Global analysis of reproductive investment in trees: effects of soil nutrients, species traits, and phylogeny

Tong Qiu, Jennifer J. Swenson and James S. Clark, Nicholas School of the Environment, Duke University, Durham, NC, Benoit Courbaud, Valentin Journe, Georges Kunstler and James S. Clark, LESSEM, University of Grenoble Alpes, INRAE, St-Martin-d'Heres, France, C. Lane Scher, University Program in Ecology, Duke University, Durham, NC

Background/Question/Methods
Seed production represents one of the most fundamental components of fitness, yet lacks any systematic quantification for forest trees, the planet's dominant life form. Specifically, which species make the largest reproductive investment (RI)? Were differences shaped by evolutionary forces in the distant past and since phylogenetically conserved? Can they be explained by gradients in productivity and species differences in their capacities to exploit nutrients? Are geographic differences driven by within-species responses or by turnover between species that are capable of high RI at different fertility levels? We hypothesize that large RI occurs predominantly in nutrient-conservative species. We also hypothesize that increased soil fertility can promote RI at low-to-medium soil nutrient level, but this effect disappears or even reverse at high fertility gradient. To test these hypotheses, we quantified seed production from the Masting Inference and Forecasting (MASTIF) network based on 18 million tree-years. By combining MASTIF with global soil cation exchange capacity (CEC), a widely used indicator for soil fertility, with climate and species traits, we determined within-species and between-species differences in RI and stand-level responses to fertility gradients. We determined species differences in RI and their connections to foliar nutrient concentrations (N and P) and to leaf habit after controlling for the effects of climate, soil nutrients, phylogeny.
Results/Conclusions
Deciduous, broad-leaved species have the highest RI, taken as mass fecundity per individual tree basal area per year (restricting the analysis to trees that are reproductively mature). Phylogenetic conservatism is shown by a strong phylogenetic signal in RI (Pagel’s λ = 0.63, p < 10-7, N = 478 species), with angiosperms (e.g. Quercus, Carya, Betula, Fagus, Diospyros) exhibiting high RI compared to gymnosperms. Species with high RI tend to have low foliar nutrient levels (both N and P). Between-species responses to soil fertility show a divergent pattern, with a negative tendency for gymnosperms along their evolutionary histories. RI plateaus or declines at high soil nutrients at the scale of individual tree in 80% of the species group we examined. At the stand level, RI remains relatively high at low-to-medium soil nutrient levels, but then declined at the highest fertility, which could be related to intense competition. This study provides a first global analysis of reproductive effort in trees and the effects of soil nutrients, with implications for the capacity to respond to fertility gradients, depending on both foliar nutrient concentrations and leaf habit.