Laboratory Evaluation and Analysis of Low-Cost NDIR CO2 Measuring Platforms for Determining Ambient CO2 Concentrations

High density CO₂ measurements are important for quantifying localized greenhouse gas emissions. These observations can improve the estimation of emissions from urban areas and improve regional scale emissions inventories. Many modern non-dispersive infrared (NDIR) sensors offer a platform for measuring CO₂ at a lower cost than traditional research-grade measurement systems and can thus be deployed in high density networks. To determine the accuracy and reliability of the low-cost NDIR measurement technology, a series of lab tests were performed on two different low-cost sensors. A Picarro Cavity Ring-Down Spectrometer (CRDS) is used as a reference alongside two SenseAir NDIR sensors of varying size, cost, and precision: a SenseAir K30 and a SenseAir High Performance Prototype (HPP). Ambient air is pumped into both the CRDS and into a sealed box containing the two NDIR sensors. To maintain the accuracy of the reference CRDS, periodic calibrations are done on the instrument using air containing a known CO₂ concentration. A time series analysis is then performed to correct for time offset and zero/span offset of the NDIR sensors. A regression analysis is also performed for each NDIR sensor against the CRDS to determine the error of the sensors and to correct for bias in the NDIR sensor data due to cross sensitivities to temperature, pressure, and humidity. The regression analysis improves the accuracy of the low-cost sensors when they are compared to the CRDS. Observations were recorded periodically over several months to characterize long-term drift in the NDIR instruments.