An integrated hydrologic model for multi-process simulation using semi-discrete finite volume approach

This thesis presents (1) a strategy to build an integrated hydrologic model using the semi-discrete finite volume approach; (2) the development of the Penn State Integrated Hydrologic Model (PIHM), i.e. an implementation of the strategy; (3) code verification of PIHM and a field application at the Shale Hills watershed in central Pennsylvania.; Numerical simulation of coupled non-linear hydrologic processes requires an efficient and flexible approach to solve a mixture of governing partial differential equations (PDE) and ordinary differential equations (ODE). A new strategy for integrated hydrological modeling is proposed in this thesis. First, those PDEs are reduced to ODES using the semi-discrete finite volume method (FVM). This leads to a local ODE system, also called the model kernel. The model kernel is distributed on an unstructured triangular irregular network (TIN) constructed from domain decomposition using Delaunay triangulation. The global ODE system is formed by combining all local ODE systems over the entire domain and the system is solved with an efficient ODE solver. This strategy is designed to capture "dynamics" in multiple processes while maintaining the conservation of mass at all cells, as guaranteed by the FVM. A hypothetical sample is presented to demonstrate the flexibility and utility of this model.; The implementation and code verification of the above strategy, named Penn State Integrated Hydrologic Model (PIHM), is also a subject of this thesis. The model is coded in C with a data structure that is widely used in geographic information system (GIS) literatures. The goal of the data structure used here is to ultimately connect the model and the solver with GIS pre and post processors in a seamless manner, e.g. without the intermediate text files. In such a seamless hydrologic modeling strategy, the numerical model and its pre/post-processor may operate on the same data-model, most likely in the form of geo-database, which enables faster access of raw data and supports advanced user inquiries and visualization of the model results. The PIHM code is verified against analytic and numerical solutions in several scenarios. For groundwater flow component, the model results converge to analytic solutions in a radial groundwater flow example. For surface overland flow and channel routing within a V-shaped catchment, PIHM agrees well with other model results using finite difference and finite element methods. (Abstract shortened by UMI.)...