546 



Fishery Bulletin 89(4), 1991 



anomalous low temperatures may have played a part 

 in the formation of the strong 1971 and 1975 year- 

 classes. Conversely, the relatively warm temperatures 

 since 1980, perhaps peaking in 1984 (Royer 1989), may 

 explain the lack of large year-classes in later years 

 (Figs. 2, 3). It is possible that surveys of larval or 

 juvenile abundance and related environmental param- 

 eters, mainly temperature, could be used to forecast 

 future abundance trends. 



Extreme variability in year-class strength can mean 

 the success or failure of the commercial fishery. The 

 1971 year-class was dominant in commercial catches 

 in Pavlof Bay during at least five fishing seasons 

 (Fig. 2B). To calculate the contribution of dominant 

 year-classes to the commercial catch, an average total 

 weight, calculated from Pavlof Bay P. borealis length- 

 weight data, W = 0.00104 CL 2 - 70160 (Anderson 1981) 

 was multiplied by the estimated number caught in 

 each year-class (Table 2). During the years 1974-78, 

 the 1971 year-class contributed about 70% and the 

 1975 year-class about 3% of the 12,384 metric tons 

 of P. borealis harvested from Pavlof Bay. Although 

 the commercial fishery was closed in Pavlof Bay from 

 1979 to 1986, little or no improvement in stock con- 

 dition occurred. In Pavlof Bay, it appears that the 

 P. borealis fishery was largely supported by a single 

 year-class. 



Estimates of growth 



Estimates of growth parameters in this study were 

 generated only from the two dominant year-classes 

 that could be followed through a time series (Fig. 2). 

 Growth estimates depend heavily on the occurrence 

 and definition of modes. Pandalus borealis have a syn- 

 chronous and relatively abbreviated hatching period 

 which gives rise to fairly well-defined size modes short- 

 ly after settlement of juveniles (Frechette and Parsons 

 1983). Survey sampling was conducted in August- 

 September, toward the end of the period of rapid sum- 

 mer growth. Studies that have continuously sampled 

 throughout the spring and summer show growth slows 

 in late summer, possibly as the result of spawning. 

 Some instar growth is possible, however, even during 

 mating and spawning, for the more frequently molting 

 young males. I interpret the double-spike top of the 

 1971 year-class depicted in the 1973 survey data (Fig. 

 2 A) as possibly representing year-class instar growth. 

 As the shrimps age and transform to females, molting 

 is reduced to perhaps two times a year, into and out 

 of breeding dress (Allen 1959). Mode definition for 

 year-classes after they become female is therefore not 

 beset with an interpretation problem resulting from in- 

 star growth. The additional problem of overlap with 

 adjacent but minor modes, especially with slower- 



O 



X 



w 



i_ 



CD 

 -Q 



E 



Z3 



0.8 

 0.6 

 0.4 



0.2 



0.8 

 0.6 

 0.4 

 0.2 



0.8 



0.6 

 0.4 



0.2 



1985 



1984 



1982 



10 15 20 26 



Carapace Length (mm) 



Figure 3 



Size-frequency polygons of Pandalus borealis 

 from Pavlof Bay, Alaska, in catch per kilo- 

 meter from annual surveys, 1982-86. 



growing females, is not so acute when using only domi- 

 nant year-classes for growth estimates. 



Separate von Bertalanffy growth curves were fit to 

 average size-at-age data for the two dominant year- 

 classes (Fig. 4). Parameters of the fitted relationship 

 were 1^ = 29.64, K = 0.16, and t (l = -1.30 for the 

 1971 year-class, and L 00 = 26.31, K = 0.29, and 

 t = -0.47 for the 1975 year-class. A Friedman two- 

 way analysis of variance by ranks (Connover 1971) 

 showed that members of the 1975 year-class were 

 significantly (P<0.001) larger for a given age, indi- 



