Abstract
The COVID-19 pandemic has emphasized the critical need for accurate disease modeling to inform public health interventions. Traditional reliance on confirmed infection data is often hindered by reporting delays and under-reporting, while widespread antigen and antibody testing can be costly and impractical. Wastewater-based surveillance offers a promising alternative by detecting viral concentrations from fecal shedding, potentially providing a more accurate estimate of true infection prevalence. However, challenges remain in optimizing sampling protocols, locations, and normalization strategies, particularly in accounting for environmental factors like precipitation. We present an integrative model that simulates the spread of serious infectious diseases by linking detailed infection dynamics with wastewater processes through viral shedding curves. Through comprehensive simulations, we examine how virus characteristics, precipitation events, measurement protocols, and normalization strategies affect the relationship between infection dynamics and wastewater measurements. Our findings reveal a complex relationship between disease prevalence and corresponding wastewater concentrations, with key variability sources including upstream sampling locations, continuous rainfall, and rapid viral decay. Notably, we find that flow rate normalization can be unreliable when rainwater infiltrates sewer systems. Despite these challenges, our study demonstrates that wastewater-based surveillance data can serve as a leading indicator of disease prevalence, predicting outbreak peaks before they occur. The proposed integrative model can thus be used to optimize wastewater-based surveillance, enhancing its utility for public health monitoring.
Introduction
Wastewater-based surveillance (WBS) emerged as a vital tool during the COVID-19 pandemic, providing a non-invasive method to monitor community health. Unlike clinical testing, which is often biased by healthcare access and reporting delays, WBS captures data from the entire population served by a sewer network. However, translating viral RNA concentrations into actionable epidemiological insights is fraught with complexity. This study addresses three primary sources of "noise" in WBS: human mobility, complex sewer hydraulics, and environmental interference such as rainfall and viral decay.
Methodology: A Coupled Digital Twin
To explore these complexities, we developed an integrative modeling framework consisting of three coupled modules:
- Infection Dynamics (ABM): Utilizing the MEmilio framework, we simulated 838 individual agents within a synthetic urban environment. Each agent follows a stochastic disease progression (Susceptible, Exposed, Infected, etc.) based on SARS-CoV-2 parameters.
- Shedding Model: This module links an agent's infection state to the quantity of viral RNA shed into the sewer system, modeled via a sigmoid function that captures the rise and fall of viral load over time.
- Wastewater Dynamics: Using the ++SYSTEMS environment, the model solves the Saint-Venant equations to simulate 1D fluid flow and reactive transport. This allows for the realistic simulation of viral "slugs" moving through a network of 16 measurement stations.
Simulation Setup and Datasets
The study utilizes a synthetic neighborhood featuring residential, recreational, university, and industrial zones to provide a controlled environment for testing. Rainfall data, sourced from the Bavarian Environment Agency (LfU), was used to simulate moderate gentle rain and moderate showers, allowing us to quantify dilution effects. Biological parameters, including viral shedding and decay rates, were calibrated using established literature on wild-type SARS-CoV-2.
Results and Discussion
The Lead Time Advantage
A primary finding of our simulation is the confirmation of WBS as a leading indicator. The peak viral load at the furthest downstream station occurred approximately 30 hours before the corresponding peak in true population prevalence. This lead time is critical for proactive public health responses.
The Normalization Trap and Environmental Factors
Rainwater infiltration significantly impacts WBS reliability. While it is common practice to normalize viral concentrations by flow rate to account for dilution, our model reveals this can be counter-productive during rain events. Rainfall increases flow velocity more rapidly than it dilutes the viral signal, leading to significant over-corrections and errors. In contrast, normalization against Pepper Mild Mottle Virus (PMMoV)—a dietary-linked virus shed at stable rates—proved far more robust, reducing median absolute error from 113 copies/L to just 8 copies/L during moderate rain.
Sampling Location and Decay Sensitivity
Sampling location is a major determinant of data quality. Upstream locations near residential areas are highly "noisy" due to the small number of contributors. Downstream stations provide smoother, more representative trends but are sensitive to viral decay. In scenarios with rapid decay (half-life of ~0.2 hours), signals from distant residential areas were largely lost, meaning downstream stations primarily monitored the immediate vicinity, such as nearby business districts.
Conclusion
Our integrative model demonstrates that while WBS is a powerful predictive tool, its interpretation must account for sewer physics and environmental conditions. By identifying the limitations of flow-rate normalization and the benefits of PMMoV normalization, this research provides a roadmap for optimizing wastewater-based surveillance to improve public health monitoring and outbreak prediction.
Summary updated