National scale water quality modelling: Challenges and successes

Freshwater quality in the future will be determined by the combined effect of climate change, land-use changes, and socioeconomic developments, with important consequences for ecosystem health and clean water availability. Robust comprehensive modelling frameworks are therefore needed to quantify the impact of these pressures across space and time, while accounting for the uncertainty in projected changes. However, simulating multiple pollutants across complex hydrological systems over multi-decadal periods at national scale presents substantial methodological and practical challenges. Here, we present an integrated modelling framework for UK surface and groundwater quality and discuss challenges and successes in producing reliable national-scale projections.

We present the modelling framework developed within the Long-Term Large-Scale Freshwater Ecosystems (LTLS-FE) project, building on an existing long-term integrated model (LTLS-IM). It dynamically couples process-based representations of surface and subsurface hydrology, agricultural and seminatural soils, sewage and septic tank emissions, and in-stream transport and fate. Using observational datasets, we evaluate modelled freshwater concentrations for a wide range of substances, including fine sediment, macronutrients, metals, and diverse micropollutants such as pesticides, antibiotics, pharmaceuticals, personal care products, industrial chemicals, and polycyclic aromatic hydrocarbons.

The framework is applied to project water quality across the UK from 1981 to 2080 under six future scenarios combining UKCP18 climate projections (RCP2.6–RCP8.5) with different Shared Socioeconomic Pathways (SSP1–SSP5). Simulations are conducted at a two-hourly timestep on a 5 km grid, producing monthly outputs of river flow, temperature, pH, biochemical oxygen demand, pollutant concentrations, and fluxes to the sea.

Challenges include limited pollutant input and validation datasets, particularly for Northern Ireland, as well as slow groundwater equilibration periods. Integrating multiple process-based sub-models while maintaining computational efficiency also required careful model design and optimisation. Despite these challenges, comparisons with national monitoring data show that the framework captures observed spatial patterns, seasonal dynamics, and long-term trends for major pollutant groups. Here we present the project outputs compared against observations and interpret future trends. These results demonstrate the feasibility and value of national-scale, multi-pollutant modelling to support future assessments of water quality risks under future change.