Spatiotemporal modeling for wastewater surveillance and epidemiology

Wastewater-based epidemiology (WBE) has emerged as a critical tool for public health surveillance, particularly during the COVID-19 pandemic, by enabling the detection of disease biomarkers such as ribonucleic acid (RNA). WBE implicates collaboration from multiple stakeholders to be effective as an early warning system to detect outbreaks and provides insights into virus and substance consumption dynamics across time and space. However, challenges remain in developing a collaborative framework for modeling to interpret infected wastewater data accurately, particularly in estimating the number of infected individuals. This complicates WBE assessments and the design of effective surveillance sampling strategies to define where, when, and how often to sample. Additionally, domestic wastewater (DW) sanitation is a fundamental human right recognized by the United Nations. Yet, access to reliable sanitation data is limited due to the high cost and complexity of collecting wastewater pollution data at multiple locations over time. There is also a need to advance spatiotemporal models for wastewater pollutants, which support sanitation and can be used for WBE. Filling the framework and modeling gaps for advancing epidemiology and sanitation can be enhanced by adopting open science (OS) principles, particularly computational reproducibility, which enhances the transparency and accessibility of outbreak and wastewater models. This research integrates WBE, DW sanitation, and the adoption of OS to advance wastewater monitoring methodologies, improve public health interventions, and promote sustainable, data-driven decision-making in global sanitation and epidemiology efforts. The development of spatiotemporal WBE and sanitation frameworks and models has been slow, as indicated by the limited literature in the field. A major contributing factor is the challenge of generating publicly accessible spatiotemporal data, which remains largely unavailable. Researching the spatiotemporal dynamics of outbreaks and pollutants in wastewater simultaneously is particularly challenging, as it involves emergent random patterns within complex systems. To address this challenge, this study adopts an Agent-Based Modeling (ABM) approach, which is well-suited for studying complex phenomena where data availability is scarce. This PhD thesis presents three studies that address these challenges. In the first study in Chapter 2, a WBE framework was designed to guide a collaborative modeling process for the development of sustainable wastewater surveillance. This study proposes that, for modeling to be effective, researchers must coordinate with public health authorities and laboratory services to ensure the relevance of model outputs for informed decision-making. The framework is demonstrated through a COVID-19 case study, which addresses the questions of when, how often, and where to sample wastewater to detect and monitor an outbreak. The second study in Chapters 3 and 4 presents a calibrated and validated spatiotemporal domestic wastewater model to analyze pollutant variability and its implications for measuring treatment efficiency. An ABM approach is employed to assign probabilistic distributions to human behaviors (specifically population mobility and water appliance usage) to simulate wastewater dynamics across time and space. The model results indicate that wastewater variability increases at higher spatiotemporal resolutions. To improve the comparability of treatment efficiencies and mitigate uncertainty caused by spatiotemporal variability, it is recommended that sewage catchment areas, population sizes, and sampling times and intervals be reported. The proposed model serves as a valuable tool for understanding wastewater variability. The third study in Chapter 5 extends the wastewater and population mobility model by integrating a generic disease model. This study addresses the technical challenges of synchronizing multiple models and validates their consistency in representing spatiotemporal pollutant and disease variability patterns. The integrated model successfully replicates epidemic curves, using COVID-19 as a case study, and estimates daily infections at the household level. Furthermore, it tracks the production of infected wastewater over time and spatially across sewage networks and treatment plants. Collectively, these studies contribute to validated and consistent models that realistically represent outbreaks and wastewater dynamics in time and space alongside a collaborative framework. The first study establishes a WBE framework that outlines stakeholder activities for sustainable surveillance programs. The second study delivers a calibrated and validated spatiotemporal wastewater model that captures the dynamics of wastewater across sewage networks and treatment plant. The third study successfully integrates a generic disease model into the wastewater model, realistically replicating epidemic outbreaks where inhabitants produce infected wastewater. A simulated COVID-19 case study demonstrates the estimation of infected inhabitants based on samples taken from multiple locations at different times and frequencies. Finally, all studies adhere to open science principles, ensuring that research software is publicly accessible to facilitate the reproducibility of results.