Figure 3.

Minimum HD allele frequency in the first 100 generations after RD release for various fitness costs, initial conditions, and release ratios. Light shades indicate higher likelihood of stochastic loss of HD, while dark shades highlight instances where removal of the HD is less likely. Axes show fitness costs of the HD (x-axis) and RD (y-axis). Initial conditions vary between panels: columns vary the HD pre-release frequency, and rows vary the RD release size, which is shown as a release ratio (e.g., “4 to 1” releases 4 RD alleles for every pre-release allele). We assume recessive fitness costs and perfect homing. The largest HD fitness cost (s _HD = 1) corresponds to a suppression HD, whereas small HD fitness costs correspond to a replacement HD. Note that maximum HD frequency varies independently from minimum HD frequency; in small RD releases (top row of panels), the HD frequency can experience large increases before dropping to the low minimum levels show here. Overall, an RD release appears least likely to eliminate a target HD when RD fitness costs are large, and when the RD release yields post-release frequencies near the equilibrium. The higher minimum frequency for larger HD fitness costs in many panels is due to the smaller amplitudes of oscillations compared to systems with lower HD fitness costs, as seen Figs 1 and 2. Smaller oscillations result in the system tending directly toward the equilibrium.