier work of Skalkin (1967), Lyubimova  19  (1968) and Brodeur and Pearcy (1984), whose extensive work in the Bering Sea and off the Oregon, Washington coast, respectively, indicate that adult perch eat primarily euphausiids.  Fish Abundance Estimation via Fishsounders Comparisons between estimates of fish abundance from trawl catches and estimates made from the fishsounder screen indicate that increased levels of perch "sign" do indeed reflect a greater abundance of perch. In all cases, a continuous signal of 1 - 3 fathoms in height off the bottom, indicated by abundance levels of 8 and 9 produced tows larger than 10 metric tons. Signals of less than 0.5 fathoms during a tow indicated by abundance levels of 1 to 2 always produced tows less than 10 metric tons. However, because fishermen target directly on perch schools, the fishsounding screen shows a consistent abundance level of fish throughout the entire haul. Fishermen either turn their vessels around or pick up the net when fish "sign" decreases. This behaviour does not allow for the comparison of differences in the physical characteristics of water masses between the locations with and without fish "sign".  Oceanographic Studies Large Spatial and Temporal Scale Pattern in the QCS Examination of all STD-12 vertical profile information for perch tows over the ten months of study revealed seasonal effects on the characteristics of the vertical water column. Seasonal wind changes, freshwater run off from large local rivers and seasonal air temperature changes are the driving forces which shape the thermocline and halocline structure in the QCS (Dodimead 1980, Thomson 1981). Seasonal changes, represented by the change in gradient steepness of both the temperature and salinity values, were examined by observing representative perch tow profiles from each of the fishing trips (Figure 2.3). The April profile had weak gradients for both temperature and salinity (Figure 2.3 a). By May there was some warming of the surface layers, down to 100 m (Figure 2.3 b). Upper layer warming, as well as  SAIJMTY (o/oo) 28  29  30  1 APRIL  u  20  SAUNTTY (o/oo) 33  34  i  i  28  29  30 "i  31 1  32  r~rr  33  ^s/X  rs  fc a cu  Q  •r ^  s  MAY  DEC  Q  JULY  Q  TEMPERATURE ( ° Q Figure 2.3.  TEMPERATURE (°C)  Temperature CT) and salinity (S) profiles taken from areas containing perch  within five different months. Vertical information is taken during the setting and hauling of the nets. Horizontal information is collected as the net is towed through the water at depth.  21  the relaxation of winds led to very steep gradients, above 150 meters, by September (Figure 2.3 d). In December the effects of winter wind mixing and cooler air temperatures have destroyed the surface gradient (Figure 2.3 e). All of these processes seem only to have had an effect on waters shallower than 150 meters. To show changes in the structure of the water column over the seasons isopleths of temperature and salinity were contoured over the year. The graphical presentation (Figure 2.4) clearly shows the upper water seasonal warming but also indicates a rise and fall of both temperature and salinity values, over the seasons, in water below depths of 150 m. The main force behind these movements in the thermocline and, to a lesser degree, the halocline, is again the action of the winds (Barber 1957). In the spring (generally in May), along the Pacific northwest coast, the winds shift directions and lessen in strength. During the summer months (May - Sept.) the winds are predominantly from the north. North winds, blowing along the west coast, cause upwelling due to the effects of the Coriolis force. The coastal water that is displaced seaward is replaced, by deeper water from offshore so that cooler, offshore water masses are more likely to be at shallower depths in summer than in winter (Barber 1957, Dodimead 1980, Thompson 1981, Figure 2.5). During winter the winds are strong and generally from the south. There is a build up of surface water along the coast which depresses the thermocline (Barber 1957, Dodimead 1980, Thompson 1981, Figure 2.5). The net effect of the changes in wind direction along the Pacific northwest coast is that water masses, characterized distinctly by temperatures of less than 7 °C, are shallower during the summer months (May - Sept) than they are during the winter months. The timing of the rise and fall of the water masses below 150 m corresponds well with the seemingly onshore/offshore, summer/winter, migrations of perch (see discussion). The movement of the halocline, generated with the same data from only perch tows (Figure 2.4), showed lower salinity values at depths of 350 m in the spring than in the fall and winter. The salinity changes seem to be influenced more by the large amounts of fresh water run-off in the spring, than just wind effects. The changes in the halocline structure over the seasons do not follow the timing of the large scale wind changes.  Figure 2.4.  Isopleths of temperature and salinity from perch tows. The contour plots  represent changes in the depth of the isopleths over the seasons.  Figure 2.5.  Pictorial representation of the wind changes along the British Columbia coast and  corresponding effects on the deeper off-shore waters. Modified from Thomson 1981.  24  Small Spatial and Temporal Scale Pattern Within Perch Sites  The STD-12 was placed directly on the headrope so that the physical data collected on each tow, included two vertical profiles of temperature and salinity as well as horizontal information on the exact location of perch catch. A striking feature of many of the tows was the large variability in salinity and temperature values. The variability can best be seen from information collected in April through July (Figure 2.3, a, b, c, f). The tows presented in Figure 2.3 were selected for two reasons; they were tows targeting on perch and each was pulled through the water in either a circular or "back and forth" configuration such that the set and haul back positions were within 1.0 km, making the vertical profile information comparable. The only available STD-12 information for a tow targeting on perch in July, (Figure 2.3 c) was incomplete due to the tow being longer than anticipated. Therefore Figure 2.3 f, which was not a perch tow, was presented in order to show the temperature and salinity variability present in July. For most trips, only one or two of the perch tows were successfully monitored by the STD-12. It was usually impossible to get information from consecutive tows because it took longer to dump the STD-12 data than it did to empty the net and send it back down. However, on the April trip three consecutive tows were monitored and provided information that best illustrates the possible reasons for the large variability of temperature and salinity values within most perch tows (Figure 2.6). The tows were circular and started and ended in approximately the same location. They were all directed around the outer-edge of a small sea mount, approximately 300 m from the surface. Each tow lasted between two and three hours, with a total of eleven hours between the set of the first and the haulback of the last. During this time there was a large change in salinity values (29.4 °/oo to 31.6 °/oo at 100 m and 29.8 °/oo to 32.3 °/oo at 300 m), with little change in temperature. Density values (24.9 a to 26.9 a at 300 m) 1  1  thus reflect changes in salinity and indicate a change in water mass characteristics from the time of the first to thetimeof the last tow. All perch tows in April, May, July and September, showed  Figure 2.6.  Three consecutive tows for perch, around a small seamount near Cape St James  on April 11,1988. Temperature, salinity and density values are plotted for each tow. The tows all start within 0.5 km of each other and are in the same direction.  26  similar shifts in water density over the duration of the tows, though not as dramatic as the example above. July and April tows were taken in the same location within area 22, but the collections in the remaining months werefromother sites. December tows showed no such shifts but it is interesting that cpue and cpuv values in December were the lowest (0.3 mt/hour and 0.9 g/m^, see Table 2.1) of the study.  Temperature, Salinity and Depth Within Perch Sites  Mean temperatures at the depths at which perch were caught, ranged from 4.8 to 6.7 °C (Figure 2.4). Mean salinity values ranged from 31.18 °/oo - 33.82 °/oo, and were lower the spring (Figure 2.4). Average depths at which perch were caught rangedfrom313 m (maximum) in the spring, to 190 m (minimum) in the summer months. Temperature/salinity values at which perch were caught showed no consistent pattern and a wide range of salinity values (Figure 2.7). Temperature and depth values were within the ranges for perch caught during Canada Department of Fisheries and Oceans biomass surveys (Westrheim 1974, Westrheim et al. 191 A, Barner et al. 1978, Harling et al. 1978, Dodimead et al. 1979a and 1979b, Carter et al. 1982, Leaman and Nagtegaal 1986) and Soviet surveys in the Bering Sea (Lyubimova 1968).  Fish Abundance versus Temperature, Salinity and Depth  Cpue, cpuv and catch showed exponential decreases with increasing temperature. Proportion of perch in the catch showed a linear decline with increasing temperature. Ln(catch) = -1.96 * temperature + 12.5, F=l 18.70, p=0.00, df=8, R =0.94. 2  Ln(cpue) = -1.57 * temperature + 9.4, F=35.83, p=0.00, df=8, R =0.84. 2  Ln(cpuv) = -1.26 * temperature + 8.2, F=13.3, p=0.01, df=8, R =0.66. 2  Proportion = -0.34 * temperature + 2.63, F=23.7, p=0.00, df=8, R =0.77. 2  The only significant relationships between perch abundance estimates and salinity was with cpuv values. Cpuv = -2.76 * salinity + 95.1, F=11.3, p=0.01, df=8, R =0.62. Partial 2  regression analysis revealed that this was not a spurious correlation due to confounding effects  27  ©  in CD  o CO CD 3  co 5.08  Q.  E  9.29  in 4.22  9.29  4.22 4.22 o in  7.22  31.0  31.5  32.0  32.5  33.0  33.5  34.0  Salinity  Figure 2.7  Average temperature and salinity values for the horizontal section of the tows  targeting on perch. The months (first number) and areas (second number) of each tow are presented on the plot.  28  from a relationship between salinity and temperature as both temperature and salinity significantly improved the relationship to a total  = 0.73.  Isopleths of fish abundance estimates with depth over time, indicate that abundance of fish is linked to depth but depths with highest abundance change over the seasons (Figure 2.9). Cpue, cpuv and to a lesser degree, proportion and catch values mimic the rise and fall of the range of temperature values perch prefer. Isopleths of higher fish abundance (3 mt and 4 g/m^) values correspond with the 5.5 °C temperature isotherm.  Discussion  To define the effects of oceanographic variables on catchability of perch, it was necessary to first identify preferred perch habitat and then to determine which abundance estimates best represented differences in fish abundance. In the first section, I will discuss biases affecting measurements of fish abundance and how they can be influenced by fish behaviour. The second section deals with oceanographic characteristics of Queen Charlotte Sound and in particular consistent patterns in the areas in which perch are targeted by fishermen. Finally, I discuss the connection between fish abundance and physical oceanographic processes.  Study Sites and Trawl Vessels The use of commercial vessels as "ships of opportunity" for the study has advantages and disadvantages. Advantages to sampling on perch highliners was an assured catch of perch each trip and the variety of ships and skippers guaranteed that more than one site would be sampled. Disadvantages were that favorite, good sites that are always tried first and most highliners tend to own larger than average trawl vessels. The possible biases due to sampling in a limited number of sites on large vessels is addressed in Chapter 3.  29  S  8  7  •  S  TEMPERATURE OC  5  •  7  TEMPERATURE OC  Figure 2.8  8  7  8  TEMPERATURE OC  8  5  8  7  8  TEMPERATURE OC  The relationship between abundance estimates and the mean temperatures at  which perch were caught. The zero values, at the bottom of each graph, indicate the mean temperatures from the randomly selected tows each trip in which no perch were caught.  9  30  400  600  800 MONTH  Figure 2.9.  1000  1200  400  600  800  1000  1200  MONTH  Isopleths of abundance estimates. The contours represent the changes in depths  of the isopleths over the seasons.  31  Measurements of Fish Abundance Catch per-tow is generally not thought to represent the abundance of the species being fished because of overriding factors such as the effort (hours-towed), the size of net/vessel, and quota/market restrictions that influence the amount of fish caught more so than actual abundance (Gulland 1964, Richards and Schnute 1986). In this study, both standardized abundance estimates, cpue and cpuv calculated per-tow, were highly correlated with catch values per-tow. This may reflect the fact that effort (hours-towed) has no relationship to any of the abundance estimates because all tows are of fairly uniform length, making cpue and cpuv scalar measures of catch. The commercial highliners, from which the data were collected, do not spend much time searching for fish. Any non-fishing interval between tows is traveltimebetween "historically known" areas and is generally done at night. The length of tow is set more by trade offs between, the time it takes to empty and re-set the nets and thetimeit takes crew members to sort and store fish for each tow. This idea is supported by the strong positive relationship between proportion of perch catch in a tow and corresponding values of catch, cpue and cpuv. These relationships indicate that perch are more efficiently caught when they are in schools that are composed of more than 85 % perch. Effort, calculated as hours-towed, has little to do with size of catch. The homogeneity and density of the targeted fish school determines the size of the catch. Fluctuations in catch per-tow are reflections of the ability to target on dense, single species schools and can not be standardized by dividing by the hours fished. Cpuv relationships to catch and proportion are similar to those of cpue, but when there were dramatic differences in school densities, the differences were picked up by cpuv values and not by cpue estimates. Using cpuv (in units of density g/m^) as a measurement of abundance standardizes effects of net size and towing speed in assessment of fish abundance. Net size and towing speed reflect boat size and horsepower and the main factors believed to influence size of catch (Treschev 1964).  