Informing COVID-19 Spatial Distribution with Wastewater Measurements in Chicago

Confirmed cases from individual testing have been the main source of information of the spread of COVID-19 during the pandemic. However, since testing is not done randomly or thoroughly, the data set is incomplete and presents biases. Other sources, such as data of hospitalization and mortality rates, have been used successfully to better estimate real prevalence. But they often lack spatial resolution and are significantly lagged.Environmental surveillance has been used successfully to track infectious diseases and pharmaceutical compounds in the past and has the potential to complement the deficiencies of other sources of data in this pandemic. To estimate prevalence of infectious diseases from wastewater measurements statistical methods have mostly been used. In this work, we use an alternative approach that considers the processes that occur within the sewer networks to obtain best estimates of spatial distributions and associated uncertainties.The basis of the model is the use of Bayes' theorem to update the best estimates. Confirmed cases and data from established models are used as a prior distribution, while a likelihood distribution is used for the wastewater measurements. The resultant posterior distribution gives us the new best approximation of the prevalence. A hydraulic model of the combined sewer system is used to determine the likelihood distribution. The model considers dry weather and wet weather flow conditions.Data and models from various sites of different scales in the Chicago area are used to answer the central question: What information do wastewater measurements provide for early warning and general diagnostic systems? Indicators of the utility of measurements and optimal sampling strategies are discussed to answer this question.