Spatial variation in abundance, size composition and viable egg production of spawning cod (Gadus morhua L.) in Icelandic waters

ICES Journal of Marine Science, 57: 824 83. 2 doi:1.16/jmsc.2.568, available online at http://www.idealibrary.com on Spatial variation in abundance, size composition and viable egg production of spawning cod (Gadus morhua L.) in Icelandic waters Gudrun Marteinsdottir, Asta Gudmundsdottir, Vilhjalmur Thorsteinsson, and Gunnar Stefansson Marteinsdottir, G., Gudmundsdottir, A., Thorsteinsson, V., and Stefansson, G. 2. Spatial variation in abundance, size composition and viable egg production of spawning cod (Gadus morhua L.) in Icelandic waters. ICES Journal of Marine Science, 57: 824 83. Changes in geographical location of spawning and/or in spatial abundance of spawners may influence offspring survival. In Iceland, the spatial dynamics of spawning cod (Gadus morhua L.) are complicated due to unequal distribution of size classes on the main spawning grounds. Information on abundance, age, and size distributions during peak spawning, potential fecundity, egg size, and estimated larval viabilities were used to contrast the reproductive potential of three adjacent spawning areas within the main spawning ground of cod in the Icelandic waters. The three areas differed with respect to age and size distribution, abundance, mean production per female and an index of total egg production. Larger and faster growing cod spawned closer to the coast (area 1). Smaller cod spawned in deeper water out on Selvogsbanki (area 2), and the smallest and youngest cod tended to spawn along the continental edge (area 3). Relative abundance was higher in area 2 than in areas 1 and 3 (64, 153, and 334 kg/net, respectively). Mean production of eggs per female was highest in area 1, while the index of total egg production was highest in area 2. The estimates of viable egg production did not change the general results. Estimates of the proportion of viable eggs produced were slightly higher in area 1 than in areas 2 and 3. However, the influence of the relative abundance of spawners in the different areas overrode any trends due to higher mean production of viable eggs. The results demonstrate that egg production, based on abundance and size and age composition of spawning cod, may vary extensively among different spawning areas. Such information, in conjunction with documentation of spatial and temporal variation in oceanic and environmental conditions, may provide an important contribution towards the understanding of recruitment processes necessary for management of commercially important fish species. 2 International Council for the Exploration of the Sea Key words: cod, egg production, spatial variation, spawning, spawning sites. Received 22 March 1999; accepted 11 January 2. G. Marteinsdottir, A. Gudmundsdottir, V. Thorsteinsson and G. Stefansson: Marine Research Institute, P.O. Box 139, 121 Reykjavik, Iceland. Tel.: +354-55224; fax: +354-562379; e-mail: runa@hafro.is Introduction Cod are widely distributed in the North Atlantic and spawn in a variety of habitats. Spawning has been reported in both inshore and offshore areas, at depths ranging from a few to several hundred metres and at temperatures from below 2 3 C up to 8 1 C (Hildebrand and Schroeder, 1972; Hutchings et al., 1993; Liem and Scott, 1966; Thorsteinsson and Marteinsdottir, 1992). Throughout its distribution range, cod are separated into groups of spawners that are sometimes well-defined populations or subpopulations, and in other cases represent less defined stock divisions or spawning aggregations (Bergstad et al., 1987; Hutchings et al., 1993; Myers et al., 1997). In each area, selection of spawning locations is likely to reflect the conditions under which offspring survival is maximized (Hutchings and Myers, 1993; Cushing, 199). Information on the relative contribution of different spawning areas or stock components to the stock s propagation are often uncertain (Dalley and Anderson, 1997). As different size (age) classes of spawners may not 154 3139//4824+7 $3./ 2 International Council for the Exploration of the Sea

