Wigley and Serchuk: Spatial and temporal distribution of juvenile Gadus morhua 



601 



distributed clumps) both generated a negative binomial 

 distribution. The second factor is that cod are not 

 fully recruited to the survey sampling gear until age 

 3 or 4 (Serchuk and Wigley 1986). For these reasons, 

 a distribution-free analysis of variance (Kruskal and 

 Wallis 1952) was employed for statistical compari- 

 sons of temperature and depth distributions by age 

 groups. 



Spatial and temporal distribution 



Commercial landings data (see Burns et al. [1983] for 

 a detailed explanation of the commercial catch sam- 

 pling program in the northeastern United States) for 

 scrod cod collected by NEFSC during 1982-86 from 

 the Georges Bank-Southern New England region 

 (NEFSC Statistical Areas 521-526 and 537-539) were 

 examined for spatial and temporal aspects of juvenile 

 cod occurrences. Biases associated with the use of 

 landings data were assumed to be negligible for this 

 highly-directed fishery. Otter trawl catches account for 

 86-90% of the annual total cod landings (Serchuk and 

 Wigley 1986); hence other gear types were excluded 

 from subsequent analysis. 



Fish in the scrod market category weigh 0.7-1.4kg, 

 measure 40-60cm TL, and are 1-3 yr of age according 

 to the growth function developed by Penttila and Gif- 

 ford (1976). 



Geographic resolution within Statistical Areas (SAR) 

 was obtained by assigning the landings within each 

 SAR to 10-min squares of latitude and longitude (~100 

 nm-) based upon information from interviewed (i.e., 

 dockside interviews of captains in which precise catch- 

 location information was obtained) trips landing scrod 

 cod. Scrod landings associated with interviewed trips 

 were prorated upward for each 10-min square by the 

 ratio, derived for each SAR, of total monthly landings 

 to monthly interviewed landings. Prorated scrod land- 

 ings were summarized quarterly for each year for each 

 10-min square. 



Scrod cod landings were partitioned into age-groups 

 1, 2, 3 and 4-i- by constructing a catch-at-age matrix 

 using a technique described by Serchuk and Wigley 

 (1986). In this method, quarterly mean weights for 

 scrod cod were calculated by applying a length-weight 

 equation to quarterly scrod length-frequency data; 

 these means, in turn, were divided into scrod landings 

 for the corresponding quarter to generate numbers 

 landed. Age compositions for scrod landings were 

 derived by applying quarterly age/length keys to 

 numbers at length landed; resulting proportions at age 

 were then applied to prorated scrod landings for each 

 10-min square to obtain estimates of landings by weight 

 for each age-group. 



Results 



Temperature and depth distribution 



Analyses of survey data were based upon 1455 and 904 

 cod collected during spring and autumn bottom-trawl 

 surveys, respectively, during 1982-86. Mean lengths- 

 at-age and associated statistics for cod ages 0-10 are 

 presented in Table 1. Geographically, juvenile cod (de- 

 fined as fish <37cm, the minimum size at first matur- 

 ity; Morse 1979) exhibited seasonal patterns of distribu- 

 tion. In spring, juveniles are dispersed throughout the 

 Georges Bank-Southern New England region (Fig. 

 2A), while in autumn they are concentrated along the 

 100 m contour west of the Great South Channel and 

 the Northern Edge and Northeast Peak of Georges 

 Bank (Fig. 2B). 



Differences in distribution with respect to both mean 

 temperature and depth were noted for all age-groups 

 of cod (Table 2). Mean temperatures were approx- 

 imately 5.3°C in spring and 9.2°C in autumn for all 

 age-groups, despite the fact that there was consider- 

 able overlap in the temperatures for the two seasons 

 (Table 2). However, within seasons, differences in mean 



