VIWA

Taihoro Nukurangi

Movement patterns and stock mixing of blue cod in Southland (BC05)

G. Carbine, J. McKenzie

Final Research Report for Ministry of Fisheries Research Project BCO9702 (Revised)

National Institute of Water and Atmospheric Research

February 2001 Final Research Report

Report Title: Movement patterns and stock mixing of blue cod in Southland (BC05)

Authors: Glen Carbines and Jeremy McKenzie

Date: 30 September 2000

Contractor: National Institute of Water and Atmospheric Research Limited

3. Project Title: Determination of movement of blue cod in Southland

4. Project Code: BCO9702

5. Project Leader: Glen Carbines

6. Duration of Project: Start Date: 1 October 1997 Completion Date: 30 September 1999

7. Executive Summary:

Movement patterns of blue cod are currently unclear, and it has not been established if populations can be regarded as a single entity within some of the present fish stock areas. The current study sought to determine movements of blue cod in domestic return area 025 within the Southland fish stock BC05.

During March 1998, 9368 blue cod were tagged within domestic fishing return area 025 (Foveaux Strait). The tagging programme used a balanced stratified series of replicated sites (n=9) within three latitudinal and three longitudinal strata spread throughout area 025 (approximately 1000 fish per stratum). Three different bottom types were also identified throughout most sites (sand flats, biogenic reefs, and rocky reef fringe).

Blue cod were tagged using t-bar tags in a ventral position. Tagged fish ranged in size from 140-510 mm total length, with a mean of 324.5 + 0.5 mm. Length frequency distributions were derived for all sites and bottom types. The larger numbers of small fish caught on sand flats and biogenic reefs suggest these habitats may be important to recruitment. CPUE results also indicated that blue cod abundance might differ between bottom habitats.

After 20 months, 743 (7.9%) blue cod had been returned from all sites and all bottom types present within those sites. The largest distance moved was 156 km, however the median was 790 m, with 60.2% of fish moving less than 1 km. It therefore appears that densities of blue cod in Foveaux Strait are temporally stable on a median-term scale. No size distance relationship was evident in tagging data, and habitats blue cod were released into also had no effect on the distance moved.

Spatial location was an important determinant of distance travelled. However, a significant interaction between latitudinal and longitudinal release strata shows that factors influencing distance travelled are more complex than the latitudinal or longitudinal location on its own.

Mixing rate calculations showed that blue cod in area 025 had moderate mixing with neighbouring domestic return areas (up to 14.7%). However, mixing was considerably higher between adjacent sub-areas within area 025 (up to 44.1%). These results suggest that blue cod stocks are relatively stable at the domestic return area scale.

While the majority of blue cod did not move any significant distance, those that did showed a strong trend toward counter current north-west movements. Mixing rate calculations indicated higher levels of immigration to the west and emigration to the east of area 025. An increased proportion of blue cod moving significant distances in the spring of both 1998 and 1999 is suggestive of a possible return seasonal migration associated with spawning. Overall, these observations suggest that blue cod movement is complex.

8. Objectives:

1. To design a tagging programme to determine movements of blue cod in domestic fishing return area 025 in Southland.

2. To complete the tagging phase of the programme.

3. To complete the recovery phase of the programme.

Introduction

Blue cod (Parapercis colias) are endemic to New Zealand. They are not a true cod but a member of the weaver family (Pinguipedidae), of which 43 species are described throughout the Indo-Pacific region. Distributed from the shore to the shelf edge of New Zealand's entire coastline, blue cod are most abundant in Southland and the Chatham Islands. They are opportunistic benthic carnivores found on reef edges, shingle/gravel or sandy bottoms, often close to rocky outcrops (Graham 1939a, Rapson 1956, Mutch 1983). Blue cod occur primarily in coastal waters shallower than 200 m, however, they may occasionally be caught by trawlers from as deep as 350 m (Anderson 1998).

Like most of New Zealand's reef fish, blue cod are not evenly distributed throughout their range of habitats, and several biological and physical factors appear to influence their distribution (Mutch 1983, Jones 1988). The response to these factors may also vary with age and may cause younger blue cod to have a more transient existence (Mace & Johnston 1983, Mutch 1983). Larval blue cod are thought to settle into offshore habitats deeper than 120 m and then move inshore upon reaching lengths of 180-200 mm (Rapson 1956). However, scientific divers have observed juvenile blue cod shallower than 20 m in both the Marlborough Sounds (Cole, 2000 Pers. Com.) and Paterson Inlet, Stewart Island (Pers Obbs. 1998).

As adults, blue cod have been shown to remain within small-localised areas (Rapson 1956, Mace & Johnston 1983, Mutch 1983). There are also anecdotal observations that suggest blue cod off Dunedin (Figure 1) migrate off shore in late winter to spawn (Graham 1939b, Robertson 1973). However, there is no evidence of seasonal migration from studies of blue cod in the Marlborough Sounds (Figure 1) (Rapson 1956, Mace & Johnston 1983).

Members of the family Pinguipedidae are usually territorial (Stroud 1982, Sano 1990, Clark etal. 1991, Kobayashi et al. 1993a, Kobayashi et al. 1993b, Sano 1993, Ohnishi et al. 1997), and large male blue cod are have been observed to hold large and rather loose territories (Mutch 1983). Underwater observations of blue cod show that male body size and territory size are positively correlated. Social groups of 3-5 females may exist within the home range a single dominant male (Mutch 1983).

Despite being predominately territorial, blue cod have been observed to occasionally school. Many fishers have reported catches of single sexed groups of blue cod (Rapson 1956, Warren et al. 1997), and mid water aggregations have been observed by scientific divers (Duffy, 1995 pers. comm.; Francis, 1998 pers. comm.). Analysis of gut contents suggests mixed sex schooling may be a feeding response (Duffy, 1995 pers. comm.)

Blue cod are an important commercial and recreational fishery in the South Island of New Zealand (Annala 1999). The commercial blue cod catch is managed under the Quota Management System (QMS) relative to eight Quota Management Areas (QMAs). Each QMA has been assigned its own annual Total Allowable Commercial Catch (TACC). The QMA with the largest TACC is BC05 (Figure 1). At 1536 tonnes, BC05 accounts for 58% of the National TACC. The BC05 TACC is estimated to be eleven times larger than Southland's recreational take (139 tonnes) (Annala 1999). Both recreational and commercial blue cod fisheries in Southland are constrained by a minimal legal size of 330 mm total length (Annala 1999).

Within each QMA, there are several domestic fishing return areas (e.g. 7 for BC05, Figure 1). Fishers can operate in any domestic fishing return area within their QMA, and this may potentially lead to an uneven distribution of fishing effort. For example in BC05 all significant ports are located within domestic return area 025, consequently a high proportion of the annual BC05 catch (52% on average) is taken from this area (Warren et al. 1997).

For the purposes of management, it is assumed that fish in each QMA can be considered a single "stock". In reality there may be several smaller discrete stocks within each QMA, and uneven distribution of fishing effort across the QMA may result in some of these stocks becoming over-fished. Additional management measures may be necessary to ensure fishing effort is proportional to abundance within sub-areas of QMA. In order to determine the appropriate spatial scale on which to manage blue cod within a QMA, it is necessary to gain information on mixing rates between areas. The first major blue cod tagging was undertaken in the Marlborough Sounds (Figure 1) between 1940-41 (Rapson 1956). Of the 5050 fish tagged, 194 (3.8%) were recaptured within 13 months; 83 within 3 months, 62 between 3 and 6 months, and 49 after 6 months. Eighteen (9.3%) of the recaptured fish had moved more than 1.6 km, and only six (3.1%) of these had moved more than 16 km. Of the 18 migrants, 16 were larger than 300 mm when tagged, suggesting that blue cod movement is size dependent. The largest movement recorded was 48 km, but as this was at the full range of fishing effort the possibility that tagged fish moved further could not be excluded (Rapson 1956).

