Modeling the hydrologic interactions between an aging reservoir and the surrounding groundwater

Aging reservoirs that have been subjected to severe sedimentation are growing in number worldwide. Searsville Reservoir, located within Jasper Ridge Biological Preserve at Stanford University, is one example. Given the many benefits, but also the potential problems, of sediment-impacted reservoirs, there is strong motivation to better understand these hydrologic systems to enhance management and investment decisions. Field data suggest that the interactions between the reservoir and the surrounding groundwater in the deposited sediments are very complex and dynamic. Water flux can be either from the sediment to the reservoir or vice versa or both, and the interactions are not only seasonal but also diurnal, forming an integrated hydrologic system. Computational modeling is essential for understanding the various hydrologic processes and also for predicting system responses to different management scenarios. This dissertation tackles several modeling challenges, solutions to which enhance our ability to understand integrated groundwater-reservoir systems. It also examines specific features of groundwater-reservoir systems and addresses important questions underlying management decisions for sediment-impacted reservoirs, using Searsville as a case study.;This dissertation is divided into the following three parts:;(1) Enhancing modeling techniques for integrated hydrological systems. The first part of this dissertation studies the feasibility of simulating a reservoir/lake as a high-conductivity variably-saturated porous medium, and presents guidelines for choosing modeling parameters for the reservoir/lake region. It concludes that when applied in variably-saturated models, this approach is most suitable for relatively simple geometries and lakes with slower and smaller fluctuations. It is also more applicable to research that studies the overall flow pattern and system fluxes, rather than the detailed flow pattern around the interaction of the lake and land surfaces.;It also explores the use of a multiphysics model in integrated hydrological modeling by implementing and verifying three important hydrologic boundary conditions: rainfall infiltration, seepage faces, and evapotranspiration fluxes. It demonstrates that with care and creativity these boundary conditions can be implemented accurately and efficiently. Therefore, boundary condition implementation should not limit the applicability of multiphysics model to a broad set of problems of interest to the hydrologic community.;(2) Examining diurnal evapotranspiration signals in coupled groundwater surfacewater systems. The second part of this dissertation analyzes the diurnal evapotranspiration signals in fully-coupled and interacting groundwater surface-water systems. It first performs numerical modeling using simplified and generic domains to understand the characteristics (i.e., the magnitudes, the timing and the phase relationships) of the diurnal signals. Using both data analysis and numerical modeling, it then examines the diurnal signals in the Searsville Reservoir surface-water elevation, in the piezometric head in the surrounding ground water, as well as in the incoming streamflow during summer. It illustrates the interactions of surface and subsurface water bodies and demonstrates the importance of intermittent incoming stream at Searsville in shaping the characteristics of the diurnal signals. It also concludes that, to extract ecohydrologic information from diurnal signals, one must consider the couplings between surface and subsurface water.;(3) Predicting subsurface hydrologic responses to alternative management scenarios. The last part of this dissertation investigates the upstream subsurface hydrologic responses to simplified hypothetical management scenarios in sediment-impacted reservoirs, using numerical models based on conditions at Searsville. It concludes that the vegetation community upstream is not highly sensitive to downstream management decisions because its water table remains relatively high and its total transpiration remains around the same. Sensitivity analysis reveals that systems with lower precipitation and/or higher reference transpiration and/or higher hydraulic conductivity in general are more sensitive to management decisions because they rely more heavily on the reservoir in maintaining the shallow water table in the deposited sediments.;Overall, the results of this dissertation contribute to our understanding and modeling of groundwater-reservoir interactions, which benefits the management of sediment-impacted reservoirs globally. Some of the results, such as those pertaining to the modeling techniques and evapotranspiration signals, are also applicable to riparian and other groundwater-lake systems.