Exploring Terrestrial Eco-Hydrological Processes from Bottom-Up and Top-Down Perspectives
Yunfei Wang is a PhD student in the Department of Water Resources. (Co)Promotors are Prof. Dr. Z. Su and Dr. Y. Zeng from the Faculty of Geo-information Science and Earth Observation (ITC), University of Twente, together with Prof. Dr. H. Cai and Prof. Dr. Q. Yu from Northwest A&F University.
Eco-hydrological processes, including evapotranspiration (ET) and gross primary productivity (GPP), govern the coupled exchanges of water, energy, and carbon in terrestrial ecosystems. They sustain ecosystem productivity and resilience, and their accurate quantification is vital for predicting ecosystem responses to climate variability, managing water resources, and constraining global budgets. However, challenges persist due to scale mismatches and limitations in observation and modeling techniques.
Existing approaches—in-situ observations, remote sensing, and process-based models—each have strengths but also major constraints. in-situ capture detailed fluxes but remain spatially limited. Remote sensing expands coverage but suffers from coarse resolution and empirical assumptions. Process-based models within the Soil–Plant–Atmosphere Continuum (SPAC) provide mechanistic representations but are hampered by parameter uncertainties and inconsistent treatment of soil, vegetation, and atmosphere. To overcome these gaps, models must integrate advanced remote sensing, preserve SPAC consistency, and bridge scales, particularly for emerging digital twin Earth frameworks.
This thesis develops and applies SPAC-based eco-hydrological models with an emphasis on soil moisture dynamics, root growth and water uptake, and solar-induced chlorophyll fluorescence (SIF). By utilizing Bottom-up and Top-down modeling, this study reduces uncertainties in simulating ecosystem fluxes.
The first component demonstrates that incorporating dynamic root growth and vertical soil moisture heterogeneity in the coupled STEMMUS-SCOPE model improves GPP and ET simulations, particularly under drought stress, and enables the forward simulation of remote sensing signals such as SIF.
The second component applies STEMMUS-SCOPE to 170 global sites, generating a long-term, physically consistent dataset of energy, water, and carbon fluxes. Without site-specific tuning, the model reproduces observed fluxes and multilayer soil moisture, providing valuable resources for both observation networks and Earth system models.
The third and fourth components integrate tower-based and satellite SIF into eco-hydrological modeling through the STEMMUS-MLR framework. This approach directly estimates GPP, ET, and soil moisture from SIF observations, improving accuracy in contrasting ecosystems and across AmeriFlux sites. Results highlight SIF’s value not only as a proxy for photosynthesis but also as a constraint on stomatal conductance and transpiration.
Finally, model inter-comparison shows that incorporating SIF improves daily GPP and latent heat flux estimates beyond climate-based simulations. Using SHAP analysis, SIF emerges as the most influential predictor of GPP and among the leading drivers of ET, underscoring its unique physiological information.
In summary, this thesis advances eco-hydrological modeling by integrating soil moisture dynamics, root processes, and SIF observations within physically consistent SPAC frameworks. The findings provide new insights into terrestrial water and carbon fluxes under changing climates and contribute to the development of the Digital Twin Earth System and sustainable ecosystem management.