Rapson (1956) described these "migrants" as "solitary fish which for some reason have left the shoal and roved far afield". While Rapson (1956) also believed that occasional shoals of blue cod did migrate (based on a total absence of returns from some areas), in most localities there was evidence that shoals remained stationary for many months. Rapson (1956) concluded that blue cod from temporally stable groups: "... a few cod migrate, some small fish as rovers and the larger, when conditions are right, as travelling cod".

A blue cod tagging programme was also undertaken in the Marlborough Sounds between 1973-76 (Mace & Johnston 1983). Anchor tags were used to mark 2430 blue cod. From these, 84 (3.5%) fish were recaptured; 43 within 3 months, 12 between 3 and 6 months, 9 between 6 and 9 months and 10 after 9 months. Of these fish, 21 (28%) had moved more than 1.6 km from their release site and only 4 (4.8%) had moved more than 16 km. The greatest movement recorded was 42 km.

In contrast to Rapson's (1956) study, Mace and Johnston (1983) observed no blue cod over 30 cm moving a significant distance from its release site, thus suggesting that smaller blue cod were more mobile. However, like Rapson (1956), Mace and Johnston (1983) concluded that populations of blue cod are sedentary in temporally stable groups.

Double tagging indicated that tag shedding rates in the 1973-76 programme may have been as high as 43% (Mace & Johnston 1983). Rapson (1956) also reported significant tag loss in his programme and attributed it to other blue cod biting off tags.

Based on the movement results of the two tagging studies, blue cod of the Marlborough Sounds are regarded as a separate stock with limited internal movement. However, Mace and Johnston (1983) noted that further work was necessary to determine if blue cod behaviour in the Marlborough Sounds differs from those in other areas and habitats.

The high tag loss rates encountered in both studies indicated a need to evaluate different types of tags for blue cod. The feasibility of tagging blue cod with a variety tag types has been investigated (Carbines 1995, Carbines 1998a). Five tag types were tested in an aquarium for site selection and 700 blue cod were released near Stewart Island (Figure 1) to examine tag retention as part of a tetracycline age validation study (Carbines 1998a). The results showed that t-bar tags inserted in a ventral position through the postcleithrum at the base of the pelvic fin avoided many problems associated with other tags and were relatively inexpensive and easy to apply (Carbines 1998a, Carbines 1998b). Each tag carried an individual identification number and a return address. T-tags were also non-cryptic and especially obvious to fishers who invert blue cod to cut the ventral aorta for bleeding. The retention rate of t-tags was estimated to be 69.8%, and a return rate of 5.23% over two years was achieved without using rewards.

The movement results from returned fish in the 1995 study indicated that blue cod in Southland behaved differently to those in the Marlborough Sounds, moving more often, and significantly further (Carbines 1998b). Consequently, to address the concern that variable fishing pressure across BC05 (area 025 in particular) may result in the over fishing of discrete blue cod sub-stocks, the Ministry of Fisheries commissioned a new tagging study. This report presents results from that study (BCO9702).

9. Methods

Experimental Design

The purpose of the experimental design was to describe the extent of movement for blue cod within domestic return area 025 (local scale), and between area 025 and adjacent domestic return areas (broad scale) (Figure 1).

Domestic return area 025 was subdivided into nine possible tagging strata, the boundaries of which where determined by a combination of three latitudinal strata (i.e. near-shore mainland New Zealand (north), mid Foveaux Straight, and near-shore Stewart Island (south)) and three longitudinal strata (i.e. West, Central, and East) (Figure 2)l.

After consultation with fishers, a tagging site (approximately 5 km2 in area) was chosen within each possible tagging stratum, as defined by stratum combinations (Figure 2).

Fishing was undertaken using cod pots within tagging sites until 1000 blue cod had been tagged within each site. The target of 1000 tags per site had a predicted return of approximatly 50 tags, potentually making returns from each site comparable with previous blue cod tagging programmes (Rapson 1956, Mace & Johnston 1983).

As blue cod may hold large and relatively stable territories (Mutch 1983), all efforts were made to return fish to the specific area caught. Pots were therefore set in a systematic manner in several discrete fishing locations (approximately 300 m2) throughout each tagging site until the target number of blue cod were tagged and released within each site (n=1000). While the number of fish tagged in each site was carefully balanced, the number of fish tagged within each fishing location was done opportunistically based on the catch of blue cod.

1 Due to the NW/SE orientation of Foveaux Strait, "latitudinal" strata are offset by approximately 18- 35°and "longitudinal" strata are offset by approximately 35-65°. Each fishing location was also characterised by depth, location (autonomous Global Positioning System (GPS)), and bottom habitat (using a commercial sonar). Bottom habitat was clasified as being either;

1) Sand flats: Open sand/gravel flats with low spatial hetrogenaity and little or no macro-algae cover.

2) Biogenic reef: Moderate spatial hetrogenaity consisting of mainly organic substrate such as bryozoans, sponges etc.

3) Rocky reef fringe: Sand/gravel reef edge, adjacent to broken rocky topography with boulders, often supporting large brown macro-algae.

A small dredge was used to confirm sonar observations of bottom habitat type. However, it was not practical to balance release numbers across the three habitat types within each site.

Tag release

Blue cod were caught using two chartered commercial vessels. Tagging was carried out between 23 January and 30 March 1998 by NIWA staff. Blue cod were caught using commercial cod pots covered in a 1 cm soft mesh, a technique shown to cause no detectable mortality of returned fish, even when handled poorly (Carbines 1999b).

After each pot was lifted, fish were quickly removed from pots with wet cotton gloves and placed in onboard holding tanks fed with constant running seawater. Blue cod were then measured to the nearest centimetre below total length, tagged at the base of the pelvic fin and placed in a recovery tank.

Yellow 15 mm Hallprint type TBA-2 t-bar anchor fish tags were inserted ventrally through the postcleithrum (at the base of the pelvic fin), using a Dennison tag-fast HI tag gun with No. 08941 needles. All tags were printed with an individual identification number, with the words "REWARD" and a contact address.

Once catches declined to a low level, all fish in the recovery tank were released at the centre of each fishing location. One mass central release per fishing location ensured that; a) blue cod were released no further than approximately 150 m from their point of capture, b) all blue cod had sufficient time to recover from tagging, and c) losses from predators such as mollymawks (Diomedea sp.) and barracouta (Thyrsites atun) were avoided.

Tag returns

The blue cod fishery is considered seasonal as many fishers are involved in the rock lobster fishery over the spring and early summer period (Warren et al. 1997). However, the timing of the tagging phase allowed returns to be collected over two seasons of the fishery. The recovery period reported here was from 23 January 1998 to 30 November 1999. Rewards were provided for returned tags to encourage a high rate of returns and to promote integrity in the catch data. Several types of rewards have been used in past studies but fishers have been particularly encouraged to participate by offering T- shirts and prize lotteries (Francis 1988, Beentjes 1999). For the current tagging program a "Blue Cod - Tagging Program" T-shirt was offered to all participants who returned a tag and supplied relevant catch information. However, as most returns were anticipated to come from commercial fishers returning several fish, a prize draw for an Anset New Zealand mystery weekend for two was used to encourage their continued participation (Carbines 1999a). All fishers returning tags were also provided with details of fish movements.

Fishers were requested to provide fish length, date of capture, and location (GPS preferably) details for all recovered tagged fish. Notice of the tagging programme was publicised in Seafood New Zealand magazine (Carbines 1998a) and the Southland Times newspaper in an attempt to notify both recreational and commercial fishing sectors. To further publicise the tagging programme, NTWA also sponsored a prize for the largest blue cod at the 1998 and 1999 Stabi-craft southern marine open water- fishing tournament in held in Bluff Harbour.

Data analysis

The nett distance travailed by blue cod was calculated as the straight line distance between release and recapture locations. Only occasionally did the calculated route intersect land, where distance travelled was then estimated by a series of seaward straight lines. For estimates of direction moved, intersection with land was ignored.

