On the objective characterization of the rainfall-runoff transformation

Abstract

The transformation of rainfall to runoff can lead to flooding under many different combinations of catchment state, rainfall intensity and duration, and spatio-temporal structure of rainfall. Hence, there is great interest in improving our understanding of physical processes which lead to flooding and in better characterizing the key features (e.g. timing, magnitude…) of the phenomena, with the final aim of advancing hydrological forecasts and flood frequency methods. However, due to the complexity in storm structure and in the interactions between surface water inputs and landscape properties, characterizing the rainfall-runoff processes is not trivial. Consequently, hydrologists develop methodologies to characterize the rainfall-runoff transformation which make a number of a priori assumptions and require multiple user decisions, so that they can at least partially answer rainfall-runoff related questions. As a result, the retrieved information may be biased and inconsistent, also making it difficult to bring together findings from different studies.
In this thesis we present novel methodologies to characterize the rainfall-runoff transformation, which reduce the number of a priori assumptions and user decisions, providing more objective tools. In particular, we present: 1. a novel timeseries-analysis-based method to estimate the catchment response time; 2. a novel timeseries-analysis-based method to identify rainfall-runoff events; 3. a novel application of metrics of rainfall spatial variability to gain a priori knowledge about the hydrograph response. Each presented method is compared to a traditionally used approach for the same scope to explain differences and highlight advantages and any limitations. Although tools which retrieve similar information are available in the literature, in this thesis our contribution to knowledge is in the more objective nature of the new methodologies, which allows an easier transferability of the methods across places and temporal resolutions of the data, thus making their application in different regions an interesting future research direction.

Date of Award	22 Mar 2022
Original language	English
Awarding Institution	University of Bristol
Supervisor	Ross A Woods (Supervisor) & Miguel A Rico-Ramirez (Supervisor)