Array of Things: A high-density, urban deployment of low-cost air quality sensors
Abstract
The Array of Things (AoT) is a collaborative effort among leading scientists, universities, local government and communities in Chicago to collect real-time data on the city's environment, infrastructure, and activity for research and public use. The AoT is composed of nodes that will measure and sense the urban environment of Chicago and provide openly accessible data in near real time. One component of each node is the ChemSense board, which uses chemical sensors to measure five gas-phase species: ozone, nitrogen dioxide, carbon monoxide, sulfur dioxide and hydrogen sulfide. In addition, the ChemSense board provides information on total reducing gases and total oxidizing gases. The nodes also include meteorological information and cameras that will provide pedestrian and traffic counts using computer vision algorithms. Because the ChemSense boards rely on low-cost sensors, characterizing the sensor responses is critical to understanding the applicability of the AoT for urban air quality issues. Initial results presented previously are promising for ozone, but much less so for sulfur dioxide. Previous results for nitrogen dioxide have unexplained spikes not observed in the EPA data that drive a poor fit. Current work focuses on using spatial techniques and meteorological information to verify the manufacturer's calibration algorithms and constants. Importantly, the zero current of the low-cost sensors is a strong function of temperature and is also potentially subject to drift. Using relatively clean air events, as determined from the sparse EPA measurements and meteorological information, the low-cost sensor calibrations can be characterized. For example, CO measurements show a strong exponential relationship with ambient temperature that is not present in a reference instrument. Once corrected, the relationship of the low-cost sensor to a research-grade carbon monoxide sensor improves markedly. This indicates that the current calibration algorithm and constants need to be adjusted, and this adjustment algorithm should be usable across the network of sensors.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2019
- Bibcode:
- 2019AGUFM.A24G..04P
- Keywords:
-
- 3305 Climate change and variability;
- ATMOSPHERIC PROCESSES;
- 3322 Land/atmosphere interactions;
- ATMOSPHERIC PROCESSES;
- 1632 Land cover change;
- GLOBAL CHANGE;
- 1637 Regional climate change;
- GLOBAL CHANGE