560 



Fishery Bulletin 100(3) 



e^-A/'^^V 



155°W 



150^/V 



145°W 



Figure 1 



Bathymetric and geographic features of the Gulf of Alaska between 130°W and 170°W. The black vertical lines separate the 

 five areas used in the survey design and analysis. Thin lines indicate 100 m, 200 m, and 500-m isobaths. The large shaded lines 

 summarize major circulation patterns. 



Materials and methods 



Study area 



Our study area included the continental .shelf and the 

 upper continental slope along the perimeter of the GOA to 

 a depth of 500 m (Fig. 1). The total study area is approxi- 

 mately 300,000 km-, extending over a distance of 2600 km 

 from Southeast Alaska to the western Aleutian Islands. 



Data sources 



All fisheries data for this study were obtained from the 

 Alaska Fisheries Science Center (AFSC), National Marine 

 Fisheries Service (NMFS) in Seattle, WA. Personnel from 

 AFSC have executed bottom trawl surveys of gi-oundfish 

 stocks in the GOA on a triennial basis since 1984 (Table 

 1). All surveys were conducted during the summer with 

 chartered fishing vessels and were based on a stratified 

 random sampling design. The GOA was divided into five 

 areas, three in the western GOA (Shumagin, Chirikof, 

 Kodiak) and two in the eastern GOA (Yakutat, Southeast, 

 see Fig. 1) and four depth strata (0-100 m, 100-200 m, 

 200-300 m, 300-500 m). Depth strata within each area 

 were further subdivided into 2-5 strata based on the type 

 of geographical area (e.g. banks, gullies, shelf, and slope), 

 resulting in a total of 49 strata. Some modifications were 

 made to the stratification scheme after 1987. Our analysis 

 was based on current stratum boundaries. Details of the 



sampling design, data collection, and sample process- 

 ing are described in Brown (1986), Martin and Clausen 

 (1995), Stark and Clausen (1995), and Martin (1997). In 

 our analysis we excluded data from the 0-100 m stratum 

 in the southeast area because it is mostly untrawlable and 

 was not sampled during three out of five surveys; thus 

 we used a total of 48 strata in the analysis. We further 

 excluded all hauls that were classified as unsatisfactory 

 in the database. 



Several tjrpes of fishing gear were used in the bottom 

 trawl sui-veys between 1984 and 1996. A poly-Nor'eastem 

 high opening bottom trawl equipped with rubber bobbin 

 roller gear has become the standard gear for GOA bottom 

 trawl surveys (Munro and Hoff, 1995) and was used for at 

 least part of the survey in all years (Table 1). In addition, 

 an older Nor'eastern trawl (gear type 160) was used during 

 parts of the 1984 and 1990 survey and two different Japa- 

 nese trawls (Table 1) were used during the 1984 and 1987 

 joint Japanese-American surveys. Fishing effort for each 

 gear type was spread out over most of the sampling region 

 within each year and one gear type (gear t3fpe 172) was used 

 during all surveys, allowing us to estimate gear differences. 



Data analysis 



Our analysis was based on standardized catch per unit of 

 effort of 72 fish taxa collected in the bottom trawl surveys. 

 Catches were standardized to weight per unit area (kg/ 

 km-), where the area swept by the trawl was estimated 