A two-factor randomisation test (Manly & Francis 1999) was used to investigate the effect of longitudinal and latitudinal stratum at release on distance moved. A similar test was used to investigate the effect of release-site bottom habitat type on distance moved.

The Manly and Francis test (Manly & Francis 1999) uses a randomisation procedure to approximate the "F' distribution used in the analogous parametric analysis of variance test (ANOVA). Unlike the standard ANOVA, the Manly and Francis procedure is not biased by the presence of unequal variances in the data (Manly & Francis 1999).

Length frequency distributions were compared (pair-wise) using the Kolmogorov- Smirnov parametric test.

Estimates of mixing within and between Domestic Return Area 025

A modification of the Peterson mark-recapture model used tagging to estimate mixing rates between areas within domestic return area 025, and between area 025 and surrounding domestic return areas (See Appendix 1).

As tagging studies can be biased by spatial variability in catch, an estimate of total catch during the recovery period was required for each area investigated. The effective blue cod habitat area (seabed shallower than 200 m) was then used to scale CPUE data and produced an estimate of relative stock size of each area (ratBA) that was then used to calculate mixing rates (See Appendix 1).

CPUEs for domestic return areas were calculated from catch data extracted from the Ministry of Fisheries Catch-Effort-Landing-Returns data base, and mixing rate calculations used all tags returned between 1 April 1998 and 30 November 1999 (n=674). CPUE within domestic return area 025 was calculated from catch data collected by 15 volunteer commercial fishing diarists (See Appendix 5). Diarists collected data from 14 sub-areas in domestic return area 025 (See Appendix 5), nine of which were also tagging strata (Figure 2). Mixing rates within area 025 were then calculated using only tags returned by those diarists from 1 April 1998 - 30 November 1999 (n=461). Diarists were recruited from the top 15 tag returnees in February 1998.

Mixing rates within area 025 were estimated relative to each of the nine tagging strata (Figure 2). The nine tagging strata allowed nine pairwise comparisons, namely the stratum of interest and all the remaining area 025 combined. Two independent estimates of mixing where therefore derived for each pairwise comparison based on:

1. movement of fish tagged outside stratum; 2. movement of fish tagged inside stratum.

The null hypothesis (a closed population consisting of two areas) requires mixing rates between the two areas to be small and reciprocal. If the mixing rates between the two areas are found to be dissimilar we must reject the null hypothesis. Mixing rates between area 025 and surrounding domestic return areas were derived from all recaptures of fish tagged within area 025. However, it was not possible to estimate reverse-movement mixing as no fish were tagged outside area 025.

10. Results

Tag and release

From 23 January to 14 March 1998, 9368 blue cod were tagged at 139 release stations throughout nine sites in domestic fishing return area 025 (Table 1, Figure 2). Tagging took 15 fishing days for each of the two vessels, and was achieved over three distinct periods of fine weather. Mean bottom depth of release stations showed a slight trend of shallower sites in the NZ strata and an increase in depth towards the western strata (Table 1). Release station depths ranged from 13 to 54.3 m, the mean being 33.5 (SE = 0.07) m.

Most blue cod were tagged on biogenic reefs and although the habitat variable was not balanced in the design, it was incorporated opportunistically (Table 1).

The size distribution of blue cod varied between bottom habitats, with comparatively more small fish caught on sand flats and biogenic reefs than on rocky reef fringe (Figure 4). However, the mean size of blue cod was largest on sand flats (Figure 4). Three paired comparisons of release length composition between habitat types were possible. Each comparison produced a statistically significant Kolmogorov-Smirnov test result (P<0.001).

There were differences in the length distributions of blue cod across latitudinal and longitudinal strata. Fish were largest in the Foveaux Strait latitudinal stratum (Figure 5), and smallest in the Stewart Island stratum (Figure 7) (means 331 and 319 mm respectively). Relative to the longitudinal strata, mean length progressively increased from east to west (Figure 5-7).

Recoveries

Between 24 January 1998 and 30 November 1999, 7.9% (743) of the 9368 tagged blue cod were recaptured. As expected most recaptures came from the commercial fishery. Recaptures came from all sites and all bottom habitats present within sites (Table 1, Figure 3). However, the number of recaptures varied greatly between sites and strata (Table 1), with most coming from the New Zealand (n=413) and Stewart Island stratum (n=215). Recapture rates ranged from 28.4% from sand flats in site NZl, to 1.5% on sand flats in site FS3 (Table 1). Recapture rates also varied between bottom habitat types (See Table 1), with most coming from rocky reef fringe (11.3%), and least from biogenic reefs (5.6%).

Length information produced by fishers was deemed to be unreliable as many fish had apparently shrunk. Consequently, all analysis involving length were based on the total length at tagging. The size distributions of returned blue cod from each bottom habitat type, and each site, are given in Figures 4—7. Return lengths ranged from 220 to 485 mm total length, however, few were below 300 mm causing a significant difference between the size distributions of tagged and recaptured blue cod (Kolmogorov-Smirnov test, P<0.001). This was the result of commercial fishers being restricted to a pot mesh size of 48 mm in order to reduce the capture of fish below the minimum legal size (330 mm TL). When blue cod below 305 mm were removed from the analysis, the difference between tagged and recaptured fish was not significant (Kolmogorov-Smirnov test, P>0.05), indicating that fish above this size are vulnerable to capture. All blue cod above 305 mm are considered sexually mature (male and female blue cod from Foveaux Strait mature at 260 mm and 280 mm respectively - Carbines 1998c).

Distance moved

Most fishers were able to provide specific details of recapture location, usually including GPS co-ordinates accurate to within 100 m. This provided a high level of confidence in calculating distances travelled (Tables 2 & 3). However, the distribution of distances travelled was highly asymmetrical, the largest being 156.1 km, but 87.1% moving less than 10 km (Figure 8). Due to the presence of a few extreme distances moved in the recovery data, the median distance of 790 m provided a better measure of average movement than the mean of 5.06 km (SE = 0.47) (Table 2).

Most blue cod were recovered near where they were released, with 60.2% moving less than 1 km (Figure 8). Results of the 2-way randomisation test (Manly & Francis 1999) confirmed a highly significant effect of release stratum on distance travelled. Although this effect was clear for both latitudinal (P=0.012) and longitudinal 9 (P=0.013) strata, a significant interaction effect (P=0.001) showed that distance travelled from each latitudinal strata was also dependent upon its longitudinal strata location, and vice versa. Differences between median distances moved occurred among latitudinal strata, but were more obvious between longitudinal strata, with notably lower movements occurring in western sites (Figure 3 & Table 2).

The median distance travelled by recaptured blue cod tagged on rocky reef fringe was less (0.46 km) than those tagged on biogenic reef (0.62 km) or sand flats (0.69 km) (Table 3). A randomisation test on the effect of bottom habitat type and release site on distance moved was undertaken using data from sites NZl, NZ3, and SI3 so as to exclude areas with zero or few returns (See Table 3). The results again showed a highly significant release site effect (P=0.002). However, bottom habitat type was not significant either on its own or as an interaction (P=0.446).

Effects of size on distance travelled

Despite a large number of smaller fish from the eastern strata moving large distances, the relationship between fish size and distance travelled was poor (R2 = 0.060). Blue cod travelling more than 20 km covered 52% of the total recapture size range (280 to 420 mm), and those moving less than 1 km include the entire size range (Figure 9).

Catch and mixing rates between domestic return areas

Mixing rates were calculated over 19 months of the tagging programme (1 April 1998-30 November 1999), in which 644 tagged blue cod were recaptured in area 025, 27 in area 030, three in area 027, and none from area 026 (Figure 3). Over this period, 896.25 tonnes of blue cod was caught by commercial fishers in domestic return area 025, 557.51 tonnes from area 030, 289.94 tonnes from area 027, and 25.09 tonnes from area 026. CPUE (mean catch/pot lift) was greatest in area 027 (26.17 kg/lift), followed by areas 030 (21.01 kg/lift) and 026 (14.14 kg/lift), CPUE was lowest in area 025 (13.54 kg/lift). Areas of available blue cod habitat were calculated using the area of sea floor shallower than 200 m (Anderson 1998) shown in Figure 1.