Egg production of spawning cod 825 IX VIII VII 66 N 65 I 1.2 1.1 1.3 VI 2.1 64 II 24 2.5 2.2 III Reykjavík 3.3 3.1 3.4 3.2 3.5 4.3 IV 4.2 4.1 22 2 18 16 14 W Figure 1. Areas surveyed for spawning cod during the annual gill net survey. The total survey area is divided into five regions (I V). Each region is divided into three to five sub-areas. 5.3 5.2 5.1 5.5 V 5.4 contribute equally to reproduction (Marteinsdottir and Thorarinsson, 1998), egg production of the stock is likely to change depending on the size structure of the populations. Similarly, the relative contribution of spawning areas, characterized by different size distribution of spawners, is likely to differ. Little information exists on temporal and spatial variation in size or age distribution of spawning cod. The effects such variation may have on the production of eggs and larvae, which are consequently subjected to environmental conditions specific for each spawning location, has received little attention. Disaggregating the spawning stock into smaller units that represent geographical locations characterized by a different biological and physical environment may facilitate our understanding of recruitment variability. We present a first attempt to estimate spatial variation in relative abundance, size composition, and production of eggs by spawning cod in three adjacent spawning areas located at the southwest coast of Iceland. Material and methods Spawning areas and data collection Data used to estimate size and age of spawners, sex ratios and catch per unit of effort (c.p.u.e.) were obtained from the annual gill net survey conducted on the spawning grounds located along the south and west coasts of Iceland. In 1997, five spawning regions were sampled with a total of 282 nets of four different mesh sizes (Fig. 1). Each region was divided into several areas based on depth, distance from the shore and historically established fishing grounds for spawning cod. In a first attempt to estimate spatial variation in egg production, four areas (3.1, 3.2, 3.4, and 3.5) representing the main spawning grounds in Icelandic waters (Jónsson, 1982) were selected for the analysis. As the two nearshore areas (3.1 and 3.2) were located next to each other within the same depth range, the possibility of combining them was initially tested by comparing the length and weight of spawners. Because no significant differences were detected (p<.1), the two areas were combined, resulting in a total of three areas, hereafter named areas 1, 2, and 3 (Fig. 2). The boundaries correspond to distinct lava fields located at different depths. Area 1 is more or less within the 75 m depth range, area 2 in the depth range of 75 1 m and area 3 is located on the slopes of the continental edge in deeper water, mostly below 1 m. All spawning areas are characterized by a rough bottom formed by patches of lava or rocky elevations. The bottom structure makes sampling with bottom

826 G. Marteinsdottir et al. 64 3' N 64 15' Reykjavík 64 ' 63 45' 63 3' 63 15' III 23 3' 23 ' 22 3' 22 ' 21 3' 21 ' 2 3' 2 ' W Figure 2. Location and size of the spawning areas studied. Area 1 is closest to the shore, area 2 in the middle and area 3 at the continental edge. The size of each area was estimated as the numbers of square minutes at which gill nets had been placed more than three times during the last six years (black boxes). The 5-, 1- and 2-m depth contours are shown. trawls difficult, and the fishery in the areas uses predominantly gill nets. Available observations (e.g., Thorsteinsson and Marteinsdottir, 1998) indicate that cod remain on and presumably spawn above these elevations. Gill net fishermen have accumulated extensive experience in finding the spawning aggregations and have placed their nets at these locations for centuries. The estimated size of each area was based on this information, using the distribution of commercial gill nets during the last six years as obtained from logbooks (Fig. 2). The number of valid gill nets employed during the survey in 1997 with other relevant information is given in Table 1. The gill nets were placed in series containing 12 nets each and representing four mesh sizes (6, 7, 8, and 9 ). The first 25 cod removed from each net were measured for length and the total number caught was recorded. Weight, sex, maturity stage, and other biological measurements and otoliths were collected, from the first cod removed from each net. Comparisons of mean length and weight were based on Schéffe s multiple comparison test. Age distributions were compared by means of a Chi square test. The c.p.u.e. was calculated both as the average number of fish and as the average weight of the catch in each net, estimated from the length weight relationship. Sex ratios were obtained from the number of sexed fish, pooled into 2 cm length groups. Egg production Average egg production (E i ) per female