Bayesian Spatio-temporal Modeling of Multi-Site Wastewater Surveillance Data
Abstract
Wastewater surveillance provides population-level evidence of infectious disease activity. As surveillance expands across multiple sites, repeated measurements must be analyzed jointly while accounting for differences among sites and the public health regions they represent. We develop a Bayesian spatio-temporal framework that decomposes wastewater measurements into a shared epidemic curve, spatially correlated regional deviations, persistent site effects, and within-site temporal dependence. This hierarchical decomposition accommodates unequal observation windows and sampling schedules, allowing regional trajectories to be compared while preserving site-level heterogeneity. A direct marginal spatial covariance specification further separates the magnitude of regional variation from the strength of association between neighboring regions, providing distinct interpretations of these two sources of spatial variation. We also introduce region-specific time shifts to distinguish regional leads and lags in epidemic timing from differences in signal levels and local fluctuations. Simulation studies demonstrate the recovery of these components and show that omitting true time shifts can lead to increased estimates of residual temporal variation. Application to Ontario SARS-CoV-2 wastewater surveillance data reveals persistent site heterogeneity, within-site temporal dependence, and spatial association among neighboring regions. Regional timing differences are more pronounced during the first year of Omicron than in the XBB-and-later period.
Related
- ISBO: Scalable Spatio-Temporal Bayesian Optimization with Log Gaussian Cox Process Models via the INLA-SPDE Approach
- Three Case Studies of Two-stage MCMC for Fast Bayesian Inference of Large Spatio-temporal Data
- Scalable generative modeling of non-Gaussian spatio-temporal fields via autoregressive Gaussian processes
- Spatio-temporal Latent Denoising Diffusion Probabilistic Models for Reduced-order Modeling of Parametrized Dynamical Systems
- STCFormer: Adaptive Spatio-Temporal Modeling with Dynamic Cluster Transformer for Station-based Weather Forecasting
BibTeX
@misc{peng2026bayesian,
title = {{Bayesian Spatio-temporal Modeling of Multi-Site Wastewater Surveillance Data}},
author = {K. Ken Peng and Chengqian Xian and Charmaine B. Dean and Robert Delatolla and X. Joan Hu and Elizabeth M. Renouf},
howpublished = {arXiv},
year = {2026},
month = nov,
doi = {10.48550/arxiv.2611.07602},
eprint = {2611.07602},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2611.07602},
}
RIS
TY - UNPB
TI - Bayesian Spatio-temporal Modeling of Multi-Site Wastewater Surveillance Data
AU - K. Ken Peng
AU - Chengqian Xian
AU - Charmaine B. Dean
AU - Robert Delatolla
AU - X. Joan Hu
AU - Elizabeth M. Renouf
PY - 2026
DA - 2026/11/05
PB - arXiv
DO - 10.48550/arxiv.2611.07602
UR - https://arxiv.org/abs/2611.07602
AB - Wastewater surveillance provides population-level evidence of infectious disease activity. As surveillance expands across multiple sites, repeated measurements must be analyzed jointly while accounting for differences among sites and the public health regions they represent. We develop a Bayesian spatio-temporal framework that decomposes wastewater measurements into a shared epidemic curve, spatially correlated regional deviations, persistent site effects, and within-site temporal dependence. This hierarchical decomposition accommodates unequal observation windows and sampling schedules, allowing regional trajectories to be compared while preserving site-level heterogeneity. A direct marginal spatial covariance specification further separates the magnitude of regional variation from the strength of association between neighboring regions, providing distinct interpretations of these two sources of spatial variation. We also introduce region-specific time shifts to distinguish regional leads and lags in epidemic timing from differences in signal levels and local fluctuations. Simulation studies demonstrate the recovery of these components and show that omitting true time shifts can lead to increased estimates of residual temporal variation. Application to Ontario SARS-CoV-2 wastewater surveillance data reveals persistent site heterogeneity, within-site temporal dependence, and spatial association among neighboring regions. Regional timing differences are more pronounced during the first year of Omicron than in the XBB-and-later period.
ER -