32  There are consistent increases of catch, cpue and cpuv and proportion with increasing depth, indicating that even though absolute depth values change, deeper tows tend to produce more fish.  Fish Abundance Estimation - Effect of Biological Factors Stage of reproduction may affect fish catchability and influence estimates offish abundance (Laevastu and Hayes 1981). April was the only month of the study in which ripe, larvae-releasing females were encountered. Estimates of fish abundance were consistently higher in April than in other months (Table 2.1). The cpuv values indicate that fish were more concentrated at thattimethan at any other. The same area was revisited in July and the mean cpuv value was approximately half the April value. There was no clear pattern of change in catchability during the remainder of the study. It appears that the reproductive stage of females must be taken into account in estimates of abundance. The degree of concentration related to larvae-releasing behaviour would lead to a false overestimation of abundance compared with estimates during other seasons. Feeding behaviour can also influence catchability. Stomach analysis of those fish whose stomachs were not everted, showed that, except for the month of December, stomachs were always moderately to completely full. The consistent fullness of the fish in at least the spring and summer months indicates that fish are being caught in areas where food is readily available. The shrunken, but not empty stomachs, in winter indicate fish are still feeding but that food is not as available as in other seasons. The lack of clear pattern between fullness and abundance does not allow suggestions on how feeding behaviour may be effecting catchability.  33  Oceanographic Studies Large Spatial and Temporal Scale Pattern in the Q C S  Seasonal effects are dramatic, with winds and fresh water run-off being the main causes of change in the QCS (Dodimead 1980, Thomson 1981). There is a reversal of winds from southerly in winter, to northerly in summer. The winter winds are strong and cause water to pile up along the coast, which in turn causes deep water masses (> 150 meters) to move downwards (Barber, 1957). These deeper water masses are characterized by temperatures less than 7 °C. In summer, when winds are generally from the north, coastal upwelling occurs and deep water masses are pulled upwards and towards the shore. Movement of the deep water masses causes temperatures to be cooler at shallower depths in summer than in winter (Barber, 1957).  Small Spatial and Temporal Scale Pattern within Perch Sites  Positioning an STD-12 direcdy on the headrope of a net provides two vertical, as well as a horizontal profile of the water masses in which perch were caught. The large changes in temperature and salinity over short distances (within 1.0 km) in approximately 3 hours towing time were not expected results, especially not at depths down to 300 meters. The same pattern was observed so regularly that equipment failure was suspected. However, the STD-12 was tested throughout the study and found to function normally. Therefore, other explanations were investigated for these apparent, but persistent anomalies. A plausible answer for the repeating pattern of large changes in density values, best presented in Figure 2.6, is that a frontal system had passed through the location when the readings were being taken. Work in the Queen Charlotte Sound region by Thomson (1987) indicates the presence of mesoscale eddies working down the outside coast of the Queen Charlotte Islands into the Sound. Fishing on the edge of an eddy could cause the apparent large changes in salinity and temperature values. However, this same pattern was also seen in the southern region of the QCS, too far south to be affected by the eddy system studied by Thomson. A study by Huppert and Bryan (1976) shows that eddies can be generated by the interaction between temporally  34  varying currents and bottom bathymetry. The bathymetry in question consisted of an abrupt rise in the ocean floor such as a seamount. Perch have long been known to live in extremely rocky areas near pinnacles or steep canyons (Carlson and Straty 1981). Work by Pingree et al. (1974) indicates that the presence of eddies or fronts can be maintained bytidalcurrent action. Thomson's (1987) work also indicates that tidal action is responsible for the patterns of eddy formation and movement. However, Huppert and Bryan's (1976) work indicate that the interaction between strong currents and steep bathymetry, would cause cooler water to reside on the surface above or near the submerged obstacle or cliff. If tidal currents are responsible for the formation of fronts and these fronts are characterized by surface waters cooler than the surrounding waters, the fronts should be visible in satellite imagery. Fortunately, work on satellite imagery from the Queen Charlotte Sound region is currently being done at the University of British Columbia. Studies by Ian Jardine indicate that there are small areas in the QCS that show cooler surface water at biweekly intervals and that these areas of cool water are not related to any wind-induced upwelling process (pers com, Oceanography, Vancouver, B.C., Canada). The biweekly occurrence of the signals implicates tidal activity as the controlling force. The lack of wind events, that may also cause this type of signal, adds strength to the assumption that interactions between bathymetry and tides are responsible for producing the areas of cooler surface water. The locations of cooler surface water correspond to areas 5,8,11, 22 and 28 (Figure 2.10), which are the areas in which the highliners I sampled from, concentrated their efforts. Large changes in salinity and temperature between the start and end of a tow are a likely result of towing across frontal zones caused by the interaction of strong tidal currents and the steep, rocky topography that perch prefer. This explanation provides the reason for the regularity with which the phenomenon was recorded and for the consistency of such changes to depths of 300 meters. Other evidence that these areas may indeed be characterized by frontal zones comes from a quote from one of the skippers: "if you don't see gooney birds, don't bother putt'n your nets in the water" (J. Roach, pers. com.; gooney birds are albatross, Diodema spp.).  35  Pelagic birds are known to congregate in frontal zones, most likely because of the high concentrations of food sources (Pingree etal., 1974, Laevastu and Hayes 1981, Hamner 1989). The occurrence of birds as a visual indicator of the presence of fish and could explain how fishermen historically found the good fishing sites without the aid of modern day instruments. The perch may prefer these sites because of large aggregations of euphausiids in these areas due to the current structure (Alldredge and Hamner 1980, Simard and Mackas 1989).  Temperature Salinity and Depth within Perch Sites  It has long been accepted that perch are caught in shallower depths in the summer months than in the winter (Alverson 1960). The seasonal temperature profile (Figure 2.4) indicates that water less than 7 °C occurs at shallower depths in the summer months than the winter months. Seasonal perch movement may thus be a function of the fish following their preferred temperature range as the large scale water masses rise and sink over the seasons. It is important to notice that in tows in which perch were not targeted, temperatures were as low as 6.3 °C (Figure 2.8) and yet no perch were caught. This indicates that either temperature is not the only factor limiting perch habitat selection or that these warmer waters represent the upper range of temperature preference and the probability of catching perch in these warmer waters is low. No similar relationship was found between salinity and all perch abundance estimates. This does not imply that perch distributions depend solely on temperature changes and bathymetric preferences. Oxygen concentration for example, may have significant effects and may explain why perch are more concentrated at the deeper end of their preferred depth range.  Fish Abundance versus Temperature, Salinity and Depth  The most interesting results of the study came from relating temperature to different abundance estimates. Amount of perch caught, cpue and cpuv had similar relationships; they all dropped off exponentially with small increases in temperature. The proportion of perch caught deceased linearly with temperature. These predictive relationships between temperature and  36  Figure 2.10.  Satellite image of the Queen Charlotte Sound. Land is pink and blue water is 15  °C. Yellow areas represent water of 13 °C and correspond to DFO sub-management areas 5, 8, 11,22, and 28.  37  abundance estimates can be used to help quantify the catchability of the perch in a specific site. For example, to estimate the actual number present at a site, one must first determine the volume of water that the fish school inhabits. The range of temperature preference enables an easy calculation of the depth range. This length of water column multiplied by an area, defined by the boundaries of the local bathymetry, provides an estimate of the volume that the specific school occupies. This volume is used as the catchability term in the fish abundance equation (3) and multiplied by the catch of fish per unit of volume that was swept (cpuv) to provide an estimate of the amount of fish in that specific site, at thattime.However, if the temperature of the water that the net passed through to calculate the cpuv estimate is 5 °C, the cpuv would be considerably higher, at this same location, than the estimate would have been if the temperature was 6.5 °C. A correction factor needs to be incorporated into the equation to account for the changes in abundance with temperature. One method would be to multiply by a coefficient, (£), that represents the slope of the relationship between cpuv and temperature. The resulting equation would be as follows;  Fish abundance = [Catch (mt)/ Effort (m )] x (t x Catchability (m )) 3  3  (3)  To test the accuracy and usefulness of equation 3 research surveys need to be conducted in which both the concentration of perch and oceanographic factors are monitored simultaneously throughout a range of known and unknown locations of perch habitat over all major seasonal changes.  38  Conclusions  Fish abundance estimates, made on a per tow basis, calculated in units of volume of water swept (cpuv) reflect changes in fish abundance better than cpue estimates. Calculating the index of fish abundance in units of volume provides direct information about the density of the fish schools targeted. The main advantages to using cpuv is that it dictates that catchability has to be calculated in units of volume (m ), which makes catchability more biologically meaningful 3  as it now represents the volume of water in which a fish school/population could potentially be caught. However, assessing the volume of water that a fish school/population occupies can only be accomplished if the physical habitat of the fish is definable. The conclusion of this field portion of the study is that Pacific Ocean Perch habitat can be defined using physical variables. Perch prefer a temperature range from 6.7 °C, down to at least 4.8 °C and their movement patterns are linked to the movement of these temperatures by coastal wind patterns. Fish concentrate in predictably higher quantities, at the lower end of this temperature range. Perch prefer areas with steep, rocky bathymetry. These areas may give rise to frontal activities due to interactions between the local bathymetry and tidal currents. Therefore the catchability of perch can be linked to its habitat preference and physical oceanographic variables can be used to qualitatively define and quantify the catchability of Pacific Ocean Perch.  