Observation-Driven Mapping of Coupled Land Water, Energy, and Carbon Cycles
Open Access DepositedDownloadable Content
The land surface, located at the interface of land and atmosphere, is a key component of the Earth system. Land–water, energy, and carbon cycles govern climate through coupled physical, chemical, and biological processes. Exchanges of heat, moisture, and carbon between the land and atmosphere determine land-surface states, influence convective triggering and boundary-layer growth, and ultimately shape weather and climate from local to global scales. Because these processes are heterogeneous and nonlinear, uncertainty in one component can cascade to others and amplify through feedbacks. Understanding and quantifying these linkages is therefore essential for reliable hydroclimate prediction and for diagnosing model deficiencies that limit predictability. Despite decades of progress, knowledge gaps persist. Point-scale observations (e.g., flux towers) are invaluable but sparse and often unrepresentative of heterogeneous landscapes. Offline land-surface modeling has exposed sensitivities in coupling metrics, yet structural biases and parameter uncertainties limit transferability and scale awareness. Coarse global products can obscure sub-grid variability, while regional assessments often lack consistent uncertainty quantification. As a result, the terrestrial leg of land–atmosphere coupling remains under-constrained at the spatial and temporal scales required for water-resources management and climate services. The main objective of this thesis is to develop a unified observation-driven framework that simultaneously estimates key parameters, states and fluxes of land water, energy, and carbon cycles, and to investigate the key interactions between these cycles. This thesis employs variational data assimilation (VDA) to combine process-based models with satellite observations by estimating states and parameters that minimize a cost function subject to governing equations. Building on the Land Integrated Data Assimilation (LIDA) framework (Abdolghafoorian and Farhadi 2019, 2020), two advances are developed and implemented over the Southern Great Plains (SGP) at 5-km resolution. LIDA-2 couples parsimonious forms of the water and energy balance and assimilates SMAP surface soil moisture and GOES land surface temperature (no LAI) to produce spatially continuous maps of evapotranspiration (ET) and diffuse groundwater recharge with diagnostic uncertainty. LIDA-3 extends this system by explicitly coupling water, energy, and carbon through the addition of vegetation dynamics and assimilation of MODIS leaf area index (LAI) alongside SMAP and GOES, enabling joint estimation of key parameters, states, and fluxes, including gross primary productivity (GPP) and ET partitioning. Both LIDA-2 and LIDA-3 are validated using point-scale, regional, and independent benchmark comparisons across the SGP. The simultaneous estimation of soil moisture, LAI, evaporative fraction, specific leaf area, GPP, and related parameters enables direct study of linkages among cycles while avoiding biases from fixed empirical relationships or scale mismatches. Finally, to support rapid applications, the thesis develops symbolic-regression surrogate models (genetic expression programming) that yield closed-form equations for daily ET and monthly diffuse recharge from readily available predictors (e.g., net radiation, air temperature, vapor-pressure deficit, surface soil moisture, land surface temperature, and LAI). These surrogates reproduce the reference fields with high correspondence while preserving interpretability, enabling fast mapping and scenario screening for irrigation planning, groundwater model forcing, and aquifer sustainability analyses. The demonstrated accuracy in simulating soil moisture, recharge, GPP, and ET—together with robust uncertainty diagnostics and consistent land–atmosphere coupling patterns—establishes LIDA-2, LIDA-3, and the associated surrogate equations as effective tools for large-scale mapping and decision support, particularly in data-scarce regions.
- All rights reserved
Notice to Authors
If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.