Unified Calibration and Spatial Mapping of Fine Particulate Matter Data From Multiple Low‐Cost Air Pollution Sensor Networks in Baltimore, Maryland

Low‐cost air pollution sensor networks are increasingly being deployed globally, supplementing sparse regulatory monitoring with localized air quality data. In some areas, like Baltimore, Maryland, there are only a few regulatory (reference) devices but multiple low‐cost sensor networks. There are many available methods to calibrate data from each network individually, including the recently proposed Gaussian process filter (GP filter) method, which mitigates the underestimation issue of other calibration methods, models spatial correlation, and yields a dynamic calibration equation. However, separate calibration of each network using a GP filter or any other calibration approach leads to conflicting air quality predictions. In this manuscript, we extend the GP filter to jointly model data from multiple low‐cost networks and reference devices. The approach provides dynamic calibrations (informed by the latest reference data) and unified predictions (combining information from all available low‐cost and reference sensors) for the entire region. This method accounts for network‐specific bias and noise, as different networks can use different types of sensors, and uses a Gaussian process to capture spatial correlations. We apply the method to calibrate PM 2.5$$ {}_{2.5} $$ data from Baltimore in June and July 2023 ‐ a period including days of hazardous concentrations due to wildfire smoke. Our method helps mitigate the effects of preferential sampling of one low‐cost sensor network in Baltimore, resulting in better predictions and more precise credible intervals. Our approach can be used to calibrate low‐cost air pollution sensor data in Baltimore and other areas with multiple low‐cost networks.