Palaeodistribution modelling and genetic evidence highlight differential post‐glacial range shifts of a rain forest conifer distributed across a latitudinal gradient

Abstract

Aim We examine the range expansion/contraction dynamics during the last glacial cycle of the late‐successional tropical rain forest conifer Podocarpus elatus using a combination of modelling and molecular marker analyses. Specifically, we test whether distributional changes predicted by environmental niche modelling are in agreement with (1) the glacial maximum contractions inferred from the southern fossil record, and (2) population genetic‐based estimates of range disjunctions and demographic dynamics. In addition, we test whether northern and southern ranges are likely to have experienced similar expansion/contraction dynamics. Location Eastern Australian tropical and subtropical rain forests. Methods Environmental niche modelling was completed for three time periods during the last glacial cycle and was interpreted in light of the known palynology. We collected 109 samples from 32 populations across the entire range of P. elatus. Six microsatellite loci and Bayesian coalescence analysis were used to infer population expansion/contraction dynamics, and five sequenced loci (one plastid and four nuclear) were used to quantify genetic structure/diversity. Results Environmental niche modelling suggested that the northern and southern ranges of P. elatus experienced different expansion/contraction dynamics. In the northern range, the habitat suitable for P. elatus persisted in a small refugial area during the Last Glacial Maximum (LGM, 21 ka) and then expanded during the post‐glacial period. Conversely, in the south suitable habitat was widespread during the LGM but subsequently contracted. These differential dynamics were supported by Bayesian analyses of the population genetic data (northern dispersal) and are consistent with the greater genetic diversity in the south compared with the north. A contact zone between the two genetically divergent groups (corresponding to the Macleay Overlap Zone) was supported by environmental niche modelling and molecular analyses. Main conclusions The climatic fluctuations of the Quaternary have differentially impacted the northern and southern ranges of a broadly distributed rain forest tree in Australia. Recurrent contraction/expansion cycles contributed to the genetic distinction between northern and southern distributions of P. elatus. By combining molecular and environmental niche modelling evidence, this unique study undermines the general assumption that broadly distributed species respond in a uniform way to climate change.

Intraspecific Divergence Associated with a Biogeographic Barrier and Climatic Models Show Future Threats and Long-Term Decline of a Rainforest Conifer

“…(Podocarpaceae) found agreement between genetic and Climate Envelope Model (CEM) disjunctions across the Clarence River Corridor biogeographic barrier [17]. Also, we reported modelled palaeodistribution and differential post-glacial range shift [35]. Here we infer the future distribution of genetic diversity using investigation of the signature of past demographic processes with a probabilistic model based on molecular data.…”

“…Climatic models for Podocarpus elatus inferred that the current geographic separation of the northern and southern envelopes corresponds to the genetically differentiated population groups [17]. Model performance and assessment are outlined by Mellick et al [35].…”

“…distributions allowed ascertaining demographic trends and trajectories of populations to develop prudent conservation strategies. Here we refer regularly to, and build upon, these previous works [17,35] to understand the deeper evolutionary history and to compose appropriate conservation strategies for the species. We have sub-sampled the genetic data from these previous works [17,35] and generated an additional HapSTR marker for a coalescent-based analysis to infer demographic history and splitting time, and have projected the current climatic envelope [17] onto an ensemble of future climate scenarios [25].…”

“…334 individuals from 32 populations; [35]. In this study, twenty-eight mature individuals from three populations ( Fig.…”

“…Ten markers from nuclear and paternally inherited chloroplast genomes [47], including three types of markers: 1) sequence [35], 2) microsatellite [17] and 3) HapSTR markers were used in the coalescent-based analysis. The four microsatellite loci (PeA16BGT, PeB31BGT, PeC19BGT, PeD14BGT) used in this study were isolated and primers developed from the methodology set out by Almany et al [47].…”