Variation in intraspecific demography drives localised concordance but species-wide discordance in response to past climatic change

Abstract

Background: Understanding how species biology may facilitate resilience to climate change remains a critical factor in detecting and protecting species at risk of extinction. Many studies have focused on the role of particular ecological traits in driving species responses, but less so on demographic history and levels of standing genetic variation. Additionally, spatial variation in the interaction of demographic and adaptive factors may further complicate prediction of species responses to environmental change. We used environmental and genomic datasets to reconstruct the phylogeographic histories of two ecologically similar and largely co-distributed freshwater fishes, the southern (Nannoperca australis) and Yarra (N. obscura) pygmy perches, to assess the degree of concordance in their responses to Plio-Pleistocene climatic changes. We described contemporary genetic diversity, phylogenetic histories, demographic histories, and historical species distributions across both species, and statistically evaluated the degree of concordance in co-occurring populations.

Results: Marked differences in contemporary genetic diversity, historical distribution changes and historical migration were observed across the species, with a distinct lack of genetic diversity and historical range expansion suggested for N. obscura. Although several co-occurring populations within a shared climatic refugium demonstrated concordant demographic histories, idiosyncratic population size changes were found at the range edges of the more spatially restricted species. Discordant responses between species were associated with low standing genetic variation in peripheral populations. This might have hindered adaptive potential, as documented in recent demographic declines and population extinctions for the two species.

Conclusion: Our results highlight both the role of spatial scale in the degree of concordance in species responses to climate change, and the importance of standing genetic variation in facilitating range shifts. Even when ecological traits are similar between species, long-term genetic diversity and historical population demography may lead to discordant responses to ongoing and future climate change.

Keywords: Biogeography; Coalescent analysis; Comparative phylogeography; Percichthyidae; Temperate Australia.

Fig. 1

Contemporary distribution and sampling map for N. australis and N. obscura. Nannoperca australis sampling sites are indicated in red, and N. obscura sites in blue. The distribution of N. australis is indicated with light green shading and dashed borders, with the distribution of N. obscura (also the region of co-occurrence) in darker green. The solid black line indicates the boundary of major drainage basins, and the dotted line demonstrates the approximate shoreline during glacial maxima. Bottom left inset depicts study region and major drainage basins in Australia. Top right inset depicts the full extent of species distributions

Fig. 2

Phylogenetic histories and migration patterns in N. australis and N. obscura. A Maximum likelihood phylogenetic trees based on ddRAD loci. Populations were reciprocally monophyletic and so were collapsed to the population level for simplicity. Both trees were rooted using N. vittata as the outgroup, which was dropped for visualisation. Node values show bootstrap support. Branch colours indicate the drainage basin of origin for each population or clade. B Best supported ancestral migration patterns inferred using TreeMix based on SNP datasets. All displayed migrations were statistically significant (p < 0.05). Arrows denote the direction of inferred migrations, with the colour indicating their relative weights

Fig. 3

Demographic histories of N. australis and N. obscura populations. A Stairway plot reconstructions of demographic history. Inset stars indicate co-occurring populations which were further explored within a codemographic framework. Populations are arranged from westernmost to easternmost within each species. B Most likely individual demographic histories for co-occurring N. australis and N. obscura populations over the Pleistocene, simulated using FastSimCoal2. Thick dark lines indicate mean Ne over time, calculated based on the means of current Ne, rates of change and timing of switching rates (see Supplementary Material). Shaded areas indicate 95% confidence intervals based on the 97.5% and 2.5% probability estimates for the same parameters. C Bayes Factor matrix of the proportion of populations showing synchronised bottlenecks (ξ) within a co-demographic model using Multi-DICE. Each cell compares the model in the column with the model in the row, with brighter colours indicating greater support for the column. D Posterior distribution of mean bottleneck strength (ε) across all six populations. E Posterior distribution of dispersion index of bottleneck strength [Var(ε)/Mean(ε)] across all six populations. F Posterior distribution of the timing of the bottleneck event, in generations/years

Fig. 4

Comparisons of summaries of distributional changes over eleven time periods spanning the Plio-Pleistocene. A Distribution extent per species. Individual models are indicated by points, with SDM method indicated by colour. The 95% confidence interval across all individual models is shown by the pale blue ribbon. The ensemble model is represented by a solid black line. B Mean cell suitability across all time periods. C Variation (standard deviation) in cell suitability across all time periods