MARINE ECOLOGY PROGRESS SERIES Vol. 202: , 2000 Published August 28 Mar Ecol Prog Ser

MARINE ECOLOGY PROGRESS SERIES Vol. 0: 93 7, 000 Published August Mar Ecol Prog Ser Spawning origins of pelagic juvenile cod Gadus morhua inferred from spatially explicit age distributions: potential influences on year-class strength and recruitment Gavin A. Begg*, Gudrun Marteinsdottir Marine Research Institute, Skulagata, PO Box 390, Reykjavik, Iceland ABSTRACT: Incongruity between the reconstructed spawning period of pelagic juvenile (0-group) cod Gadus morhua and the observed spawning period for adult cod on the main spawning grounds south of Iceland have indicated that large numbers of the surviving juvenile population may, in some years, originate from other smaller inshore spawning grounds located within fjords of the west, north and east coasts. We examine this hypothesis by using age/length and temperature/growth relationships, based on aging of 0-group cod from 995 and 997, to analyze historical demographic data from 0-group fish surveys (970 to 99) and determine the spawning distribution and origin of pelagic juvenile cod. Data on early life-history characteristics in conjunction with temporal and spatial distribution of pelagic juvenile cod are used to create new abundance indices representing the inshore and offshore juvenile cod-population components. Partitioning of the historical 0-group index into these population components at a finer geographic scale than has been used previously provides new information for estimating the relative contribution of the main spawning component versus the smaller in-fjord spawning components towards variable recruitment levels in the Icelandic cod stock. Spawning diversity, or multiple spawning components, are important for recruitment success as they disperse the mortality risk of the early life-history stages, which are thought to be the principal determinants of year-class strength. KEY WORDS: Stock Recruitment 0-group Cod Gadus morhua Population dynamics Resale or republication not permitted without written consent of the publisher INTRODUCTION Fundamental to understanding the population dynamics, and subsequent assessment and management, of exploited marine fish stocks is the relationship between spawning stock biomass and recruitment. Although, spawning stock biomass and recruitment are functionally related (Ricker 95, Beverton & Holt 957, Cushing 97, Fogarty 993, Myers et al. 995, Myers & Barrowman 996), this relationship is often poorly defined; most likely due in part to the inherent variability associated with ecological and/or environmental factors (Jakobsen 996, Sparholt 996). Recently, this has resulted in alternate, potentially more *E-mail: gavin@hafro.is informative, measures of recruitment being sought, such as egg production indices (Marshall et al. 99), and the need to disentangle individual factors that may be contributing to the inherent variability of year-class strength and stock-recruitment relationships, such as age and size diversity of spawning fish (Marteinsdottir & Thorarinsson 99). Similarly, the diversity of spawning components and subsequent origins of progeny derived from these components contribute further to the inherent variability within stock-recruitment relationships, particularly as these relationships are based on a single-unit stock assumption (Begg et al. 999, Stephenson 999). However, often this assumption is violated by greater spawning diversity (i.e. multiple spawning components) than that which typically has been assessed in historically established management Inter-Research 000

9 Mar Ecol Prog Ser 0: 93 7, 000 units. In this study, we examine this assumption by investigating the origin of pelagic juvenile (0-group) cod (Gadus morhua) in Icelandic waters with an aim to determining the relative contribution of individual spawning components to the overall recruitment of the stock. In recent years, the Icelandic cod stock has been severely depleted, with recruitment levels low since the mid-90s (Schopka 99, Baldursson et al. 996). Historically, the main spawning grounds of cod in Iceland have been located along the south coast (Sæmundsson 9, Jónsson 9), although smaller inshore and in-fjord spawning grounds in other locations around the country also appear to contribute to the overall recruitment of the stock (Marteinsdottir et al. in press b). However, the relative contribution of these smaller spawning components to the main spawning component is not known. Pelagic eggs and larvae derived from the main spawning component off the south coast, during April and May, drift clockwise around the country with the westward flowing coastal currents and the northward flowing Irminger Current (Fridgeirsson 9, Astthorsson et al. 99) (see Fig. ). Generally, the majority of pelagic juveniles are found in the Arcto-boreal waters off the northwest, north, and northeast coasts of Iceland, although in some years large concentrations are also found near Greenland where the Irminger Current flows westward (Vilhjalmsson & Fridgeirsson 976, Vilhjalmsson & Magnusson 9, 95, Astthorsson et al. 99). In contrast to this expected dispersal pattern, a declining size gradient is typically found for pelagic juvenile cod from the west to the east coasts that has been suggested to derive from temperature-induced growth differences (Astthorsson et al. 99). However, a recent study by Marteinsdottir et al. (in press b), based on the 995 to 997 0-group fish surveys, found a corresponding declining age gradient for pelagic juvenile cod from the west to the east coasts, indicative of differential spawning components. Consequently, the 0-group fish surveys provide an important source of information by which to examine the life history, origin, and stock complexity of pelagic juvenile cod that may explain some of the variability in year-class strength and subsequent recruitment to the stock. Pelagic juvenile (0-group) fish surveys have been conducted in the northeast Atlantic since 965 to provide annual estimates of year-class strength and recruitment (Helgason & Sveinbjörnsson 97, Sundby et al. 99, Jakupsstovu & Reinert 99). These estimates predict recruitment at older ages, but do not correlate well in some years (Campana et al. 99, Sundby et al. 99, Astthorsson et al. 99, Anderson & Dalley 997). Some of this variation may be due to density-dependent predation after settlement (Bogstad et al. 99), while further variation may be due to the coarse geographic scale at which the 0-group abundance indices are calculated and the need within these calculations to account for multiple spawning components. In Iceland, the 0-group abundance indices are presently calculated for 5 large geographic regions that incorporate both inshore and offshore waters (Vilhjalmsson & Fridgeirsson 976, Sveinbjörnsson & Jónsson 99). However, numerous studies have shown that cod typically have separate inshore and offshore spawning components that may be reproductively isolated from one another (Hurley & Campana 99, Hutchings et al. 993, Anderson et al. 995, Marteinsdottir & Petursdottir 995, Ruzzante et al. 996, Anderson & Dalley 997, Dalley & Anderson 997). Similarily, in waters north of Iceland, an offshore component may represent juvenile cod that drifted into the northern waters from the main spawning grounds in the south, while an inshore component may to a greater extent include juveniles that originated from local spawning within fjords. Hence, there exists a need to re-examine the historical 0-group abundance indices for Icelandic cod at a finer geographic scale than has previously been examined, separating inshore and offshore components, in order to explain variable recruitment levels in the stock particularly in view of the more recent study of Marteinsdottir et al. (in press b) that suggests the existence of multiple spawning components. In this study, we examine the historical 0-group survey data to determine the origin of pelagic juvenile cod in Icelandic waters. The traditional paradigm of cod stockstructure was tested to determine whether pelagic juvenile cod originate from spawning grounds other than those of the main spawning grounds off the south coast. The findings of Marteinsdottir et al. (in press b) were extended by using length/age and temperature/growth relationships based on their aging of 0-group cod to provide information on historically reconstructed spawning times and locations. The statistical reconstruction of inter-annual variation in spatial distribution, abundance, length, and age composition of pelagic juvenile cod, modeled with environmental conditions were used to infer final locations of progeny surviving to the pelagic juvenile stage. Furthermore, this enabled predictions about the dispersal of early life-history stages and relationships between their spatial distribution and abundance to be evaluated as a function of stock and environmental conditions, while evaluating the significance of spawning diversity or multiple spawning components to year-class strength and recruitment. MATERIAL AND METHODS Historical distribution, abundance, and length composition data for pelagic juvenile cod Gadus morhua