39 »  Chapter 3  Comparisons of DFO Commercial Data Base with Field Data  Introduction  The previous chapter was concerned with simultaneously monitoring catch of Perch and characteristics of water masses in which the fish schools were caught from data collected in six different locations, on four different vessels, in five different months. To determine whether information collected from these five commercial trips is representative of the QCS region as a whole the Department of Fisheries (DFO) historical groundfish commercial data base (logbook information) was examined. The data base is known not to represent the actual per-tow catches of fish at sea due to the discarding of unwanted or undersized fish and is therefore referred to as landings per unit of effort, lpue, (Nagtegaal 1983). The data base is composed of groupings of tows (described below) which influence the manner in which the data can be interpreted. To investigate the effects of these differences, relationships between abundance estimates and tow characteristics from the groundfish commercial data base were compared to the same relationships calculated from the field data of chapter 2, for which the amount of discards were known and information was gathered for all individual tows. For the collection of information at sea (see chapter 2), I purposely chose larger boats among the trawl fleet because they consistently brought in catches of perch. This selection of larger boats could introduce a bias. The groundfish data base may be used to evaluate the effect that boat size has on abundance estimation. As mentioned in chapter 2, fishermen preferentially use some areas, and the selection of boats may have also led to a bias in areas sampled. Again, the DFO groundfish data base may be used to determine whether the areas visited in the field study were representative of the entire QCS region. Finally, the sampling at sea was designed to cover major seasonal changes in perch movement patterns. To determine whether the months  40  sampled were representative of seasonal changes, sampled months were compared to unsampled months using the larger DFO data base. There are, no doubt, differences in abundance estimates at the annual level, so long-term changes in the perch fishery were also examined.  Methods  Trawl skippers keep a tow-by-tow logbook of when and where they fish, how much of each species of fish is caught and what depth ranges are covered. To maintain a data base of groundfish catches, DFO has compiled this logbook information since the beginning of the fishery in the 1950's (pers. com. Rick Stanley, PBS, Nanaimo, B.C.). The collection of information was voluntary until 1987, when it became mandatory for the logbooks to be turned over to DFO managers (pers. com. Nevill Venables DFO, Vancouver, B.C.). The groundfish data base is the best historical record of fish catch information available, but it is inadequate for at least three reasons. 1) It does not include records of fish that were discarded at sea, such as undersized fish, unwanted fish species and fish that were dumped because quotas had been reached (Nagtegaal 1983). 2) The catch of each species is computed on a per-trip basis and not on a per-tow basis. Tows are grouped by sub-management areas per-trip before being entered into the data base (Tyler and Fargo 1990), therefore all abundance estimators, such as catch, cpue, proportion and hours-towed, are also grouped per-area, per-trip. In addition, minimum/maximum depth values are the range of depths fished in an entire area during a trip (Nagtegaal 1983). DFO sub-management areas vary in size from approximately 100 km to 2  1580 km  2  (see appendix A). 3) As enforcement measures increased, the accuracy of the  information regarding amounts of fish caught per-species, sub-management areas and depths fished, has decreased dramatically (pers. com. Deep Sea Trawlers Association, Surrey, B.C.). To determine the implications of these inadequacies, relationships between the abundance  41  estimators (perch-catch, cpue, proportion, hours-towed and depth) reported in the DFO data base were compared to the relationships calculated from the field data. The DFO groundfish data base is confidential and in order to comply with trawl skippers desires no latitude or longitudinal coordinates will be stated. DFO sub-management areas are number coded but the names and approximate sizes of the areas are given in appendix A. Unless otherwise stated all analyses were performed using SAS procedures (SAS Institute, Carey, NC) and a significance level is considered to be P <0.05. To test for differences between the DFO commercial data set (based on per-trip estimates) and the information collected at sea (based on per-tow estimates), the degree of association between abundance estimators for each data set were compared. Correlations between each of the abundance estimators (catch, cpue, proportion, hours-towed and maximum depth of catch) were calculated and correlation coefficients (r) and slopes of regression were compared using ANCOVAs (Figures 2.1 & Table 3.1). Maximum depth was used because perch are generally the deepest rockfish species targeted (Westrheim 1970) and it is more likely that maximum depth represents the actual depth at which perch were caught. DFO researchers have recognized that there may be problems associated with the pooling of information per-trip and to correct for low catch and cpue values due to pooling, only those per-trip catch records with 25% or more of the species in question are included in any analysis (Tyler and Fargo 1990). To distinguish between the actual amount of fish caught at sea and the amount brought in to port, DFO uses landings per unit of effort, lpue, instead of cpue (Tyler and Frago 1990). An alternative technique to correct for low catch and due to pooled per-trip data, is to remove those trips that involve more than 20 hours of effort in one area. If fishermen are trawling for more than 20 hours in one area, it is because they have switched to mid-water gear and are targeting on species other than perch in that area (pers. obser.). Analyses were repeated for only those trips with greater than 25% perch or less than 20 hours of effort in any one area per-trip.  42  Analyzing the relationships between the DFO (per-trip) data and the field (per-tow) data provides an opportunity for comparing two different data sets. However, to quantify differences, direct comparisons need to be made between what is actually caught and what is reported. Information on the locations and amounts of perch caught during the five sampling field trips was grouped per-area, per-fishing trip and compared to the corresponding entry in the DFO data base. Differences between abundance estimates, depths and location of perch catch were calculated. To determine whether the size of a boat signiflcandy biases abundance estimation, information from the DFO data bases was used to construct two categories of boat size; big boats (> 800 hp) and smaller boats (< 800 hp). Only information from 1972 to 1988 was used for this analysis because the average horsepower of the fleet was increasing until 1972, but has remained relatively constant since that time (Figure 3.5). Comparisons of the distributions of catch, proportion, cpue, hours-towed and depth categories between the two size groups were made to detect significant differences. To determine whether the areas and months sampled were different from the areas and months not sampled in the QCS, information in the DFO data base from 1967 to 1988 was used to create two area and two month groups; sampled areas and months versus unsampled areas and months. Categories within the two sets of groups were analyzed in the same manner as for the boat categories. Further explorations of differences between individual areas and months were performed by comparing the mean values of each of the following variables, catch, proportion, cpue, hours-towed, depth and horsepower, for each area and month. To appreciate how different or similar the year 1988 was to other years, means of all the same variables, stated above, were compared between the years.  43  Results To compare the DFO groundfish data base with the field data, correlation coefficients and slopes of relationships for abundance estimators and tow characteristics were calculated (Figure 2.2 and Table 3.1). Results from field data, collected on a per-tow basis, suggested that cpue was significantly correlated with catch (r=.89, p>0.00, n=27). Increases in the proportion were similarly associated with significant exponential increases in cpue and catch (proportion of perch catch versus ln(cpue) and ln(catch) both, r=.78, p>0.00, n=27). Effort (hours-towed) had no significant association with either catch, cpue or proportion (r<.27, p>0.17, n=27). Relationships resulting from the DFO groundfish data set, compiled on a per-trip bases, were somewhat different. There was a significant but much weaker association between cpue and catch (r=.49, p>0.00, n=2441, Table 3.1, Figure 3.1). Increases in proportion were associated with significant exponential increases in cpue and catch, but the relationships were much weaker (ln(cpue), r=.65 and ln(catch), r=50, both p>0.00, n=2441) than their field data counterparts. The most dramatic difference was a significant inverse relationship between effort and cpue (Figure 3.1 row 3, column 2). If effort was high cpue values were always small and only at relatively low values of effort were any of the cpue values large. Analysis of covariance revealed that the slopes of the following relationships were not significantly different between the field and DFO data; catch vs hours, catch vs depth, ln(cpue) vs proportion, cpue vs hours, cpue vs depth, proportion vs hours, proportion vs depth and hours vs depth. Only the slopes for catch vs cpue and In catch vs proportion were significantly different (n=2648, p>0.00 and p=0.03 respectively). Relationships from the DFO data base had either equal or larger intercept values, as well as flatter slopes than the relationships from field data (Table 3.2). Relationships resulting from the subset of the DFO data base with only those trips that had greater than 25% perch in the total catch per-area, showed a weaker positive relationship between catch and cpue than the original DFO data set (r=.46, p>0.00, n=2084, Table 3.1). Cpue and hours-towed still showed the same negative relationship (Figure 3.2). There were greater differences between the resulting intercept and slope parameters, relative to the field data, than  44 Table 3.1 Correlation coefficient (r) and their respective level of significance (p) for all combinations of relationships between abundance estimates and tow characteristics for each data set. Field data is the information that was collected at sea. DFO data is the information from the historical DFO groundfish data base (1967 - 1988). The data base labeled DFO > 25%, represents the subset of the original DFO data base with only those trips that have more 25% perch per-area. The data base labeled DFO < 20 Hr, represents the subset of the original DFO data base with only those trips that have less than 20 hours of fishing time per-area. Values that are bold face best show that the original DFO data set and the subset, >25% perch, are very different form the field data. The relationships in the DFO subset, <20 hours, are more similar to the field relationships  Field Data  DFO  n=27  n=2441  Relationship  Corr (r)p  Catch vs Cpue In Catch vs Proportion Catch vs Hours Catch vs Depth In Cpue vs Proportion Cpue vs Hours Cpue vs Depth Proportion vs Hours Proportion vs Depth Hours vs Depth  0.89 0.00  0.49  0.00 0.98 0.01 0.00 0.17 0.00 0.33 0.00 0.45  0.50  0.78  -0.00 0.49 0.78  -0.27 0.53 -0.19 0.59 -0.15  Corr (r) P  0.49 0.40 0.50  -0.13 0.38 -0.12 0.32 -0.02  0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.38  DFO > 25% D F O < 2 0 H r  n=2084 Corr (r) P 0.46 0.17  0.63 0.37 0.28  -0.11 0.34 -0.02 0.09 0.09  0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.42 0.00 0.00  n=1734 Corr (r)i P 0.74 0.48  0.34 0.41 0.57  -0.08 0.38 -0.06 0.25 0.08  0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.00  45  0  40 80  140  0  40 80  140  0  4  8 12  40 80  140  0  4  8 12  o o CM O  in  •Co Z3<=>  0  0.0  0.4  0.8  0.0  0.4  0.8  o o  CM O O  o o 40 80  CATCH  Figure 3.1  140  CO  4  8 12  CPUE  O O CO  PROPORTION  50  150  HOURS  Relationships between abundance estimates and tow characteristics for the DFO  historical groundfish data base.  46  0  40 80  140  40 80  140  40 80  140  o CO  0  0  4  8 12  o o  CD  Figure 3.2  0.2  0.6  1.0  o o  40 80 140 CATCH  4 8 12 CPUE  CO  0.2 0.6 1.0 PROPORTION  40 100 160 HOURS  Relationships between abundance estimates and tow characteristics for the subset  of the DFO data base containing only those trips with more than 25% perch in the total catch per-area.  47  4 0  0  40 80  140  0  40 80  140  0  4  8  12  0  40 80  140  0  4  8  12  0  40 80 CATCH  140 *° 0  Figure 3.3  4 8 12 CPUE  ^0.0 0.4 0.8 PROPORTION  ^0.0  15.0001 HOURS  Relationships between abundance estimates and tow characteristics for the subset  of the DFO data base containing only those trips with less than 20 hours of effort per-area.  48  for the parameters of the original DFO data set (Table 3.2). For example the slope for the relationship ln(catch) vs proportion for the field data was 0.19 and the intercept value 0.39. The slope for the same relationship for the original DFO data was 0.09 and the intercept 0.56. When catch records with the amount of perch less than 25% of the total catch were removed, the slope for the same relationship was 0.02 and the intercept 0.77, indicating that the subset was less similar to the field data than the original DFO data base. When the subset with only trips containing less than 20 hours of effort were considered, the relationships were similar to those obtained from the field data. The positive relationship between catch and cpue was stronger (r=.74, p>0.00, n=1734) and there was no obvious relationship between hours towed and cpue (Figure 3.3 & Table 3.1). Relationships, catch vs cpue and ln(catch) vs proportion had intercept and slope values in closer agreement with the field data than the complete DFO data set or the subset with only trips of at least 25% perch catch (Table 3.2). Deletion of catch records with greater than 20 hours of effort in any one area removed large catch values with correspondingly low cpue values, thereby eliminating some of the misleadingly small cpue values that are only the result of vessels targeting other species while in the same area in which they caught perch. The most direct test of the quality of information within the DFO data base is to compare differences between the actual value of fish catch per-area for the five trips that were sampled against the corresponding entries in the DFO data base. This comparison showed large differences between the two data sets (Table 3.3). Percent differences between observed and reported information showed that a range of 7 to 114 % of perch caught, was reported in the DFO data base. The range in cpue values was even larger; reported values differed from the observed by 33 to 267% and the correct area of catch was only reported 50% of the time. The indications are that, at least in recent times, the information within the DFO data base is misleading.  