A quantitative microbial risk assessment (QMRA) framework for optimizing harvestable volume in stormwater reuse systems

Urban stormwater reuse systems represent a critical ecotechnology for restoring hydrological cycles and providing mutual benefits for human populations and natural ecosystems. The design of these systems requires balancing ecosystem service delivery (maximizing harvestable volume) with human health protection (minimizing pathogen risk). This study presents a novel framework integrating Quantitative Microbial Risk Assessment (QMRA) with stochastic hydrological modelling to optimize stormwater harvesting system design. A quantitative, process-based relationship linking harvestable volume to human health risk, measured in Disability-Adjusted Life Years (DALYs), was established. Applying logistic regression to synthetic time series from long-term rainfall data created a computationally efficient predictive model for risk assessment. In an urban catchment case study, results demonstrated that increasing the harvestable volume from 3000 m3 to 6000 m3 per event increased annual water yield by 40%, but elevated mean annual risk from 10−7 to an unacceptable 10−5 DALYs. The logistic regression models (AUC > 0.90) identified an optimal design volume of 4500 m3, which maximizes water reuse while ensuring risk remains below the WHO benchmark of 10−6 DALYs. This optimal scenario concurrently prevents over 60% of untreated runoff volume from discharging into receiving waters, directly contributing to stream restoration.This approach advances ecological engineering by providing a transferable, quantitative tool for designing urban water ecosystems that restore natural hydrological processes while ensuring human safety. The methodology enables practitioners to optimize harvesting infrastructure without computationally intensive simulations, supporting sustainable urban ecosystem rehabilitation.