Mixing rate calculations (See Appendix 1) showed that between 1 April 1998- 30 November 1999, blue cod in domestic return area 025 mixed moderately with those in area 030 (S025-030 = 14.70%), and less so with area 027 (S025-030 = 7.26%). There was no mixing calculated with area 026 as no tagged blue cod were recovered there (Figure 3).

Catch and mixing within domestic return area 025

Commercial diarists recorded 22,702 pot lifts over 903 fishing days, catching a total of 224.55 tonnes of blue cod in domestic return area 025. This accounts for 71.6% of tag returns and 25% of the total commercial catch in this area from 1 April 1998- 30 November 1999.

Within 14 sub-areas of domestic return area 025, the blue cod catch (tonnes), habitat area, and CPUE (mean catch/pot lift) of diarists is shown in Figure 10A and Appendices 2-4. The area off Halfmoon Bay and Paterson Inlet (Stewart Island) yielded the largest catch (42.20%), followed by Ruapuke Island (12.71%) to the north- west (Figure 10A). Little or no catch was taken from the eastern end of domestic return area 025. CPUE was highest along the South Island coast and lowest along the Stewart Island coast (Figure 10A).

Blue cod showed significantly higher rates of mixing between sub-areas within domestic return area 025 (Figure 10B-J). As blue cod were tagged in separate areas of area 025, it was possible to derive estimates of both immigration and emigration between each tagging strata and the rest of area 025 (Figure 10). All mixing rates are expressed relative to the size of the small sub-area population at the time of tagging. Mixing rates were as high as 44.1% emigration from FS1 and 39.0% immigration into NZ3 (Figure 10). However, strata such as FS2 and NZl remained relatively isolated with no mixing rates more than 2.6% (Figure 10).

While immigration and emigration were similar in stratum FS3, in most strata there were large differences with over a hundred times more immigrations than emigration in NZ3 (Figure 10G). Thus, the null hypothesis that sub-areas of area 025 are at equilibrium is rejected.

As a rule, when grouped by latitude, eastern strata (NZl, FS1, SIl) had the least immigration (mean 1.2% ± 0.8%) and the most emigration (mean 20.8% ± 2.7%). Where as western strata (NZ3, FS3, SI3) had the most immigration (mean 16.0% ± 2.5%) and the least emigration (mean 7.0% ± 1.5%) (Figure 10). When grouped by longitude, immigration was highest among the northern strata (NZ1-3, mean 15.9% ± 2.6%), but least in mid Foveaux Strait strata (FV1-3, mean 1.8% ± 1.0%). However, emigration increased with longitude and was consistently higher along the southern strata (SI1-3, mean 21.4% ± 1.5%)

Direction of movement

When movement data is standardised to a common point of release, a clear predominance of a north-east pattern of movement is evident (Figure 11). Fishing effort was low in the eastern stratum, which may account for the lack of eastern movement (Figure 10a). However, fishing effort in central and western strata was comparable (Figure 10a), and low fishing effort in the east does not fully explain the overall westward trend in the data (Figure 11).

Seasonal effects of catch and movement

Tags were returned every month following release in summer 1998 (Figure 12). While the blue cod fishery is typically most active from March through July (Warren et al. 1997), during the recovery period it peaked later in 1998 and earlier in 1999 (Figure 12). The numbers of tags recovered roughly followed catch, although the largest individual number of recaptures came from January 1999 (Figure 12).

Seasonal trends in patterns of movement in the tagging data were evident when the data was categorised into four distance movement classes (0-2 km, 2-5 km, 5-20 km, 20+ km) (Figure 13). In the spring of both 1998 and 1999 there were a significant increases in the percentage of returns moving more than 20 km (%2(i998) = 40.61 (2),

2 2 p=0.0001; x (i999) = 27.31 (3), p<0.0001), more than 5 km (% (i998) = 34.79 (2), 2 2 p<0.0001; X(i999) = 21.34 (3), p=0.0001), and more than 2 km (%(i998) = 26.34 (2),

2 p<0.0001; % (i999) = 27.09 (3),p<0.0001).

These spring peaks, and the subsequent reduction in number of fish moving in summer, autumn and winter imply that a return seasonal migration may have occurred. In addition to the above observations, the trend of north-west movement is more pronounced over the spring periods (Figure 14).

11. Conclusions

Rapson (1956) found from 5050 blue cod tagged in the Marlborough Sounds, that 9.3% of returns had moved more than 1.6 km, and only 3.1% moved more than 16 km. Twenty seven years after Rapson's study, Mace and Johnson (1983) found considerably more movement from 2430 blue cod also tagged in the Marlborough Sounds (28.0% > 1.9km, and 4.8% > 16 km). Domestic return area 025 is almost 800 km south-west of the Marlborough Sounds and 9368 blue cod were tagged there during the current programme (Figure 1). Of the total number of blue cod recovered during the current programme, 29.1% had moved more than 1.6 km. This result is consistent with Mace and Johnson's (1983) findings. However, Mace and Johnson (1983) reported that only 4.8% of their recaptured animals had moved more than 16 km, whereas, in the current study 11.0% had mover further than this distance. These results suggest that densities of blue cod in Southland and the Marlborough Sounds have comparable temporal stability. However, the mobile portion of the area 025 population moved greater distances than the mobile portion of the Marlborough Sounds population.

Both Rapson (1956) and Mace and Johnson (1983) suggested that movements of blue cod were size dependent. However, Rapson (1956) considered larger fish were transient, while Mace and Johnson (1983) concluded it was the smaller fish that moved. Due to commercial mesh size restrictions, little can be concluded about the movements of small fish (>305 mm) in the current study, however, no size - distance relationship was evident in domestic return area 025 (Figure 9).

The blue cod habitat in area 025 could be sub divided into three distinct bottom types: sand flats, biogenic reefs, and rocky reef fringe. The size distribution of blue cod caught during the tagging phase of the programme varied significantly in relation to these broad habitat types. Larger numbers of small fish were caught on sand flats and biogenic reefs, which suggests these habitats may be more important to recruitment than rocky reef fringe. This finding is consistent with Mutch's (1983) observations of blue cod in Northland.

CPUE data analysis indicated blue cod abundance also differs in relation to habitat type. However, there was no evidence in the tagging data that fish moved to a greater or lesser degree in relation to each habitat type.

The length composition of fish tagged during the release phase of the current programme varied across the nine strata. Fish were largest in the central Foveaux Strait stratum and smallest in the southern Stewart Island stratum. There was also a progressive increase in size from east to west. The large number of small fish observed at all central longitudinal stratum is evidence that this area may be important to recruitment. The CPUE results provide further evidence that blue cod abundance might differ between areas as well as habitats. It appears that more blue cod were present in the east for both central Foveaux Strait and the southern Stewart Island strata, but in the west in the northern New Zealand stratum. The release data suggests that geographical location can influence the distribution of blue cod in domestic return area 025.

Although the habitat that blue cod were release into had no apparent effect on distance moved, spatial location was found to be an important factor influencing distance travelled. However, a significant interaction between release strata shows that factors influencing distance travelled are more complex than simply latitudinal or longitudinal location.

The tag mixing data showed that blue cod in area 025 had only little or moderate interaction with neighbouring domestic return areas. The largest rate of mixing occurred with the western domestic area 030 (S025-030 = 14.70%), and no mixing could be shown with eastward area 026 as no tagged blue cod were recovered there.

