Ground-Truthing Effective Population Size Estimators: Using Long-Term Population Data from Inland Salmonid Populations

Effective population size (N_e) is a foundational concept in conservation biology, in part due to its relationship to the adaptive potential of populations.  Although N­_e is often estimated for wild populations, particularly using genetic estimators, it is rarely calibrated against actual population estimates (N_c) other than to produce N_e/N_c ratios.  This project used demographic and genetic data for lake trout (Salvelinus namaycush) populations from two intensively-studied lakes within the Experimental Lake Area (ELA) in northwestern Ontario.  Mark-recapture efforts in these lakes over several decades have yielded remarkably precise population estimates, which provide a novel opportunity to “ground truth” demographic and genetic estimators of N_e.  Three different genetic methods of estimating N_e (linkage disequilibrium and heterozygosity excess point estimates, and temporal estimates using changes in allele frequencies over time) were compared against demographic N_e estimates and known population data.  The effect of genetic sample size and the number of microsatellite loci included for N_e estimation was compared against generated N_e values and observed for variation within and among estimators.  Changes in genetic N_e estimates over time were tracked and compared to changes in demographic structure and fluctuating census estimates.  The impact of a population bottleneck on N_e was assessed using demographic and genetic N_e estimators to examine temporal changes in these values, both during population reduction and recovery.  Performance and sensitivity of the genetic N_e estimators to the biological reality of the studied populations (polygamy, iteroparity, and overlapping generations) was also assessed.  Differences among methods of N_e estimation may be substantial when considering demographic and life history parameters of populations. Preliminary findings suggest that genetic sample size has an impact on point estimators such as linkage disequilibrium, with a significant decrease in N_e estimates with decreasing sample size (test range of 5 to 80 individuals).  This was particularly apparent for the study population with a past bottleneck.  This research provides insight into the utility of genetic N_e estimates: performance differences among the tested estimators highlight their potential biases and reliance on different assumptions, and illustrate their potential value and caveats for assessing adaptive potential of wild populations.