Little, Stefan A. , Steven, Kembel , Wilf, Peter , Royer, Dana .
Phylogenetic signal in leaf traits and its influence on leaf-climate correlations.
Leaves are the most common plant macrofossils, and fossil leaf characters are widely used to estimate paleoclimate. For example, there are significant, taxon-free positive correlations among extant communities between leaf area and precipitation, and between proportion of taxa with untoothed leaves and mean annual temperature. Several additional leaf shape traits, including area:perimeter ratios, tooth number, and tooth size, are used to estimate paleoclimate in a recently developed method, digital leaf physiognomy.
Phylogenetic signal is the tendency for close relativesí phenotypes to resemble one another, and it is known to affect the independence of species data, a requirement in many statistical analyses. Current leaf-paleoclimate methods assume that phylogenetic signal is negligible, but this has never been explicitly tested. Because leaf phenotypes respond to the environment through controlled genetic pathways, we hypothesize that phylogenetic signal in extant community data used for calibration may affect leaf-paleoclimate models and estimates.
We used a phylogenetic supertree to test for phlyogenetic signal (K* metric of Blomberg et al.) in the 2005 digital leaf physiognomy dataset. There was significant but low-magnitude signal in most traits, in agreement with conventional paleobotanical wisdom. However, K* was relatively high in most traits relating to leaf teeth, which have high nonphylogenetic correlations with temperature. We tested the influence of phylogenetic signal on correlations between leaf traits and mean annual temperature (MAT) using phylogenetic generalized least squares regression (PGLS). The PGLS analyses revealed weaker relationships between traits with high K* and climate than in uncorrected regressions; for low K* traits, the PGLS regressions were similar to the uncorrected regressions. Phylogenetic signal may explain differences in MAT estimates based on calibration floras from different biogeographic regions. The widespread assumption that leaf shape is controlled predominantly by climate can now be explicitly tested, potentially leading to improved leaf-paleoclimate models.
Log in to add this item to your schedule
1 - Pennsylvania State University, Department of Geosciences, University Park, PA, 16802, USA
2 - University of California Berkeley, Integrative Biology, 3060 Valley Life Sciences Building #3140, Berkeley, CA, 94720-3140, USA
3 - Department of Earth and Environmental Sciences, Wesleyan University, Exley Science Center 445, 265 Church St., Middletown, CT, 06459-0139, USA
Digital leaf physiognomy
Mean annual temperature.
Presentation Type: Oral Paper:Papers for Sections
Date: Wednesday, July 30th, 2008
Time: 1:15 PM