Assessing Climate- and Land-Use-Driven Water Quality Change in Irish Catchments Using Statistical and Spatial Modelling

Understanding the responses of riverine nutrient concentrations to combined land-use and climate pressures is essential for effective catchment management and water quality protection. However, water quality monitoring data are frequently sparse and irregularly sampled, particularly in regions with limited resources, presenting a widespread challenge for the robust analysis of nutrient dynamics. Generalized Additive Models (GAMs) provide a flexible statistical framework capable of capturing non-linear relationships, accounting for seasonal and interannual variability, and handling uneven temporal observations, making them well suited for analysing limited water quality datasets.

This study investigated the historical changes in nitrate and phosphate concentrations across three Irish river catchments representing contrasting land-use patterns: the predominantly rural Midleton catchment, the semi-urban Lee catchment, and the highly urbanised Liffey catchment. Observations collected between eight and fifteen times per year over an eight-year period were analysed using GAMs to quantify associations with climatic drivers, evolving land-cover characteristics, and temporal trends. The relationships between nitrate and phosphate concentrations were examined to identify how the two nutrients respond together under different environmental conditions.

To explore potential future trajectories, land-use and land-cover changes were projected using an Artificial Neural Network–Cellular Automata (ANN–CA) modelling framework. Spatially explicit land-cover scenarios were generated under two Socioeconomic Pathways: SSP4.5 and SSP8.5, representing moderate and high climate forcing and socio-economic development. These land-cover projections were integrated with corresponding climate scenario data to examine expected changes in both nitrate and phosphate concentrations, assessing how catchment characteristics modulate nutrient responses under alternative climate and land-use futures.

By applying a consistent analytical framework across rural, semi-urban, and urban catchments, the study enables a comparative assessment of how land-use intensity, hydrological context, and climate variability may influence the nutrient dynamics. Combining GAM-based statistical analysis with ANN–CA land-cover projections provide a reliable and adaptable approach for studying nutrient interactions in catchments with limited data. This framework can support evidence-based catchment management, nutrient control strategies, and the evaluation of possible future changes in water quality under different land-use and climate scenarios.