in area i was estimated from the equation: E i = 1 Σ(N 1i P 1 F 1 )/ 1 Σ(N 1i P 1 ), (1) where N 1i =number of mature fish at length 1 (cm) in area i, P 1 =proportion of females (derived for each 2 cm length interval), and F 1 =potential fecundity at length 1 (cm) for females from all three areas in 1997 (F 1 =.248 1 4.22 ;r 2 =.93, n=16). An index of relative production of eggs (G i ) by area i was estimated as: G i =[ 1 Σ(N 1i P 1 F 1 )/ 1 ΣN 1i ] C i A i, (2) where C i =mean number of fish per net in area i, and A: represents the standardized size of area i relative to area 1. The viable proportion was estimated by taking into consideration the size of the eggs produced by females of

Egg production of spawning cod 827 Table 1. Summary statistics for the three spawning areas. Spawning area 1 2 3 Size of spawning area (km 2 ) 135 157 22 Mean depth s.d. 62.8 13.7 98.1 1.8 144.3 8.2 Number of nets 18 25 6 Number measured and counted 2795 21 84 344 Number weighed and aged 171 247 57 c.p.u.e. n (mean N/net) s.d. 15.5 15. 87.4 47.8 57.3 51.1 c.p.u.e. kg (mean kg/net) s.d. 153.9 149. 64.4 276. 334.6 258.6 Mean length (cm) s.d. 96.8 9.9 88.3 8.3 82.4 8.7 Mean weight (kg) s.d. 1.7 4. 7.7 3.2 5.7 2.4 Mean condition (K) s.d..85.9.73.7.71.5 Mean age s.d. 7.4 1.9 7.3 2.1 6.8 1.7 Mean egg production/female 6.7 1 6 4.6 1 6 3.6 1 6 Mean viable egg production/female 3.7 1 6 2.3 1 6 1.7 1 6 Index of total egg production 4.3 1 6 188. 1 6 15.3 1 6 Index of total viable egg production 22.3 1 6 94. 1 6 7.1 1 6 different sizes and the influence of egg size on viability of newly hatched larvae as proposed by Marteinsdottir and Steinarsson (1998). Egg diameter in mm (D) is related to length in cm (TL) of females in stage II (i.e. females that have spawned between 3 55% of their eggs) according to: D=1.23+.2 TL (3) The viability of larvae (V) was based on the frequency of larvae (Y) that developed swimbladders by the age of ten days (Marteinsdottir and Steinarsson, 1998; Scott et al., 1999): V=exp(Y)/(1+exp(Y)), (4) and the logistic relationship between Y and egg diameter: log(y/(1 Y))= 18.3+12.9 D (5) The viable egg production is presented as the product of the viability constant (V) and the egg production (E) or the index of total egg production (G) in each area. Results Spawning cod ranged in size from 59 to 123 cm. Cod collected from area 1 were generally larger and heavier than those from area 2 and 3 (Fig. 3(a), (b), Table 1). Comparisons of length and weight distributions using ANOVA and Schéffe s F-test showed that mean size in area 1 was significantly larger than in area 2, which in turn was significantly larger than in area 3 (p<.1). Differences in age distributions among areas were less pronounced, but still significant (χ 2 =57.8, d.f.=1; age groups 3 5 and 1 15 were pooled to provide large enough sample sizes in each cell). However, cod in area 1 were not significantly older than cod in area 2 (Fig. 3(c)). Therefore, differences in mean size in areas 1 and 2 were not due to differences in mean age, but rather to cod in area 1 being generally larger (and heavier) at age than cod in area 2 (Fig. 4; p<.5; Schéffe s multiple based on an ANOVA on log-transformed gutted body weight between different regions for each age among 6 9-year-old spawners). The proportion of females among spawning cod increased with age (Fig. 5). Only 2 3% in the smallest size group in areas 2 and 3 were females. Among those larger than 1 cm, 5% were females in areas 1 and 2 and 1% in area 3 were females. The c.p.u.e. differed considerably between areas, both in numbers and in weights per net (Table 1). In area 2, c.p.u.e. was four to fivefold higher than in area 1, while values for area 3 were intermediate. Similarly, egg production differed between areas (Table 1). Because spawners in area 1 were on average larger, the mean production of eggs per female was significantly larger than in the other two areas. Nevertheless, the index of total egg production per area, derived from the relative abundance (c.p.u.e.) and the relative size of the spawning area, was four times higher in area 2 than in area 1 (Table 1). The index for area 3, the smallest area, was nearly three times lower than in area 1, and 12 times lower than in area 2. Similar differences between spawning areas were obtained when the egg production estimates included only the proportion of eggs considered viable (Table 1). The proportion of viable eggs produced in area 1 (55%) were slightly higher than in areas 2 (5%) and 3 (47%).