Begg & Marteinsdottir: Spawning origins of juvenile cod 95 collected from the 0-group fish surveys between July and September 970 to 99 were analyzed to determine spawning origins of cod in Icelandic waters. Sampling was conducted using a Harstad pelagic trawl ( m opening; 0.5 0.5 cm cod-end mesh) at <90 stations between 970 and 97, and >50 stations since 973. Since 90, a fixed survey route has been used, although the extent of the survey into the Irminger Sea has fluctuated. Standardised survey methods were used throughout the survey period, with trawls generally made at depths of 0 to 50 m. Acoustic records were used to monitor changes in fish distribution and abundance and to determine sampling frequency. A sample was taken whenever acoustic records changed or approximately every 0 nautical miles. Abundance measures recorded at each station were the number of cod captured per nautical mile towed. Total lengths (TL) were adjusted (based on a measured growth rate of 0.65 mm d ; Marteinsdottir et al. in press b) to the mean cruise date of the surveys (0 August) estimated for the entire time series, enabling a direct comparison of historical data between years. The 0-group data were separated into inshore (a to 7a) and offshore (b to 7b, ) statistical regions on the basis of known spawning locations and dispersal patterns of eggs and larvae (Marteinsdottir et al. in press a) and the 00 m depth contour, enabling a more accurate representation of spawning and spatial stock structure than previously achieved for the Icelandic cod stock (Table, Fig. ). Temperature data were obtained from fixed survey stations located along standardized transects used in annual hydrographic research cruises to characterize ambient environmental conditions experienced by surviving pelagic juvenile cod (Fig. ). Water temperatures in the upper 00 m of the water column were averaged for each day (in each year) that data were recorded to represent annual daily temperature cycles. Annual daily temperature cycles were derived for waters north (combined Statistical Regions to 6) and south (combined Statistical Regions and 7) of Iceland, respectively, on the basis of known oceanographic conditions (Malmberg et al. 996, Anonymous 999), and concerns associated with low sample sizes in some years for any given statistical region (Fig. ). Although, it was desirable to estimate daily temperature cycles Table. Annual number (n) of 0-group cod Gadus morhua measured in each region (a to ), 970 to 99. na = no data available Year n (region) Total a b a b 3a 3b a b 5a 5b 6a 6b 7a 7b 970 na 30 0 536 3 99 39 9 6 na na na na na 395 97 0 70 07 7 3 3 7 na na na na na 7 97 79 6 7 63 3 9 na 00 na na na na na 97 973 3 0 65 9 50 30 556 67 67 7 305 95 9 7 0 97 39 na 5 00 9 390 57 na na na 03 975 3 5 5 3 3 5 6 5 0 7 39 3 na na 63 53 976 33 99 6 59 99 5 060 6 00 659 3 5 5 9 6707 977 na 73 75 75 9 66 5 99 37 55 56 7 na 5 97 3 760 63 66 66 9 5 7 30 9 na na 6 36 979 39 9 69 5 3 5 0 5 na na 70 90 0 6 35 76 6 37 70 0 7 3 na 7 7 9 0 35 6 5 5 6 na na na na na 53 930 9 6 na 3 9 5 na na na na 60 93 60 33 5 37 5 75 97 na na na 5 3 9 09 57 66 90 3 90 33 35 360 3 30 73 79 95 0 5 363 75 36 96 375 5 6 59 na na na 375 33 96 6 5 3 95 39 6 37 73 6 6 na 7 3 6 50 97 na 3 6 79 0 37 7 77 5 na na 3 037 9 9 7 39 6 6 9 5 3 na na na na 500 99 5 63 36 6 0 09 5 0 3 6 na na 0 990 7 9 6 5 3 5 9 5 05 na na 37 03 99 3 na 3 3 3 0 5 3 na na na na 37 99 3 7 0 39 3 na 9 5 3 na na 96 993 9 6 505 379 6 670 9 6 5 9 5 na na 7 99 na na 6 5 33 6 6 na 3 970 995 6 6 3 3 03 6 35 0 90 9 na 5 996 5 3 3 03 69 9 53 3 na na 5 na 90 997 50 0 67 593 7 66 79 99 55 33 59 na 0 750 99 56 375 963 9 59 700 33 9 77 39 na na 6 96 Total 709 06 07 76 0035 39 57 7677 59 37 9 5 30 90

96 Mar Ecol Prog Ser 0: 93 7, 000 Latitude N Latitude N 6 66 6 6 67 66 65 6 63 Greenland Irminger Current b a Main spawning grounds b 3b b a 3a 3 * * * * * Iceland 36 3 0 6 Longitude W 6 0 6 Longitude W a * 7a * * 7b 5 7 5a * 5b 6 6a 6b Fig.. General ocean circulation patterns (Malmberg et al. 996), and inshore (a to 7a) and offshore (b to 7b, ) statistical regions of 0-group cod Gadus morhua sampled in Icelandic waters. : Locations of hydrographic stations used to * characterize yearly temperature data for northern (hatched Regions to 6) and southern (Regions and 7) areas. Main spawning grounds located in Region for each statistical region separately, it was not possible because of sample limitations. However, previous studies have shown that there are relatively well defined horizontal gradients between waters north and south of Iceland (corresponding to the above combined regions), whereby the northern waters are colder and more variable in any given year than the southern waters (Malmberg & Kristmannsson 99, Kristmannsson 99). Consequently, to reconstruct historical daily temperature cycles and estimate temperature at capture, hatching and spawning of pelagic juvenile cod, fourth order polynomial regressions were fitted to the daily temperature data for each year and area (i.e. north or south). For those cod found in Region, the daily temperature cycles for the southern area were used, as the conditions there were assumed to be most similar because of known ocean circulation patterns (Malmberg et al. 996). Also, for some years (northern area: 970, 97, 973; southern area: 970), daily temperature cycles could not be estimated directly because of limited data; hence, temperature cycles for other years approximating similar environmental conditions were used instead (based on Astthorsson et al. 99, Anonymous 999). Furthermore, for some years (northern area: 99; southern area: 979, 9, 93), daily temperature cycles were based on the mean monthly data for the mean day of sampling for each month rather than the raw daily temperature data, because of an improved fit of each regression. For similar reasons, the daily temperature cycle for 97 (northern area) was based on data also collected from adjacent offshore stations. Spatial distribution and relative abundance (number per nautical mile) of pelagic juvenile cod were initially compared on an annual basis, irrespective of region, to determine the general strength and direction of progeny dispersal following spawning. Annual abundance indices of pelagic juvenile cod were then calculated for each statistical region (a to ) following the methods of Astthorsson et al. (99). Prior to these calculations the abundance data were first log-transformed to normalize the distributions and reduce the effects of the relatively few large catches. An abundance index (A r ) of pelagic juvenile cod in each region was then calculated for each year according to the following equations: X r = N r ln(x i ) () P r = N r N t () A r = X r P r (G r G t ) (3) where, in each region (a to ), X r = logarithmic mean abundance of pelagic juvenile cod; N r = number of stations where catch was >0; X i = number of cod captured per nautical mile towed at Stn i where catch also was >0; P r = proportion of non-zero tows; N t = total number of stations; G r = geographic area (km ) of the region; and G t = total geographic area of all the regions. Adjusted (0 August) total length distributions of pelagic juvenile cod were determined for fish in each year and region (a to ) to assess differences in demographic characteristics indicative of differential spawning times and locations. Likewise, a growth model was derived for pelagic juvenile cod based on 995 and 997 aging data from Marteinsdottir et al. (in press b), and applied across years and regions (a to ) to provide historical age distributions of pelagic juvenile cod. Analysis of covariance (ANCOVA) was initially used to examine differences in growth rates of cod from the different regions. Multiple regression was then used to examine the effect of temperature at capture because of the known influence temperature has on growth (Campana & Hurley 99) and the widely differing temperatures between waters north and south of Iceland (Anonymous 999).

