Immature chum salmon were collected by the F/V Northwest Explorer between September 5 and October 8, during the 2002 BASIS survey across the eastern Bering Sea shelf and Aleutian Islands (for details, see Murphy et al. 2003). Approximately 1,600 fish were aged, checked for the presence of hatchery thermal marks, and genotyped for allozyme loci. Scale aging and otolith mark identification were done by the Alaska Department of Fish and Gameb??s Mark, Tag, and Age Laboratory in Juneau, Alaska. Otoliths with thermal marks were compared with voucher specimens to verify hatchery of origin. Heart, liver, and muscle tissues were extracted and then analyzed with protein electrophoresis to identify genotypes for the 20 allozyme loci in the chum salmon coastwide genetic baseline (Kondzela et al. 2002). Genetic data were pooled into one of four geographic areasb??western Aleutian Islands, eastern Aleutian Islands, southeastern Bering Sea shelf, and northeastern Bering Sea shelf. In the eastern and western Aleutian Islands, the catches were large enough to further stratify the data by ocean age. Regional origin estimates were made for each mixture collection using a conditional maximum likelihood method (Pella and Masuda model in SPAM v. 3.7, ADF&G 2001) and the full 356-population genetic baseline. The 95% nonsymmetric confidence intervals were determined from 1000 bootstrap estimates in which the baseline and mixture were re-sampled.

Genetic data were collected and prepared with the use of protein electrophoresis from 52 spawning locations in southeastern Alaska and northern British Columbia. Genetic relationships were examined from principal components analysis and unrooted trees constructed from genetic distances between collections. These descriptive analyses suggest a geographic basis to genetic divergence among populations. This geographic basis was confirmed using log-likelihood-ratio analysis and analyses of variance. Three groups of populations were observed: one from systems that drain into the inside waters of northern and central southeast Alaska; another from the far southeastern islands (including Prince of Wales Island); and the third in systems of the southern inside waters. Although the geographic structure was a statistically significant component of the overall genetic structure, gene diversity analysis indicates that only about 4.7% of the total genetic variability was attributable to genetic differences among those regions, whereas about 8.4% of the total was due to differences among populations within each region. The other 87.0% of the variation occurred, on average, within each collection.

Allozyme variation was used to examine population genetic structure of adult chinook salmon, Oncorhynchus tshawytscha, collected between 1988 and 1993 from 22 spawning locations in Southeast Alaska and northern British Columbia. Thirty-five loci and two pairs of isoloci were variable, and of these, 25 loci and one pair of isoloci expressed the most abundant allele with a frequency of less than or equal to 0.95 in at least one collection. Aneighbor-joining (NJ) tree of genetic distances defined five regional groups: (1) King Salmon River (the only island collection), which has large allelic frequency differences from other populations in this study; (2) heterogeneous coastal populations from southern southeast Alaska; (3) transmountain collections from the Taku and Stikine Rivers on the eastern side of the coastal mountain range; (4) Chilkat River in northern Southeast Alaska; and (5) northern coastal Southeast Alaska, which consists of the Situk River and the Klukshu River, a tributary of the Alsek River. A second NJ tree that included collections from the Yukon River and British Columbia did not reveal any strong genetic similarity between Southeast Alaska and the Yukon River. The data suggest that Southeast Alaska may have been colonized from both northern and southern refugia following the last glaciation b?? a period of sufficient time to allow for isolation by distance to occur.

Summer surveys (Julyb??August) of juvenile salmon ecology along the continental shelf of the Gulf of Alaska are conducted annually by scientists from the Ocean Carrying Capacity program of the National Marine Fisheries Serviceb??s Auke Bay Laboratory. These surveys are an effort to link changes in salmon production to biological and physical factors in the ocean environment. An improved understanding of salmon distribution is one objective of this research. We identified the origin of juvenile chum salmon collected in transects from around the Gulf of Alaska in 2000 and 2001, using the presence of thermal marks in hatchery fish and the divergence of genetic characteristics among regional groups of populations.

