Quantifying and modelling the responses of leaf gas exchange to drought
thesisposted on 2022-03-28, 03:08 authored by Shuangxi Zhou
Global climate change is expected to increase drought duration and intensity in certain regions while increasing rainfall in others. The quantitative consequences of increased drought for ecosystems are not easy to predict. Process-based models must be informed by experiments to determine the resilience of plants and ecosystems from different climates. The thesis provides quantitative information aimed at improving land surface models (LSMs). It includes four papers. (1) Responses of leaf-atmosphere gas exchange to short-term drought were analysed, across plant functional types and climates, based on a synthesis of previous experiments. Explicit and consistent definitions of stomatal versus non-stomatal responses were adopted. Both types of response were shown to be important, and plants adapted to arid climates responded very differently from others. (2) Parallel responses of stomatal conductance, mesophyll conductance, and photosynthetic capacity were found in two glasshouse experiments with tree species from Australia and Europe, revealing a common,coordinated pattern of increasing tolerance in plants from drier environments. (3) Xeric and riparian species of Eucalyptus were subjected to short- and long-term drought. The species were found to differ not only in their tolerance for short-term drought, but also in the extent to which they could acclimate to long-term drought. (4) Experimentally based drought responses were used to define new, plant type-dependent relationships of stomatal sensitivity and photosynthetic capacity to soil water potential in the Community Atmosphere Biosphere LandExchange (CABLE) LSM. Comparison with CO2 and latent heat flux measurements from eddy covariance flux measurement sites in Europe during the 'heatwave' year of 2003 showed that discrepancies between model results and observations were not substantially improved by the inclusion of the more realistic functions, due to a persistent positive bias in the model's simulation of evapotranspiration which overshadowed the differences between different representations of drought response functions.