Begg & Marteinsdottir: Spawning origins of juvenile cod 97 Hatch and spawning (day-of-the-year, DOY) distributions of pelagic juvenile cod were then examined to determine spatial origins of cod in Icelandic waters. Hatch distributions of pelagic juvenile cod in each region (a to ) were determined directly by subtracting the age of each fish from the day at capture, and were not corrected for mortality effects because all samples were collected in the low-mortality juvenile stage (Campana & Jones 99, Campana 996). In contrast, spawning distributions were determined by subtracting an incubation period (T hatch ) from the estimated hatch day of each fish. Consequently, we estimated the potential distribution of pelagic juvenile cod that may have originated from the main spawning grounds in the south (Region ) by using incubation periods and reconstuctured spawning distributions based only on the southern area (i.e. null hypothesis, H 0 : all pelagic juvenile cod originate from the main spawning grounds in the south). The incubation periods, based on the estimated daily temperature cycles for the southern area in each year, were estimated for the mean mid-incubation day the mid-day between the estimated annual mean hatch day (this study) and the presumed mean spawning day (Jónsson 9, Marteinsdottir et al. in press b). Accordingly, the incubation period was related to the ambient water temperature following Pepin et al. (997): T hatch = 6.e 0.7 temperature () where temperature was that estimated from the daily temperature cycles for the southern area in each year for the estimated annual mean mid-incubation day. The proportion of pelagic juvenile cod estimated to originate from spawning later than the 99 and 95% spawning distribution percentiles for those cod from the main spawning grounds (Region ) were then assumed to originate from spawning grounds other than the main spawning grounds. RESULTS Daily temperature cycles Historically, lower daily temperatures were consistently experienced by those pelagic juvenile cod, Gadus morhua, inhabiting waters off the northwest, north, and northeast coasts (Regions to 6) than those off the southwest, south, and southeast coasts (Regions and 7: Table, Fig. ). In each area, water temperatures steadily increased from April to June (presumptive spawning period), showing a consistent seasonal heating cycle across all years. Throughout this period, water temperatures ranged from.65 to 5.60 C in the northern area, and from.07 to. C in the southern area. Generally, water temperatures varied by C between the areas on any given day during the presumptive spawning and hatching periods, with mean temperatures across years in the northern area approximating.70 C ( April),.7 C (0 May), and 3. C ( June), and in the southern area 5., 6.5, and 7. C, respectively (Fig. ). Distribution and abundance In August/September, pelagic juvenile cod were distributed around the west, north, and east coasts of Iceland (Fig. 3). In most years, large aggregations of pelagic juvenile cod were typically found off the northern coast of Iceland (Regions to 5), although in some years large aggregations were also found near Greenland (Region ), where they had probably drifted from the main spawning grounds in the south (Region ) with the northern and western branch of the Irminger Current. Few pelagic juvenile cod were found in waters off the southeast coast (Region 7), indicating that eggs and larvae are not retained in this region during the summer months following hatching. Likewise, few pelagic juvenile cod were typically found on the main spawning grounds as no doubt prior to sampling they had already drifted north with the coastal and Irminger currents following spawning (Fig. 3). Generally, there has been a progressive decline in relative abundance of pelagic juvenile cod throughout the years (F = 5.6, df =, 7, p < 0.09), although there have been intermittent increases, particularly in more recent years (Fig. ). Abundance indices tended to be greater for regions in the north than for those in the south, with high abundance indices estimated near Greenland (Region ) in some years. Similarly, abundance indices tended to be greater in the offshore regions (b to 7b) than in their corresponding inshore regions (a to 7a) (Fig. ). Notably, in those years where abundance indices were greatest (i.e. 973, 976, 9, 95, 997, 99) large aggregations of pelagic juvenile cod appeared to be more evenly distributed along the entire northern coast (Regions to 6) (Fig. 3). Length composition Pelagic juvenile cod tended to decrease in length from the west to the east coasts (Regions to 7) (Fig. 5). This decreasing, clockwise, length gradient of pelagic juvenile cod around the country was evident for all years combined, and for nearly all individual years (except 976) throughout the survey period, indicative of progeny originating from differential spawning or hatching times and locations and/or the result of spa-