828 G. Marteinsdottir et al. Frequency (%) Frequency (%) Frequency (%) 5 4 3 2 1 (a) 4 49 6 5 4 3 2 1 5 4 3 2 1 (b) 1 3 (c) 3 However, these differences did not change the general picture; the total production of viable eggs in area 2 exceeded the production in the other two areas by a factor four up to 13 (Table 1). Discussion 6 69 8 89 1 19 Length interval (cm) Area 1 Area 2 Area 3 12 129 4 6 7 9 1 12 13 15 16 18 19 21 22 24 25 28 4 5 6 Weight interval (kg) 7 8 9 1 11 12 13 14 Age (years) Figure 3. (a) Length, (b) weight and age (c) frequency distribution of sexually mature cod by area. The data indicate considerable spatial variation in size and age of spawners among adjacent geographical locations within the region generally referred to as the main spawning ground of cod. The size differences were highly significant; the mean weight of spawning cod in shallow water closest to the shore was nearly twice as high as on the edge of the continent. Differences in age distributions were also significant; these findings are consistent with previous reports. Based on research trawl data and commercial samples from gill nets, Marteinsdottir and Petursdottir (1995) showed that both Total length (cm) 12 11 1 9 8 7 6 Area 1 Area 2 Area 3 5 6 7 8 9 1 Age (years) Figure 4. Mean length ( s.d.) at age of sexually mature cod by area. Percentage 1 8 6 4 2 4 59 Area 1 Area 2 Area 3 6 79 8 99 1 119 Length interval (cm) size and age of spawners in the near shore area were significantly greater than on the bank and along the continental edge. Spatial segregation of size and age classes has been a common feature in the annual gill net surveys. Many studies report spatial segregation of size or age classes in relation to environmental variables such as depth and temperature (e.g., Sinclair, 1992; Shackell et al., 1997). However these refer to distributions outside the spawning season while the fish are on the feeding ground, rather than to spawning aggregations. Mean egg production and the total egg production index were influenced by the spatial differences in the size structure of the spawning population. As the potential and relative fecundity (number of eggs per unit body size or age) increased with size, the mean egg production per female differed significantly between areas. However, the total egg production index was largely determined by the relative abundance and the size of the areas. The index assumes that all sexually mature cod present in each area at the time of sampling (i.e. during the peak spawning season) spawn locally. The estimates of viable egg production, based on an assumed general relationship between viability of larvae and size of the female spawners through variation in egg diameter, did not change the general picture. However, 11 12 14 Figure 5. Proportion of sexually mature females by 2-cm length interval and area.