Table 3.2 Regression parameters for three relationships between thefielddata, DFO historical groundfish data and two subsets from the DFO data base. Values of the intercepts and slopes are presented along with the probabilities that these values are significantly different from zero (p), the standard error (SE) and the number of observations in each data set (N).  Relationship Catch vs Cpue  InCatch vs Proportion  InCpue vs Proportion  Data Set  Intercept  P  SE  Slope  P  SE  N  Field data  0.39  0.16  0.27  0.36  0.00  0.04  27  DFO data  0.51  0.00  0.03  0.03  0.00  0.00  2441  DFO > 25%  0.62  0.00  0.03  0.03  0.00  0.00  2084  DFO < 20 to-  0.38  0.00  0.03  0.09  0.00  0.00  1734  Field data  0.39  0.00  0.05  0.19  0.00  0.03  27  DFO data  0.56  0.00  0.00  0.09  0.00  0.00  2441  DFO > 25%  0.77  0.00  0.01  0.02  0.00  0.00  2084  DFO < 20 In-  0.62  0.00  0.01  0.10  0.00  0.00  1734  Field data  0.55  0.00  0.04  0.18  0.00  0.03  27  DFO data  0.82  0.00  0.00  0.14  0.00  0.00  2441  DFO > 25%  0.83  0.00  0.00  0.06  0.00  0.00  2084  DFO < 20 hr  0.81  0.00  0.01  0.13  0.00  0.00  1734  Table 3.3. Comparisons between the actual amount of perch caught at sea per-area, per-trip, to the corresponding entries in the DFO groundfish data base. The percent differences are calculated by dividing the reported (DFO) values by the observed (Field) values and subtracting 100%. Negative values indicate that the reported values underestimate the observed values by x%. Positive values indicate the reported values overestimate the observed values by x%. When the correct area is reported a "T" is used to denote the truth and when the incorrect area is reported an "F" is used to denote false information.  Field Data  DFO Data  Percent Difference  Catch  Cpue  Hours  Area  Catch  Cpue  Hours  22  25.9  5.6  4.6  22  3.3  6.6  0.5  T  -87%  +18%  -89%  22  36.0  4.1  9.0  22  19.3  1.5  12.5  T  -48%  -67%  +38%  22  34.4  4.9  7.1  22  27.5  4.1  6.8  T  -20%  -16%  -4%  21  4.3  5.7  0.8  21  4.9  2.5  2.5  T  +14%  -56%  +213%  18  1.8  0.6  3.0  Not Reported  11  4.6  0.4  11.0  -93%  -25%  -89%  08  3.9  0.4  9.2  05  6.7  0.3  19.3  -69%  +167%  -87%  18*  0.6  Area  Perch  Cpue  Hours  Vrea  F 0.3  2.3  F F  04  2.1  0.8  2.5  F  (*) both areas 11 and 08 are assumed to have been combined in the entry for area 18 because no other entries were found for the trip.  51  Effects of Boat, Area, Month and Year Boat  Comparisons of the distribution of catch, proportion, cpue, hours-towed and depth categories between the larger (> 800 hp) and smaller (< 800 hp) vessels did not reveal any significant differences for the distributions of catch, proportion and cpue. There were, however, significant differences between the hours-towed and depth of catch (Table 3.4 a,b). Larger boats had significantly more trips with less effort per-area (0-10 hours) than expected and fewer trips with larger amounts of effort per-area (50> hours) than smaller boats (Table 3.4 a). Larger boats had fewer shallower tows (150 - 225 m) and more deeper tows (226 - >350 m) than smaller boats (Table 3.4 b). The only differences when using the subset of data less than 20 hours was a significant difference (p>0.00) between catch values, with larger boats catching more perch. This indicates that smaller vessels are making the high catches with low cpue values that were eliminated from this data subset. Area  DFO sub-management areas, 5, 8, 11, 18, 21 and 22 were sampled during the field work. Unsampled were areas 2, 6, 7,17,19, 20, 26, 27, 28 (see appendix A for names and sizes of areas). Comparisons of the distribution of catch, proportion, cpue, hours-towed and depth categories between the two groups, sampled and unsampled areas, revealed significant differences between groups for all variables (Table 3.5 a,b,c,d,e). In general, sampled areas had lower than expected frequencies at the lower values and higher than expected frequencies at the higher values. Only amount of hours-towed showed higher than expected frequencies at the lowest value (0 -10 hours) and lower than expected frequencies at the highest value (>50 hours) for the sampled area than for the unsampled areas (Table 3.5 e). This lesser degree of effort in the sampled areas indicates that these areas were more likely used for targeting on perch. When the subset of data containing only those trips with less than 20 hours of effort in any one area was analyzed, the only change was that effort was no longer significandy different between areas.  Table 3.4 Comparisons of categories between the boat sizes used for sampling (> 800 hp) with the boat sizes not used for sampling (< 800 hp). Information from the DFO groundfish data base from 1967 to 1988..  A) Comparison of effort (hr) categories by boat size  Boats > 800 hp  0->10  ll->20  21->30  31->40  41->50  Frequency  255.0  76.0  41.0  36.0  12.0  15.0  Expected  222.5  89.3  49.4  33.1  14.3  26.5  4.8  2.0  1.4  0.3  0.4  5.0  Frequency  727.0  318.0  177.0  110.0  51.0  102.0  Expected  759.5  304.7  168.6  112.9  48.8  90.5  1.4  0.6  0.4  0.1  0.1  1.5  Cell Chi-sq  >50  Boats < 800 hp  Cell Chi-sq  Chi-square = 17.797, df = 5, P>0.00  r<">  B) Comparison of depth (m) categories by boat size  151-175 176-200 201-225 226-250 251-275 276-300 301-325 326-350 >350  Boat > 800 hp  <150  Frequency  18.0  19.0  24.0  70.0  58.0  96.0  61.0  39.0  34.0  16.0  Expected  19.5  25.8  62.3  79.8  57.1  70.2  50.8  26.5  28.5  14.5  0.1  1.8  23.6  1.2  0.0  9.5  2.1  5.9  1.0  0.2  Frequency  68.0  95.0  251.0  282.0  194.0  214.0  163.0  78.0  92.0  48.0  Expected  66.5  88.2  212.7  272.3  194.9  239.8  173.3  90.5  97.5  49.5  0.0  0.5  6.9  0.3  0.0  2.8  0.6  1.7  0.3  0.0  Cell Chi-sq Boat < 800 hp  Cell Chi-sq  Chi-square = 58.545, df = 9, P>0.00  54  o C3  27 ?3 28, !27 28 28 • 2626! 26  m  CM  20 20 20 20 20151515151515  o  < UJ  CM  1  5,2 22 22 22 1818  1?  in  11 11 m  -  7 5  8  7 5  8 8 8  § 5  2 2 2 2 65  11  70  2 2 2 75  80  85  90  YEAR  Figure 3.4  a) Mean per-trip catch of perch versus the mean depth of catch for each area. The  area. Area numbers are indicated on the graph, b) Areas fished in each year.  m  Table 3.5 Comparisons of categories between sampled (6,8,11,18, 21,22) and unsampled areas (2,5,8,17,19,20,26, 27, 28). Information from the D F O groundfish data base from 1967 to 1988.  A) Comparison of catch (mt) categories by area  41-50  51-60  >60  103.0  52.0  28.0  84.0  142.9  80.4  40.2  22.4  67.2  2.3  4.0  6.4  3.5  1.4  4.2  631.0  103.0  50.0  19.0  9.0  6.0  18.0  526.5  126.6  74.0  41.6  20.8  11.6  34.8  20.7  4.4  7.8  12.3  6.7  2.7  8.1  Sampled Areas  0-10  11-20  21-30  31-40  Frequency  912.0  268.0  167.0  Expected  1016.5  224.4  10.7  Frequency Expected  Cell Chi-sq Unsampled Areas  Cell Chi-sq  Chi-square = 95.28, df=6, P>0.00  B) Comparison of proportion (%) categories by area  Sampled Areas  0-10  11-20  21-30  31-40  41-50  51-60  61-70  71-80  81-90  90-100  Frequency  95.0  56.0  45.0  58.0  66.0  81.0  96.0  132.0  240.0  745.0  Expected  134.4 .  71.8  55.3  59.9  67.9  89.6  95.5  135.7  216.7  687.1  11.5  3.5  1.9  0.1  0.1  0.8  0.0  0.1  2.5  4.9  Frequency  109.0  53.0  39.0  33.0  37.0  55.0  49.0  74.0  89.0  298.0  Expected  69.6  37.1  28.7  31.0  35.1  46.4  49.5  70.3  112.3  355.9  Cell Chi-sq  22.2  6.7  3.7  0.1  0.1  1.6  0.0  0.2  4.8  9.4  Cell Chi-sq Unsampled Areas  Chi-square = 74.359, df = 9, P>0.00  C) Comparison of cpue (mt/hr) categories by area  Sampled Areas  0.0-1.0  1.1 - 2.0  2.1 - 3.0  3.1 - 4.0  4.1 - 5.0  5.1 -10  >10  Frequency  1034.0  382.0  99.0  33.0  18.0  37.0  11.0  Expected  1127.8  316.2  80.4  28.3  16.5  32.9  U.9  7.8  13.7  4.3  0.8  0.1  0.5  0.1  Frequency  678.0  98.0  23.0  10.0  7.0  13.0  7.0  Expected  584.2  163.8  41.6  14.7  8.5  17.1  6.1  15.1  26.4  8.3  1.5  0.3  1.0  0.1  Cell Chi-sq  Unsampled Areas  Cell Chi-sq  Chi-square = 79.966, df=6, P>0.00  D) Comparison of effort (hr) categories by area  Sampled Areas  0-10  11-20  21-30  31-40  41-50  >50  Frequency  784.0  336.0  198.0  138.0  53.0  105.0  Expected  809.6  338.6  187.8  117.3  53.4  107.3  0.8  0.0  0.6  3.7  0.0  0.1  Frequency  445.0  178.0  87.0  40.0  28.0  58.0  Expected  419.4  175.4  97.2  60.7  27.6  55.6  1.6  0.0  1.1  7.1  0.0  0.1  Cell Chi-sq Unsampled Areas  Cell Chi-sq  Chi-square = 14.987, df=5, P>0.01  E) Comparison of depth (m) categories by area  Sampled Areas <150  151-175 176-200 201-225 226-250 251-275 276-300 301-325 326-350 >350  Frequency  40.0  79.0  174.0  266.0  134.0  137.0  110.0  55.0  54.0  32.0  Expected  42.8  64.0. •  181.3  240.5  141.6  158.4  115.6  52.5  56.0  28.2  0.2  3.5  0.3  2.7  0.4  2.9  0.3  0.1  0.1  0.5  Cell Chi-sq  Unsampled Areas  Frequency  57.0  66.0  237.0  279.0  187.0  222.0  152.0  64.0  73.0  32.0  Expected  54.2  81.0  229.7  304.5  179.4  200.6  146.4  66.5  71.0  35.8  0.1  2.8  0.2  2.1  0.3  2.3  0.2  0.1  0.1  0.4  Cell Chi-sq  Chi-square = 19.661, df= 9, P>0.02  60  Further investigations into differences between individual areas using regression analysis between means of each of the following variables, revealed that the strongest relationship was a linear increase in per-area mean catch of perch with increasing per-area mean depth of catch (catch = -0.17 * depth + 28.7, R =0.81, F=56.4, p>0.00, df=14, Figure 3.4). The number of 2  areas fished over the years was not constant. Some of the deeper areas were not fished until the late 1970's (Figure 3.4).  Month The months sampled in the field were April, May, July, September and December. Comparing distributions of D F O data for these months with those for unsampled months revealed no significant differences between catches, proportion or hours-towed. There were significant differences between distributions of cpue categories, as well as depth categories (Table 3.6 a,b). The cpue category representing values from 1-2 mt/hr was the only one with large deviations from expected frequencies, with sampled months having fewer than expected (Table 3.6 a). The depth categories that had large deviations were distributed throughout the range of depth values, with sampled months having more shallow tows (150-175 m) and less deep tows (250-275 m) than the unsampled months (Table 3.6 b). The only difference using the data subset with only trips containing less than 20 hours, was that cpue values were not significantly different between months. Mean values for all abundance estimates and tow characteristics showed definite seasonal variation (Figure 3.5). In summer months (May, June, July and August) amount of effort per-trip was greater than for the remaining months. Mean catch, cpue and boat size (horsepower) all exhibited greater values in Jan, Feb, March and April than for the remaining months. Mean depth of catch per-trip was at least 50 m deeper in winter months (Jan, Feb, March, April, Nov and Dec). Proportion of perch catch remained approximately the same (62% to 76%) over the seasons. Because differences in areas were so clearly linked to depth (see results section "Areas"), I re-calculated all of the mean monthly values after categorizing the D F O sub-management areas  Table 3.6 Comparisons of categories between sample (April, May, July, September and December) and unsampled months (January, February, March, June, August, October and November). Information from the DFO groundfish data base from 1967 to 1988.  A) Comparison of cpue (mt/hr) categories by month  Sampled Months  0.0 - 1.0  1.1 - 2.0  2.1 - 3.0  3.1 - 4.0  4.1 - 5.0  5.1 -10  >10  Frequency  775.0  181.0  60.0  20.0  9.0  24.0  12.0  Expected  775.4  211.8  53.8  19.0  11.0  22.1  7.9  0.5  4.5  0.7  0.1  0.4  0.2  2.1  Frequency  937.0  299.0  62.0  23.0  16.0  26.0  6.0  Expected  956.6  268.2  68.2  24.0  14.0  27.9  10.1  0.4  3.5  0.6  0.0  0.3  0.1  1.6  Cell Chi-sq Unsampled Months  Cell Chi-sq  Chi-square = 14.972, df=6, P>0.02  B) Comparison of depth (ni) categories by month  Sampled Month<150  151-175  176-200 201-225 226-250 251-275 276-300 301-325 326-350 >350  Frequency  43.0  49.0  133.0  329.0  263.0  318.0  239.0  86.0  100.0  54.0  Expected  63.9  95,5  270.8  359.0  211.5  236.5  172.6  78.4  83.6  42.1  6.8  22.7  70.1  2.5  12.6  28.1  25.5  0.7  3.2  3.3  Cell Chi-sq  Unsampled Month Frequency  54.0  96.0  278.0  216.0  58.0  41.0  23.0  33.0  27.0  10.0  Expected  33.1  49.5  140.2  185.9  109.5  122.5  89.4  40.6  43.3  21.8  Cell Chi-sq  13.2  43.7  135.3  4.9  24.2  49.3  1.4  6.2  6.4  Chi-square = 514.427, df=9, P>0.00  54.22  63  450  -400  HORSEPOWER 550 650 750  -350  DEPTH (m) -300 -250  -200  -150  0  5  HOURS-TOWED 10 15 20 25 30 35  0  5  PERCH CATCH (mt) 10 15 20 25 30 35  4>> 0.0  o  n  PROPORTION PERCH 0.2 0.4 0.6 0.8 1  1  1  1  0.0  1.0 r  Q  0.5 1  CPUE (mt/hr) 1.0 1.5 2.0 1  1  1  2.5  3.0  I  I  ro  X T  QOi  z  jr _^  o  ro  Figure 3.5  y  *-  Mean monthly values and standard error bars for all abundance estimates and tow  characteristics. All areas and years (67-88) are combined.  