The mediation of terrestrial vegetation in water-carbon coupling: from ecophysiology to land-atmosphere interactions
Embargoed until 2026-03-27
Author
Date
2024Type
- Doctoral Thesis
ETH Bibliography
yes
Altmetrics
Abstract
This thesis investigates the relationship between terrestrial vegetation and water availability, exploring how this interaction influences land-atmosphere exchanges. The focus is on understanding the ecophysiological responses of plants to water stress and the broader implications for water-carbon coupling in different ecosystems. This research is partic- ularly pertinent in the context of climate change, where alterations in water availability significantly impact climate feedback mechanisms and human livelihood.
The first study uses deep learning to analyze evapotranspiration responses to water deficit across various biomes. By training neural networks with eddy-covariance mea- surements, the research defines a water stress factor (fET) and explore its relationship with cumulative water deficits. This approach reveals diverse responses of ecosystems to water stress, ranging from significant declines in evapotranspiration rates in savannahs and grasslands to milder reductions in forests. This variability highlights the role of plant hydraulic resilience and access to groundwater or deep soil moisture, suggesting a complexity in plant responses that is often oversimplified in standard land surface models.
The second paper shifts the focus to the global scale, investigating the role of ground- water in modulating photosynthesis. Using remote sensing data and explainable machine learning, the study demonstrates that groundwater significantly regulate photosynthesis, both spatially and temporally, across different ecosystems. The findings emphasize that the control of groundwater on ecosystem productivity is comparable to that of aridity, highlighting its critical but previously underappreciated role in global vegetation dynamics.
The third study addresses biases in the representation of soil moisture limitation in Earth System Models. By comparing model outputs with observational datasets, the research highlights significant underestimations in soil moisture reductions in these models. These biases illuminate the need for improved soil moisture representations in Earth System Models, which is key for accurately projecting land-atmosphere interactions under changing climatic conditions.
Overall, this thesis leverages and develops novel methods to understand terrestrial vegetation dynamics under water stress, and explores the representation of these dynamics in different types of models. By integrating advanced computational methods with different data types, it offers insights into how vegetation mediates water-carbon coupling. While the primary focus is on hydroclimatology and water-vegetation interactions, the implications of this work extend to broader ecological and climatological contexts, including the carbon cycle. The findings illuminate the complex nature of plant-water interactions in the Earth system and offer potential opportunities for refining climate models and guiding ecosystem management strategies, particularly in the context of global environmental changes. Show more
Permanent link
https://doi.org/10.3929/ethz-b-000666415Publication status
publishedExternal links
Search print copy at ETH Library
Contributors
Examiner: Seneviratne, Sonia I.
Examiner: Stocker, Benjamin
Examiner: Gentine, Pierre
Examiner: Papale, Dario
Publisher
ETH ZurichOrganisational unit
03778 - Seneviratne, Sonia / Seneviratne, Sonia
More
Show all metadata
ETH Bibliography
yes
Altmetrics