9 Mar Ecol Prog Ser 0: 93 7, 000 Table. Annual temperature relationships represented by fourth order polynomial regressions used to reconstruct historical day-of-the-year (DOY) temperatures at capture, hatching, and spawning for 0-group cod Gadus morhua in northern (Regions to 6) and southern (Regions and 7) areas, 970 to 99 Year n r Parameter estimate Intercept DOY DOY DOY 3 DOY Northern 970 a 0 0.67 0.050 0.05 0.000 0.000005 0.00000000 97 a 0 0.67 0.050 0.05 0.000 0.000005 0.00000000 97 b 0.95 0.339 0.63 0.00 0.0000 0.0000000376 973 c 0.77 7.0 0.67 0.007 0.00003 0.00000006 97 0.77 7.0 0.67 0.007 0.00003 0.00000006 975 0 0.67 0.050 0.05 0.000 0.000005 0.00000000 976 0.0 0.6 0.5 0.005 0.00007 0.00000005 977 6 0.76. 0.7 0.00 0.000003 0.00000000 97 0.76.5 0.0 0.0003 0.00000 0.000000003 979 5 0.6.63 0.0 0.000 0.000007 0.0000000 90 0.9 6.07 0.75 0.009 0.000006 0.0000000055 9 0. 0.960 0.36 0.0037 0.0000 0.0000000 9 0.9 0.3 0.06 0.00 0.000007 0.0000000 93 6 0.7.90 0.0 0.000 0.000005 0.000000009 9 0.5.05 0.9 0.0093 0.000039 0.000000055 95 5 0.5.3 0.7 0.009 0.000009 0.000000039 96 7 0.6 7.50 0. 0.007 0.000005 0.000000005 97 7 0. 5.7 0.70 0.0076 0.00007 0.000000035 9 6 0.9 3.577 0.73 0.00 0.0000 0.0000000 99 3 0.97 0.793 0.0 0.0006 0.000005 0.00000000 990 0.5 9.59 0.5 0.007 0.000003 0.000000000 99 5 0.70 0.59 0.30 0.006 0.00000 0.00000009 99 6 0.6 0.59 0.60 0.00 0.000006 0.000000009 993 0.7 0.7 0.3 0.006 0.0000 0.000000039 99 0.7.9 0.3 0.00 0.000005 0.000000007 995 0.3. 0.7 0.00 0.00000 0.0000000035 996 0.7. 0.53 0.006 0.00007 0.000000003 997 0. 3.05 0.0 0.0039 0.00003 0.000000039 99 d 7 0.7 9.095 0.550 0.0056 0.0000 0.00000006 Southern 970 e 9 0.7 0.7 0.7 0.00 0.000006 0.000000006 97 9 0.7 0.7 0.7 0.00 0.000006 0.000000006 97 6 0.99 7.07 0.03 0.0003 0.00000 0.00000000 973 7 0.36 5.90 0.3 0.003 0.0000 0.000000035 97 0 0.0.79 0.39 0.006 0.0000 0.0000000 975 0.73.0 0.070 0.0003 0.00000 0.00000000 976 5 0.6.3 0.5 0.009 0.0000 0.00000005 977 0.69 3.7 0.306 0.003 0.00003 0.000000057 97 0 0.73 6. 0.05 0.0005 0.00000 0.0000000006 979 d 7 0.9 7.670 0.036 0.000 0.000003 0.0000000059 90 5 0.95.0 0.006 0.0000 0.00000 0.000000009 9 0.7.766 0.006 0.0003 0.00000 0.000000000 9 d 6 0.99 7.033 0.0 0.006 0.000007 0.00000003 93 d 7 0.7 6.95 0.073 0.0009 0.000003 0.0000000037 9 0.6.33 0.07 0.0007 0.00000 0.0000000007 95 7 0. 6.0 0.06 0.0005 0.00000 0.000000000 96 0.7.737 0.0 0.00 0.000007 0.0000000 97 3 0.0.66 0.0 0.000 0.000003 0.000000006 9 3 0.70 3.73 0.7 0.007 0.0000 0.00000005 99 0 0.6 0.69 0.6 0.007 0.00000 0.0000000 990 0 0..09 0.9 0.007 0.00007 0.000000000 99 0.73 6.793 0.360 0.0035 0.0000 0.000000039 99 3 0.7.09 0.3 0.00 0.000007 0.000000007 993 7 0.6 9. 0.5 0.00 0.00000 0.0000000033 99 0.77 7.7 0.06 0.003 0.000005 0.000000005 995 0.63 5.973 0.09 0.00 0.00000 0.0000000036 996 9 0.69.97 0.6 0.006 0.0000 0.000000057 997 0.76.70 0.063 0.000 0.000003 0.000000000 99 0.6.5 0. 0.006 0.000006 0.0000000065 a Polynomial for 975 (northern area) used instead, as environmental conditions were most similar b Polynomial for 97 (northern area), also based on data from offshore stations c Polynomial for 97 (northern area) used instead, as environmental conditions were most similar d Polynomial based on mean monthly data for mean day of sampling for each month e Polynomial for 97 (southern area) used instead, as environmental conditions were most similar tially differential growth rates. Furthermore, when the mean lengths of pelagic juvenile cod were estimated for individual survey stations in each year this length gradient was even more pronounced (Fig. 6). Interestingly, the length composition of pelagic juvenile cod comprising the Greenland component (Region, mean [± SD] all years combined = 59 ± 0 mm) was very similar to that of those cod originating from the main spawning component (Region, 59 ± mm), providing further support for pelagic juvenile cod from these regions having similar spawning origins. Growth and age composition Growth rates did not appear to differ between pelagic juvenile cod from the northern (Regions to 6) and southern (Region ) areas (995 and 997 data combined: ANCOVA, F =.9, df =, 5, p > 0.0); thereby justifying the assimilation of these data for statistical purposes (Fig. 7). Consequently, we decided to combine all the data into a single growth model for purposes of consistency and because of sample limitations, to be used across all the years for historical reconstruction of age at capture, hatching, and spawning of pelagic juvenile cod. In addition, temperature at capture was found to have a significant effect on growth, and so was also incorporated into the final growth model (Table 3). Pelagic juvenile cod tended to decrease in age from the west to the east coasts (Regions to 7: Table, Fig. ). Similar to length, this decreasing, clockwise, age gradient of pelagic juvenile cod around the country was evident for all years combined and for nearly all of the individual years throughout the survey. Once again, the age composition of pelagic juvenile cod comprising the Greenland component (Region, mean [±SD] all years combined = 93 ± d) was very similar to that of cod originating from the main spawning component (Region, 9 ± 6 d) (Table ).

