Recent ecological divergence despite migration in sockeye salmon (Oncorhynchus nerka)

Abstract

Ecological divergence may result when populations experience different selection regimes, but there is considerable discussion about the role of migration at the beginning stages of divergence before reproductive isolating mechanisms have evolved. However, detection of past migration is difficult in current populations and tools to differentiate genetic similarities due to migration versus recent common ancestry are only recently available. Using past volcanic eruption times as a framework, we combine morphological analyses of traits important to reproduction with a coalescent-based genetic analysis of two proximate sockeye salmon (Oncorhynchus nerka) populations. We find that this is the most recent (approximately 500 years, 100 generations) natural ecological divergence recorded in a fish species, and report that this divergence is occurring despite migration. Although studies of fish divergence following the retreat of glaciers (10,000-15,000 years ago) have contributed extensively to our understanding of speciation, the Aniakchak system of sockeye salmon provides a rare example of the initial stages of ecological divergence following natural colonization. Our results show that even in the face of continued migration, populations may diverge in the absence of a physical barrier.

Figure 1

Aniakchak National Monument and Preserve (ANMP) showing Aniakchak Caldera and the two study populations.

Figure 2

Size distribution of substrate for the two breeding locations.

Figure 3

Comparison of body depth in beach and outlet males. Body depth was standardized by fitting regressions of body depth as a function of midorbital to hypural length for each population for each year. Then, the residual of body depth for each fish from the appropriate regression line was calculated, and that residual was used to calculate a standardized body depth at the average size. Beach males show consistently deeper bodies each year, though the difference is smaller in 2002.

Figure 4

Average and standard error of standardized egg weight for beach and outlet female sockeye salmon from Surprise Lake. The residuals are taken from regressions of egg weight against midorbital to hypural length.

Figure 5

Posterior probability distributions for time since divergence in years. All three runs of the IM program are illustrated; run 1 is black dashed, run 2 is black, run 3 is gray. The inset is an enlargement of years 1–1000.

Figure 7

Posterior probability of effective population sizes of beach, outlet, and ancestral populations. Solid lines are the average probability for all three runs of the IM program. The fields represent the standard error of the runs for each parameter. The inset shows detail.

Figure 6

Posterior probability of migration rates between populations. The composite figure of all three runs of the migration rate posterior probability distributions shows the beach to outlet migration as a gray line on a black field. The line is the average probability for the three runs, and the black field is the standard error field. The outlet to beach average probability is the thick white line, and the standard error space around the line is depicted by the thin white lines.