HORSEPOWER 550 650 750  450  -350  0.0  6  0  DEPTH (m) -300 -250 -200 -150  PROPORTION PERCH 0.2 0.4 0.6 0.8  1.0  T  Figure  3.6  HOURS-TOWED 5 10 15 20 25 30 35  5  o  0.0  4  PERCH CATCH (mt) 10 15 20 25 30 35  CPUE (mt/hr) 0.5 1.0 1.5 2.0 2.5 3.0  n -  Mean monthly values for all abundance estimates and tow characteristics for three  groups of areas. The groups are made up of deep areas (average depth of areas are greater than 300 meters, shown as solid line  ), intermediate areas (average depth of areas are between  200 and 300 meters, shown as the spaced line —) and shallow areas (average depth of areas less than 200 meters, shown as dotted line  ).  65  into three depth groups. Areas with mean monthly depths above 200 m were considered shallow, between 300 and 200 m were intermediate and greater than 300 m deep. Mean monthly values for all abundance estimates and tow characteristics for the three depth categories did not show the same seasonal variation as the overall mean monthly information (Figure 3.6). Instead, mean monthly values of the categories indicate that some of the seasonal variation in Figure 3.5 is an artifact of differences between the areas. Individual area categories, especially the deeper ones, did not change as drastically over the seasons as the overall means would imply. Effort values showed no consistent patterns within the different area categories, over the seasons. Mean perch catch, proportion catch, cpue and mean horsepower values in the deepest areas were consistently higher, over the entire season, than in areas of intermediate and shallow depths. Years Mean values for abundance estimates and tow characteristics showed dramatic changes over the years. There was a large increase in effort from 1967 to 1971 of as much as 28.6 hours per-trip (Figure 3.7 a), coinciding with an increase in the average horsepower of the fishing vessels. After 1972 there was a steady decline of effort to the present level of 11.3 hours pertrip. Average vessel horsepower has remained between 590 and 650 since 1972. The average size of boats increased substantially in 1978 and 1979 (Figure 3.7 b). This effect can be ignored since it is due to the addition of one vessel, the "Calastratus", which was taken out of the fishery when purchased by DFO and is now known as the "Ricker". The apparent downward trend in horsepower since 1985 is due to the addition of 42 smaller sized boats to the QCS fishing region. Between 1978 and 1985 the fleet fishing in the QCS was made up of approximately 21 vessels; in 1988 it was 63. Annual per-trip perch catch values showed two distinct, three year time periods of high catches (Figure 3.7 c). Catches from 1972-74 and 1981-83 showed greater than 20 mt per-trip. Corresponding to the sametimeperiods were two groups of deeper than average mean annual depths of catch (Figure 3.7 d). Mean annual cpue values tracked only the second period of maximum catch (1981-83) (Figure 3.7 e). High effort values per-trip may have counterbalanced the corresponding increases in catch values, leaving cpue values constant during  66  HORSEPOWER 550 650 750  ^450 cn I  i—•—i  i  1  1  — i  0  OT  r  cn  HOURS-TOWED 10 15 20 25 30 35  5  p  1  '  1  1  1  n  1  co o  CO  o ^00  -350  DEPTH (m) -300 -250  -200  -150  CJl  CO  PERCH CATCH (m) 25 30 35  o>  0  ui  p  '  0.0  0.5  5  1  0  1  1  5  2  0  1  1  n  1-  1  co  o ^ 0^0 oi n  PROPORTION PERCH 0.2 0.4 0.6 0.8 1  1  1  .  1.0 n  m  CJI  p  '  CPUE (mt/hr) 1.0 1.5 2.0 1  1  1  2.5 1  3.0  n  CO  o  Figure 3.7 characteristics.  Annual means and standard error bars for all abundance estimates and tow  67  the period of high catches in 1972-74. The proportion of perch catch per-trip has steadily declined from a high of 85% in 1968 to the present value of 63% in 1988 (Figure 3.7 f). Because the values are calculated on a per-trip basis, this decline could be due to changes in populations of perch or to fishermen targeting on a greater variety of species per-trip for present day markets.  Discussion  Information collected at sea, on board commercial trawlers, shows that catch and cpue are strongly linearly related. Both abundance estimators show exponential increases with an increasing proportion of perch in the catch and neither show any relation with effort calculated as hours-towed. These results reflect the behaviour of both the fishermen and the fish. Because fishermen tow for approximately the same length of time during daylight hours, effort values are nearly constant, making catch and cpue scalar values of one another. Perch form schools, and fishermen target directly on the schools of perch. Larger catches are thus not due to increases in effort, but are a reflection of the better fishermen, who are more able to find and target on the schools of perch. In contrast, the D F O data base shows no relationship between catch and cpue values and indicates that cpue has a strong inverse relationship with effort. Presumably fishermen only target on perch for 3-6 tows per-area, per-trip due either to market, or more recently, quota restrictions (pers. obser.). Increasing effort in a DFO sub-management area is representative of boats that stay within an area and target on species other than perch. The fishermen can do this by switching to mid-water gear and fishing.at shallower levels in the water column, above the perch population. Hence the catch of perch remains constant while the total effort value for that area increases giving rise to misleadingly small cpue values for the size of perch catch and in turn, the lack of relationship between catch and cpue. This also explains why such small cpue values are seen for only those trips with the larger effort values.  68  Eliminating trips with less than 25% perch does not change the inverse relationship between cpue and effort, nor does it improve the relationship between catch and cpue. However, eliminating the trips with greater than 20 hours of effort in one area, removes catch values with disproportionately small cpue values. After this removal, the remaining information shows the expected strong linear relationship between catch and cpue as well as a lack of relationship between cpue and effort. A drawback to the removal of trips with greater than 20 hours is that it represents throwing away about 25% of the data base. Also it is really only the cpue and proportion estimates that lose their accuracy when the information is grouped per-trip. Catch and maximum depth information remain unaltered. An alternative procedure is to use only catch values to look for changing trends in abundance, while being aware of changes in either market demands or quota restrictions. If a standardized index of abundance is desired for a multi-species fishery, the data must be collected in a manner that allows distinguishing between targeting and non-targeting fishing. The indications provided by this work, point to abundance estimates being calculated on a pertow basis rather than a per-trip basis. Use of hours-towed as effort should be reassessed since differences in the amounts of catch have been shown to be more dependant on the proportion of the target species in the schools rather than on the amount of time spent fishing the schools. Direct comparisons between field data, grouped per-area, per-trip and the corresponding entry in the DFO data base indicate that misleading logbook information is being given by the fishermen. Concerns voiced by the Deep Sea Trawlers Association of B.C. that the introduction of enforcement measures for the mandatory submission of logbooks in 1987 has lead to increases in the amount of incorrect information and their fears that erroneous data may be used for stock assessment appear justified. Differences between the two sets of data can be as great as a 93% under-representation of perch catch, to as great as a 167% over-representation of cpue values. The differences stem from a combination of poor estimation of species composition at sea as well as on shore, discards that were not brought into the calculations, and incorrect information  69  in logbooks. No simple correction factor can be implemented to improve the DFO data base and it is possible that information after 1987 should not be used for any analysis.  Effects of Boat, Area , Month and Years Selection of larger boats (> 800 hp) on which to collect samples, did not directly bias abundance estimators. Larger boat owners tend to target on only one species per-area and therefore fish for fewer hours within an area. They also tend to fish deeper than the smaller boats. However, these differences failed to show any significant effect on catch, proportion or cpue values, indicating that even though there may be behavioral differences between the operators of larger and smaller vessels, the end results are similar. The sampled areas generally have higher catches, higher proportions of perch in the catch, higher cpue values, tows are directed deeper and have fewer fishing hours than areas that were not sampled. Significant differences between catch, proportion and cpue values between the areas, but not between the vessel groups, indicates that the areas are very different in perch abundance and these differences are not due to artifacts of boat size. Mean depth of perch catch within an area predicts that the deeper areas consistently produce higher catches. Therefore, field estimates of abundance, from sampled areas which include all but 3 of the deepest areas, represent the upper end of the range of abundance values. The sampled and unsampled months show significant differences within the variables cpue and depth. The unsampled group contains more of the months during which perch are at the deepest range of their annual migration (Jan., Feb., and March). It may be that perch are in denser schools at thesetimes,especially in March when females begin releasing larvae. This counterbalances the notion that the field abundance estimates represent the upper end of the range of values. It, instead, suggests that there may have been under-representative sampling during the seasons when perch are found in their deepest range and most condensed schools which has led to a lack of samples containing the highest possible abundance estimates.  70  Before any good analysis of a historical data set is complete, one should be aware of changes, especially human changes, that have occurred over the time period to insure that the results are interpreted correctly. The perch fishery shows a drastic increase in mean horsepower size of the fleet up until 1972. The steady decline in effort since 1972 can be interpreted as an increase in efficiency due to increased local knowledge, as well as advances in technology since a majority of the vessel skippers that are presently fishing, were fishing in 1972 (pers. com. Deep Sea Trawlers Association, Surrey, B.C.). Two periods of high catches are thought to be caused by two large cohorts becoming available to the fishery at thosetimeperiods (Archibald et al. 1983). However at those sametimeperiods, annual mean depths of catch were at historical extremes. Indicating that estimates of perch abundance are influenced by the absolute depth at which the fish are caught. Cpue values do not track the occurrence of the large cohorts because of changing levels of effort over the decades. There is a continuous decrease in proportion of perch catch but it is difficult to say whether it is due to either actual decreases in abundance of perch or to increases in the catch of other species.  Conclusions  In the BC coast multi-species trawl fishery, catch and proportion of the catch of perch, together, are good indicators of abundance levels. Effort, calculated as the amount oftimewith fishing gear in the water, has no effect on catch values and is really just a reflection of the individual fishermen's choice of optimal tow length. Therefore effort, as hours-towed, is a poor choice of value with which to standardize catch estimates. Grouping tow information, as it is done for the DFO historical groundfish data base, further confuses abundance estimation by mixing together data from tows that target on different species. This, in turn, renders the use of proportion, effort and cpue estimates on a per-species basis, meaningless. However, even after grouping the data, the catch estimates of individual species are still accurate and in this case, the maximum depth values are representative of the  71  depth of perch capture, because perch are the deepest species that is targeted in the Queen Charlotte Sound region. Recent changes in the accuracy of logbook information indicate that it should not be used as a source of abundance estimation at all. DFO sub-management areas are very differentfromeach other. In fact greater differences are seen between areas than between either seasons or boats. Larger catches of perch are found in the deeper areas and in the winter months, when the fish are in their deepest range. This indicates the abundance estimates from the field data were over-representative of the deeper areas but under-representative of the true winter months.  