Faecal pollution from rural septic tank systems: Modelling environmental risk and understanding human perceptions

Abstract

Septic tank systems (STS) are widely used in rural areas without centralised sewerage, but they are increasingly recognised as potential sources of faecal pollution, especially under changing climate conditions. Such pollution, commonly measured by reporting levels of faecal indicator organisms (FIOs), typically Escherichia coli (E. coli), arises when effluent containing high loads of FIOs bypasses adequate attenuation processes and reaches surface and groundwater. Quantifying STS-related pollution is difficult in large catchments with multiple sources, as traditional FIO monitoring lacks source specificity. This challenge is compounded by fragmented STS-specific data (e.g., condition, treatment performance), limited understanding of user perceptions of STS-related pollution risks and maintenance practices, both critical for pollution management. This thesis applied a mixed-methods approach using a Bayesian Network (BN) and user surveys to provide a holistic understanding of STS-related faecal pollution. The specific objectives were to 1) adapt an existing phosphorus risk BN model and extend it to account for FIO pollution from STS; 2) elicit expert judgements to inform BN structure and aid our understanding of FIO losses from STS to the wider environment; 3) validate a BN model against observed data to characterise FIO losses from STS in rural catchments, and 4) understand user perceptions and attitudes towards STS-related faecal pollution risks. Findings show that BNs can be effectively adapted for pollutants such as phosphorus and FIOs, applied from catchment to national scales, and be informed by expert judgement to assess STS pollution risks. Validation confirmed that the BN presented in this thesis is fit for purpose and provides credible risk assessments of STS contributions to watercourses. Self-reported STS user data proved valuable for refining model outputs and capturing awareness and perceptions of pollution risks. This research advances the application of BNs in modelling faecal pollution from STS and highlights the importance of integrating user-reported data. Future work should explore the extent to which such data can enhance model performance, supporting both scientific understanding and effective management of STS-related pollution risks.