Genetic stock identification techniques were used to identify the origin and provide stock-specific migration and distribution patterns of juvenile chum (Oncorhynchus keta) salmon caught during annual fall surveys (2002) along the eastern Bering Sea (Fig. 1). Preliminary results indicate that: 1) Yukon River Fall chum salmon are widely distributed from offshore of the Yukon River, eastward to 62B0N, 172B0W, and as far south as Nunivak Island (60B0N), suggesting a southwesterly migration pathway along the Bering Sea shelf; 2) juvenile chum salmon from the Kuskokwim River are narrowly distributed south of Nunivak Island from the mouth of the Kuskokwim River, south to 58B0N, and as far west as 168B0W, suggesting a westerly migration pathway along the Bering Sea shelf; and 3) northern Russia juvenile chum salmon stocks (mainly stocks from rivers draining into the Gulf of Anadyr) are distributed as far east as 62B0N, 171B0W (Fig. 2). These results are unique in that they represent the first attempt to identify early marine distribution and migration of juvenile chum salmon stocks on the eastern Bering Sea shelf.

A genetic analysis of samples from the chum salmon (Oncorhynchus keta) bycatch of the 2005 Bering Sea walleye pollock (Theragra chalcogramma) trawl fishery was undertaken to determine the overall stock composition of the sample set. Samples were genotyped for eleven microsatellite markers and results were estimated using the current chum salmon microsatellite baseline. In 2005, genetic samples were collected opportunistically as part of a special project and supplemented with archived scales from the Observer Program. Sample biases have the potential to affect stock composition analysis results; consequently, stock composition estimates apply to the sample set and may not represent the entire chum salmon bycatch. Based on the analysis of 1,084 chum salmon bycatch samples collected throughout the 2005 Bering Sea trawl fishery, East Asian (29%), North Asian (29%), Pacific Northwest (19%) and Western Alaska (16%) stocks dominated the sample set with smaller contributions from Southwest Alaska (<2%) and the Upper/Middle Yukon River (5%) stocks. The estimates for the 2005 chum salmon bycatch sample set were similar to the 1994-1995 chum salmon bycatch estimates, suggesting consistency of the regional stock contributions across years. Analysis of temporal groupings within the groundfish B season revealed changes in stock composition during the course of the season. Whether the decreasing proportional contributions of Western Alaska and Upper/Middle Yukon stocks and increasing proportional contributions from Asia over time are due to temporal or spatial differences in the sample set are unknown.

A genetic analysis of samples from the chum salmon (Oncorhynchus keta) bycatch of the 2007 Bering Sea walleye pollock (Theragra chalcogramma) trawl fishery was undertaken to determine the overall stock composition of the sample set. Samples were genotyped for eleven microsatellite markers and results were estimated using the current chum salmon microsatellite baseline. In 2007, genetic samples were collected opportunistically as part of a special project, but sample biases have the potential to affect stock composition analysis results. Consequently, stock composition estimates apply to the sample set and may not represent the entire chum salmon bycatch. Based on the analysis of 1,279 chum salmon bycatch samples collected throughout the 2007 Bering Sea trawl fishery, North Asian (35%), East Asian (34%), and Eastern Gulf of Alaska/Pacific Northwest (16%) stocks dominated the sample set, with smaller contributions from Western Alaska (10%) and Upper/Middle Yukon River (5%) stocks. The estimates for the 2007 chum salmon bycatch sample set were similar to the 2005, 2006, and 2009 chum salmon bycatch estimates, suggesting consistency of the regional stock contributions across years. Analysis of temporal groupings within the groundfish B season revealed changes in stock composition during the course of the season with decreasing contribution of Upper/Middle Yukon stocks over time, but leaves unanswered whether these changes are due to temporal or spatial differences in the sample set.