The tagging results suggest diffusion (mixing) of blue cod within BC05 is likely to be slow. Therefore, if fishing pressure is disproportionate to stock abundance there is a potential to locally deplete areas at the domestic return area scale. Movement of blue cod within domestic return areas, as indicated by the within area 025 mixing results, is complex. The movement patterns observed were possibly driven by broader scale stock movement processes. Certain sub-areas appeared to be acting as sinks collecting animals in from other areas of 025, while most sub-areas acted as drains losing fish to other areas of 025. As a rule, most drains for area 025 were to the east or along the coast of Stewart Island, with the only major sink was in the north-west corner of area 025 (Figure 10). However, as no small scale mixing rates could be calculated between 025 sub-areas and sub-areas of domestic return area 030, it is possible that the north­ west sink of area 025 drains into some adjacent area of 030.

Although the majority of tagged blue cod recovered during the current study had not moved greater than 1 km, there was a strong bias toward the north-west by those that did. This north-west trend could not be explained by spatial differences in fishing pressure. The within area 025 mixing rate calculations also show a higher level of mixing into the north-west.

An increased proportion of blue cod moving significant distances in the spring of both 1998 and 1999 further suggests a return seasonal migration had occurred. The trend of north-west movement was also enhanced over the spring periods. As Southland blue cod spawn between September and November (Carbines 1998b), it is probable that such a migration is associated with spawning.

In spring there also seems to be a congregation of recaptures near Centre Island and the northern tip of Stewart Island. These results were consistent with the 1995 pilot study, in which blue cod tagged off Halfmoon Bay (Stewart Island) were recapture west near Centre Island and Te Waewae Bay (Carbines 1998b). The concept of southern blue cod undertaking seasonal spawning migrations is not new (Graham 1939a, Rapson 1956, Robertson 1973, Warren et al. 1997), however, it has been dismissed for populations in the Marlborough Sounds (Rapson 1956, Mace & Johnston 1983). As Southland and Otago waters are mainly open coasts, they are subject to the influence of the Tasman and Southland currents (Houtman 1966, Heath 1972, Heath 1981). Counter current seasonal movements of blue cod in Southland may play crucial roles in the larval recruitment process by countering the effects of displacement in the direction of the currents transporting pelagic eggs and larvae. However, to determine the extent and nature of this behaviour in blue cod some form of continuous tracking of fish would be required.

The management implications of a blue cod spawning migrations are not likely to be urgent at this time, as there is only limited fishing effort in spring. However, should this situation alter in future, additional research and possibly some form of area protection may be required.

Acknowledgements

All those who have returned tags are thanked. Thanks also go skippers Milton Roderique, Alan Whiting (FV Rex) and Sid Bull (FV Rosane), to technicians Derrick Parkinson and Jill Parkyn for assistance in the tagging phase; and to Barry Thomas and Andrew Willsman for helping with the data base. A special thanks to Malcolm Francis for guidance and inspiration, without whose help this project would not have happened.

References

Anderson, O. F., N.W. Bagley, R.J. Hurst, M.P. Francis, M.R. Clark, & P.J. McMillan 1998: Atlas of New Zealand fish and squid distributions from research bottom trawls. Annala, J. H., Sullivan, K.J., & O'Brien, C.J. (Comps.) 1999: Report from the Fishery Assessment Plenary, April 1999 stock assessments and yield estimates. Anon 1998: Prize for catching tagged cod. Invercargill. Beentjes, M. P. F., M.P. 1999: Movement of hapuku (Polyprion oxygeneios) determined from tagging studies. New Zealand Journal of Marine and Freshwater Research 33: 1-12. Carbines, G. 1998a: How far do blue cod move? Seafood NZ 6: 44-47. 7. Carbines, G. D. (1995). Blue cod (Parapercis colias), tagging for age validation in the Southland region. New Zealand Marine Sciences Society Annual Conference, Wellington. Carbines, G. D. 1998b: Blue cod age validation, tagging feasibility and sex-inversion. Carbines, G. D. 1999a: Blue cod tag winner 7: 36. Carbines, G. D. 1999b: Large hooks reduces catch-and-release mortality of blue cod Parapercis colias in the Marlborough Sounds of New Zealand. North American Journal of Fisheries Management 19: 992-998. 4. Clark, E., Pohle, M., & Rabin, J. 1991: Spotted sandperch dynamics. National geographic research & exploration 7: 138-155. 2. Francis, M. P. 1988: Movement patterns of rig (Mustelus lenticulatus) tagged in southern New Zealand. New Zealand Journal of Marine and Freshwater Research. 22: 259-272. Graham, D. H. 1939a: Breeding habits of the fishes of Otago Harbour and adjacent seas. Graham, D. H. 1939b: Fishes of Otago Harbour and adjacent seas with additions to previous records. Heath, R. A. 1972: The Southland current. New Zealand Journal of Marine and Freshwater Research 6: 497-533. 4. Heath, R. A. 1981: Oceanic fronts around southern New Zealand. Deep-sea research 28a: 547-560. 6. Houtman, J. 1966: A note on the hydrological regime in Foveaux Strait. New Zealand Journal of Science 9: 472—483. Jones, G. P. 1988: Ecology of rocky reef fish of north-eastern New Zealand: a review. New Zealand Journal of Marine and Freshwater Research 22: 445—462. Kobayashi, K., Suzuki, K., & Shiobara, Y. 1993a: Reproduction and hermaphroditism in Parapercis snyderi (Teleostei, Parapercidae) in Suruga Bay, Central Japan. J. Fac. Mar. Sci. Technol. 35: 149-168. Kobayashi, K., Suzuki, K., & Shiobara, Y. 1993b: Studies on the gonadal sex succession in Parapercis snyderi cross examined in captivity. Bull. Inst. Oceanic Res. & Develop. 14: 83-91. Mace, J. T. & Johnston, A. D. 1983: Tagging experiments on blue cod ((Parapercis colias) in the Marlborough Sounds, New Zealand. New Zealand Journal of Marine and Freshwater Research 17: 207-211. Manly, B. F. J. a. C. R. I. C. Francis. 1999: Analysis of variance by randomization when variances are unequal. Australian and New Zealand Journal of Statistics 41: 411-429.4. Mutch, P. G. 1983: Factors influencing the density and distribution of the blue cod (Paraercis colias) (Pisces: Mugiloididae). Master of Science thesis, University of Auckland, Auckland. 76 p. Ohhishi, N., Yanagisawa, y., & Kohda, M. 1997: Sneaking by harem masters of the sandperch, Parapercis snyderi. Environmental biology of fishes 50: 217-223. 50. Rapson, A. M. 1956: Biology of the blue cod (Parapercis colias forsteri) of New Zealand. Doctor of Science thesis, Victoria University, Wellington. 53 p. Robertson, D. A. 1973: Planktonic eggs and larvae of some New Zealand marine teleosts. thesis, 33-35. 341-350 p. Sano, M. 1990: Patterns of habitat and food utilisation in two coral-reef sandperches (Mugiloididae): competitive or non-competitive coexistence. J. Exp. Mar. Biol. 140: 209-223. Sano, M. 1993: Foraging activities and diets of males and females in a haremic sandperch (Pisces: Pinguipedidae. Marine ecology progress series 98: 55-59. 172. Stroud, B. G. J. 1982: The taxonomy and biology of fishes of the genus Parapercis (Teleostei: Mugiloididae) in the Barrier Reef waters. Doctor of Philosophy thesis, James Cook University, North Queensland. 428 p. Warren, E., Grindley, R., Carbines, G., & Teirney, L. 1997: Characterisation of the Southland blue cod fishery. (1991-1996). 12. Publications

Carbines, G.D. (1998). How far to blue cod move? Paper presentation to the New Zealand Marine Sciences Society Annual Conference. July 1998, Christchurch. Carbines, G.D. (1998). How far do blue cod move. Seafood New Zealand. August 1998. Vol. 6, No. 7: 44-47. Carbines, G.D, (1998). Blue cod age validation, tagging feasibility and sex-inversion. Research Report for Ministry of Fisheries Project SOBC04. 77 pp. Carbines, G.D. (1999). Blue cod tag winner. Seafood New Zealand. December 1999. Vol. 7: 36 p. Carbines, G.D. (2000). Blue cod - the movement. Oral presentation to the New Zealand Marine Sciences Society Annual Conference. August 2000, Hamilton.

