Adult shoots and seedings of the ribbon weed (Posidonia australis Hook.f.), a widespread temperate seagrass, were sampled from four meadows around Rottnest Island, Western Australia. The data set contains multilocus genotypes for adult shoots from four meadows and seedlings from three meadows over two consecutive years. The metadata file contains: Pop number (1 – 10), Sample site, Latitude (S), Longitude (E), individual sample code, year of sampling, life stage (adult shoot or seedling), genotypes (2 columns per locus). The seven polymorphic microsatellite loci are: PaA1, PaA105, PaA120, PaB6, PaB8, PaB112, PaD113. Most genotypes are diploid, however, 3N genotypes are included.

Fifty adult P. australis shoots were collected from randomly generated GPS coordinates within a 50 m diameter circle from four meadows around Rottnest Island: Thomson Bay (RTB), Stark Bay (RST) and Nancy Cove (RNC) and Parker Point (RPP). Approximately 30 one year old P. australis seedlings were also collected for genotyping following annual surveying for two consecutive years (2014 an 2015). All adult shoot and seedling samples were genotyped for seven polymorphic DNA loci (PaA1, PaA105, PaA120, PaB6, PaB8, PaB112, PaD113). Laboratory and genotyping methods are described in full in Sinclair et al. (2014b).

Clonal and genetic diversity indices were estimated for all adult shoots and seedling cohorts using Genalex version 6.5 (Peakall & Smouse, 2012). The ‘Find Clones’ option was used to identify shared multilocus genotypes (MLGs) within and among meadows. P_ID was then used to determine whether shared MLGs were considered to come from the same vegetative clone or a separate recruitment event (from seed). Clonal richness (R = (G-1)/(N-1)), where G = number of MLGs, and N = number of samples, was estimated for each meadow (Dorken & Eckert, 2001). Genetic diversity parameters were based on the complete data set: the total number of alleles (Na), number of private alleles (p[i]) estimated as those alleles that only occur in a single sampled meadow, observed (H_o) and expected heterozygosity (H_e), and Fixation Index (F) using Genalex. Tests for Hardy-Weinberg equilibrium (HWE) were conducted using GENEPOP v4.2 (http://genepop.curtin.edu.au/).

A Bayesian population assignment method was used to determine whether seedlings were recruited locally or from a nonlocal meadow using GeneClass2 (Piry et al., 2004). Simulations suggest that 100% correct assignments can be made using Rannala and Mountain’s (1997) assignment method with sample sizes of 30-50, 10 microsatellite loci, and an F_ST = 0.1 (Waples & Gagliotti, 2006). These parameters were closely met by our data for established meadows (49-50 shoots per meadow, 7 polymorphic loci, and an overall estimate of F_ST = 0.143). We used data from our mating system study at Parker Point and Stark Bay where the maternal parent was known (Sinclair et al., 2020), to test the power of assignment tests and set confidence thresholds to assign seedlings to a local or nonlocal meadow of origin. ‘Seeds of known origin’ were assigned to one of the four sampled meadows. The probability of a meadow being a seed source (by cohort and individual) was established using the Monte Carlo re-sampling algorithm using 10,000 replicates resampling permutations following Paetkau et al. (2004). The log-likelihood (-log(L)) of originating from each source meadow was then calculated following the Bayesian-based method (Rannala & Mountain, 1997). The proportion of correctly assigned seeds was used to set two assignment thresholds, one based on the correct assignment of the cohort (all seeds) to the meadow of origin, and a second for the correct assignment of individual seeds within each cohort to the meadow of origin. The thresholds reflect a compromise between adopting too low a stringency level, which increased the risk of an incorrect assignment due to genotyping errors and/or inclusion of interpopulation hybrids resulting from pollen flow (see Roques et al., 1999), or too high a stringency level in which too few individuals were assigned. We then applied the threshold to our seedling. An underlying assumption for a cohort assignment is that all seedlings sampled within a cohort originated from the same source meadow. An unambiguous assignment was accepted when the ‘% score’ was above the confidence threshold set using ‘seeds of known origin’. A cohort or individual that was assigned above the defined threshold to a meadow other than where it was sampled from was regarded as successful recruitment following a dispersal event.