4.7 Further developments of indicators
For the indicator ΔH, recommendations for threshold values are based on the general guideline of retention of 95% heterozygosity over 100 years (Allendorf & Ryman, 2002). It should be noted, however, that this guiding principle refers to short term genetic conservation and may need to be revised for long term genetic resilience. Further, detecting changes of the magnitude proposed – i.e. a reduction of 5% over 100 years – can be statistically challenging. Sampling over extended periods (several generations) is warranted, but even then, it may be difficult to obtain a reasonable statistical power for detecting statistically significant changes.
Further work is needed to refine the indicators. The statistical power for detecting various levels of heterozygosity change from typically used sets of markers are needed. Also, N eestimation in non-isolated populations is complex. For instance, no method for assessing inbreeding effective population size (N eI) from genotypic data is currently available for structured populations. In such situations, different types ofN e differ - N eV andN eLD reflect inbreeding rates of local population in isolation, but fail to provide estimates of inbreeding rates for populations in the face of migration (Ryman et al., 2019). If we are able to sample a full metapopulation with all subpopulations at proportions reflecting their contribution to the total system, we can use N eV to get a good estimate of inbreedingN eI. However, if we fail in identifying the full metapopulation, estimates of N eV will underestimate the N eI of the metapopulation (Ryman et al., 2019). As discussed above, we do not think we have been able to completely identify all subpopulations in the present case.
Further, to implement the ΔF ST indicator we used the simplifying assumption of an island model in migration-drift equilibrium to translate temporal changes of F STamong subpopulations into migration rates. Clearly, important future work includes investigating the effects of violating this assumption. The limiting values used here are also highly subjective, and it is unclear if they are sufficient to detect biologically important changes of connectivity. A temporal aspect of the change is also missing. The national indicators are now being applied to more species in Sweden including herring, cod, salmon, and eelgrass (Maria Jansson, SwAM, pers. comm.) and this work will provide more empirical information that can aid in modifying and improving the indicators and limiting values applied for them.