13. Data Storage

The data collected in this project is stored on the MFish recreational database housed at Greta Point. Table 1: Number and length (cm) of blue cod tagged within each site by bottom habitat type. Also shown are the numbers of stations (Sin.), CPUE (fish per station), mean size of tagged fish, depth, recaptures, and percent returned. (SE, standard error)

Site and No. No. CPUE Mean SE Mean SE No. % habitat tagged Stn. length (mm) depth ( m) recapt. Recapt.

NZl (East) 1093 16 68.3 312 1.09 35.3 0.19 198 18.1 Sand flat 102 4 25.5 315 3.62 40.7 0.00 29 28.4 Biogenic reef 102 5 20.4 311 3.39 41.5 0.00 22 21.6 Rocky reef 889 7 127 311 1.21 34.0 0.21 147 16.5

NZ2 (Central) 1030 36 28.6 324 1.67 29.8 0.18 48 4.7 Sand flat 276 14 19.7 327 3.09 26.2 0.39 8 2.9 Biogenic reef 706 15 47.1 324 1.94 32.4 0.06 32 4.5 Rocky reef 48 7 6.9 249 11.79 13.0 0.00 8 21.1

NZ3 (West) 1045 11 95.0 338 1.13 27.9 0.11 167 16.0 Sand flat 119 4 29.75 350 2.86 29.6 0.00 19 16.0 Biogenic reef 262 1 262 339 2.34 32.9 0.00 50 19.1 Rocky reef 664 6 110.7 335 1.41 25.7 0.10 98 14.6

FS1 (East) 993 6 165.5 323 1.52 37.0 0.28 43 4.3 Sand flat 0 Biogenic reef 263 2 131.5 314 3.01 . 51.2 0.00 11 4.2 Rocky reef 730 4 182.5 327 1.74 31.9 0.09 32 4.4

FS2 (Central) 1047 20 52.4 339 1.85 34.0 0.04 56 5.3 Sand flat 514 13 39.5 332 2.87 33.9 0.04 22 4.3 Biogenic reef 533 7 76.1 345 2.39 34.0 0.07 34 6.4 Rocky reef 0

FS3 (West) 1034 13 79.53 331 1.41 34.8 0.11 16 1.5 Sand flat 199 4 49.8 362 2.72 33.4 0.11 3 1.5 Biogenic reef 835 9 92.8 324 1.51 35.1 0.14 13 1.6 Rocky reef 0

SIl (East) 1198 12 99.8 309 1.31 36.6 0.31 75 6.3 Sand flat 0 Biogenic reef 641 8 80.1 305 1.95 28.6 0.26 22 3.4 Rocky reef 557 4 139.3 313 1.69 45.7 0.26 53 9.5

SI2 (Central) 1024 12 85.3 313 1.87 31.6 0.14 37 3.6 Sand flat 404 5 80.8 321 3.16 35.0 0.14 9 2.2 Biogenic reef 620 7 88.6 308 2.28 29.4 0.15 28 4.5 Rocky reef 0

SI3 (West) 904 13 69.5 336 1.51 34.2 0.21 103 11.4 Sand flat 350 4 87.5' 321 2.05 42.1 0.00 42 12.0 Biogenic reef 74 3 24.7 347 4.82 30.0 0.00 12 16.2 Rocky reef 480 6 80.0 346 2.16 29.1 0.08 42 8.6

Overall Total 9368 139 67.4 325 0.51 33.5 0.07 743 7.9 Sand flat 1964 48 40.9 330 1.24 34.5 0.13 139 7.1 Biogenic reef 4036 57 70.8 322 0.80 33.6 0.10 224 5.6 Rocky reef 3368 34 99.1 324 0.76 32.8 0.14 380 11.3 Table 2: Descriptive statistics of recaptured blue cod tagged at mainland New Zealand (NZ 1-3), mid Foveaux Strait (FS 1-3) and Stewart Island (SI 1-3) sites (SE, standard error; N, sample size)

Variable Mean Median SE N Min Max

NZl (East) Tagged length (mm) 323.13 320 1.95 198 225 , 440 Release depth (m) 36.29 39.4 0.42 198 25.6 41.5 Period at liberty (days) 300.22 313 7.52 198 35 564 Distance travelled (km) 3.43 0.66 0.66 196 0.08 75.36 NZ2 (central) Tagged length (mm) 340.10 340 5.56 48 220 460 Release depth (m) 27.55 31.1 1.13 48 13 35 Period at liberty (days) 272 261.5 21.85 48 23 623 Distance travelled (km) 12.68 1.87 3.54 47 0.34 156.09 NZ3 (West) Tagged length (mm) 344.88 345 2.50 166 225 440 Release depth (m) 28.35 27.7 0.28 167 23.8 32.9 Period at liberty (days) 293.44 340 13.03 167 86 612 Distance travelled (km) 1.04 0.45 0.37 167 0.03 58.63 FS1 (East) Tagged length (mm) 343.26 335 5.95 43 280 485 Release depth (m) 36.99 34.7 1.33 43 29.2 51.2 Period at liberty (days) 263.09 332 21.02 43 3 577 Distance travelled (km) 15.15 1.57 3.17 43 0.30 66.68 FS2 (Central) Tagged length (mm) 361.07 365 4.72 56 225 440 Release depth (m) 34.02 34.6 0.16 56 31.1 35.2 Period at liberty (days) 231.13 209 22.36 56 12 602 Distance travelled (km) 2.90 1.19 0.66 56 0.08 26.90 FS3(West) Tagged length (mm) 346.88 342.5 5.61 16 320 380 Release depth (m) 32.71 31.1 0.62 16 31.1 38.4 Period at liberty (days) 252.44 303 34.63 16 2 464 Distance travelled (km) 1.79 1.23 0.35 16 0.29 3.79 SIl (East) Tagged length (mm) 325.87 325 3.30 75 275 390 Release depth (m) 39.02 32.9 1.17 75 18 54.3 Period at liberty (days) 265.21 314 17.88 75 5 600 Distance travelled (km) 5.84 1.30 1.88 75 0.30 109.23 SI2 (Central) Tagged length (mm) 352.43 '' 350 5.27 37 295 400 Release depth (m) 29.86 31.8 0.69 37 25.6 36.6 Period at liberty (days) 256.30 237 24.75 37 25 573 Distance travelled (km) 10.68 1.76 2.80 37 0.40 83.05 SI3 (West) Tagged length (mm) 342.28 335 2.98 103 285 470 Release depth (m) 35.45 31.1 0.64 103 26.6 42.1 Period at liberty (days) 195.67 96 18.43 103 3 630 Distance travelled (km) 6.46 1.00 1.03 99 0.04 43.91 Overall Total Tagged length (mm) 338.12 335 1.22 741 220 485 Release depth (m) 33.57 32.9 0.27 743 13 54.3 Period at liberty (days) 268.27 310 5.67 743 2 630 Distance travelled (km) 5.06 0.79 0.47 736 0.03 156.09 Table 3: Descriptive statistics of recaptured blue cod tagged and released on sand flats, biogenic reef and rocky reef habitat of all sites in domestic return area 025 (SE, standard error; N, sample size)

Variable Mean Median SE N Min Max

Sand flats Tagged length (mm) 339.35 335 2.79 139 225 440 Release depth (m) 36.88 40.7 0.51 139 20 42.1 Period at liberty (days) 247.87 283 14.21 139 2 574 Distance travelled (km) 5.36 0.90 0.96 137 0.13 75.36 Biogenic reef Tagged length (mm) 343.94 340 2.18 222 220 430 Release depth (m) 33.64 32.9 0.40 224 18 51.2 Period at liberty (days) 256.70 261 10.67 224 5 623 Distance travelled (km) 6.01 1.05 0.89 223 0.08 109.21 Rocky reef Tagged length (mm) 334.26 330 1.72 380 225 485 Release depth (m) 32.32 31.1 0.41 380 13 54.3 Period at. liberty (days) 282.56 313 7.45 380 3 630 Distance travelled (km) 4.41 0.66 0.67 376 0.03 156.09 Overall Total Tagged length (mm) 338.12 335 1.22 741 220 485 Release depth (m) 33.57 32.9 0.27 743 13 54.3 Period at liberty (days) 268.27 310 5.67 743 2 630 Distance travelled (km) 5.06 0.79 0.47 736 0.03 156.09 Appendix 1: Tag mixing models used to estimate mixing rates in the Blue cod The net flow of animals from one area (A) to another area (B) per unit time is termed the mixing rate. Giving mixing only occurs between two stock areas (A and B), then animals tagged in area A area will eventually be distributed between A and B in proportion to relative ratio of the populations, i.e. they will have attained equilibrium distribution. For example, if the populations in A and B are the same size the equilibrium distribution of tagged animals between A and B will be 1:1. If population A is twice as large as population B the equilibrium distribution of tagged animals will be 2:1.