Egg production of spawning cod 829 no attempt was made to include information on variable spawning time among areas in the viability estimates. Spawning in area 1 extends generally over a longer period compared to areas 2 and 3 (Marteinsdottir and Petursdottir, 1995). As a longer spawning period may influence the average survival of eggs and larvae (Mertz and Myers, 1994), the production of viable eggs in this area may have been underestimated. We have chosen a relative total egg production index because estimates of the absolute egg production in each area would be influenced by errors caused both by inaccurate estimates of the size of the spawning area and sampling efficiency (the area effectively cleared of fish in one net setting). The latter has been shown to vary from a few hundred square meters up to a few thousand square meters during a 24-h setting (Dickson, 1989). According to the experience of local fishermen, spawning cod select particular spots on rocky elevations and therefore form an extremely patchy distribution within a specific area. For example, spawning aggregations may be observed repeatedly on a single peak of lava or rocky elevation rising from the sea floor, while no fish are observed at another similar peak close by. These spots may be no more than a single square mile in size. The size and age structure of spawners along with the associated egg production in each area will change over time. For instance, the frequency of 1+ cod during the spawning seasons decreased in area 1 between 1994 and 1995 (Marteinsdottir and Petursdottir, 1995) due to declining numbers of the strong 1983 and 1984 year classes. At the same time, changes were observed in areas 2 and 3 when the 199 year class started to recruit to the spawning population, resulting in relatively high frequencies of four and five-year-old cod in 1994 and 1995, respectively. The importance of the nearshore area may, therefore, become more prominent again when strong year classes have aged, while production in the offshore areas may peak at times when strong year classes are recruiting to the spawning stock. Eggs and larvae produced in different spawning areas may experience different environmental scenarios, including timing and amplitude of plankton productivity and/or advection due to spatial differences in oceanographic conditions. Spawning of both cod (Marteinsdottir and Petursdottir, 1995) and Calanus finmarchicus (Gislason and Astthorsson, 1991, 1995, 1996) starts earlier in nearshore waters compared to offshore areas, suggesting that the timing of the production cycle and production of fish larvae are more or less synchronized within each area. Nevertheless, considerable variation exists both with respect to the timing of the spring bloom (Thordardottir, 1986; Gislason et al., 1994) and the peak spawning of C. finmarchicus (Gislason et al., 1994). However, in the long run, the mean probability of survival may be expected to be higher in the nearshore area than in the other two areas. First of all, the extended spawning period in the nearshore area may increase the probability of egg and larval survival (Mertz and Myers, 1994). Secondly, primary production and the formation of a pycnocline is probably more predictable due to the influence of freshwater runoff by two main rivers (Fig. 2) on the nearshore spawning sites (Thordardottir, 1986; Olafsson, 1985). Thirdly, the west- and northward flowing coastal current induced by freshwater runoff (Olafsson, 1985) may provide a more predictable transport mechanism for eggs and larvae into the nursery areas west, north and east of the country than the offshore current. The extensive variation that exists between geographically adjacent spawning locations includes both biological parameters related to reproduction and oceanic conditions. Detailed analysis of historical data on temporal and spatial changes in relative abundance and structure of the spawning aggregation in relation to available environmental information is required to improve our understanding of the recruitment processes in cod. Acknowledgements We wish to thank all colleagues and fishermen participating in the gill net survey conducted in 1997. References Bergstad, O. A., Jorgensen, T., and Dragesund, O. 1987. Life history and ecology of the gadoid resources of the Barents Sea. Fisheries Research, 5: 119 161. Cushing, D. H. 199. Plankton production and year-class strength in fish populations: and update of the match/ mismatch hypothesis. Advances in Marine Biology, 26: 249 293. Dalley, E. L., and Anderson, J. T. 1997. Age-dependent distribution of demersal juvenile Atlantic cod (Gadud morhua) in inshore/offshore northeast Newfoundland. Canadian Journal of Fish and Aquatic Sciences, 54: 168 176. Dickson, W. 1989. Cod gillnet effectiveness related to local abundance, availability and fish movement. Fisheries Research, 7: 127 148. Gislason, A., Astthorsson, O. S., and Gudfinnsson, H. 1994. Phytoplankton, Calanus finmarchicus, and fish eggs southwest of Iceland, 199 1992. ICES Marine Science Symposia, 198: 423 429. Gislason, A., and Astthorsson, O. S. 1991. Distribution of zooplankton across the coastal current southwest of Iceland in relation to hydrography and primary production. ICES CM 1991/L:17. Gislason, A., and Astthorsson, O. S. 1995. Seasonal cycle of zooplankton southwest of Iceland. Journal of Plankton Research, 17: 1959 1976. Gislason, A., and Astthorsson, O. S. 1996. Seasonal development of Calanus finmarchicus along an inshore-offshore gradient southwest of Iceland. Ophelia, 44: 71 84. Hildebrand, S. F., and Schroeder, W. C. 1972. Fishes of Chesapeake Bay. Smithsonian Institute Press, Washington, D.C.

