Hydrologic modeling using triangulated irregular networks : terrain representation, flood forecasting and catchment response

Abstract

Numerical models are modern tools for capturing the spatial and temporal variability in the land-surface hydrologic response to rainfall and understanding the physical relations between internal watershed processes and observed streamflow. This thesis presents the development and application of a distributed hydrologic model distinguished by its representation of topography through a triangulated irregular network (TIN) and its coupling of the surface and subsurface processes leading to the catchment response. As a research tool for hydrologic forecasting and experimentation, the TIN-based Real-time Integrated Basin Simulator (tRIBS) fully incorporates spatial heterogeneities in basin topography, surface descriptors and hydrometeorological forcing to produce dynamic maps of hydrologic states and fluxes. These capabilities allow investigation of theoretical questions and practical problems in hydrologic science and water resources engineering. Three related themes are developed in this thesis. First, a set of methods are developed for constructing TIN topographic models from raster digital elevation models (DEM) for hydrologic and geomorphic applications. A new approach for representing a steady-state estimate of a particular watershed process within the physical mesh is introduced. Hydrologic comparisons utilizing different terrain models are made to investigate the suitable level of detail required for capturing process dynamics accurately. Second, the TIN-based model is utilized in conjunction with a rainfall forecasting algorithm to assess the space-time flood predictability. For two hydrometeorological case studies, the forecast skill is assessed as a function of rainfall forecast lead time, catchment scale and the spatial variability in the quantitative precipitation forecasts (QPF). Third, the surface and subsurface runoff response in a complex basin is investigated with respect to changes in storm properties and the initial water table position.The partitioning of rainfall into runoff production mechanisms is found to be a causative factor in the nonlinearity and scale-dependence observed in the basin hydrograph response. The model applications presented in this thesis highlight the advantages of TIN- based modeling for hydrologic forecasting and process-oriented studies over complex terrain. In particular, the multi-resolution and multi-scale capabilities are encouraging for a range of applied and scientific problems in catchment hydrology.