Spatio-Temporal Changes in Effective Population Size in an Expanding Metapopulation of Eurasian Otters

Conservation efforts are leading to demographic growth and spatial expansion of some previously endangered species. However, past population bottlenecks or population size fluctuations can have lasting effects on effective population size (N_e), even when census size (N_c) appears large or recovered. The UK metapopulation of Eurasian otters (Lutra lutra) has a well-documented history of population recovery over recent decades, with indicators of presence (faeces and footprints) increasing in distribution and number over successive national surveys. To determine whether this increase in N_c is reflected in increased N_e, we analysed a large-scale microsatellite dataset (21 years: 1993–2014; 407 individuals) for signals of recent N_e change using BOTTLENECK and LDNe, and evaluated potential biases associated with unaccounted spatial genetic structuring and inclusion of admixed genotypes. We obtained clear bottleneck signals in East England, and signals of recent population expansion in Wales and South West England in some analyses, consistent with national otter surveys and recent findings from whole-genome sequencing. Analyses that did not account for spatial genetic structuring yielded strong spurious signals of United Kingdom-wide population expansion, and N_e estimates from these analyses were suppressed by a factor of 3–4. Inclusion of admixed individuals had weaker impacts on N_e estimates, with overlapping 95% confidence intervals from different analyses. Notably, total N_e summed across regions was small and well below the N_e = 500 size deemed necessary for long-term population viability (sum of river basin district groups: 170.6, 95% C.I.: 102.1–348.3). Conclusions drawn from UK otter surveys, which had suggested a robust population close to panmixia, are therefore not supported by our genetic evidence. Our study highlights the value of including genetic monitoring of endangered or recovering species in monitoring plans, while also providing methodologically important information about N_e estimation from real-world datasets.