Predicted longer-term genetic consequences of linear transport infrastructure on the impacted population.
Figure 1 shows, for each phase of the two disturbance processes (natural vs linear transport infrastructure), the simulated predicted speed at which genetic diversity will be lost over the next 30 generations from genetic drift alone. Together, it illustrates how rapidly the expected heterozygosity (He) of the small remaining population of koalas located above the linear transport infrastructure could erode from genetic drift alone. After five generations, the predicted loss of expected heterozygosity (He) of the resident koala population impacted by ‘natural’ mortality events is 4.1% (Dataset 2) in comparison to 41.38% (Dataset 3above) and 5.75% (Dataset 3below) for the remaining two subdivided koala populations located above and below the linear transport infrastructure. After ten generations, these predictions are 8.99%, 69.9% and 12.48% respectively.
Predicted effect of longer-term dispersal post-habitat fragmentation on the genetic connectivity and genetic diversity of the impacted population.
Forward time dispersal simulations showed the predicted impact of different dispersal treatments on the genetic connectivity and diversity of the now subdivided koala populations located above and below the linear transport infrastructure (Figure 2 and 3, Table S2-S4). Results from these simulations can be summarised as follow. First, any dispersal treatments ≤ 6 koalas in each direction (e.i. = ≤ 12 koalas per generation in total) would require remediation after 10 generations, given that they would result in Fst greater than 0.05 (Figure 2; e.g. upper 95% CI for 0 koala dispersing treatment = 0.249 and lower 95% CI for 6 koala dispersing treatment = 0.055, Table S2) and an extensive expected loss in genetic diversity (Figure 3 and TableS3). After 10 generations, for instance, the He ± 95% confidence interval (CI) of any dispersal treatments ≤ 6 koalas above (0 koalas dispersing 95% CI = 0.167-0.174; 6 koalas dispersing 95% CI = 0.226-0.231) and below (0 koalas dispersing 95% CI = 0.237-0.241; 6 koalas dispersing 95% CI = 0.245-0.248) were much smaller and showed no overlap with He of the population prior to its subdivision (He = 0.271, Dataset 2, Table 1). An increase to 8 dispersing koalas in each direction per generation (16 koalas in total per generation) would be the minimum required number of dispersing koalas to maintain Fst below 0.05 (95% CI-Fst = 0.045-0.049, Table S2) and a zero change in Shannon’s mutual information (MI) which is weighted more strongly by rare variants. Genetic diversity, however, would still show a sharp decline for the koala population located above the linear transport infrastructure with its He ± 95% confidence interval (0.229-0.233, TableS3) smaller and not overlapping with He of the population prior to its subdivision (He = 0.271, Dataset 2, Table 1). This would still result in an average loss of more than 15% when compared to the genetic diversity (He) prior to the population subdivision. An increase to 16 dispersing koalas in each direction per generation (32 koalas in total per generation) would be best case scenario given that both genetic connectivity and genetic diversity would remain unchanged.