The total time taken for a group of marked animals, released in area A, to achieve equilibrium distribution across areas A and B is a function of the mixing rate and the relative size of the populations in A and B. The length of time needed for two stock areas to achieve equilibrium mixing has implications for stock management. For example, if the equilibrium time is long, the populations in the two areas are relatively independent.

Given the following assumptions we can use tagging to estimate the population mixing rates.

1. Tag mortality/loss constant over all areas

2. Under-reporting constant over all areas

3. All movement over the period of interest occurs before the second capture sample is taken («2)

4. Tagged and untagged animals have equal catchability

5. The two populations are at equilibrium in respect to movement, i.e. an equal number of animals move from A->B as from B->A. Implicit also, is that the rate of interchange between the two stocks is constant from year to year.

6. We also assume the two populations are at equilibrium in respect to recruitment and total mortality. If this is not true then we need to assume that the rates at which the two stocks are changing- (increasing/ decreasing) are the same.

The size of a tagged population can be approximated by the Petersen estimator given as:

N = Equation 1

m2

Where: N = estimate of population size ri] = number of tagged animals in first sample ri2 = number of animals examined for tags in second sample

m2 = number of tagged animals observed in second sample Given all animals were originally tagged in area A, at the end of 19 months the proportion of animals from area A moving to area B (mixAB) is given by: n, „ IB mixAB = Equation 2 n +n iB u

We can use Equation 1 to derive an estimate of mixAB as follows:

_ NAm2A n \A n 1A Equations 3

_ NBm2B nlB ~ n 2B

While NA or NB are not known explicitly, they can be substituted in Equation 3 by an independent measure of the relative size of the two populations (ratBA):

NB CPUER * AreaB

= ratBA = ^ B NA CPUEA * AreaA Equation 4

Dividing the equations (3) for nlA and nlB by NA gives:

*n ~mu n\A ~ n2A Equations 4

ratBAm2B

L2B

Substituting Equations 4 into Equation 2 gives

n n ratBAm2B 2A 2B

l2B n2Bm2A +n2AratBAm2B

n2AratBAm2B

n2Bm2A+n2AratBAm2B Appendix 2: Area and diary derived CPUE data for mixing calculation

Tagging Strata Diary Area Inside area size Outside area Inside area CPUE Outside area CPUE* Inside/outside area (See Figure 2) (See Appendix 2) size (km2) Size (km2) (kg/pot lift) (kg/pot lift) population ratio

1 330.70 3228.10 34.13 9.72 2.78 SI3 2 195.60 3363.20 8.83 11.51 22.41 FS3 3 226.80 3332.00 11.15 11.37 14.98 NZ3 4 310.20 3248.60 16.18 11.07 7.17 SI2 5 131.80 3427.00 9.26 11.62 32.63 FV2 6 230.40 3328.40 9.05 11.44 18.26 NZ2 7 348.50 3210.30 11.78 11.35 8.87 SIl 9 426.40 3132.40 8.41 13.51 11.80 FS1 10 603.30 2955.50 10.38 11.50 5.43 11 163.90 3394.90 12.90 11.34 18.21 NZl 12 444.70 3114.10 11.96 11.33 6.63 13 146.50 3412.30 6.00 11.38 44.19 * weighted mean of all other areas Appendix 3: Estimate of the percent of movement from each tagging stratum into the remainder of domestic return area 025 from April 1998 until November 1999

Tagging Strata Diary Area No. tagged No. stratum tags No. stratum tags Inside stratum catch Outside stratum Outside stratum/inside Percent of inside (See Figure 2) (See Appendix 2) in stratum recovered in stratum recovered out of stratum (tonnes) catch (tonnes) stratum pop" ratio stratum pop" emigrating

1 SI3 2 904 46 7 12.73 211.82 22.41 17.01% FS3 3 1034 9 1 8.49 216.06 14.97 6.14% NZ3 4 1045 114 1 12.65 211.9 7.165 0.37% SI2 5 1024 20 2 24.92 199.63 32.63 28.94% FV2 6 1047 24 1 7.6 216.95 18.26 2.60% NZ2 7 1030 17 5 5.35 219.2 8.874 5.99% SIl 9 1198 39 1 94.76 129.79 11.80 18.10% FS1 10 993 6 6 28.53 196.02 5.428 44.14% 11 NZl 12 1093 158 1 10.48 214.07 6.633 0.21% 13 Appendix 4: Estimate of the percentage movement from outside stratum into each stratum

Outside Inside Ratio of Initial no. No. outside stratum No. outside stratum stratum stratum Inside stratum Percent of outside Percent of stratum predicted Tagging strata Diary area tagged outside tags recovered tags recovered catch catch /outside stratum stratum pop" mix expected if observed (see Fig. 2) (see App. 2 of stratum out of stratum in stratum (tonnes) (tonnes) pop" ratio emigrating pop" closed movement 1 SB 2 8464 457 1 211.82 12.73 0.044 0.16% 3.64% 0.21 FS3 3 8334 457 1 216.06 8.49 0.066 0.37% 5.55% 0.90 NZ3 4 8323 447 11 211.9 12.65 0.139 5.44% 38.98% 104. SI2 5 8344 458 0 199.63 24.92 0.030 0.00% 0.00% 0.00 FV2 6 8321 458 0 216.95 7.6 0.054 0.00% 0.00% 0.00 NZ2 7 8338 457 1 219.2 5.35 0.112 1.00% 8.88% 1.48 SIl 9 8170 446 12 129.79 94.76 0.084 0.31% 3.67% 0.20 FS1 10 8375 458 0 196.02 28.53 0.184 0.00% 0.00% 0.00 11 NZl 12 8275 458 0 214.07 10.48 0.150 0.00% 0.00% 0.00 13 Appendix 5. Fishing diary used by volunteer fishers. SUMMER (1 JAN - 31 MARCH)

Skippers Name: Occti Beach (north enlrance) Please write in each of the 15 areas, the estimated number of days fished in that area from 1 January to Vessel Name: 31 March.

