How different factors including demography, recombination or genome duplication may impact selection efficacy and the burden of deleterious mutations in different populations is a central question of evolutionary biology and genetics. Here we show that key evolutionary processes, including variation in effective population size through postglacial change in demography and recombination rates have affected the efficacy of selection and impacted the genetic load in Coho salmon (Oncorhynchus kisutch), a widely distributed salmonid species on the west coast of North America. Using whole genome resequencing data from 14 populations at different travelled migratory distances from their southern glacial refugium, we found support for postglacial gene surfing, with reduced Ne at the range recolonization front, thus inducing both a reduction in the efficacy of selection and a surf of deleterious alleles in populations evolving at low Ne. This inference was robust to various proxies of the load. In addition, comparing residual tetrasomic and re-diploidizing regions of the salmon genome, we found support for a prime role of recombination rates in shaping the within-genome variation of the load. Overall, our empirical results are remarkably consistent with expectations under the nearly neutral theory of molecular evolution. We discuss the fundamental and applied implications of these findings for evolutionary and conservation genomics.