Begg & Marteinsdottir: Spawning origins of juvenile cod 99 Estimated temperature ( C) Apr (9 DOY) 0 70 75 0 5 90 95 0 May (30 DOY) 0 70 75 0 5 90 95 Jun (5 DOY) 0 70 75 0 5 90 95 to 7: Table 5, Fig. 9). Those cod found off the southwest coast, near the main spawning grounds, tended to hatch around May (Region, mean [SD] all years combined = 3 ± 6 DOY), while those further north and east hatched progressively later throughout May, June and early July (Region, ± 7; Region 3, 55 ± 7; Region, 60 ± 7; Region 5, 6 ± ; Region 6, 63 ± ; Region 7, 6 ± 6 DOY). Hatch distributions of pelagic juvenile cod found near Greenland (Region, 39 ± DOY) tended to closely reflect the hatch distributions of those cod found near the main spawning grounds in the south (Region : Fig. 9). 0 70 75 0 5 90 95 Year Hatch distribution Pelagic juvenile cod tended to hatch progressively later from the west to the east coasts for all years combined and for most of the individual years (Regions Aug (3 DOY) Fig.. Annual estimated temperatures derived from temporal and spatial specific fourth order polynomial regressions for specific days-of-the-year (DOY) experienced by 0-group cod Gadus morhua in northern (H) and southern (J) areas, 970 to 99 Table 3. Growth relationship (total length/age) of 0-group cod Gadus morhua used to reconstruct historical age at capture for determination of hatching and spawning distributions. Relationship based on 995 and 997 ageing data from Marteinsdottir et al. (in press b) for Regions to 6. [TL = 5.73 + 0.56(age).50(temperature) + 0.0307(age temperature) (r = 0.3)] Source df MS F p Model 3 7.9 36.9 0.000 Error 5.56 Age 39.06 5.7 0.000 Temperature 00.06.3 0.0357 Age Temperature 9.9.37 0.037 Parameter Estimate SE p Intercept 5.73.5965 0.065 Age 0.56 0.0 0.000 Temperature.50.06 0.0357 Age Temperature 0.0307 0.07 0.037 Spawning distribution Pelagic juvenile cod assumed to originate from the main spawning grounds in the south (Region ) were estimated to incubate prior to hatching for approximately 5 ± d (all years combined). This incubation period was based on the estimated annual daily temperature cycles for the southern area for the mean mid-incubation day (range to DOY) in each year, which was the mid-day between the annual mean hatch day (3 to 67 DOY) and the presumed mean spawning day (05 DOY). Based on these assumptions, pelagic juvenile cod were estimated to spawn progressively later from the west to the east coasts for all years combined, and for most of the individual years (Regions to 7: Table 5, Fig. 0). Those cod found off the southwest coast, near the main spawning grounds, were estimated to originate from spawning periods around 3 May (Region, mean [SD] all years combined = 3 ± 6 DOY), while those further north and east spawned progressively later throughout May (Region, 9 ± 7; Region 3, 0 ± 7; Region, ± 7; Region 5, 6 ± ; Region 6, 7 ± ; Region 7, 6 ± 5 DOY). Estimated spawning distributions of pelagic juvenile cod found near Greenland (Region, 3 ± DOY) closely reflected the spawning distributions of those cod found near the main spawning grounds in the south (Region ); indicative of similar spawning origins (Fig. 0). Hence, the overall estimated spawning distribution of pelagic juvenile cod from the main spawning grounds in the south (Region ), across all years, ranged from March to June (99% percentiles: Fig. 0). Assum-

00 Mar Ecol Prog Ser 0: 93 7, 000 Fig. 3. (Above and following pages). Annual relative distribution and abundance (numbers nautical mile ) of 0-group cod Gadus morhua, 970 to 99

Begg & Marteinsdottir: Spawning origins of juvenile cod 0

0 Mar Ecol Prog Ser 0: 93 7, 000 Fig. 3. continued

Begg & Marteinsdottir: Spawning origins of juvenile cod 03 0-group abundance index 0. 0.3 0. 0. 0.0 70 75 0 5 90 95 Year 0-group abundance index 0. 0.3 0. 0. 0.0 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 70 75 0 5 90 95 Year 0-group abundance index 0. 0.3 0. 0. 0.0 5 5 5 5 6 6 5 5 5 5 6 5 5 5 6 6 5 5 5 5 6 5 5 5 5 5 5 65 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 6 6 6 5 5 5 5 5 5 5 6 6 6 6 6 6 6 6 6 6 6 6 6 5 6 6 6 6 56 6 6 5 70 75 0 5 90 95 Year 0-group abundance index 0. 0.6 0. 0. 0.0 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 70 75 0 5 90 95 Year Fig.. Annual abundance indices estimated for 0-group cod Gadus morhua in inshore (continuous lines) and offshore (dashed lines) regions, and all regions combined (S), 970 to 99. Numbers on individual lines = indices for each region ( to ). Statistics describe relationship between abundance index and year 0-group abundance index.5.0.5.0 0.5 0.0 Index =.67 0.03 (Year) ( F = 5.6, p < 0.09, r = 0.6) 70 75 0 5 90 95 Year ing this period to be indicative of cod originating from the main spawning grounds (Region ), considerable proportions of pelagic juvenile cod from the other regions ( to 7) were spawned outside this period, both overall and for most of the individual years, indicating that they most likely originated from other spawning grounds around the country (Table 6). The relative proportion of pelagic juvenile cod estimated to originate from spawning grounds other than the main spawning grounds in the south progressively increased from the west to the east coasts (all years combined: Region, 3%; Region 7, 37%). However, based on a less conservative, although probably more realistic estimate (95% estimated spawning distribution percentiles), the proportion of cod estimated to originate from spawning grounds other than the main spawning grounds in the south was even greater (Region, %; Region 7, 5%: Fig. ). DISCUSSION Historically, variable but in some years relatively large proportions of pelagic juvenile Gadus morhua in Icelandic waters originate from spawning grounds other than those of the main spawning grounds off the

