If'" __ ,, International Council tor C.M. 1988/G:5 the Exploration of the Sea Demersal Committee Ret: Mariculture Committee THE MIGRATION OF COASTAL COD (Gadus morhua L.) TAGGED IN A FJORD OF WESTERN NORWAY by J.T. Nordeide1 and A.G.V. Salvanes2 11nstitute of Marine Research. Box 1870, N-5024 Bergen-Nordnes, Norway 2Haugesund Maritime College, Skäregt. 103, N-5500 Haugesund, Norway d • j . ' ABSTRACT 'This stlJdy is part of a research program in~estigating th~ possibility of enhancing the production of c~astal cod via a large scale release of pond prodlJced fish iri ,.. ,. ,_., .. :.i,. • < ,." '., " I: I Masfjorden, a fjord of western NoiWay. An analysis of parts cif the ecosystem are being carricid out prior to the iarge' scale experiment with a special focus on the cod population of the' fjord. The paper deals with the migration pattern of wild cod which were ca~ght, tagged arid re~aptured and aiso, p6nd produced cod rele~sed in small scale tag-release-recaptlJre experiments. We appiied Spearman's' rank correlation test arid,' ". Koimogrov-Smirnov two-sampie test to assess whether larger individuals migrate longer " distances than smaller. coda . T~e major proportion of the immature wild and pond produced cod were neady stationary' < " I. ' .. ,', , arid only 15.0 alld 7.0 percent, respectively were recaptured more than five kilometers '. from the release site. However, a larger proportion, 54 percent, of mature cod were recaptured outside the reiease' sile;' of which the iargest individuals we're found 10 , , inigrate the longest distances, with some of the latter intermingling with cod from contiguous fjords. INTRODUCTION This study is apart of a research program investigaiing the possibility of enhancing the production of coast~1 cod by release of pond produced fish in Masfjorden. This , '. .' . study was made feasible, in 1983 when Norwegian scielltists succeed to rrias~ produce '. jlJveniie cod (0iestad et al. 1985)., Masfjorden, a fjord of western NOrWay, was . ~ . '. selected as the site tor a large scale tag-release-recapture experiment. A large . scale release program is planned beginning in the autuinn of 1988. Prior to this large scale reiease, parts of the ecosystem of the fjord are being.examined with research fO~USSing on ~igration patterns.. A central qlJestion for.. itie research program i~ wh'ether the released individuals wili remain in the area from which they ware . '.'' . ' .' ,I.. ' ., ' • ,,' released. Similar smaJl scale tag-release-recapture experiments of both wild, and pond .produced cod tiava previously been conducted in öther regions of western and souther-n Norway. The juveniles are reported to be nearly staiiciriary in the area whero they 3 were released (Dahl 1905, Lßversen 1946, Dannevig 1953, Danielssen 1969, Godß 1984, ,I I ' Moksness and 0iestad 1984, Sväsand and GOd0 1987). However, mature coastal cod'have been found to migrate 'Ionger'distances than the iuveniies (Hylen 1964, GOd0 1984, G6d0 et al. 1986). This paper: analyses the' migration pattern of cod fram experiments with a smail scale , . , ' release" program of tagged pond produced cod and wild cod, focusing on the question of . ..' '. ..,' . stationarity.' We also utilize Spearman's rank correlation' test and Kolmogrov-Smirnov two-sample test to determine whether larger individuals migrated' longer distances' thari' small cod. ," MATERIALS AND METHODS' . , , ) ; .,', .' Masfjorden is a typical fjord on the west coast of Norway, located north of Bergen at 600- 50'N, 50 25'E. Both Masfjorden and Austfjorden, a contiguous fjord, are connected to the outer coast and continental shelf through Fensfjorden (Fig. 1). A further , , ~ description of the fjords is given by Salvanes et al. (in prep). Tag-release~recapture experiments ware conducted för pond produced and wild cod. Pond produced cOd were released in Nove~ber, 1985 and December, 1986 (Fig.1). Wild cod were divided' into two groups: mature cod caught, tagged and rele~~'ed 8t the spawning • grounds during the spawning season (Win'ter Group), and cod caugtit~ tagged and released during the summer period (Summer Group)~ Fig.:3 ilhistrates the distri'b'ution of migration distance fram release tCl recapture for all three groups. The wild cod were • .' - '. r·' • captured by trap-net, trammel-net o'r gilt-net. Length-distribLitions at release for , '"1 . • .1 the groups are given in Fig. 2. Datei pertaining to the ta.g-release-recapture experiments are given in Tab. 1 and'Tab. 11. 1 i ; . ,- . {. , 4 l .. All cod were individually tagged with an external "Floy Anchor Tag". The experiments were announced in the local newspapers and by posters at local post-offices and l stores. These outlets also provided catch torms to record intormation of recaptured cod. .Areward of Nkr. 25 (Le., 4.00 $U.S) was paid tor every tag returned. Recaptures were also recorded during our experimental fistiery in Masfjorden. Below we will provide' more detailed information of the three different' groups of cod. ........ Pond Produced Group The cod in this group were produced in Heimarkspollen, Austevoll, as described by • 0iestad et aJ. (1985) based on' broodstock fram' the Austevoll region. ' These juveniles were tagged and released' aso o-g'roup in Masfjord~n and araund .the small islands . outside the sill which separates Masfjorden tram Fenstjorden. Sub-groups of cod were released along the shoreline in sub-areas which were of unequal size. Density ot , . released cod did not exceed 3 tenths per meter of shoreline. Untortunately, the 'exact position of release is not known tor each individual within the su~-area.· We therefore excluded data-sets tram sub-areas in which the ·Iength .ot the shoreline exceeded tive kilometers. We turthermore defined the migration distance as· the . distance· tram the point of recapture to' the nearest point where. cod. tram the • respective sub-group were released. Summer Group . This group contains wild cod 'caught, tagged and released mostly' in Masfjorden, tram May to Nov'ember 1987. In addition, a tew individuals were also released around the small islands outside Masfjorden. The exact position of capture before they were tagged and released is known for each individual of this group. Cod recaptured less . ., , '. ' .. than 0.4 kilometers from the' release site, were defined as'recaptured within the sub- . , 5 , , I i • i : II I I i I I, :I area of release. .. , ". " " I I I I I ,I Winter Group ! I I . .. The cod in this group, were caught by 106al fishermen during March, 1987 and February, .1988' in Masfjorden, outside, but elose to the sill of Masfjorden and in Austfjorden. Detailed information about position of capture. before tagging of individual fish was .' not provided. However, most were caught at or near the spawning localities in Masfjorden' and Austfjorden. Since the fishermen caught the cod over large areas, we defined the size of the sub-areas of release to 11 kilometers of shoreline. This '. group is considered mostly mature cod, since only recaptured cod larger than 45 cm arc included in our analysis and 50 percent were found to be, mature and above this size • , , (Nordeide et al. 1988). ," " ... , ~. ", I Statistical tests I ' I ! t~::~ Two different nonparametric tests were applied to determine whether larger cod migrated longer dista-nces than smaller cod1. First, the Kolmogorov-Smirnov two-sample ­ test' was applied to assess' if there was any difference in length-distribution between cod which were .recaptured within and outside the sub-areas of release. Secondly, for i those cod recaptured outside the sub-areas of release, Spearman's coefficient of rank i I I, correlation (rsl was used to test the correlation between migration distance and size • of the fish at recapture (j=A) and between migration distance and number of days from ,I release to recapture (j=8).' The tests were applied separatelyon the Pond Produced II Group (P),; Summer.Group (S) and Winter, Group (W). ; The theory is described in , 1Nonpa~amet~ic tests wem ap~lied because assumptions of' norm~lity tor our data sets were questionable. since the dominating part of the recaptures were reported by local fishermen which had fished with several gear types. 6 L Sokal and Rohlf (1981). First, we applied the Kolmogrov-Smirnov two-sample test to assess the following null hypotheses: HoA: The cod recaptured outside the release sub-area comes from the same length­ distribution as the cod recaptured within. H1A: The cod recaptured outside the release sub-area come trom different length­ distributions than the cod recaptured within. HoB: No difference exist in the number of days in the gord for the cod recaptured outside the release sub-area compared to those recaptured within. H1B: The cod recaptured outside the release sub-area spant significantly different number of days in the gord before recapture than those recaptured within. The test statistic (ON) is the maximum difference between the cumulative relative frequency distributions for cod recaptured within (f;; and outside (f;/) release sub-area, separately for fish length Ü=A); and days in the fjord before recapture Ü=8); for each of the groups P, Sand W. The test statistic is given by the expression: (SokaI and Rohlf 1981) • The null hypothesis is rejected if ON > critical value at a given significance level. The second test applied is the Spearman's rank correlation. It was applied on the part of the sampies which were recaptured outside the release sub-area for the groups 7 l I P, Sand W to test the possible association between migration distance and fish length (C) and between migration distance and days elapsed in the fjord before recapture (0). We tested the following hypotheses: HoC: No correlation exist between migration distance and length of the cod (rsC=O), against the altemative thata correlation exist (rsc,.or H : No correlation exist between migration distance and number of days ,elapsed oO in the fjord before recapture (rsO=O) against the altemative that a correlation exist (rs010).