Abstract
Rainfall-runoff models play a vital role in understanding hydrologic processes, estimating streamflow, and predicting flood and drought risks across various scales. However, hydrological modelling still faces significant uncertainties and challenges. Difficulties arise in identifying and characterizing hydrological processes (e.g. subsurface losses), selecting and evaluating model structures, and dealing with uncertainties in observational data and model structures. These problems become even more complex in large-sample hydrology due to the heterogeneity of catchments, the abundance of catchment types, and the variability in data quality and human influence.In this thesis, we address three challenges in rainfall-runoff modelling across a large sample of Great Britain catchments. Firstly, we assess the role of catchment location in understanding water balance issues in highly permeable catchments when available catchment descriptors are insufficient. We find that catchment location relative to the coast and within a wider river basin shed light on water balance issues in highly permeable catchments. Secondly, we explore the importance of prior model selection through a comparison of two modular modelling frameworks. By selecting model structures consistent with expected hydrologic variability, we demonstrate the possibility of observing meaningful performance differences between model structures in specific catchments. Lastly, we develop a signature-based hydrologic efficiency metric that proves comparable to traditional statistical evaluation metrics. This metric shows promise for model evaluation in ungauged catchments if its signatures can be well-regionalized.
All three contributions pave the way for follow-up research on new location-based catchment descriptors, hydrologically tailored efficiency metrics in gauged and ungauged basins, and identifying appropriate components for modular modelling system. We end by defining two specific ideas for future research. Firstly, quantifying hydrologic ecosystem services through new signatures to assess benefits and understand spatial-temporal variations. Secondly, coupling national-scale groundwater modelling across Great Britain with catchment-scale modelling to estimate inter-catchment groundwater flow between neighbouring catchments.
| Date of Award | 3 Oct 2023 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | A S M Mostaquimur Rahman (Supervisor) & Gemma Coxon (Supervisor) |