Abstract
Genetic diversity is rapidly lost from small, isolated populations by genetic drift. Measuring the level of genetic drift using effective population size (N (e)) is highly useful for management. Single-cohort genetic N (e) estimators approximate the number of breeders in one season (N (b)): a value < 100 signals likely inbreeding depression. Per-generation N (e) < 1000 estimated from multiple cohort signals reduced adaptive potential. Natural populations rarely meet assumptions of N (e)-estimation, so interpreting estimates is challenging. Macquarie perch is an endangered Australian freshwater fish threatened by severely reduced range, habitat loss, and fragmentation. To counteract low N (e), augmented gene flow is being implemented in several populations. In the Murrumbidgee River, unknown effects of water management on among-site connectivity impede the design of effective interventions. Using DArT SNPs for 328 Murrumbidgee individuals sampled across several sites and years with different flow conditions, we assessed population structure, site isolation, heterozygosity, inbreeding, and N (e). We tested for inbreeding depression, assessed genetic diversity and dispersal, and evaluated whether individuals translocated from Cataract Reservoir to the Murrumbidgee River bred, and interbred with local fish. We found strong genetic structure, indicating complete or partial isolation of river fragments. This structure violates assumptions of N (e) estimation, resulting in strongly downwardly biased N (b) estimates unless assessed per-site, highlighting the necessity to account for population structure while estimating N (e). Inbreeding depression was not detected, but with low N (b) at each site, inbreeding and inbreeding depression are likely. These results flagged the necessity to address within-river population connectivity through flow management and genetic mixing through translocations among sites and from other populations. Three detected genetically diverse offspring of a translocated Cataract fish and a local parent indicated that genetic mixing is in progress. Including admixed individuals in estimates yielded lower N (e) but higher heterozygosity, suggesting heterozygosity is a preferable indicator of genetic augmentation.