Supporting river water-quality management by high-resolution modelling: a case study in a lowland permeable chalk catchment

: (Alternative Format Thesis)

Abstract

Case for action: There is a pressing need to develop the next generation of modelling tools to assess river water-quality deterioration – a major threat to water security and key driver of global freshwater biodiversity decline. Water-quality risk assessment models have thus far lagged behind advances in high-frequency monitoring technologies and lack the spatiotemporal resolution (particularly temporal) to capture significant variability in contaminant concentrations, limiting their accuracy in exposure assessments and effectiveness in understanding the fine-scale processes governing water-quality dynamics.

Aim and broad approach: The overarching aim of this thesis was to develop a high resolution hydrodynamic river model that produces spatially and temporally explicit
concentration profiles, enabling more accurate assessment of contaminant transport and fate to support targeted river management. To achieve this, a high-resolution hydrodynamic nitrate model was first developed and tested, leveraging available high-frequency data to predominantly evaluate the spatiotemporal dynamics caused by continuous wastewater discharges. The model was then adapted to speculatively predict emerging contaminant risk during baseflow conditions. The River Frome, Dorset, UK, served as the case study, providing a well-characterised setting to demonstrate the model’s application.

Summary of methods, results, and conclusions: Chapter 3 draws upon local geological, hydrogeological, and hydrochemical data to infer a perceptual model of the surface/groundwater interactions within the study reach, highlighting their influence on spatial and temporal patterns in hydrochemistry. Chalk springs were shown to dominate flow accretion and Sankey diagrams were used to describe the flow source-apportionment, providing a solid conceptual framework to support the development of hydrological models. Chapter 4 describes the development of a high-resolution hydrodynamic surface water model, calibrated using groundwater level as a proxy for spring discharges. This parsimonious approach accurately simulated dry-weather flows over a five-year period (Nash-Sutcliffe efficiency = 0.97, mean relative error = 2.86%). The model’s simplicity enhances its transferability to other groundwater-dominated catchments and offers a practical tool to assess point-source pollution risk while avoiding the complexity of fully integrated groundwater-surface water models. Chapter 5 presents the results from a high resolution water-quality monitoring campaign focused on reactive nitrogen dynamics during extreme baseflow conditions. A mass balance approach, illustrated as a Sankey diagram (building on Chapter 3), quantified source contributions, showing that the majority of fluvial nitrate load originated in the first half (ca. 8 km) of the study reach, mainly attributed to a sewage treatment discharge (37% of accreted load) and Chalk spring discharges (55%). Chapter 6 builds on these findings by developing a high-resolution nitrate, ammonium, and dissolved oxygen model that integrates hydrodynamic processes with biogeochemical transformations. A model-sample fusion approach enhanced parameter calibration and enabled more accurate quantification of nitrate transformative processes by disentangling them from transport-related effects. Results showed that wastewater discharges dominate diel nitrate signals, even 18 km downstream, and highlighted the importance of model sample fusion to leverage the standalone value of high-frequency monitoring data. Finally, Chapter 7 adapts the water-quality model to speculatively predict EC risk, focusing on ibuprofen, carbamazepine, metformin, and steroid oestrogens as exemplary compounds. Using postcode-level prescription data, and explicit simulation of in-stream hydrodynamic attenuation processes, the model generated spatiotemporal risk profiles of each compound under baseflow conditions. By resolving diurnal flow and removal variability, the approach addressed limitations of existing steady-state approaches. This research demonstrates the potential for high-resolution tools to enhance water-quality risk assessment and improve river management, whilst also highlighting key areas for future research.

Date of Award	7 May 2025
Original language	English
Awarding Institution	University of Bath
Supervisor	Jan Hofman (Supervisor), Nicholas Howden (Supervisor), Ruth Barden (Supervisor) & Barbara Kasprzyk-Hordern (Supervisor)