Vessel Number:

Seasons catch:

Days Fished:

Average number of pot lifts/day:

Bails used: Marlborough Sounds

N

S

(784 sq km)

032 (BC05)

(1215 sq km)

031 (BC05)

(7066 sq km) [) (5628 sq km) 030 (BC05) 026,(8003)

(5303 sq km) (12050 sq km) BC05 029 (BC05) 027 (BCj05)

/ r (9162 sq km) 100 200 028 JBC05),/ kilometers

Figure l. South Island blue cod quota management areas BC03, BC05 and BC07. The seven domestic fishing return areas of BC05, and area 026 (BC03), are shown with blue cod habitat areas shallower than 200 m shaded. Figure 2. The nine tagging sites used in Foveaux Strait (domestic return area 025). Tagging sites are stratified within three longitudinal strata and release locations are identified by symbols within each tagging site (mainland New Zealand (A NZ 1-3), mid Foveaux Straight (• FS 1-3), Stewart Island (• SI 1-3)). The three latitudinal strata release locations are identified by colour (West (blue), Central (green), and East (red)). Neighbouring domestic return areas are labelled. A NZ1 release NZ1 recapture A NZ2 release NZ2 recapture A NZ3 release NZ3 recapture g

FS1 release • FS1 recapture FS2 release • FS2 recapture FS3 release FS3 recapture 026 (25.09t) • SI1 release O 511 recapture 025 • SI2 release O 512 recapture (896.25t) © SI3 release 513 recapture

027 (289.94t)

Figure 3. The nine tagging sites and recapture locations of blue cod in Foveaux Strait (domestic return area 025). Sites are stratified within three longitudinal strata identified by symbols (mainland New Zealand (A NZ 1-3), mid Foveaux Straight (• FS 1-3), Stewart Island (• SI 1-3)) and three latitudinal strata are identified by colour (West (blue), Central (green), and East (red)). Recapture locations are identified as open symbols in the shape and colour combination of their release site. Domestic return areas are labelled and the 400 -I All sites : Sand flats 350 • N = 1964, Mean 330 ± 12 mm 300 - n= 139, mean 339 ±28 mm

W 250 • <5 .Q 200 - E z 150 -

100 -

50 - 0 - ^-rriTfTfTfT TTTT-n 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 Length (cm)

400 -i All sites : Biogenic reef

350 • N = 4036, Mean 322 ± 8 mm n = 223, mean 344 ± 22 mm 300 •

at 250 - a> xi 200 - E Z 150 -

100 -

50 - 0 - r-r-r-rT 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 Length (cm)

400 -i All sites : Rocky reef

350 - N = 3368, Mean 324 ± 7 mm 300 - n = 380, mean 335 ± 17 mm

(0 250 •

_

100 -

50 • 0 • • • • • . i^-tTtT 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 •E44 ttt46-t- 48i 50 52 Length (cm)

Figure 4: Length frequencies of blue cod from different bottom types throughout all the sites. N = the number of blue cod tagged (white) and n = the number recaptured (black). Mean ± standard error for each are shown. 140 -

120 - N = 993, Mean 322 ± 15 mm n = 43, mean 343 ± 60 mm 100 •

aj 80 - £> E 60 - Z 40 -

20 - 0 - "TTTTT-n n i i 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 Length (cm)

140 -J Total site : FS2 (Central) N = 1046, Mean 339 ± 19 mm 120 • n = 56, mean 361 ± 47 mm 100 • w a> 80 - -Q E 60 - Z 40 -

20 - 0 - 1111111 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 Length (cm)

140 -i Total site : FS3 (West)

120 • N = 1034, Mean 331 ± 14 mm n= 16, mean 347 ±56 mm 100 • f/t <5 80 - E 3 60 • Z 40 -

20 •

0 - r-rnT -i—i i i h-rTh^n I I I I 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 Length (cm)

Figure 5: Length frequencies of blue cod from sites in latitudinal stratum Foveaux Strait. N = the number of blue cod tagged (white) and n = the number recaptured (black). Mean ± standard error for each are shown. Longitudinal strata are shown in brackets. N = 1093, Mean 312 ± 11 mm n= 198, mean 323 ±19 mm

Length (cm)

180 - Total site : NZ2 (Central) 160 - 140 - N = 1030, Mean 323 ± 17 mm 120 • n = 48, mean 340 ± 56 mm 100 - E 3 80 • Z 60 • 40 - 20 • 0 • -r-r • i r-rrT ~b~r-r-T-r i—i—i—i—i—i 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 Length (cm)

180 - Total site : NZ3 (West) 160 - 140 - N = 1045, Mean 338 ±11 mm n= 167, mean 346 ±25 mm 120 • w _o 100 • E 80 • Z 60 • 40 - 20 -

I I I I I i—r 0 • i i i i i i i few W 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 Length (cm)

Figure 6: Length frequencies of blue cod from sites in latitudinal stratum mainland New Zealand. N = the number of blue cod tagged (white) and n = the number recaptured (black). Mean ± standard error for each are shown. Longitudinal strata are shown in brackets. 140 -

120 - N = 1198, Mean 309 ± 13 mm 100 • n = 75, mean 327 ± 33 mm (ft a 80 - xEi 3 60 - Z 40 -

20 •

0 - -• rfftf 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 Length (cm)

140 - Total site : SI2 (Central)

120 • N = 1024, Mean 313 ± 19 mm n = 37, mean 352 ± 53 mm 100 -

VI a 80 - XI E- 60 • z 40 -

20 •

0 - , , r-rrlT ITh-T. 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 Length (cm)

140 i Total site : SI3 (West)

120 - N = 904, Mean 336 ± 15 mm 100 • n = 103, mean 342 ± 29 mm vt a> 80 • XI E 3 60 - z 40 •

20 -

0 •• , rrrTTT l~m-i 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 1150 152 Length (cm)

Figure 7: Length frequencies of blue cod from sites in latitudinal stratum Stewart Island N = the number of blue cod tagged (white) and n = the number recaptured (black). Mean ± standard error for each are shown. Longitudinal strata are shown in brackets. 140

120

• FS1 (n=43) O FS2 (n=56) S 100 - • FS3 (n=16) A NZ1 (n=196) o £ 80 ANZ2(n=47) ANZ3(n=166) • A OS11 (n=75) b 60 •O O SI2 (n=37) °0 O O SI3 (n=99)

A 40 OA _ 2 o • ° a A • 20 • AA °o^loai' °° -Qffl^laftaiiftililiaiiagBaBisaas ^^w -AQ- 200 250 300 350 400 450 500 Total length (mm)

Figure 9: Relationship between blue cod length and distance moved by release sites in Foveaux Strait (domestic return area 025). Longitudinal strata identified by symbols (mainland New Zealand (A NZ 1-3), mid Foveaux Strait (• FS 1-3), Stewart Island (# SI 1-3)) and latitudinal strata identified by colour (West (blue), Central (green), and East (red)). Figure 10. A) Diarist's catch (tons), <200 m deep area (km2), and mean CPUE (kg/pot lift) in 15 sub areas of domestic return area 025. B-J show mixing rates (See Appendix 1) of release strata (hatched) with the rest of area 025. A ratio of net gain is marked in each stratum and all percentages are in terms of each stratum's relative population size. North (km)

• FS1 • FS2 • FS3 60 A NZ1 A NZ2 A NZ3

40 OSI1 O °o OSI2 ° °_r> OSI3 B ° °2° cPo© o West (km) East (km)

-150 -100 -50 50 100

O A •

A°

-40

-60 South (km)

Figure 11: Plot showing the direction and magnitude of travel standardised from the point of origin for all tagging sites. 50

40

• >2km ^ 30 » - ->5km o -* >20km

10 !

Autumn 98 Winter 98 Spring 98 (61) Summer 99 Autumn 99 Winter 99 (75) Spring 99 (25) (82) (144) (224) (100) Season

Figure 13: The percentage of blue cod moving > 2km, > 5km and > 20km in each of the seven seasons of the tagging programme.

80

60

40

O o • FS1 (n=6) o • FS2 (n=14) ft % • FS3 (n=9) O20 O ANZ1 (n=7) O ANZ2 (n=11) • A o A NZ3 (n=22) OSI1 (n=5) -150 -100 -50 A 50 100 A3 O SI2 (n=3) O SI3 (n=9) -20 A

-40

-60

Figure 14: Plot showing the direction and magnitude of travel standardised from the point of origin for all tagging sites for spring 1998 and 1999 only. Figure 15. The nine tagging sites and recapture locations of blue cod in Foveaux Strait (domestic return area 025) for spring 1998 and 1999 only. Sites are stratified within three longitudinal strata identified by symbols (mainland New Zealand (• NZ 1-3), mid Foveaux Straight (• FS 1-3), Stewart Island (• SI 1-3)) and three latitudinal strata identified by colour (West (blue), Central (green), and East (red)). Recapture locations are identified as open symbols in the shape and colour combination of their release site. Domestic return areas are labelled.