Hoff Biodiversity as an index of regime shift in tlie eastern Bering Sea 



227 



Survey area in the 

 stations from 20 

 outer domains. 



eastern Bering Sea continental 

 shelf survey have been standard- 

 ized since 1982. Prior to 1982, a 

 400-mesh eastern trawl (smaller 

 than the current 83-112 trawl) was 

 used (Bakkala, 1993). Although 

 the biomass estimates produced 

 by the two trawls are not directly 

 comparable because of unknown 

 catchability differences between 

 the nets, survey data from 1975 

 and 1979-81 were included in the 

 present analysis because of the 

 importance in the timing of the 

 1970s regime shift. The biodi- 

 versity indexes are based on the 

 relative proportion of the species 

 and therefore are an indicator of 

 assemblage structure. 



Data used for this study were 

 collected by the Resource Assess- 

 ment and Conservation Engineer- 

 ing (RACE) division of the AFSC, 

 which surveys the EBS shelf each 

 summer (May- August). The sur- 

 vey area extends from the Alaska 

 Peninsula north to Nunivak Is- 

 land and St. Matthew Island, and 

 west to the 200 m shelf break. 

 Trawl hauls were conducted on a 



grid of 356 fixed stations (20 nmi by 20 nmi) (Fig. 1) 

 during daylight hours. Hauls were towed for 30 min at 

 3 knots and ranged in depth from 15 m in Bristol Bay 

 to nearly 200 m near the shelf edge. Most trawling 

 was conducted with the AFSC 83-112 eastern trawl 

 (1982-2002), which is a low-opening two-seam trawl 

 with a 26.5-m headrope and 34.1-m cable footrope 

 wrapped with rubber striping and chain hangings that 

 contact the bottom while the trawl is towed (Rose and 

 Walters, 1990). 



Height and width of the net were measured with an 

 acoustic SCANMAR (Scanmar, Asgardstrand, Norway) 

 or NETMIND (Northstar Technical Inc., St. John's, NF, 

 Canada) net mensuration system, or estimated by us- 

 ing a function that relates trawl widths to tow depths 

 from measured hauls. Each haul was measured with 

 GPS or LORAN to record latitude and longitude data 

 at the start and end of the trawl in order to determine 

 distance fished. 



Processing of the catch was done entirely in the field 

 at the time of capture. The entire catch was sorted to 

 species, enumerated, and weighed, or a weighed sub- 

 sample was used for very large catches. 



Catch per unit of area (CPUE) was determined for 

 each trawl station completed during each survey. Bio- 

 mass estimates for each species were then calculated 

 by expanding the average CPUE for each stratum and 

 then summing over all strata of the total survey area 

 to obtain a biomass estimate in metric tons for each 

 species. 



Figure 1 



eastern Bering Sea. Crosses represent annual trawl survey 

 200 m. Bathymetry lines delineate the inner, middle and 



EBS shelf domains and study area 



The EBS is composed of three well-defined regions des- 

 ignated as the inner, middle, and outer domains from 

 east to west, respectively, across the shelf (Fig. 1). The 

 three domains are distinct regions characterized by 

 depth, water temperature, current flow, summertime 

 primary production, and species composition (Bakkala, 

 1993; Schumacher and Stabeno, 1998; Stabeno et al., 

 2001). Briefly, the inner domain is a relatively shal- 

 low (<50 m) and well-mixed warm basin with strong 

 influences from a large coastal area, several major 

 river systems, and a current flow northward along the 

 coast from the Aleutian Islands. The middle domain is 

 a relatively stagnant area of deeper water (50-100 m) 

 that has strong summer water-column stratification and 

 low current flows. The bottom water mass is relatively 

 cold, overlaid by a layer of warmer wind-mixed water. 

 Although characterized as a cold region, the water 

 mass in the middle domain varies from year to year. 

 The outer domain is influenced by the EBS slope region 

 and by upwelling and northward current flow from the 

 Aleutian Islands. This domain is relatively warm when 

 compared to the middle domain (see Hunt et al., 2002, 

 for review). For the purposes of this study the inner 

 domain designation consisted of trawl stations less 

 than 50 m in depth; the middle shelf designation was 

 for trawl stations between 51 and 100 m in depth; and 

 outer domain designation was for trawl stations between 

 101 and 200 m in depth. 



