Research

Rainfall variability and estimation for hydrologic modeling

A remote sensing based study at the source basin of the Upper Blue Nile river

PhD thesis (3.22 MB)

Quantifying affects of rainfall sampling and understanding the implications in rainfall driven processes requires good understanding of the space-time patterns of rainfall variability. Such understanding is often missing in many geographical regions due to sparse rain gauge networks and low observation frequency but also due to the highly dynamic nature of rain events.

Studying patterns of rainfall variability and the mechanisms that affect such variability may be considered relatively easy when the study area has a flat topography. The study area for this research, however, is characterized by mountainous and adjacent lake areas. Besides geographic and geometric settings of the area is rainfall by a number of interacting processes that cause high spatial and temporal variability.

The study area in this research is the Lake Tana basin which is the source of the Blue Nile River. The basin is characterized by a large lake and various topographic features that include a range of mountains.

The main objective in this study is to determine the spatio-temporal patterns of rainfall variability and its effects on hydrological model simulations. Three sub-objectives are:

• Determine spatial and temporal patterns of rainfall variability,
• Develop a parsimonious and reliable model approach for rainfall estimation and rainfall detection from remotely sensed data,
• Determine the sensitivity of a physically based rainfall-runoff model to rainfall representation.

For rainfall-runoff modeling the Representative Elementary Watershed Approach (REW) model is selected. The REW model constitutes an approach which can be placed in between a lumped and a physically-based spatially-distributed conceptualizations of watersheds. Point scale conservation equations for mass, momentum and energy are upscaled to the scale of representative control volumes. By integrating, spatial gradients of the conserved physical properties are converted into exchange terms for the property across the boundaries of defined control volumes. The exchange terms for mass, forces and energy are closed in terms of the gradients of the physical property across the control volume boundaries. In contrast to the lumped-conceptual models, the exchange terms are driven by physically quantifiable quantities, such as piezometric head differences or REW-scale velocities. The REW approach, formulated in terms of REW-scale integral balance laws precludes the solution of partial differential equations, as it is the case for other (true) physically-based methods.