the major target species of foreign trawl fisheries there. It 

 supports the largest single species fishery in the North 

 Pacific. 



The modem history of groundfish exploitation in the 

 Bering Sea began in 1954, but the fishery for pollock did 

 not begin in earnest until the early 1960s. At that time 

 yellowfin sole, Limanda aspera, declined substantially in 

 abundance (Forrester et al. 1978) and in 1964 

 mechanized techniques for processing pollock into suri- 

 mi (a minced fish product) were successfully imple- 

 mented on Japanese motherships and large factory stern 

 trawlers (Bakkala et al. In press). As a result pollock 

 catches increased more than tenfold between 1964 and 

 1972 (from 175,000 t to nearly 1.9 million t. Table 1). 

 Catches declined steadily thereafter to <1 million t in 

 1977-78 due to declining stock abundance and fishery re- 

 strictions on the catch. This trend in stock abundance 

 was generally monitored by CPUE data from the fish- 

 ery, but the ADI procedure appeared to be a good 

 measure of abundance as well. They both depended on 

 data compiled from the Japanese trawl fisheries in the 

 Bering Sea. 



Table 1. — Annual catch of walleye pollock, in thousand metric tons, 

 in the eastern Bering Sea, 1964-78. 



Data fnr 1964-76 were reported by respective nations to the U.S. Na- 

 tional Marine Fisheries Service. 



Data for 1977-78 were estimated by the U.S. National Marine Fisheries 

 Service. 



"Union oi" Soviet Socialist Republics. 



DATABASE 



ADI = (KA X On/iZA 



Major nations participating in the pollock fishery are 

 Japan, Union of Soviet Socialist Republics, Republic of 

 Korea, and Taiwan (Table 1). Japan has accounted for 

 more than 85"^ c of walleye pollock catches (Forrester et al. 

 1978) and since 1964 has collected more detailed and 

 complete catch-effort records on her fisheries than any 

 other nation. It is this data base of catch-effort records 

 where we place great reliance for monitoring stock abun- 

 dance. The fundamental data collected were daily catch- 

 effort statistics by species or groups, within statistical 

 blocks (1° longitude by '2° latitude, Fig. 1), and by in- 

 dividual fishing vessels and gear type. Vessels were 

 further classified according to 10 tonnage classes from 1 

 to 4,500 gross registered tons (GRT). Major gear types 

 were pair trawls, Danish seines, side trawls, longlines, 

 gill nets, and stem trawls. 



For 1964-72, many daily catch-effort records for in- 

 dividual vessels were not incorporated by the Fisheries 

 Agency of Japan into the data file for economic reasons. 

 Rather, every third record of such records was selected 

 for key-punching. These subsampled records were then 

 extrapolated to represent the total fishery. Since 1973 all 

 catch-effort records have been included in the data base. 

 The total data base is maintained by the Fisheries 

 Agency of Japan. Monthly summaries of the catch-effort 

 information by statistical blocks by fishing gear type and 

 vessel size class were subsequently made available to 

 the United States and Canada as member nations in 

 INPFC. 



AVERAGE DENSITY INDEX 



The ADI equation presented at the 1974 INPFC 

 meeting by Japanese scientists^ was as follows: 



where i = a statistical block number, 



A = area (km-) of each statistical block , 

 = catch per unit effort in metric tons per hour 

 trawled within each statistical block. 



The application of this procedure for walleye pollock 

 involved some selection of catch-effort data from pair 

 trawlers which they determined to be the most consistent 

 gear type directed to pollock fishing activities in each 

 statistical block. Statistical blocks used in the calcula- 

 tions were those where the daily pollock catch was 

 greater than that of any other species. From these 

 selected "pollock-majority" catch-effort data an ADI was 

 computed for an area made up of several statistical 

 blocks. 



TOPICS OF ANALYSIS 



Objectives of our study were to evaluate the data base 

 and procedures used for calculating ADIs for pollock. 

 Specifically, we chose four topics to: 



1) determine if bias towards high or low catches was 

 introduced in the data base by subsampling of daily 

 data records during 1964-72; 



2) conduct variability studies in pair trawl data used 

 for ADI and CPUE computations; 



3) compare pollock ADIs calculated from all-data and 

 pollock-majority-data; and 



4) identify area-time cells for more accurate assess- 

 ment of pollock abundance. 



'Fishery Agency of Japan. 1974. Pollock stocks in the eastern Bering 

 Sea. Unpubl. manuscr.. 38 p. Far Seas Fish. Res. Lab., 1000 Orido, 

 Shimizu 424, Japan. 



