Comparing the use of mobile device and flow data to characterize dynamic populations in wastewater surveillance monitoring areas

Two wastewater monitoring areas connecting to a graph. On the x-axis is the month of the year; on the y-axis is population. Mobile device counts and flow rates both show fluctuations, but mobile device counts have larger swings. Mobile device data has the potential to capture fluctuations in populations contributing to a wastewater catchment area, which can aid in interpreting pathogen concentrations. We evaluated the use of mobile device data to characterize population dynamics as an alternative to flow rates, which are commonly used to normalize pathogen concentrations. Analyses were conducted at the daily and monthly scale across 66 sewersheds in Colorado from August 2020 to December 2024. Using K-means clustering, we identified two distinct patterns of population dynamics using mobile device data, separating high-tourism areas from other areas. There were significant fluctuations in daily and monthly device counts, with more pronounced patterns at the monthly level. Device counts and flow rates were not consistently well-correlated over time. Consistent with well-documented patterns of urban drainage, we found that rainfall and snowmelt explained periods of decoupling between device counts and flow rates. Mobile device data can be used to describe population dynamics over multiple geographical areas and over short and long timescales independent of the environmental processes that impact flow. It should be considered in efforts to improve normalization algorithms for wastewater-based epidemiology.