72  Chapter 4  Analysis of DFO Research Survey Information  Introduction  DFO Research Surveys Soviets researchers conducted intensive surveys off the BC and Alaska coast in the 1950's. Lyubimova (1968), stated that perch live in waters between 4.0 and 6.5 °C, but did not present data to substantiate this claim. DFO has also long recognized the need to collect survey data independently of commercial catch for biomass estimates. Research surveys were designed to investigate patterns of fish distribution in relation to depth and temperature (Harling et al. 1970,71, 73, 77, Westrheim etal. 1972,73, 74,76). DFO surveys targeting on rockfish, especially perch, were initiated in 1960's. Only information after 1967 and before 1984 is used in the following analysis because another rockfish, Sebastes reedi commonly known as reedi; similar in appearance to perch, Sebastes alutus. was not identified as a separate species until 1967. After 1984, temperature readings were taken less often than in previous surveys because of a change in research procedures. Earlier surveys took temperature readings at the end of almost every tow, but more modern surveys repeatedly sampled standardized locations for temperature information. The standard locations were not visited daily or even weekly and as a consequence there have been large temperature changes over short time periods that went undetected. For an example of such changes, information from the 1984 rockfish survey (Nategaal et al. 1986) showed a 2°C shift in bottom temperature which occurred sometime within a two week period in the Goose Island Gully region. This amount of change, unrecorded, makes detailed comparisons between perch abundance and temperature impossible. No seasonal comparisons are possible because the  73  research surveys were generally conducted once a year in approximately the same season. At present they are run only once every 2 or 3 years and usually in September.  Method  D F O survey data provides an opportunity to directly compare abundance estimates with temperature information. Other oceanographic information from the Queen Charlotte Sound is too sparse in temporal and spatial coverage to be useful for this type of analysis (Dodimead 1980, Thomson 1981). Cpue, proportion and the average depth of catch of perch per tow can be compared to the temperature taken at the end of each tow. Catch values were not investigated because effort was constant (30 minutes) for each tow so that cpue and catch values are simply scalar values of one another. The relationships of abundance estimators are compared to corresponding relationships from the field data. Daily upwelling values were compared with the depth of catch to identify the lag time as well as to quantify the effects that winds might have had on the vertical movement of perch schools. Wind events may affect abundance estimation by causing fish school movement which could result in an apparent change in abundance (Rose and Leggett 1989, also see section "upwelling index"). The survey information is used to map water mass movement as indicated by temperature characteristics and changes in cpue values at given locations and depths over time. Information from five research surveys conducted in the Goose Island Gully region (Harling et al. 1970, 71, 73, 77, Westrheim et al. 1976) was analyzed on a daily basis. The position and depth of each tow was examined, such that patterns of cpue and temperature change could be compared with changes in wind strength and direction reported in the daily upwelling index (Bakun 1973).  74  Results DFO research survey information indicates that Perch were always caught between 4-7 °C. The highest cpue values were found between 5 -6.7 ° C (Figure 4.1). Temperature was taken only once at the end of a tow and there can be more than a 30 m difference between the depth at which the temperature reading was taken and the average depth of catch. For example, a 30 m difference in depth at ranges between 150 m to 300 m could mean a change in temperature as great as 0.5 °C. There is no significant relationship between proportion of perch caught and temperature (p=0.18, F=1.8, df=293, Figure 4.1). Proportion of perch verse catch showed the same significant exponential relationship that was seen in the field data and commercial data (p>0.00, F=238.3, df=293, Figure 4.1). Perch are most commonly caught between 200 and 300 m but proportion of perch caught compared to depth of capture, showed no significant relationship (p=0.12, F=2.5, df=293, Figure 4.1). Temperature values compared to depth ranges indicate that any one temperature can be found at a wide range of depths. For example water of 5 °C can be found from 225 m to 450 m. Conversely, at a given depth, 300 m, there is a wide range of temperatures, 4 to 6.8 °C (Figure 4.1. Detailed analysis of five DFO surveys revealed that the primary priority of the survey design - to repeatedly sample set depth ranges throughout the Goose Island Gully region, eliminated the possibility of detecting changes in fish abundance that might only be due to fish school movement. To collect information at set depth ranges, the survey vessel "hops" back and forth across the Goose Island Gully, rarely returning to the same locations daily. This technique, of continually moving around the gully, makes it all but impossible to track changes in winds and how they may be affecting movements of particular water masses and movements of particular schools of fish. The one survey (Ffarling etal. 1977) that did repeatedly sample the same locations and depths daily, did so for only 3 sequential days and did not record temperature values after the initial day.  75  -500  -400  DEPTH (m) -300 -200  -100  0  PERCH CPUE (mt/hr) 1 2 3 4 5  C*3 —f  -a  m  > m  CO  -400  DEPTH (m) -300 -200  -400  DEPTH (m) -300 -200  -100  0.0 « p-  0.0  PROPORTION PERCH 0.2 0.4 0.6 0.8  1.0  PROPORTION PERCH 0.2 0.4 0.6 0.8  1.0  1  i  .  i f  * "* ' *  *  *  •* >*  •  i -  . . .  •  Figure 4.1  Survey information from the Goose Island Gully region. The same relationships  from the field data are also presented and are represented as "1" on the plots.  76  Discussion Survey information indicates that perch will occupy a range of temperatures from 4 7°C, with the highest cpue values corresponding to water of 5 - 6.7°C. This is in very close agreement with the range of 4 - 6.5°C described by the Soviet studies (Lyubimova 1968). The temperature information I collected during commercial operations also fits well within this range. However, my data show atight,exponentially decreasing relationship between cpue and increasing temperature (Figure 2.8), whereas the DFO surveys indicates a great scatter of cpue values over the range of temperatures. This scatter can partially be explained by the lack of accuracy in the location of temperature measurements but also indicates that temperature is not the only factor effecting the position of perch within the water column. The relationship between cpue and the proportion of perch in the catch, exhibits the same exponential increase in perch abundance with increasing purity of perch in the catch, that the field data revealed. This reaffirms the notion that increases in catch are due entirely to fishing in purer schools of perch, not to increases in effort since effort was constant for all DFO survey tows. The range of depths (225 m - 450 m) that exhibit the same temperature value (5°C) in one region - Goose Island Gully - indicates that one can not use depth as a description of perch habitat without first identifying the temperature of the water mass in which they were found. The corresponding range of temperatures (4 -6.8°C) that can be found at a single depth (300 m) indicates that the ocean waters are capable of a great degree of large scale vertical movement. A detailed understanding of thetimingand forces responsible for these water mass movements is needed before abundance estimates from surveys can be assumed to represent an index that is standardized for location and season.  77  Conclusions  By analyzing information from DFO surveys in detail it was hoped that one could begin to understand the effects of winds on the water mass movement and corresponding perch school movements in at least one area (Goose Island Gully). This understanding would provide the information needed to correct for large changes in species abundance that are due solely to the movement and hence, availability of fish schools, rather than some true changes in abundance. However, the surveys were not designed to capture this type of information and therefore not very useful in this particular quest. To identify and quantify the variables that are most important in describing perch habitat, the research surveys need to be designed to follow fish schools over time with simultaneous and continuous monitoring of their surrounding marine environment, including wind directions and strengths.  78  Chapter 5  Comparisons of the Upwelling Index with the Depth of Catch  Introduction  The depth ranges that perch are most likely to be caught within, change over the year (Alverson 1960, Gunderson 1972). The change in depth may be linked to the direction and strength of the coastal winds. The coastal winds cause vertical movement of deep water masses (> 150 m) that are characterized by temperatures of less than 7 °C (Barber 1957, Dodimead 1980). Perch prefer water masses of these temperatures (chapter 2), so that changes in depth ranges that perch occupy over a season may be linked to wind strength and direction. To test this idea, I compared an index of wind effects to the depth of catch in both the DFO historical commercial data and the DFO research survey data. The comparisons to survey data are presented in the previous chapter. Bakun (1973) compiled an upwelling index that represents the amount of water transported away from the coast due to the wind stress over a unit width of coast line. The wind stress is estimated using atmospheric pressure data and the index is calculated as the Ekman transport perpendicular to the coast in units of m^/sec per 100 m of coast line (Bakun 1973). When water is transported offshore, deep water masses are pushed upwards, towards the shore and the index is positive. When water is transported towards shore, deep water masses are pushed downwards, away from the shore and the index is negative. Therefore the sign of the upwelling index (positive or negative) represents the direction of the transported upper waters (offshore or onshore, respectively) and can be use as an a priori predictor of the vertical movement of the deeper water masses (up or down, respectively) where the perch population reside.  79  Methods  The upwelling index is calculated on a 6-hourly basis and can be compiled daily, monthly and annually. The index is not accurate on a spatial scale smaller than 200-400 km (Bakun 1973). Single positions, in this case 5 LOON 131.00W, are used to determine over which segment of the coast the perpendicular Ekman transport will be calculated over a 400 km segment. This large spatial scale, combined with the lack of knowledge concerning the lag time between the wind event and effects at different depths, allows only a rather crude interpretation of the effect of coastal winds. To test the hypotheses that the direction and strength of the winds can affect the depth of catch, a regression analysis was performed between annual and monthly upwelling index and the corresponding mean annual and monthly depth of commercial catch, calculated from the DFO data base (1967-1988). The expected results were shallower catches of perch when the upwelling index was positive and deeper catches of perch when the upwelling index was negative. To attempt to compare daily upwelling values with the daily depth of fish catch would be preferable, but the dates available in the DFO data base are the landing dates for the fishing trips. The actual date of fish capture can be up to 10 days prior to landing and the grouping of tow information into trips, leaves no means of determining the actual date of capture. There is reason to believe that different sub-management areas react differently to wind events depending on their distance to the shore and maximum depth. For these reasons, the per-trip depth of perch catch information from only the most heavily used sub-management area (18 - South East Edge Goose Island Gully) was compared to the upwelling index calculated as the mean for the 10 days previous to each landing date from this area.  Figure 5.1  Mean monthly depth of commercial perch catch versus monthly Bakun upwelling  index (1967-1988).  81  Results  Analysis of mean annual depth of catch versus the mean annual upwelling values did not reveal any significant relationships. Mean monthly depth of catch was significantly and positively (shallower depths of catch with increasing positive upwelling values) related to monthly upwelling values with an R =0.16, p>0.00. No relationship was found between 10-day 2  upwelling values and the depth of catches from 10-day trips within area 18 - SE edge of Goose Island Gully.  Discussion and Conclusions  The relationship between mean monthly values, even though it is significant, does not explain much of the variability in the depth of capture. These results can be interpreted three ways. 1) That there is no connection between the two variables. 2) That averaging over too long or short of a time scale masks any relationship that may be present. 3) The lack of relationships is due to the fact that oceans have "memory". By "memory", I mean that the oceans have stored the previous days and months worth of wind stress energy and are in a state of constant readjustment. A storm with strong onshore winds will have downwelling values typical of winter storms, but can occur in the middle of the summer. The corresponding depth of perch catch during this storm will, however, be more representative of shallower summer depths because of previous more typically, offshore (upwelling) summer winds. Therefore the oversimplified, relationship of winds to depth of catch may be an inappropriate test of the hypothesis. A modeling exercise, incorporating the predicted water mass movement with the corresponding wind forces is needed to truly test this hypothesis. However, the modeling efforts for this region of the coast are not yet appropriate for such a test (pers. com. Charles Hanna, Dept. Physics UBC, Vancouver, B.C.).  