0 Mar Ecol Prog Ser 0: 93 7, 000 Table. Annual mean adjusted (0 August) age (d) of 0-group cod Gadus morhua for each region (a to ), 970 to 99. na = no data available Year Region Total a b a b 3a 3b a b 5a 5b 6a 6b 7a 7b 970 55 na 5 7 6 7 73 7 6 67 na na na na na 75 97 69 69 9 7 7 7 69 na na na na na 9 73 97 03 0 75 96 7 6 73 66 na 6 na na na na na 973 97 93 9 9 67 9 53 69 9 5 5 5 5 5 00 69 97 na 9 76 79 69 77 6 na 57 na na 7 975 97 7 9 6 56 59 60 6 67 60 na na 9 69 976 96 9 7 6 9 3 90 5 9 90 7 3 77 977 na 06 9 75 7 69 67 6 59 63 5 0 na 77 75 97 07 0 9 9 7 93 7 76 75 73 69 7 na na 6 3 979 7 6 63 5 5 50 5 5 na 50 na 7 55 90 73 77 6 6 73 7 5 3 76 na 76 7 75 9 93 96 0 77 7 7 na na na na na 9 9 9 9 9 70 na 7 79 76 67 6 na na na na 7 7 93 95 9 6 75 3 6 0 7 76 7 na 5 na na 0 9 95 75 00 96 9 75 6 7 5 3 6 67 9 9 9 95 95 9 06 6 70 76 69 73 na na na 77 95 96 06 95 90 6 63 79 79 na 7 9 00 97 na 0 77 6 66 67 65 6 67 6 na na 9 73 9 93 0 6 70 6 7 65 7 na na na na 79 7 99 6 5 75 7 7 70 66 69 7 6 na na 5 76 990 79 93 9 0 7 73 66 67 6 6 na na 6 75 99 5 na 73 75 65 6 63 69 na na 0 9 na na 96 67 99 73 56 7 77 69 69 7 60 na 59 6 na na 6 6 993 90 00 73 7 66 6 73 69 6 59 6 6 na na 75 75 99 na na 67 50 5 5 56 63 66 77 7 na 97 0 55 995 7 0 70 5 6 6 66 66 6 5 5 06 6 na 65 996 7 63 73 6 7 6 5 5 na na 5 35 na 7 997 0 9 9 6 79 3 75 79 73 70 73 66 5 na 96 79 99 99 00 9 7 69 7 6 6 56 57 na na 0 75 Total 95 9 6 9 75 0 7 73 69 7 6 7 7 6 93 south coast. Our results contrast with traditional paradigms of cod stock structure and spawning diversity (i.e. Jónsson 9), while advancing the findings of Marteinsdottir et al. (in press b), who first indicated multiple spawning origins of pelagic juvenile cod in Icelandic waters. Consequently, each year class is comprised of progeny originating from multiple spawning components, whose relative contribution varies each year dependent upon stock and environmental factors that influence dispersal and subsequent survival of the early life-history stages (Marteinsdottir et al. in press b). Incorporating spawning diversity and stock complexity into the functional stock and recruitment relationship will explain some of its inherent variability, while reducing the variability associated with biological reference limits and other management strategies that are based on this relationship (i.e. Cook 997). Generally, pelagic juvenile cod are distributed around the west, north, and east coasts of Iceland following spawning of adult cod from March to June. In years when the relative abundance of pelagic juvenile cod was high (i.e. 973, 976, 9 and 95), large aggregations were also found near Greenland, where they no doubt dispersed from the main spawning grounds in the south with the flow of the northern and then western branch of the Irminger Current (Malmberg et al. 996). Pelagic juvenile cod comprising these aggregations are thought to remain in Greenland waters for several years before returning to Icelandic waters as mature individuals to spawn (Schopka 99). In the latter years of the 0-group fish survey (997 and 99), when the relative abundance of pelagic juvenile cod was also high, the apparent lack of cod near Greenland was probably due to inadequate station coverage as opposed to the unlikely absence of fish (i.e. Sveinbjörnsson & Jónsson 99). Concurrent length, age, hatch and spawning distributions of pelagic juvenile cod from the Greenland component (Fig. : Region ) with those of cod from the main spawning component (Region ), provide strong evidence for similar spawning origins. As a result, the Greenland component provides a demographic template that can be compared to the other components (Regions to 7) and used as a measure of natal congruity with those pelagic juvenile cod originating from the main spawning component in the south. Consequently, in any given year, significant proportions of pelagic juvenile cod found off the north and east coasts

Begg & Marteinsdottir: Spawning origins of juvenile cod 05 Inshore Offshore 50 a 00 b 600 a 00 b 00 3a 300 3b 500 a 600 b Frequency 500 5a 300 5b 00 0 6a 6b 30 7a 5 7b 0 0 0 60 0 00 Adjusted TL (mm) 50 Fig. 5. Adjusted (0 August) total length (TL) (mm) distributions of 0-group cod Gadus morhua for each region (970 to 99 data combined). Dashed vertical lines = mean length for each region 0 0 0 60 0 00 Adjusted TL (mm)

06 Mar Ecol Prog Ser 0: 93 7, 000 Fig. 6. (Above and following pages). Annual distribution of 0-group cod Gadus morhua mean-adjusted (0 August) total lengths (mm), 970 to 99

Begg & Marteinsdottir: Spawning origins of juvenile cod 07

0 Mar Ecol Prog Ser 0: 93 7, 000 Fig. 6. continued

Begg & Marteinsdottir: Spawning origins of juvenile cod 09 Table 5. Annual mean hatch and predicted spawning DOY of 0-group cod Gadus morhua for each region (a to ), 970 to 99. na = no data available Year Region Total a b a b 3a 3b a b 5a 5b 6a 6b 7a 7b Mean hatching DOY 970 77 na 7 60 6 5 59 5 6 65 na na na na na 57 97 63 63 3 6 6 60 63 na na na na 0 na 3 59 97 9 3 57 36 5 6 59 66 na 6 na na na na na 5 973 35 39 0 3 65 3 79 63 3 7 7 7 0 3 63 97 na 3 56 53 63 55 70 na 75 na na 5 50 975 35 5 3 6 76 73 7 6 65 7 na na 63 976 36 3 5 6 9 7 3 5 9 55 977 5 na 6 3 57 60 63 65 6 73 69 7 na 55 57 97 5 53 3 5 39 5 56 57 59 63 6 na na 6 9 979 5 50 70 69 7 0 7 7 9 na 5 na 6 77 90 59 55 6 6 59 60 7 5 9 56 na 56 5 57 9 39 36 30 55 5 50 50 5 5 na na na na na 3 3 9 3 6 na 5 5 53 56 65 6 na na na na 60 5 93 37 3 6 57 9 6 5 5 56 6 na na na 50 5 9 37 5 3 36 57 6 5 7 9 6 65 3 3 95 37 3 6 6 6 56 63 59 na na na 56 37 96 6 37 6 69 5 53 53 50 na 60 3 3 50 97 na 50 55 6 66 65 67 7 65 7 na na 3 59 9 5 39 3 6 6 6 60 67 60 na na na na 53 5 99 6 7 50 57 5 60 6 66 63 6 6 na na 7 56 990 53 39 9 5 5 60 59 66 65 7 6 na na 6 57 99 7 na 59 57 67 6 69 63 na na 35 na na 36 65 99 59 76 6 55 63 63 6 7 na 73 7 5 na na 7 6 993 3 59 5 66 7 59 63 6 73 6 6 na na 57 57 99 na na 65 0 0 76 69 66 55 59 na 35 30 77 995 5 6 70 7 66 66 6 7 0 50 6 6 na 67 996 0 5 5 69 59 6 5 70 7 7 na na 7 97 na 5 997 30 0 6 53 9 57 53 59 6 59 66 7 na 36 53 99 33 3 3 6 5 63 60 70 6 76 75 na na 57 Total 37 0 6 57 5 60 59 63 60 6 6 6 6 39 Mean spawning DOY 970 6 na 3 7 5 5 6 5 55 5 na na na na na 97 50 35 50 5 7 50 na na na na 97 na 30 6 97 6 3 5 5 6 53 na 5 na na na na na 3 973 3 7 6 6 66 6 57 6 57 63 5 6 97 05 na 3 5 3 35 0 50 57 na 6 na na 3 37 975 30 3 50 3 6 59 5 5 5 5 na na 7 9 976 9 3 3 7 35 33 30 9 0 07 35 30 977 36 na 9 5 50 53 5 5 59 3 na 0 97 0 3 3 6 39 39 6 na na 3 3 979 6 3 5 50 59 6 63 59 6 7 na 63 3 na 5 90 33 9 0 9 53 5 3 36 3 na 30 9 5 0 36 35 35 3 3 na na na na na 9 7 0 na 0 37 39 5 7 na na na na 6 93 7 30 33 30 36 0 5 na 65 na na 3 36 9 6 0 5 30 9 3 33 30 9 5 7 95 03 3 33 3 7 na na na 9 96 7 3 33 5 36 3 3 35 na 5 3 7 35 33 97 na 0 36 50 5 5 53 57 5 57 na na 5 9 37 5 7 3 3 6 3 53 6 na na na na 39 0 99 3 9 30 33 0 3 5 9 6 7 na na 30 39 990 36 3 3 35 3 9 5 5 na na 9 0 99 3 na 3 5 53 7 na na 06 9 na na 0 9 99 3 60 5 39 7 7 5 56 na 57 55 35 na na 55 993 5 5 9 5 6 7 56 5 7 na na 0 0 99 na na 5 6 66 66 6 7 55 5 5 na 6 63 995 3 6 65 5 55 50 50 5 5 6 3 0 30 na 5 996 95 5 3 56 6 55 5 57 6 65 na na 0 na 5 997 3 35 3 39 35 69 na 35 99 7 6 7 5 35 7 5 5 60 59 na na Total 5 3 6 36 5 3 7 5 7 3

