A Coupled Numerical Model of Vadose Zone Hydrology and Subsidence in the Everglades Wetlands

Peatlands, a type of wetlands, cover less than 3% of the earth’s surface but are responsible for nearly 5% of global carbon emissions when drained. Alterations to the natural hydrology of the Everglades, one of the largest peatlands in the United States, have resulted in the presence of vadose zones and the consequent subsidence of peat. Everglades restoration efforts are guided by hydrological models that neglect unsaturated water flow, which limits their ability to quantify critical wetland processes. Large datasets of the spatially varying soil hydraulic parameters (SHPs) like the soil water retention curves (SWRCs) are necessary to develop distributed, deterministic models. Furthermore, conventional models of unsaturated flow parameterization and transport ignore the effect of volume-change (VC) commonly observed in peat. The main objectives of this study were to characterize the SHPs of Everglades soil, investigate the effect of VC on the SWRC parameterization process, and develop and test an improved finite-difference model, REVC (Richards Equation Volume Change), which incorporates subsidence in the Richards equation transport mode. Using laboratory methods, a large dataset of SWRC, shrinkage, organic content, fiber content, and saturated hydraulic conductivity from 53 sites across the Everglades was generated. Agglomerative clustering resulted in three clusters with distinct SWRCs - marl, mixed marl-peat and peat. Application of volume-correction to the van Genuchten Mualem (vGM) model resulted in deviations from typical SWRC behavior (attributed to the collapse of macropores). Mesh convergence analysis was conducted to guide REVC spatial and temporal discretization. Sensitivity of REVC to vGM parameters with three parameter-transport model combinations for shallow vadose zones resulted in lower surficial pressure heads and volumetric water contents (VWCs) when VC was ignored in both parameterization and transport models. REVC simulated VWC and subsidence in a lysimeter experiment with constant boundary conditions (BCs) and a field subsidence study with variable BCs produced excellent to fair model fit. The practical application of the REVC model is demonstrated through the case scenarios of multiple accretion, reversible shrinkage models. Although the generated SWRCs with REVC can be implemented to model distributed flow, further testing at field-scale with site-specific calibration by practitioners is recommended