82  Chapter 6  General Discussion  A main concern of fisheries science has been the development of accurate estimates of fish abundance. Catch per unit of effort (cpue) has been widely used as an index of abundance but has been proven to be misleading (Winters and Wheeler 1985, Crecco and Overholtz 1990). The shortcomings stem from the assumption that the coefficient of catchability (the probability of catching afishper unit oftimespent fishing) remains constant as population size changes. However, catchability is intimately related to population size through the physical structure of fish schools and selection by fishermen of larger or denser schools and has little to do with the actual time spent fishing. How then can catch records be standardized to provide a reliable index of abundance? One solution that has been offered, is to use catch per unit of volume of water fished (cpuv) as an index of abundance (Treschev 1964). Changes in cpuv values would reveal changes in fish density within schools and the definition of the coefficient of catchability becomes the probability of catching afishwithin a unit of volume of water. The cpuv index has not been used, because in order to calculate actualfishabundance, the total volume of water occupied by each fish species has to be known. This is not an easy task and it is understandable that gathering knowledge of this type has been avoided. However, it is becoming clear that cpue is not a reliable index of abundance and that detailed knowledge of fish behaviour and habitat use needs to be studied before accurate predictions of abundance can be made. This study set out to see if it was possible to link changes in apparent abundance to habitat preferences of Pacific Ocean Perch (Sebastes alutus). by using physical oceanographic variables to define habitat. This was done by simultaneously monitoring temperature, salinity and depth values in the exact water masses in which the perch were caught during commercial fishing operations. The physical variable were then compared to a variety of abundance estimates (catch, proportion, cpue, cpuv) and tow characteristics (hours-towed, depth) to identify  83  which values best related to changes in physical characteristics of sites in which fish were caught. The purpose was to identify variables needed to define the 3-dimensional space that would accurately describe perch habitat. The study indicates that perch habitat can be defined in volumetric terms. Perch prefer a range of temperature from 4 to 7°C, which provides a predictable range of depth values. Deep water masses characterized by the same temperatures that perch prefer (less then 7°C), are forced upwards in the water column in the summer and downwards in the winter months by coastal winds. The annual migration pattern of perch follows the same pattern and it seems reasonable to assume that perch are physiologically adapted to that particular range of water temperatures and that there is a direct link between the depth of perch habitat and the strength and direction of the coastal winds. Before a priori predictions of expected seasonal depth ranges can be made detailed physical modeling needs to be done. Perch prefer areas of steep, rocky bathymetry where the interaction between strong tidal currents and rocky walls give rise to local small scale (10 km) frontal activity. The frontal activity may concentrate main food items, euphausiids, (Alldredge and Hamner 1980, Simard and Mackas 1989), which would explain the preference for these areas. The volume of water that perch will be found in can therefore be calculated on a site-specific basis as the distance from a canyon wall or rock outcropping,timesthe vertical depth range, (defined by temperatures from 4 to 7°C), times the distance along the wall that the fish occur. Larger scale abundance estimation for the QCS population of perch could be calculated by summing up the total volume of likely areas of perch habitat predicted by analyzing topographical maps and models of tidal currents in that region. Collecting information on board commercial vessels has advantages and disadvantages. The main advantage is the ability to monitor fishermen and fish behaviour simultaneously which allows better understanding of both sources of bias in abundance estimates. Catch, cpue and cpuv estimates of abundance made on a per-tow basis aboard commercial boats are similar to each other, but cpuv more accurately reflects changes in densities of schools than cpue.  84  Analyzing catch, proportion and effort for each tow, reveals that effort (hours per-tow) has nothing to do with catch and that catch values only increase when the proportion of perch in the catch increases, indicating that larger catches are due to fishermen targeting on denser or purer schools of perch. The main disadvantage to using operators and commercial vessels which are specialists on rockfish species, is that they only fish in the "best" perch areas and they tend to own larger vessels. Information collected from only this source may lead to biases in abundance estimates and as well, may not reveal areas of perch habitat that do not contain large, concentrated schools. The effect of such possible biases were investigated by comparing abundance estimates and tow characteristics between sampled and unsampled vessels, areas and seasons using historical catch records from commercial vessels operating in the QCS. Unfortunately, DFO's method of grouping individual tows by area per-trip, leads to misleading cpue and proportion values. Low values for both estimates can indicate either low abundance or increased targeting on species other than perch within that area. DFO's attempt to correct this situation by only including those trips with more than 25% of the species in question per-area does not address either of the problems. Removing those trips with greater than 20 hours of effort in any one area per-trip at least eliminates cpue and proportion estimates which are low due to large amounts of effort being directed at species other than perch, but may also throw out some valuable catch information. Therefore both the original DFO data set as well as the subset with all trips having greater than 20 hours of effort per-are removed were used to analysis the effects of vessels, areas and months. Comparisons between the larger vessels and the remaining fleet, reveals that operators of larger vessels are more likely to target only on perch in any one area and they do so in deeper waters. The DFO defined sub-management areas within the QCS are very different from each other, however, there is a very clear trend that indicates more fish are caught in deeper areas and attimesof the year when perch are at the deepest end of their depth range. Therefore it is likely that the areas sampled on board rockfish highliners are bias towards areas of larger concentrations.  85  Government research surveys are another source of abundance estimates and they have the advantage of not being commercially oriented such that areas of lesser concentrations of fish are as likely to be monitored as the areas of high concentration. However, abundance estimates calculated from the surveys should not be used exclusively until there is a better understanding of the effects that the physical environment can have on apparent abundance estimates. When the environmental factors that affect fish behaviour and movement patterns are understood and incorporated into abundance estimation, only then can an index of abundance be considered to be standardized.  Conclusions  Incorporating behaviour and movement pattern is common practice for the abundance estimation of most terrestrial organisms. Habitat mapping, to identify probable distributions, is used for most mammals which range over large distances that make individual counts nearly impossible (Siniff and Skoog 1964, Norton-Griffiths 1978). Within fisheries science many appeals have been broadcast for information on the distribution of fish habitats and for physical oceanographic and fisheries researches to work together on these projects (Bakun et al. 1982, Sharp 1988). To quote from the review by Bakun et al.; "The importance of data accessibility and multidisciplinary research activities, beginning with dialogues between the ocean research and fishery resource scientists is paramount to successful completion of these tasks (identification and accounting for causal climate-driven ambient variations)". For some of the more commercially important species, such as the Atlantic skipjack tuna (Katsuwonus pelamis). this cooperative type of research has been attempted with some success (Evans etal. 1981). Historical records of catches, sea surface temperatures, depths of isotherms  86  and dissolved O2 where used to make maps of possible areas and depths of skipjack tuna habitat. However, as Evans et al. point out, the large spatial (1000's of km) and temporal (annual) scales generally used in these types of reviews are not appropriate for the actual behaviour and movement responses that fish have to seasonal, perhaps daily events. To understand the changes in fish behaviour and movement in response to environmental factors at a level that enables this information to be incorporated into abundance estimation, studies have to be conducted at the appropriate spatial and temporal scales. A study by Cury and Roy (1989), illustrate how successful it is to first understand the small scale causal mechanisms between upwelling intensity, turbulence and food availability to larval fish before making large scale predictions on subsequent numbers of recruits to the fishery. Studies on the east coast of Canada by Rose and Leggett 1988, Perry et al. 1988 and Smith et al. 1990, show promise for informative investigations when biological and physical oriented researchers have joined forces. The works by Perry et al. (1988) and Rose and Leggett (1988) have demonstrated the ability to correlate the abundance of individual fish species with water mass type. The paper by Smith et al. (1990) presents a methodology to incorporate this type of knowledge into estimates of population abundance. An intriguing study by Jones and Scholes (1980) illustrates the point that each species reacts to its environment in a unique way. In a series of deductive steps, they describe how and why catches of plaice (Pleuronectes platessa) and cod (Gadus morhua) change so drastically and differendy when the winds shift in the North Sea. A common theme and indication from these successful studies is that it is not possible to make generalizations for different fish species concerning habitat preferences and the effects of environmental variables; each species needs to be considered separately. Before groans of "budgets" are voiced, I suggest that a vast amount of untapped knowledge on individual species habitat, exists within the community of people that make a living out of accurately predicting where fish are; the fishermen. It is up to the scientific community to tap those resources and combine interdisciplinary forces. A next step, to understand causal links between environmental factors and fish abundance, is to identify regions and times where there is enough biological and  87  physical information in the appropriate spatial and temporal scales to illuminate any causal mechanisms. Simultaneously, realistic physical models need to be constructed such that they can be used to test hypotheses of the effects that different environmental factors have on fish behaviour and movement. When causal mechanisms become apparent or if there is a lack of appropriate information, research activities should concentrate on studies performed at sea which are directed towards identifying causal mechanisms between changes in abundance and oceanographic factors.  88  Literature Cited Alldredge, A. L. and W. M. Hamner. 1980. Recurring aggregation of zooplankton by a tidal current. Estuar. Coast. Mar. Sci. 10:31-37. Alverson, D. L . 1960. A study of annual and seasonal bathymetric catch patterns for commercially important groundfishes of the Pacific northwest coast of north America. Pac. Mar. Fish. Comm. Bull. 4: 66 pp. Archibald, C. P., D. Fournier and B. Leamen. 1983. Reconstruction of stock history and development of rehabilitation strategies for Pacific Ocean Perch in Queen Charlotte Sound, Canada. North Amer. J. Fish. Manage. 3:283-294. Bakun, A. 1973. Coastal upwelling indices, west coast of north America, 1946 - 71. Dep. Commer. NOAA Tech. Rep. NMFS SSRF-671,103 pp. Bakun, A„ J. Beyer, D. Pauly, J. G. Pope and G. D. Sharp. 1982. Ocean sciences in relation to living resources. Can. J. Fish. Aquat. Sci. 39:1059-1070. Barner, L. W., J. Selsby and F. Mo'ttl. 1978. Rockfish survey off the west coast of the Queen Charlotte Islands make on Arctic Harvester Hydroacoustic Cruise 78-2 May 11-June 2,1978. Fish. Mar. Ser. Data Rep. No. 110. pp. 86. Barber, F. G. 1957. The effect of the prevailing winds on the inshore water masses of the Hecate Strait region, B.C. J. Fish. Res. Bd. Can. 14:945952. Beverton, R. J. H. and S. J. Holt. 1957. On the dynamics of exploited fish populations. Fish. Invest., Lond., Ser. 2,19,533pp. Brandt, S. B., J. J. Magnuson and L. B. Crowder. 1980. Thermal habitat partitioning by fishes in Lake Michigan. Can. J. Fish. Aquat. Sci. 37:1557-1564. Brodeur, R. D. and W. G. Pearcy. 1984. Food habits and dietary overlap of some shelf rockfishes genus Sebastes from the northeastern Pacific Ocean. Natl. Mar. Fish. Serv. US Fish. Bull. 82:269-294.  89  Byrne, C. J., T. R. Azarovitz and M. P. Sissenwine. 1981. Factors affecting variability of research vessel trawl surveys. In: Bottom trawl surveys. Can. Spec. Publ. Fish. Aquat. Sci. No. 58. pg 258-273,273 pp. Carlson, H. R. and R. E. Haight. 1976. Juvenile life of Pacific Ocean Perch, Sebastes alutus. in coastal fjords of southeastern Alaska: their environment, growth, food habits, and school behavior. Trans. Amer. Fish. Soc. 105 :191-201 Carlson, H. R. and R. R. Straty. 