Uncertainty about the magnitude, frequency, location, and timing of the nonlocal harvest of sockeye and chum salmon was the impetus for the Western Alaska Salmon Stock Identification Program. The program was designed to use genetic data in mixed stock analysis to reduce this uncertainty. A baseline of allele frequencies in spawning populations is required for use in mixed-stock analysis to estimate the stock of origin of harvested fish. This report describes the methodology used to understand the population genetic structure among chum salmon populations and to build and test a baseline for use in mixed stock analysis of chum salmon. Of the 35,921 fish from 434 collections selected to be genotyped, the final baseline was composed of 32,817 fish from 402 collections representing 310 populations. Average population sample size was 106 fish. Reporting groups were determined through a combination of stakeholder needs and identifiability using genetic information, as measured using proof tests. The final reporting groups included Asia, Kotzebue Sound, Coastal Western Alaska, Upper Yukon River, Northern District (Alaska Peninsula), Northwest District (Alaska Peninsula), South Peninsula (Alaska Peninsula), Chignik/Kodiak, and East of Kodiak.