Temperature and Rainfall Are Separate Agents of Selection Shaping Population Differentiation in a Forest Tree

Abstract

Research highlights: We present evidence indicating that covariation of functional traits among populations of a forest tree is not due to genetic constraints, but rather selective covariance arising from local adaptation to different facets of the climate, namely rainfall and temperature. Background and Aims: Traits frequently covary among natural populations. Such covariation can be caused by pleiotropy and/or linkage disequilibrium, but also may arise when the traits are genetically independent as a direct consequence of natural selection, drift, mutation and/or gene flow. Of particular interest are cases of selective covariance, where natural selection directly generates among-population covariance in a set of genetically independent traits. We here studied the causes of population-level covariation in two key traits in the Australian tree Eucalyptus pauciflora. Materials and Methods: We studied covariation in seedling lignotuber size and vegetative juvenility using 37 populations sampled from throughout the geographic and ecological ranges of E. pauciflora on the island of Tasmania. We integrated evidence from multiple sources: (i) comparison of patterns of trait covariation within and among populations; (ii) climate-trait modelling using machine-learning algorithms; and (iii) selection analysis linking trait variation to field growth in an arid environment. Results: We showed strong covariation among populations compared with the weak genetic correlation within populations for the focal traits. Population differentiation in these genetically independent traits was correlated with different home-site climate variables (lignotuber size with temperature; vegetative juvenility with rainfall), which spatially covaried. The role of selection in shaping the population differentiation in lignotuber size was supported by its relationship with fitness measured in the field. Conclusions: Our study highlights the multi-trait nature of adaptation likely to occur as tree species respond to spatial and temporal changes in climate.