0 Mar Ecol Prog Ser 0: 93 7, 000 Table 6. Proportion of 0-group cod Gadus morhua in each region (a to ) predicted to have originated from main spawning grounds in south (Region ) for main spawning period, 970 to 99. Proportions are based on number of cod spawned at later DOY than that estimated for 99% percentile (59 DOY) and 95% percentile (50 DOY). na = no data available Year Region Total a b a b 3a 3b a b 5a 5b 6a 6b 7a 7b 99% percentile (59 DOY) 970 0.5 na 0.99 0.7 0.6 0.95 0.6 0.9 0.5 0. na na na na na 0. 97 0.9.00 0.9.00 0.95 0.9 0.9 0.90 na na na na.00 na.00 0.93 97.00.00 0.9.00 0. 0.7 0.76 0.67 na 0.7 na na na na na 0. 973.00.00 0.99.00 0.7 0.96 0.39 0.6 0.33 0. 0.5 0. 0. 0.37.00 0.67 97.00 na.00.00 0.95 0.90 0.95 0.73 0.75 0.67 na 0.5 na na.00 0.9 975.00.00 0.96 0.99 0.60 0.90 0.36 0.9 0.5 0.6 0.97 0.6 na na.00 0.6 976.00 0.99 0.99 0.9.00 0.9 0.99 0.9.00 0.9.00 0.6.00 0.9 0.95 0.9 977.00 na.00 0.99 0.9 0.7 0.0 0.65 0.75 0.5 0.75 0.5.00 na.00 0. 97.00.00.00.00 0.9 0.99 0.97 0.99 0.99 0.95 0.90 0.93 na na.00 0.9 979.00 0.9 0.75 0.5 0.9 0. 0.35 0.50 0.0 0.0 na 0.00.00 na 0. 0. 90.00 0. 0.95 0.9 0.6 0.75 0. 0.9 0.99 0.97.00.00 na.00.00 0.9 9.00.00.00 0.9 0.9 0.0 0.9 0.93 0.9 na na na na na 0.99 0.9 9.00.00 0.9 na 0.3.00 0.6 0. 0.67 0.00 na na na na.00 0.7 93.00.00.00 0.0 0.99 0.93 0.7 0.7 0.90 0.55 na 0.5 na na.00 0. 9.00.00.00.00.00 0.3 0.97 0.96 0.9 0.96 0.99 0.9 0.9.00.00 0.9 95.00.00 0.99 0.99 0.9 0.99 0.5 0.77 0.9 0.90 na na na.00.00 0.93 96.00.00 0.9 0.96 0.97 0.75 0.90 0.9.00 0. na 0.9.00.00.00 0.95 97 na.00.00 0.96 0.9 0.7 0.75 0.76 0.6 0.5 0.0 0.75 na na.00 0.0 9.00.00.00 0.3 0.9.00 0.7 0. 0.7 0.7 na na na na.00 0.6 99.00 0.95.00 0.97 0.96.00 0.96 0.0 0.79 0.3.00.00 na na.00 0.95 990.00.00.00.00 0.99.00 0.93 0.9 0.9 0.93 0.3.00 na na.00 0.95 99.00 na.00 0.67 0.6.00 0.7 0.7 na na.00.00 na na.00 0.77 99 0.50 0.5 0. 0.3 0.79 0.7 0.0 0.55 na 0.59 0.66.00 na na.00 0.77 993.00.00 0. 0.95 0. 0.69 0.9 0. 0.0 0.5 0.79 0.0 na na 0.93 0. 99 na na 0.59 0.00 0. 0.0 0.5 0.0 0.5 0.50.00.00 na.00.00 0.3 995 0.99.00 0.90 0.00 0.67 0.6 0.75 0.7 0.75 0.5 0.9 0.9.00.00 na 0.7 996.00.00 0.97 0.67 0.9.00 0.93 0.67 0.0 0.33 na na.00.00 na 0.6 997.00.00.00 0.99 0.99 0.95 0.96 0.96 0.9 0. 0.79 0.6 0.00 na.00 0.9 99.00.00 0.99 0.9 0.7 0.96 0. 0.7 0.6 0.7 0.7 0.6 na na.00 0.5 Total 0.99 0.99 0.97 0.97 0.6 0.9 0. 0.5 0.76 0.77 0.7 0.76 0.63 0.63 0.99 95% percentile (50 DOY) 970 0.00 na 0.9 0.5 0.36 0.7 0.57 0.53 0. 0.3 na na na na na 0.60 97 0.5 0.7 0. 0.9 0.63 0.6 0.66 0.9 na na na na.00 na.00 0.6 97.00.00 0.57.00 0.6 0.5 0.60 0.35 na 0. na na na na na 0.6 973.00.00 0.9 0.99 0.5 0.9 0.0 0.65 0. 0.6 0. 0. 0.9 0.9.00 0.5 97.00 na.00.00 0.63 0. 0. 0. 0.50 0.0 na 0.00 na na.00 0.7 975.00.00 0.7 0.9 0. 0.7 0. 0. 0.6 0.3 0.9 0.3 na na 0.9 0.5 976.00 0.96 0.96 0.9 0.99 0.9 0.96 0.9 0.95 0.9 0.96 0.66.00 0.67 0.79 0.93 977.00 na.00 0.95 0.7 0.63 0.5 0. 0.30 0.6 0.3 0.09.00 na 0.0 0.60 97.00.00 0.99 0.9 0.0 0.97 0. 0. 0. 0. 0.67 0.73 na na 0.9 0. 979.00 0.9 0.39 0.3 0.9 0. 0.07 0.6 0.00 0.00 na 0.00.00 na 0.6 0. 90.00 0.5 0.70 0.79 0.55 0.3 0.77 0.66 0.9 0.6 0.93 0.9 na 0.9 0.9 0.75 9 0.9.00.00 0. 0.9 0.0 0.9 0.7 0.7 na na na na na 0.97 0.9 9.00.00 0.66 na 0.6.00 0.6 0.6 0.0 0.00 na na na na.00 0.6 93 0.9 0. 0.95 0.0 0.9 0.93 0.75 0.65 0.73 0. na 0.5 na na.00 0.7 9 0.96 0.75.00 0.96 0.99 0.67 0.9 0.90 0.5 0.9 0.7 0.7 0.3 0.90.00 0.9 95 0.9.00 0.9 0.95 0.3 0.9 0.60 0.63 0. 0.60 na na na.00 0.9 0.0 96.00.00 0.95 0.93 0.90 0.5 0.0 0.3.00 0.75 na 0.70.00.00 0.97 0.6 97 na.00.00 0.6 0.73 0.6 0.3 0.9 0. 0.39 0.0 0.00 na na.00 0.59 9 0.69.00.00 0.3 0.6.00 0. 0. 0.7 0. na na na na.00 0.6 99 0.96 0.95 0.9 0.9 0..00 0.79 0.60 0.5 0.70 0.50.00 na na.00 0. 990.00.00.00.00 0.95.00 0.6 0.76 0.53 0.73 0.9 0.50 na na 0.97 0.77 99.00 na 0.67 0.67 0. 0.50 0.35 0.57 na na.00.00 na na.00 0.7 99 0.50 0.00 0.60 0.66 0.55 0.6 0.5 0.9 na 0. 0.3.00 na na 0.00 0.50 993.00.00 0.70 0.7 0.5 0.3 0.7 0.5 0.55 0. 0.3 0.60 na na 0. 0.70 99 na na 0.39 0.00 0. 0.00 0.6 0.00 0.5 0. 0.75.00 na.00.00 0. 995 0.9.00 0.5 0.00 0.3 0.3 0.9 0.9 0. 0.5 0.07 0.9.00.00 na 0.5 996.00.00 0.9 0.33 0.7 0.00 0.7 0. 0.06 0.00 na na.00.00 na 0.65 997.00.00 0.99 0.95 0.9 0.5 0.79 0.3 0.7 0.66 0.6 0.59 0.00 na.00 0.3 99 0.99.00 0.95 0.9 0.63 0.5 0.57 0.6 0.35 0.6 0.0 0.5 na na.00 0.67 Total 0.95 0.96 0. 0.90 0.70 0.7 0.6 0.65 0.55 0.60 0.9 0.55 0. 0.7 0.97

