Electronic Theses and Dissertations

Babak Azari

Date

2023

Document Type

Dissertation

Degree Name

Doctor of Philosophy

Department

Civil Engineering

Committee Chair

Brian Waldron

Committee Member

Farhad Jazaei

Abstract

Integrating existing surface water (SW) and groundwater (GW) models like SWAT, MODFLOW, and HEC-RAS has been explored to simulate the complexities of SW-GW interactions. Still, challenges arise from temporal and spatial scale disparities. To tackle the temporal scale issue, this study introduces the novel explicit solver (EXP1) for MODFLOW 2005, enabling daily GW modeling similar to SWAT and HEC-RAS by reducing runtime and computational burden. The proposed solver incorporates a stability criterion to assess the stability of the proposed solver. The study compared the EXP1 solver to the widely used Preconditioned Conjugate Gradient (PCG) solver in three scenarios: a 1D model, a 2D model, and the real-world model, which was MERAS. The results showed the efficiency and accuracy of the EXP1 solver in the 1D and 2D models, with minimal deviations in head and water budget compared to PCG and shorter runtimes. However, when applied to the complex MERAS model, some modifications were necessary to ensure stability and accuracy. Stability analysis identified the main culprits of instability, including extremely small cell thicknesses, specific storages, and large external sources/sinks. Remarkably, unconfined cells exhibited high stability when a 1-day time step was chosen, attributed to the fact that the specific yield in unconfined aquifers is several orders of magnitude larger than that of confined cells. While a 1-day time step was preferred and increasing cell sizes impractical, unstable cells were converted to constant heads to achieve stability. The EXP1 solver demonstrated a 57% faster speed than PCG while maintaining comparable head accuracy. A water budget comparison showed over 10% discrepancy due to many constant head cells. To address the 1-day comparison issues, an additional assessment was conducted using a 0.01-day time step, where the EXP1 solver still performed faster and accurately in terms of GW head and water budgets. These findings indicate that the EXP1 solver is an excellent choice for modeling unconfined aquifers, which are of great interest in SW-GW model couplings due to the characteristics of unconfined cells. Conversely, implicit solvers like PCG should be the preferred option for standalone groundwater modeling to avoid stability issues.

Comments

Data is provided by the student.

Library Comment

Dissertation or thesis originally submitted to ProQuest

Notes

Open Access

Download

Share

COinS