Fritz and Brown: Interactions between the Pacific cod fishery and Steller sea lions 



505 



Bottom Trawl Survey - Pacific cod 

 WGTCPUE 



o Cod CPUE=0 



° Cod CPUE < Mean 



* Mean < Cod CPUE < Mean + 2 SDs 



• Mean + 2 SDs < Cod CPUE < Mean + 4 SDs 

 #Cod CPUE > Mean + 4 SDs 



Figure 3 



Catch per unit of effort (CPUE=kg/km 2 l of Pacific cod {Gadus macrocephalus) during the February- 

 March 2001 bottom trawl survey of the southeastern Bering Sea. "Wgtcpue" refers to the CPUE of 

 Pacific cod from individual hauls (Table 2). Area shading is the same as that in Figure 1. 



the net are captured) and that it is constant over the 

 course of the survey. This assumption is also made in the 

 Leslie analyses of fishery data. In addition, each haul is 

 assumed to be a random, normally distributed estimate 

 of the density of cod within the stratum. Therefore, the 

 average of the haul CPUEs of cod was assumed to be an 

 unbiased estimate of the true density of cod, allowing 

 linear extrapolation from the CPUE within the area 

 swept to a biomass estimate for each stratum. 



Analysis of fishery data 



Fishery observers record a wide variety of information 

 about each haul taken by a fishing vessel, including 

 retrieval location, depth, date and time of catch, and 

 total catch weight (all referred to hereafter as "haul 

 data"). In addition, the catch of a randomly chosen subset 

 of hauls was sampled to determine the species composi- 

 tion of the haul and the length distribution of the target 

 species (see Nelson et al. 1981 and NMFS 12 for observer 

 sampling methods). Observer data were queried for any 



12 NMFS. 1996. Manual for biologists aboard domestic 

 groundfish vessels, 431 p. U.S. Dep. Commer., NOAA. 

 NMFS, Alaska Fisheries Science Center, 7600 Sand Point 

 Way, NE, Seattle, WA 98115. 



hauls with any gear in which Pacific cod were caught in 

 the eastern Bering Sea and Aleutian Islands region in 

 2001. The geographic distribution of the observed Pacific 

 cod catch was used to estimate the distribution of the 

 actual catch of Pacific cod from January-April 2001 

 in four areas of the southeastern Bering Sea (Fig. 1): 

 the high and low sampling-effort areas surveyed in 

 February-March 2001, and two areas outside of the 

 area surveyed — one to the east, and one to the south. 

 To account for Pacific cod catches in both unsampled 

 hauls and on unobserved vessels, the observed catch 

 of cod was multiplied by the ratio of total-to-observed 

 catch by processing sector and gear type (Table 1). For 

 this procedure, the catch of the unobserved portion of 

 the fleet is assumed to be similar to the observed por- 

 tion. Ratios of total-to-observed catch by sector and gear 

 ranged from 1.02 to 33.94, but for the majority of the 

 catch, the ratios were less than 2 (Table 1). 



A simple Leslie analysis of fishery catch and effort 

 data was conducted on data collected by observers on- 

 board vessels targeting groundfish. For the basic Les- 

 lie model (Ricker, 1975; Hilborn and Walters, 1992; 

 Gunderson, 1993) a deterministic linear relationship 

 between CPUE and cumulative catch is assumed: 



C t 



■■qB -qK t , 



