Skip to main content
Dryad logo

Weak founder effects but significant spatial genetic imprint of recent contraction and expansion of European beech populations.


Oddou-Muratorio, Sylvie; Lander, Tonya; Klein, Etienne; Roig, Anne (2020), Weak founder effects but significant spatial genetic imprint of recent contraction and expansion of European beech populations., Dryad, Dataset,


Understanding the ecological and evolutionary processes occurring during species range shifts is important in the current context of global change. Here, we investigate the interplay between recent expansion, gene flow and genetic drift, and their consequences for genetic diversity and structure at landscape and local scales in European beech (Fagus sylvatica L.) On Mont Ventoux, South-Eastern France, we located beech forest refugia at the time of the most recent population minimum, approximately 150 years ago, and sampled 71 populations (2042 trees) in both refugia and expanding populations over an area of 15,000 ha. We inferred patterns of gene flow and genetic structure using 12 microsatellite markers. We identified six plots as originating from planting, rather than natural establishment, mostly from local genetic material. Comparing genetic diversity and structure in refugia versus recent populations did not support the existence of founder effects: heterozygosity (He = 0.667) and allelic richness (Ar = 4.298) were similar, and FST was low (0.031 overall). Still, significant spatial evidence of colonization was detected, with He increasing along the expansion front, while genetic differentiation from the entire pool (βWT) decreased. Isolation by distance was found in refugia but not in recently expanding populations. Our study indicates that beech capacities for colonization and gene flow were sufficient to preserve genetic diversity despite recent forest contraction and expansion. Because beech has long distance pollen and seed dispersal, these results illustrate a ‘best case scenario’ for the maintenance of high genetic diversity and adaptive potential under climate-change related range change.


This data set combines the genotypes, geographical locations and basic measurement of 2532 adult beech trees in 71 plots covering five different regions of Mont Ventoux, SouthEast France. We combined 3 different data sets consisting in:

  1. Three “intensively” studied plots (West-N2, West-384 and West-257-2), within the NW_REF region, with exhaustive sampling of adult trees (579 trees in total). This data set is previously described and analyzed in Oddou-Muratorio, Gauzere, Bontemps, Rey, & Klein (2018) and in Lander, Oddou-Muratorio, Prouillet-Leplat, & Klein, (2011).
  2. A set of 48 plots where a total of 1353 adult trees were sampled non-exhaustively (see detail protocol below), previously described and analysed in Lander, Oddou-Muratorio, Prouillet-Leplat, & Klein, (2011).
  3. A set of 20 plots where a total of 600 adult trees were sampled non-exhaustively specifically for this study, following the protocol of Lander, Oddou-Muratorio, Prouillet-Leplat, & Klein (2011).

For data sets B and C, the same sampling protocol was applied. We selected 28 adult beech trees on average (up to a maximum of 40 individuals) in an area of ~50 m radius so that all trees were separated by at least 3 meters. Circumference at breast height and coordinates were recorded for each tree. Moreover, half of the trees were chosen because they had the largest circumference in the stand (“Old” trees) and half had the smallest circumference (but > 160 mm; “Young” trees).

In the three intensive study plots of data set A, all adult trees were exhaustively sampled, and their circumference at breast height and coordinates were recorded. This intensive sampling effort allowed us to estimate the probability of vegetative reproduction. Indeed, beech is known to have the ability to reproduce vegetatively through resprouting around cut or fallen trees. Note that in data set A, if a tree obviously had multiple stems, only the largest stem was sampled. After careful elimination of clonal individual (see paragraph 2  below), on each of the three plots, we selected 20 small (“Young”) and 20 large (“Old”) trees based on their  circumference (40 trees in total per plot), ensuring a minimum distance between them >3 m.

For data sets A and B leaf samples were collected in 2008 and 2009 and stored at INRA Avignon laboratory at –20°C. For data set C, leaf samples were collected in 2015 and stored at the INRA Avignon laboratory at ambient temperature (using desiccant for conservation).

The maximal age of a tree within each plot was estimated based on two approaches. In 44 plots among the 71 studied, two trees per plot were selected to be cored at 1.30 m (targeting the largest and the smallest one). Cores were read to estimate tree age from ring profile. The oldest age was considered as a rough estimate of the maximal age of a tree in 2015 in a given plot. In the other plots, this maximal age of a tree was estimated based on detailed dendroecological studies (F Jean & P Dreyfus, pers comm) which combined tree ring profile, size distribution of trees, and historical record of management operation within each plot.


Agence Nationale de la Recherche, Award: ANR-07-JCJC-0117

Sixth Framework Programme, Award: GOCE-016322