Aim: Southern conifers have evolved under different evolutionary pressures to northern lineages, but in both regions these plants have undergone extensive extinction and range alteration over the Cenozoic (the last 66 million years). It is not possible to directly observe the ecology of fossils, but indirect evidence of changes in bioclimatic envelopes can be derived from no-analogue assemblages – groups of co-occurring fossils that have climatically incongruous living relatives. We identify and examine the specific pairs of no-analogue fossils within assemblages to disentangle the effects of climatic factors on past conifer extinctions, and suggest which of these factors are likely to threaten southern conifer biodiversity in the future.

Location: Southern Hemisphere

Time period: Cenozoic

Major taxa studied: Conifers

Methods: We use a recently developed method, ‘hyperoverlap’, to identify no-analogue pairs of southern conifers. We characterise each pair in terms of temperature and precipitation, and evaluate temporal patterns in no-analogue pairs. These analyses represent a novel approach to studying both no-analogue fossils, and past changes in bioclimatic envelopes.

Results: We identified 240 no-analogue pairs in the Cenozoic record of southern conifers. Most (75.4%) observations of no-analogue pairs are likely to result from a change in the thermal (rather than hydrological) distribution between the fossil taxa and their extant counterparts, regardless of region or assemblage age. Thus, the fossil record shows some thermal lability, but strong hydrological stability in the tolerances of these plants. This implies that most southern conifers have inhabited wet climates through the Cenozoic, which is consistent with physiological evidence suggesting strong conservatism in drought tolerance.

Main conclusions: Southern conifers have successfully adapted to a wide range of temperatures, but future changes in rainfall are likely to pose the greatest threat to these plants, either directly or indirectly (e.g., through increased incidence of fire in mesic areas).

This dataset was collected from the Global Biodiversity Information Facility (GBIF; https://www.gbif.org/) and curated to exclude fossil, cultivated or suspected erroneous records. We extracted five environmental variables that are linked to physiological stresses in plants (Table 1) from WorldClim V2 at a resolution of 30 seconds (approximately 1km2). Precipitation variables were fourth-root transformed to an approximately normal distribution to account for the non-linear physiological responses to precipitation. The environmental variables were then normalised to the global means and standard deviations to ensure that axes were comparable. See README file for more details. 

Missing values denoted by empty cells or NA values. See README file for explanations of each variable.