How many and when? Optimising targeted gene flow for a step change in the environment

Abstract

Targeted gene flow is an emerging conservation strategy that involves introducing individuals with particular traits to places where these traits are of benefit. One obvious application is to adapt a recipient population to a known threat, but questions remain as to how best to achieve this. Here, we vary timing and size of the introduction to maximise our objective – survival of the recipient population's genome. We explore a generic population model as well as a specific example – the northern quoll, an Australian marsupial predator threatened by the toxic cane toad. We reveal a trade‐off between preserving the recipient genome and reducing population extinction risk, but key management levers can often optimise this so that nearly 100% of the recipient population's genome is preserved. Any action was better than none but the size of the benefit was sensitive to outbreeding depression, recombination rate, and the timing and size of the introduction.

Effects of learning and adaptation on population viability

“…The model was customized to reflect the life history of quolls and contained a simple quantitative genetic model, allowing for the evolution of a continuous trait that, once beyond a threshold value, confers toadsmart behavior on an individual. The initial additive genetic and environmental variances are identical to those used by Kelly & Phillips (2019b) (Appendix S1). The complete description of the model (including code) is available in Supporting Information (Appendix S1), on GitHub (https://github.com/benflips/CTAquolls), and in Kelly and Phillips (2019b).…”

“…The initial additive genetic and environmental variances are identical to those used by Kelly & Phillips (2019b) (Appendix S1). The complete description of the model (including code) is available in Supporting Information (Appendix S1), on GitHub (https://github.com/benflips/CTAquolls), and in Kelly and Phillips (2019b). The model is motivated by empirical observations that a proportion of the quoll population is innately toad smart and that this trait is heritable (Kelly & Phillips 2017, 2019a) (Table 1).…”

“…For the PVA, we used the discrete-time, individualbased population model described in Kelly and Phillips (2019b). The model was customized to reflect the life history of quolls and contained a simple quantitative genetic model, allowing for the evolution of a continuous trait that, once beyond a threshold value, confers toadsmart behavior on an individual.…”

“…Our base scenario considered a carrying capacity (K) of 1000 individuals (though we also ran a set of scenarios with K = 100), and we initiated our scenarios with an initial fitness (W ) of 0.38, which is the observed proportion of animals that would not take cane toad baits in the field and are, thus, likely to exhibit an innate (genetically based) aversion to consuming cane toads (Indigo et al 2018;Kelly & Phillips 2019a). The genetic heritability of this toad-smart behavior is unknown; although as with many behavioral traits, it likely ranges from 0.1 to 0.3 (Kelly and Phillips 2019b;Roff 2012). We explored outcomes across heritabilities from 0 to 0.5.…”

“…Modeling can provide guidance for successful delivery of management techniques within complex biological systems, saving time and money (McLane et al 2011;Bode & Brennan 2011;Restif et al 2012). A recently developed evolutionary model (Kelly & Phillips 2019b) allows for genetic adaptation and has already been customized to describe the quolls' life history and evolutionary response to toads. We added the possibility of rapid learning and cultural transmission to that model and used it in a PVA framework to determine the conditions under which CTA might enable quoll populations to adapt to the presence of toads.…”