Influence of phylogenetic relatedness on paleoclimate estimation using fossil wood: Vessel and fiber-related traits

Abstract

Wood anatomical traits have been historically used as a climate inference tool. During the last 20years, the use of more formal statistical approaches has allowed the field to start moving from general qualitative paleoclimate inference to quantitative estimates. These estimates are derived from the correlation between wood anatomical traits and climate, however the role of shared evolutionary history in this relationship has not been assessed. Greater influence of phylogenetic relatedness on trait variation (phylogenetic conservatism) would entail a reduction of wood trait variation due to environmental convergence, the central premise of the paleoclimate inference approach using fossil wood. Here, we analyzed the role of phylogenetic relatedness in the relationship between wood anatomical traits and climate. We also assessed the strength of phylogenetic signal in wood traits and climate preferences, as well as their most likely mode of evolution. Although we found similar trends in the phylogenetically uninformed and informed regressions, the fit of the latter models was better for every trait. We also found moderate but significant non-random phylogenetic signal in all traits but potential hydraulic conductivity and vulnerability index. Our results suggest that despite adaptive evolution is driving the relationship between wood anatomy and climate, a significant role of phylogenetic signal indicates that paleoclimate estimates using fiber and vessel traits are subject to broader errors than initially anticipated.