Sampling Strategies for Brook Trout Effective Population Size

@font-face { "Times"; }@font-face { "Cambria"; }p.MsoNormal, li.MsoNormal, div.MsoNormal { margin: 0in 0in 10pt; font-size: 12pt; ""; }div.Section1 { page: Section1; }

The goal of many conservation efforts for headwater fishes is to maintain robust (genetic and demographic) populations by increasing connectivity and habitat patch size. Collecting unbiased estimates of effective population size (N_e) is critical for monitoring these conservation actions.  When the species of concern is iteroparous and has overlapping generations, single population samples of young-of-the-year (YOY) provide estimates of the effective number of breeders (N_b) that gave rise to that cohort.  Recommendations for appropriate sample sizes and sampling strategies to obtain precise and unbiased estimates of N_b are currently unavailable.  We used empirical microsatellite data from three headwater populations of brook trout in Virginia along with a simulation-based approach to test the effects of different combinations of sample sizes and subsampling strategies on estimates of N_b. The empirical data were obtained from comprehensive samples of YOY (mean N = 480) collected throughout the entire headwater reaches of three streams.  Stream-specific simulations were parameterized from the empirical genetic data.  We conducted subsampling routines to obtain multiple N_b estimates for both the empirical and simulated data. We considered subsample sizes of 25, 50, 75, 100, 125, 150, and 200.  We considered the following subsampling strategies: (1) sample continuously from one starting point, (2) divide the stream reach of interest in half and sample continuously from two starting points, (3) divide the stream reach of interest into thirds and sample continuously from three starting points, (4) divide the stream reach of interest into quarters and sample continuously from four starting points. The optimal amount of sampling effort (number of sampling locations within a stream and sample size) that led to precise and unbiased estimates of N_b occurred with samples sizes of 75 and division of the stream into thirds. This strategy provides a large enough sample size while also ensuring that samples contain individuals from a high proportion of available families. We recommend this level of sampling effort and a focus on estimating the effective number of breeders for genetic monitoring of iteroparous species with overlapping generations. Our results for headwater eastern brook trout populations should apply to other headwater fish species with similar population and family structures.