484 



Fishery Bulletin 97(3), 1999 



egon fishery, although CaHfornia still requires a mini- 

 mum codend mesh of 34.9 mm. Prior to 1978, shrimp 

 growth was slower (Hannah and Jones, 1991) and 

 age-1 shrimp in the study area (Fig. 1) were clearly 

 not fully recruited to trawl gear until very late in 

 the season (ODFW^ ). After 1978, shrimp growth in- 

 creased, possibly owing to reduced density from 

 heavy fishing, although other factors were also in- 

 volved, notably accelerated sex change that may have 

 been a response to reduced numbers of older shrimp 

 (Charnovetal., 1978; Hannah and Jones, 1991 1. Av- 

 erage codend mesh size in the Oregon fieet also de- 

 creased sometime during the 1970s, averaging 30.2 

 mm by 1981, decreasing to 28.7 mm by 1991-92 

 (Jones et al., 1996 ). Accordingly, in this study I chose 

 the August-September time period to index the age-1 

 shrimp population, and restricted my analysis to the 

 years after 1979. With the months of August-Septem- 

 ber, the population was indexed eight months earlier 

 than in the previous study, but at a time period late 

 enough that age-1 shrimp could reasonably be assumed 

 to be fully recruited to the trawl gear being fished. 



To index shrimp recruitment in the present study, 

 I used the average August-September CPUE for 

 age-1 shrimp only as an index of density and multi- 

 plied this index by stock area (ha) for the same year 

 class (Hannah, 1997). In this case, CPUE was ex- 

 pressed as shrimp per hectare trawled, using an av- 

 erage estimate of 5.93 ha trawled per single-rig 

 equivalent hour fished (Hannah, 1995). To account 

 for years in which early season catches of age-1 

 shrimp were high, I added in the catch of age-1 

 shrimp for the months of April-July of the same year. 

 Thus, recruitment was calculated as 



R, = iD,A^) 



+ C, 



(3) 



where R 

 D 



recruitment of age-1 shrimp in year /; 



the average fishei-y CPUE of age-1 shrimp 



in August and September of year t\ 

 A = the stock area for the age-1 year class in 



year t\ and 

 C = the summed fishery catch of age-1 



shrimp in the months April to July of 



year t. 



Use of these additional catches for April-July does 

 rely on an assumption of constant natural mortality, 

 although the importance of this assumption is clearly 

 reduced. Age-1 shrimp CPUE for Pacific States Ma- 

 rine Fisheries Commission (PSMFC) statistical ar- 

 eas 82-88 was obtained from Zirges et al. ( 1982) and 



' ODFW (Oregon Department of Fish and Wildlife i. 

 present. 2040 SE Marine Science Dr., Newport, OR. 

 data. 



Stock 

 Unit 



Washington 



Oregon 



Ocean shrimp stock area - 1984 

 Ocean shrimp stock area - 1988 



California 



1981 to 

 Unpubl. 



Figure 1 



Location of commercial concentrations of ocean shrimp (Pandalus 

 jordani) off coastal Oregon, in Pacific States Marine Fisheries 

 Commission statistical areas 82-88. Dark areas show the approxi- 

 mate minimum areal extent of the shrimp grounds (1984) and 

 the lighter shaded areas show the largest areal extent observed 

 from 1980 to 96 1 1988). 



Hannah et al. ( 1997). The collection and analysis of 

 biological samples from the commercial catch has 

 been described by Hannah and Jones ( 1991 ). 



The data used to construct the recruitment and 

 spawning stock indices developed in this study dif- 

 fer in some respects from the data used for the prior 

 recruitment study. In the prior study, logbook data 

 were available from Washington, Oregon, and Cali- 

 fornia, such that a complete accounting of catch and 

 effort was possible for PSMFC areas 82-92 (Fig. 1). 

 Beginning in 1992, logbook data from the states of 

 California and Washington became unavailable. Ac- 

 cordingly, the data used in this study was based on 

 Oregon landings for the entire study period. To mini- 

 mize the impact of the missing information on the 

 indices developed, the study area was limited to ar- 

 eas 82-88 (Fig. 1). I believe that the missing infor- 

 mation will create very minimal error in the new 

 indices for several reasons. First, Oregon landings 

 comprise the great majority of the removals of shrimp 

 from statistical areas 82-88. The average percent- 