83 G. Marteinsdottir et al. Hutchings, J. A., and Myers, R. A. 1993. Effect of age on the seasonality of maturation and spawning of Atlantic cod, Gadus morhua, in the Northwest Atlantic. Canadian Journal of Fisheries and Aquatic Sciences, 5: 2468 2474. Hutchings, J. A., Myers, R. A., and Lilly, G. R. 1993. Geographic variation in the spawning of Atlantic cod Gadus morhua, in the Northwest Atlantic. Canadian Journal of Fisheries and Aquatic Sciences, 5: 2457 2467. Liem, A. H., and Scott, W. B. 1966. Fishes of the Atlantic Coast of Canada. Fisheries Research Board of Canada, Ottawa. Jónsson, E. 1982. A survey of spawning and reproduction of the Icelandic cod. Rit Fiskideildar, VI(2): 1 45. Marteinsdottir, G., and Steinarsson, A. 1998. Maternal influence on the size and viability of Iceland cod (Gadus morhua L.) eggs and larvae. Journal of Fish Biology, 52: 1241 1258. Marteinsdottir, G., and Thorarinsson, K. 1998. Improving the stock recruitment relationship in Icelandic cod (Gadus morhua L.) by including age diversity of spawners. Canadian Journal of Fisheries and Aquatic Sciences, 55: 1372 1377. Marteinsdottir, G., and Petursdottir, G. 1995. Spatial and temporal variation in reproduction of Icelandic cod at Selvogsbanki and nearby coastal areas. ICES CM 1995/G:15. Mertz, G., and Myers, R. A. 1994. Match/mismatch predictions of spawning duration versus recruitment variability. Fisheries Oceanography, 3(4): 236 245. Myers, R. A., Barrowman, N. J., and Hutchings, J. A. 1997. Inshore exploitation of Newfoundland Atlantic cod (Gadus morhua) since 1948 as estimated from mark-recapture data. Canadian Journal of Fisheries and Aquatic Sciences, 54: 224 235. Olafsson, J. 1985. Recruitment of Icelandic haddock and cod in relation to variability in the physical environment. ICES CM 1995/Q:59. Scott, B., Marteinsdottir, G., and Wright, P. 2. The potential effects of maternal factors on spawning stock recruitment relationships under varying fishing pressure. Canadian Journal of Fisheries and Aquatic Sciences. In press. Sinclair, M. 1992. Fish distribution and partial recruitment: The case of Eastern Scotian Shelf cod. Journal of Northwest Atlantic Fishery Sciences, 13: 15 24. Shackell, N. L., Stobo, W. T., Frank, K. T., and Brickman, D. 1997. Growth of cod (Gadus morhua) estimated from markrecapture programs on the Scotian Shelf and adjacent areas. ICES Journal of Marine Science, 54: 383 398. Thordardottir, T. 1986. Timing and duration of spring blooming south and southwest of Iceland. In The role of freshwater outflow in coastal marine ecosystems, pp. 345 36. Ed. by S. Skreslet. Series G: Ecological Sciences, Vol. 7. Springer, Berlin. Thorsteinsson, V., and Marteinsdottir, G. 1992. Tagging and recapture of spawning cod in two fjords at the East coast of Iceland in 1991 and 1992. Ægir, 2: 3 1 (in Icelandic). Thorsteinsson, V., and Marteinsdottir, G. 1998. Size specific time and duration of spawning of cod (Gadus morhua) in Icelandic waters. ICES CM 1998/DD:5.