1981. Habitat and nursery grounds of Pacific rockfish, 'Sebastes' spp, in rocky coastal areas of southeastern Alaska. Mar. Fish. Rev. 43:13-19. Carscadden, J. E., K. T. Frank and D. S. Miller. 1989. Capelin (Mallotus villosus) spawning on the southeast shoal: influence of physical factors past and present. Can. J. Fish. Aquat. Sci. 46:1743-1754. Carter, E. W., D. A. Nagtegaal, B. M. Leaman, C. P. Archibald and B. J. Westmen. 1982. Catches and trawl locations of the M/V Tenacious during the rockfish biomass survey in southern Hecate Strait (Moresby Gully) June 1981. Can. Data Rep. Fish. Aquat. Sci. No. 321, 106 pp Christie, G. C. and H. A. Regier. 1988. Measures of optimal thermal habitat and their relationship to yields for four commercial fish species. Can. J. Fish. Aquat. Sci. 45:301-314. Crecco, V. A. and W. J. Overholtz. 1990. Causes of density-dependent catchability for Georges Bank Haddock Melanogrammus aeglefinus. Can. J. Fish. Aquat. Sci. 47:385-394. Crecco, V. A. and T. F. Savoy. 1985. Density-dependent catchability and its potential causes and consequences on ConnecticutriverAmerican Shad, Alosa sapidissima. Can. J. Fish. Aquat. Sci. 42:1649-1657. Cury, P. and C. Roy. 1989. Optimal environmental window and pelagic fish recruitment success in upwelling areas. Can. J. Fish. Aquat. Sci. 46:670-680.  90  Dodimead, A. J. 1980. A general review of the oceanography of the Queen Charlotte Sound - Hecate Strait - Dixon Entrance region. Can. MS. Rep. Fish. Aquat. Sci. No. 1574, 248 pp. Dodimead, A. J., A. Ballantyne and M. Douglas. 1979a. Oceanographic observations during fisheries research surveys off the British Columbia coast in 1977. Fish. Mar. Ser. Data Rep. No. 144, 41 pp. Dodimead, A. J., A. Ballantyne and M. Douglas. 1979b. Oceanographic observations during fisheries research surveys off the British Columbia coast in 1978. Fish. Mar. Ser. Data Rep. No. 160, 136 pp. Doubleday, W. G. and D. Rivard, (ed). 1981. Bottom trawl surveys. Can. Spec. Publ. Fish. Aquat. Sci. No. 58, 273 pp Drinkwater, K. F. and R. A. Myers. 1987. Testing predictions of marine fish and shellfish landings from environmental variables. Can. J. Fish. Aquat Sci. 44:1568-1573. Evans, R. H., D. R. McLain and R. A. Bauer. 1981. Adantic skipjack tuna: influences of mean environmental conditions on their vulnerability to surface fishing gear. Mar. Fish. Rev. 43:1-11. Frank, K. T. and J. E. Carscadden. 1989. Factors affecting recruitment variability of capelin (Mallotus villosus) in the Northwest Atlantic. J. Cons, int Explor. Mer. 45:146-164. Freeland, H. J. and K. L. Denman. 1982. A topographically controlled upwelling center off southern Vancouver Island. J. Mar. Res. No. 40:1069-1093. Gulland, J. A. 1956. On the fishing effort in English demersal fisheries. Fish. Invest., Lond., Ser 2,20(5), 41 pp. Gulland, J. A. 1964. Catch per unit effort as a measure of abundance. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 155:8-14.  91 Gunderson, D. R. 1971. Reproductive patterns of Pacific Ocean Perch (Sebastodes alutus) off Washington and British Columbia and their relation to bathymetric distribution and seasonal abundance. J. Fish. Res. Bd. Can. 28:417-425. -  Gunderson, D. R. 1972. Availability, size composition, age composition and growth characteristics of Pacific Ocean Perch (Sebastes alutus) off the northern Washington coast during 1967-72. J. Fish. Res. Bd Can. 31:21-34. Gunderson, D. R. 1974. Evidence that Pacific Ocean Perch (Sebastes alutus) in Queen Charlotte Sound form aggregations that have different biological characteristics. J. Fish. Res. Bd. Can. 29:1061-1070. Hamner, W. M. 1988. The lost year' of the sea turtle. Trends Ecol. Evol. 3:116-118. Harting, W. R., D. Davenport, M. S. Smith, U. Kristiansen and S. J. Westrheim. 1970. G. B. Reed groundfish cruise No. 70-1, March 5 - June 18,1970. Fish. Res. Bd. Can. Tech. Rep. No. 205, 82 pp. Harting, W. R., D. Davenport, M. S. Smith, R. M. Wowchuk and S. J. Westrheim. 1971. G. B. Reed groundfish cruise No. 71-3, October 1-29, 1971. Fish. Res. Bd. Can. Tech. Rep. No. 290, 35 pp. Harting, W. R., D. Davenport, M. S. Smith, A. C. Phillips and S. J. Westrheim. 1973. G. B. Reed groundfish cruise No. 73-2, September 5-25,1973. Fish. Res. Bd. Can. Tech. Rep. No. 494, 37 pp. Harling, W. R. and D. Davenport. 1977. G. B. Reed groundfish cruise No. 77-3, August 22 to September 8, 1977. Fish. Mar. Ser. Data Rep. No. 42, 46 pp. Harling, W. R., R. M. Wallis, M. R. Fretwell and F. W. Mottl. 1978. Artie Harvester groundfish cruise No. 78-3 (February 27-March 10,1978). Fish. Mar. Ser. Data Rep. No. 78, 35 pp. Huppert, H. E . . and K. Bryan. 1976. Topographically generated eddies. Deep-Sea Res. 23:655-679.  92  Jones, F. R. H. and P. Scholes. 1980. Wind and the catch of a Lowestoft trawler. J. Cons. int. Explor. Mer. 39:53-69. Ketchen, K. S. 1981. Reconstruction of Pacific Ocean Perch (Sebastes alums) stock history in Queen Charlotte Sound. Part 2. Catch per unit of effort as a measure of abundance, 1959-79. Can. MS. Rep. Fish. Aquat. Sci. No. 1599,72 pp. Laevastu, T. and M. L. Hayes, (ed). 1981. Fisheries oceanography and ecology. Fishing News Books Ltd. England, 199 pp. Leaman, B. M. and D. A. Nagtegaal. 1986. Biomass survey of rockfish stocks in the Dixon Entrance-Southeast Alaska region, July 5-22,1983 (R/V G. B. Reed and M/V Free Enterprise No.l). Can. Tech. Rep. Fish. Aquat. Sci. No. 1510, 63 pp. Leggett, W. C , K. T. Frank and J. E. Carscadden. 1984. Metrological and hydrographic regulation of year-class strength in Capelin (Mallotus villosus). Can. J. Fish. Aquat. Sci. 41:1193-1201. Longhurst, A. 1984. Heterogeneity in the ocean - implications for fisheries. Rapp. P-v. Reun. Cons. int. Explor. Mer. 185: 268-282. Lough, R. G., P. C. Valentine, D. C. Potter, P. J. Auditore, G. R. Bolz, J. D. Neilson and R. I. Perry. 1989. Ecology and distribution of juvenile cod and haddock in relation to sediment type and bottom currents on eastern Georges Bank. Mar. Ecol. Prog. Ser. 56:1-12. Lyubimova, T. S. 1968. Basic aspects of the biology and distribution of the Pacific Ocean Perch (Sebastodes alutus Gilbert) in the Gulf of Alaska. In: Soviet fisheries investigations in the northeast Pacific. Part 1. Clearinghouse Fed. Sci. Tech. Inform., Springfield, VA. TT67-51203:308-318. Lyubimova, T. S. 1968. Main stages in the life cycle of the rockfish Sebastodes alutus Gilbert in the Gulf of Alaska. In: Soviet fisheries investigations in the northeast Pacific. Part IV. Clearinghouse Fed. Sci. Tech. Inform., Springfield, VA. TT67-51206:85-111.  93  MacCall, A. D. 1976. Density dependence of catchability coefficient in the California Pacific Sardine, Sardinops sagax caerulea. purse siene fishery. Cal. Coop. Ocean. Fish. Invest. 18:136-148. Nagtegaal, D. A. 1983. Identification and description of assemblages of some commercially important rockfishes {Sebastes spp) off British Columbia. Can. Tech. Rep. Fish. Aquat. Sci. No. 1183 82 pp. Nagtegaal, D. A., B. M. Leamen and R. D. Stanley. 1986. Catches and trawl locations of R/V G.B. Reed and MIV Eastward Ho. during the Pacific Ocean Perch assessment cruise to Queen Charlotte Sound, August - September, 1984. Can. Data Rep. Fish. Aquat. Sci. No. 611,109 pp. Norton, J. G. 1987. Ocean climate influences on rockfish recruitment in the California current. In: Proceedings of the international rockfish symposium October 20-22,1986, Anchorage, Alaska. Alaska Sea Grant, University of Alaska, Fairbanks, Alaska, 393 pp. Norton-Griffiths, M. 1978. Counting animals, 2nd ed. African Wildlife Leadership Foundation, Nairobi. 110 pp. Paloheimo, J. E. and L. M. Dickie. 1964. Abundance and fishing success. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 155:152-163. Perry, R. I., J. S. Scott and R. J. Losier. 1988. Water mass analysis and groundfish distributions on the Scotian Shelf, 1979-84. Northwest Ad. Fish Organ. Serial No. N1533, SCR Doc. 88/81. 14 pp. Peterman, R. M . and G. J. Steer. 1981. Relation between sportfishing catchability coefficients and Salmon abundance. Trans. Amer. Fish. Soc. 110:585-593. Pingree, R. D., G. R. Forster and G. K. Morrison. 1974. Turbulent convergenttidalfronts. J. Mar. Biol. Ass. U.K. 54:469-479.  94  Ponomareva, L. A. 1966. Euphausiids of the north Pacific; their distribution and ecology. Isreal program of Scientific translations, Jerusalem, 154 pp. Radovich, J. 1976. Catch-per-unit-of-effort: fact, fiction, or dogma. Cal. Coop. Ocean. Fish. Invest. Rep. 18:31-33. Richards, L. J. and J. T. Schnute. 1986. An experimental and statistical approach to the question: Is CPUE an index of abundance? Can. J. Fish. Aquat. Sci. 43:1214-1227. Ricker, W. E. 1944. Further notes on fishing mortality and effort. Copeia, No. 1:23-44. Rivard, D. 1981. Catch projections and their relation to sampling error of research surveys. In: Bottom trawl surveys. Can. Spec. Publ. Fish. Aquat. Sci. No. 58. pg 93-109, 273 pp. Rose, G. A. and W. C. Leggett. 1988. Atmoshpere-ocean coupling and Adantic cod migrations: Effects of wind-forced variations in sea temperatures and currents on nearshore distributions and catch rates of Gadus morhua. Can. J. Fish. Aquat. Sci. 45:1234-1243. Sharp, G. D. 1988. Fish populations and fisheries. Their perturbations, natural and man-induced. In H. Postma and J. J. Zijlstra (eds.), Ecosystems of the world - Continental shelves. Vol 27. Elsevier science publishing company inc. New York, N.Y. 421 pp. Shepherd, J. G., J. G. Pope and R. D. Cousens. 1984. Variations in fish stocks and hypotheses concerning their links with climate. Rapp. P.-v. Reun. Cons. int. Explor. Mer. 185:255-267. Simard, Y. and D. L. Mackas. 1989. Mesoscale aggregations of eupshausiid sound scattering layers on the continental shelf of Vancouver Island. Can. J. Fish. . Aquat. Sci. 46:1238-1249.  95  sound scattering layers on the continental shelf of Vancouver Island. Can. J. Fish. Aquat. Sci. 46:1238-1249.  Siniff, D. B. and Skoog, R. O. 1964. Aerial censusing of caribou using stratified random sampling. J. Wildl. Manage. 28:391-401. Skalkin, V. A. 1967. Diet of rockfish in the Bering Sea. In: Soviet fisheries investigations in the northeast Pacific. Part 11. Clearinghouse Fed. Sci. Tech. Inform., Springfield, VA. TT67159-51204:159-174. Smith, S. J, R. I. Perry and L. P. Fanning. In press. Effects of trends in enviromental variation on estimates of fish population abundance. Environmetrics. Thomson, R. E. and R. E. Wilson. 1987. Coastal countercurrent and mesoscale eddy formation bytidalrectification near and oceanic cape. J. Phys. Oceanogr. 17:2096-2126. Thomson, R. E. 1981. Oceanography of the British Columbia coast. Can. Spec. Publ. Fish. Aquat. Sci. No. 56,291 pp. Treschev, A. I. 1964. Onfishingintensity. Rapp. P.-V. Reun. Cons. Int. Explor. Mer 155:19-20. Treschev, A. I. 1975. Fishing unit measures. Rapp. P-V. Reun. Cons. Int. Explor. Mer 168:54-57. Tyler, A. V., and J. Fargo (ed). 1990. Groundfish stock assessments for the west coast of Canada in 1989 and recommended yield options for 1990. Can. Tech. Rep. Fish. Aquat. Sci. No. 1732. 343 pp. Tyler, A. V., R. P. Foucher and J. Fargo (ed). 1986. Groundfish stock assessments for the west coast of Canada in 1985 and recommended yield options for 1986. Can. MS. Rep. Fish. Aquat. Sci. 1897:108 pp.  96  Westrheim, S. J. 1970. Survey of rockfishes, especially Pacific Ocean Perch, in the northeast Pacific Ocean, 1963-66. J. Fish. Res. Bd. Can. 27:17811809. Westrheim, S. J. 1974. Echo-sounder and trawl survey of Queen Charlotte Sound and Southern Hecate Strait, 1972-73. Fish. Res. Bd. Can. MS Rep. No. 1307,43 pp. Westrheim, S. J. 1975. Reproduction, maturation, and identification of larvae of some Sebastes (Scorpaenidae) species in the northeast Pacific Ocean. J. Fish. Res. Bd. Can. 32:2399-2411. Westrheim, S. J. 1983. A new method for allotting effort to individual species in a mixed-species trawl fishery. Can. J. Fish. Aquat. Sci. 40:352360. Westrheim, S. J., D. Davenport, M. S. Smith and W. R. Harling. 1972. G. B. Reed groundfish Cruise No. 72-1 June 13-29,1972. Fish. Res. Bd. Can. Tech. Rep. No. 328,17 pp. Westrheim, S. J., W. R. Harling, D. Davenport, M. S. Smith and A. C. Phillips. 1973. G. B. Reed groundfish Cruise No. 73-1 June 5 - July 26, 1973. Fish. Res. Bd. Can. Tech. Rep. No. 410, 51 pp. Westrheim, S. J., W. R. Harling, D. Davenport and M. S. Smith. 1974. G. B. Reed groundfish cruise No. 74-4 September 4-25,1974 (data record). Fish. Mar. Ser. Tech. Rep. No. 497, 43 pp. Westrheim, S. J., B. M. Leamen, W. R. Harling, D. Davenport, M . S. Smith and R. M. Wowchuk. 1976. G. B. Reed groundfish Cruise No. 76-3 September 8-27, 1976. Fish. Mar. Ser. Data Rep. No. 21,45 pp. Winters, G. H. and J. P. Wheeler. 1985. Interaction between stock area, stock abundance, and catchability coefficient. Can. J. Fish. Aquat. Sci. 42:989-998.  97  Appendix A  Estimates are calculated made from legends found in Ketchen 1981 and from maps used by the groundfish division at the Pacific Biological Station, Nanaimo, B.C., Canada.  5 = Triangle  5 A 1101  675 k m  2  6= = Cape Scott Spit  5 A 1102  310 k m  2  7 = Mexicana (Stump Ranch)  5 A 1103  220 k m  2  8 = Topnot  5 A 1104  150 k m  2  9 = Pine Island  5 A 1105  100 k m  2  10 = South Scott Island  5 A 1106  ?km  11 = Outside West Triangle  5 A 1107  235 k m  2  14 = South Triangle  5 A 1110  150 k m  2  15 = Pisces Canyon  5 A 1111  100 k m  2  17 = North East Corner  5B801  330 k m  2  18 = South East Edge  5B802  660 k m  2  19 = North West Corner  5B803  510 k m  2  20 = South West Corner  5B 804  890 km  2  21 = Mitchells  5B805  1580 km  22 = South of Cape St James  5B 806  750 k m  2  26 = Outside Goose and Mitchell  5B 810  285 k m  2  27 = South West Middle Bank  5B 811  700 k m  2  28 = Outside Cape St James  5B812  375 km  2  29 = South Moresby  6  650 km  2  30 = Horseshoe  6  ?km  2  2  2  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0302382/manifest

Comment

Related Items