Estimating the extent of dispersal required to maintain or restore genetic connectivity and diversity between the subdivided populations.
Forward time simulations were programmed in PYTHON using the package SIMUPOP (Peng & Kimmel 2005), http://simupop.sourceforge.net/]. The aim of these simulations was to estimate what amount of movement would be needed to maintain genetic connectivity between the above and below koala populations located on either side of the newly built rail line. These simulations randomly selected two koalas (one male, and one female) from within each of the two populations (above and below) and bred them to create an offspring for the next generation. This process was then repeated until the number of offspring matched the number of parents in each population (therefore, our simulations maintained a stable population size) for every generation (Peng & Kimmel 2005). The mating structure of the simulations was set to mimic the mating system of the impacted koala population where 30% of the males (Schultz et al. 2020) and 90% of the females (Beyeret al. 2018) mate per generation. To mimic this, each generation of the simulation only allowed a random 30% of male koalas and 90% of female koalas to breed and contribute to the genetic make-up of the next generation of koalas. The simulations had overlapping generations, with each koala given a 20% change to persist through to the next generation (up to a maximum of 2 generations). Ten generations were simulated. The simulations imported the genetic data from the two post-construction datasets (Dataset3above = 27 koalas and Dataset3below = 75 koalas) as a starting point. This shared starting point of the simulations means that genetic differentiation between each side of the linear transport infrastructure (above and below), and the genetic diversity within each of those, was the same at the start of every simulation. Genetic differentiation was measured at each simulated generation using Fst, Gst and Mutual Information (Sherwin et al. 2006; Jost 2008; Sherwin 2010; Sherwin et al. 2017; Jost et al. 2018). Based on the imported genetic data, the starting point of these connectivity measures were: Fst = 0.005, Gst = 0.006 and Mutual Information = 0.003. Genetic diversity was measured at each simulated generation using expected heterozygosity (He) and Shannon’s information. We then further simulated 12 dispersal treatments by increasing the number of koalas that dispersed from each side of the linear transport infrastructure per generation: (1) zero koalas = no dispersal (2) 1 koala, (3) 2 koalas, (4) 4 koalas, (5) 6 koalas, (6) 8 koalas, (7) 10 koalas, (8) 12 koalas, (9) 14 koalas, (10) 16 koalas, (11) 18 koalas, and (12) 20 koalas. In the simulations, dispersal occurred before breeding. The simulated dispersal between the subdivided populations caused the population size of each subpopulation to become equal over time. Each dispersal treatment was simulated for 10 generations and replicated 1000 times.
RESULTS