A genetic analysis of samples from the chum salmon (Oncorhynchus keta) bycatch of the 2006 Bering Sea walleye pollock (Theragra chalcogramma) trawl fishery was undertaken to determine the overall stock composition of the sample set. Samples were genotyped for eleven microsatellite markers and results were estimated using the current chum salmon microsatellite baseline. In 2006, genetic samples were collected opportunistically as part of a special project, but sample biases have the potential to affect stock composition analysis results. Consequently, stock composition estimates apply to the sample set and may not represent the entire chum salmon bycatch. Based on the analysis of 1,367 chum salmon bycatch samples collected throughout the 2006 Bering Sea trawl fishery, North Asian (31%), East Asian (29%), and Pacific Northwest (25%) stocks dominated the sample set with smaller contributions from Western Alaska (8%) and Upper/Middle Yukon River (6%) stocks. The estimates for the 2006 chum salmon bycatch sample set were similar to the 2005 and 2009 chum salmon bycatch estimates, suggesting consistency of the regional stock contributions across years. Analysis of temporal groupings within the groundfish B season revealed changes in stock composition during the course of the season with decreasing contribution of Upper/Middle Yukon stocks over time, but leaves unanswered whether these changes are due to temporal or spatial differences in the sample set.

A genetic analysis of samples from the chum salmon (Oncorhynchus keta) bycatch from the 2011 Bering Sea walleye pollock (Theragra chalcogramma) trawl fishery was undertaken to determine the overall stock composition of the sample set. Samples were genotyped for 11 microsatellite markers and results were estimated using the current chum salmon microsatellite baseline. In 2011, genetic samples were collected systematically as part of a special project to reduce sample biases that exist in collections from previous years that have the potential to affect stock composition analysis results. One genetic sample was collected for every 31.1 chum salmon caught in 97% of the midwater trawl fishery that was sampled. Evaluation of sampling based on time, location, and vessel indicated that the genetic samples were representative of the total bycatch. Based on the analysis of 1,472 chum salmon bycatch samples collected throughout the 2011 Bering Sea trawl fishery, the Eastern Gulf of Alaska (GOA)/Pacific Northwest (PNW) stocks dominated the sample set (38%), with moderate contributions from East Asian (17%), North Asian (18%), and Western Alaska (16%) stocks, and smaller contributions from Upper/Middle Yukon River (9%) stocks. The estimates for the 2011 chum salmon bycatch sample set differed from the 20052010 estimates, indicating a change in the consistency of the regional stock contributions across the previous 6 years, possibly due to the larger proportion of bycatch caught later in the season and in the more southeastern NMFS reporting areas in 2011. There were significant spatial differences in stock distribution, with the Asian stocks dominating the central Bering Sea area and the Eastern GOA/PNW stocks dominating the southeastern Bering Sea. Analysis of temporal groupings revealed changes in stock composition during the course of the season with decreasing contribution of East Asia and Upper/Middle Yukon stocks and increasing contribution of Eastern GOA/PNW stocks over time.

A genetic analysis of samples from the chum salmon (Oncorhynchus keta) bycatch of the 2009 Bering Sea walleye pollock (Theragra chalcogramma) trawl fishery was undertaken to determine the overall stock composition of the sample set. Samples were genotyped for eleven microsatellite markers and results were estimated using the current chum salmon microsatellite baseline. In 2009, genetic samples were collected opportunistically as part of a special project, but sample biases have the potential to affect stock composition analysis results. Consequently, stock composition estimates apply to the sample set and may not represent the entire chum salmon bycatch. Based on the analysis of 1,437 chum salmon bycatch samples collected throughout the 2009 Bering Sea trawl fishery, North Asian (30%), East Asian (34%), and Eastern Gulf of Alaska/Pacific Northwest (18%) stocks dominated the sample set, with smaller contributions from Western Alaska (13%), Upper/Middle Yukon River (2%), and Southwest Alaska (4%) stocks. The estimates for the 2009 chum salmon bycatch sample set were similar to the 2005 chum salmon bycatch estimates.