Analysis and prediction of plant functional traits in contrasting environments: information for next-generation ecosystem models
thesisposted on 2022-03-28, 02:38 authored by Henrique Furstenau Togashi
Three empirical plant-trait studies are presented, based on new field data plus extensive data analysis. All aim to provide information to improve the basis of dynamic global vegetation models (DGVMs), by exploring neglected ecophysiological processes. The first study investigates a key model parameter, the leaf area (LA) to cross sectional sapwood-area (SA) ratio, in Australian evergreen woody plants. It confirmsthe pipe model, which states that SA at any point should scale isometrically with LA distal to that point. But although the LA:SA ratio is reported to vary with hydroclimate, aridity determined only upper and lower bounds of LA:SA. Great variation among species in similar environments exists and may reflect variation in stem hydraulics. The second study refutes the common model assumption that photosynthetic traits of evergreen plants are fixed in time. If it were true, carboxylation capacity (Vcmax) should increase with temperature following Rubisco kinetics. The study employs a theoretical framework to predict, instead, that Vcmax should acclimate to temperature : increasing, but less steeply. Repeat sampling of plants in the semi-arid Great Western Woodlands, during the warm and cool seasons, revealed that Vcmax (and other photosynthetic parameters) show thermal acclimation as predicted. The third study focuses on tropical rainforests in Australia and China, occupying a relatively narrow range of climates, to examine the association of plant functional traits (including growth traits such as maximum height as well as field-measured photosynthetic traits) with species' successional `roles'. Consistent relationships were found, providing key information for modelling forest dynamics. Each study tackles an aspect of DGVMs that so far has involved simplifying assumptions,such as assigning fixed trait values to functional types without considering variability across environments, seasons, and species. To take account of such variation, DGVM development will require targeted empirical studies like those presented here.