A systems-based approach to address pollution sources and impacts
Conference
65th ISI World Statistics Congress 2025
Format: IPS Abstract - WSC 2025
Keywords: bayesian, decision-making, probabilistic, systems
Session: IPS 766 - Modelling and Monitoring Water Resilience and Ecosystem Services
Wednesday 8 October 2 p.m. - 3:40 p.m. (Europe/Amsterdam)
Abstract
Diffuse and point pollution remain a significant source of water quality impairment, responsible for microbial, nutrient, and chemical contamination, including emerging contaminants such as antibiotics, biocides, persistent chemical pollutants, and pharmaceuticals. Some of these contaminants arise from the disposal of sewage sludge on land, others originate from slurry applications, farmyards and septic tank systems. However, quantifying these sources and assessing the impacts on human and ecological health is often hampered by lack of available data. Future climate, land use and demographic changes, including more frequent extreme hydrological events such as floods and droughts, will exacerbate the pressure on both water quality and water resources. Hence, addressing these pressures will require a holistic system-based approach to address multiple pollution sources and impacts.
Here, we introduce a suite of system-based probabilistic modelling tools, based on Bayesian Belief Networks (BBNs), to simulate the likely critical source areas and delivery pathways of faecal indicator organisms, pesticides, pharmaceuticals and nutrients from multiple sources (diffuse, farmyards, septic tank systems, sewage treatment works) from catchment to regional scales. These graphical models capture the uncertainty associated with both data and knowledge and can be applied at river network scale and as probabilistic maps to simulate the likely effectiveness of a range of mitigation options – both technological and nature-based. Challenges associated with the application of BBNs in water quality assessment are also discussed.
The findings to date suggest that an integrated system-based approach focussing on the mitigation from all sources will be needed, including a significant reduction in inputs from agricultural land use to achieve ecological and raw drinking water targets (e.g. up to 50% reduction in pesticide application rates and reduction in P inputs to or below the current agronomic optimum). Ongoing research focusses on simulating additional interventions (e.g. sustainable urban drainage systems, domestic sewage treatment, constructed wetlands and retention ponds) and the impact of pollutant mixtures on river ecology under future climate and socio-economic scenarios under uncertainty.