AAAR 37th Annual Conference October 14 - October 18, 2019 Oregon Convention Center Portland, Oregon, USA
Abstract View
Gas and Particle Observations from Sensor Packages Deployed in Rural Malawi and a Near-Road Monitoring Location in North Carolina
ASHLEY BITTNER, Eben Cross, Carl Malings, Eric Lipsky, Andrew Grieshop, North Carolina State University
Abstract Number: 484 Working Group: Air Quality Sensors: Low-cost != Low Complexity
Abstract Malawi, like many other Sub-Saharan African countries, is a low-income nation with a near-complete dearth of ambient air quality (AQ) observations. Reference-grade monitors have historically been cost-prohibitive in these settings, but the advent of low-cost sensor packages has created an opportunity for AQ monitoring in previously inaccessible regions. Consequently, since June 2017 we have deployed three low-cost, low-power sensor packages (Aerodyne Research Inc.’s ‘ARISense’) to three rural sites in Malawi to collect data on gases (CO2, CO, NO/NO2, O3), particles (size/count) and meteorological parameters. Before and after deployment to Malawi, the packages were collocated with regulatory monitors at a near-highway site in Durham, NC. Temperature, RH, and concentration-span data collected in NC and Malawi suggest the collocation and deployment settings are environmentally similar. Using the two collocation data sets, we compare the performance of a set of hybrid random-forest (RF), regression, and neural network calibration models to reproduce accurate concentration values and account for cross-sensitivities in the electrochemical gas sensors. We apply the calibrations to 13 months of ambient data collected in Malawi and find consistent diurnal trends (indicative of peak household cooking periods) and similar annual trends (consistent with regional agricultural-burning periods). We also compare measurements from the optical particle counter (OPC) in the sensor packages to in-situ optical nephelometer and integrated filter measurements collected using RTI’s MicroPEM devices under ambient (N = 3) and near-source conditions (N = 14). We compare the real time data to investigate the reproducibility of particle mass estimates across the two instruments and evaluate the MicroPEM response relative to varying OPC size distributions. Additionally, we examine the influence of relative humidity and other environmental factors on the OPC response. Lastly, we discuss the real-world applications and feasibility of using low-cost sensor packages in remote and challenging environments.