Begg & Marteinsdottir: Spawning origins of juvenile cod TL (mm) 90 75 60 5 30 5 30 50 70 90 0 30 Age (d) TL = 5.73 + 0.56(age).50(temperature) + 0.0307(age x temperature) Fig. 7. Total length (TL) (mm)/age (d) relationship of 0-group cod Gadus morhua used to reconstruct historical age at capture for determination of hatching day and spawning origin. Growth relationship based on Marteinsdottir et al. (in press b) for combined 995 (H, J) and 997 (S, D) data for northern [Regions a to 6a: (H, S)] and southern [Region a: (j, D)] areas appear to have originated from spawning components other than those of the main spawning component in the south. In contrast to the expected dispersal patterns of the coastal and Irminger currents (Malmberg et al. 996), the clockwise demographic gradient of pelagic juvenile cod from the west to the east coasts, with progressively decreasing length and age distributions, has previously been considered to be the result of temperature-induced growth and/or incubation differences (Astthorsson et al. 99, Marteinsdottir et al. in press b). However, although there are distinct temperature differences between waters off the north and south coasts of Iceland (Fig., and Malmberg et al. 996), pelagic juvenile cod from the different regions appear to grow at similar rates (Fig. 7, and Marteinsdottir et al. in press b). Certainly, distinct differences in incubation temperatures would prevail between those cod spawned in waters off the north and east coasts and those cod spawned on the main spawning grounds in the south, but is in itself implies the existence of differential spawning components. Those pelagic juvenile cod spawned in the colder waters off the north and east coasts would be assumed to incubate for longer periods, with spawning and subsequent hatching occurring later than in those cod spawned on the main spawning grounds in the south. Our results supported these assumptions, with smaller localized spawning occurring within fjords and inshore waters off the north and east coasts of Iceland. Similarly, Suthers & Sundby (993) indicated that regions of later-localized spawning were responsible for smaller pelagic juvenile cod than those found near the main spawning grounds off Norway. Spawning of cod has also been reported throughout inshore and offshore waters all around Iceland (Sæmundsson 9, Jónsson 9, Marteinsdottir et al. in press a,b). Moreover, results of a mid-summer ichthyoplankton survey conducted in 99 found large numbers of newly hatched larvae in several in-fjord areas west and north of the country (Marteinsdottir et al. unpubl. data). Differential spawning components in regions other than the main spawning grounds in the south have been suggested previously, but have always been considered to be insignificant (Sæmundsson 9, Jónsson 9), and the relative proportions of the different components have never been estimated. In contrast, our study, one of the first to attempt an historical reconstruction of pelagic juvenile-cod spawning-distributions, has provided evidence for differential spawning components and estimated the relative proportions of the different regional spawning components compared to the main spawning component in the south (Table 6, Fig. ). These proportions, however, may have been overestimated if egg drift extended through the colder waters off the northwest coast, which would delay hatching but which was not accounted for in our calculations. Future studies need to collect independent evidence of spawning times and locations around the country in order to confirm our estimates of relative proportions of pelagic juvenile cod originating from the different regional spawning components. In any given year, the relative proportions of pelagic juvenile cod originating from the different regional spawning components fluctuates widely, most likely in response to the inflow of Atlantic water from the main spawning grounds in the south to the North Icelandic shelf. Considerable variation in the Atlantic inflow to the North Icelandic shelf was demonstrated between 90 and 990 (Malmberg & Kristmannsson 99, Kristmannsson 99), while more recently satellite-tracked drifters have also shown considerable inter-annual variation in the strength of the inflow (Valdimarsson & Malmberg 999). Undoubtedly, dispersal and the resultant distribution of cod early life-history stages fluctuates each year, partly in response to variable oceanographic conditions (Anderson & Dalley 997). Future studies need to examine the strength and timing of the Atlantic inflow on an annual basis relative to the proportions of pelagic juvenile cod originating from the different spawning components. In addition, variable stock size and mortality rates of pelagic juvenile cod originating from the different spawning components each year, dependent on temperature and other environmental factors, may also be responsible. Variable mortality rates of progeny originating from the different spawning components most likely also

Mar Ecol Prog Ser 0: 93 7, 000 50 Inshore a 0 Offshore b 00 a 00 b 00 3a 300 3b 500 a 600 b Frequency 500 5a 300 5b 00 6a 0 6b 30 7a 30 7b 0 50 00 50 Adjusted age (d) 00 Fig.. Adjusted (0 August) age (d) distributions of 0-group cod Gadus morhua for each region (970 to 99 data combined). Dashed vertical lines = mean age for each region 0 50 00 50 Adjusted age (d)