A220-0009
Field evaluation of low-cost particulate matter sensors in Beijing
Wednesday, 16 December 2020
Poster
Han Mei1, Pengfei Han2, Yinan Wang3, Ning Zeng Dr.4, Di Liu5, Qixiang Cai6, Zhaoze Deng7, Yuepeng Pan1, Xiao Tang1 and Yinghong Wang1, (1)Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China, (2)Corresponding Author, Beijing, China, (3)Corresponding Author, LAGEO, Beijing, China, (4)Institute of Atmospheric Physics, Chinese Academy of Sciences, State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics, Beijing, China, (5)LASG, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China, (6)Insititute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China, (7)IAP Insititute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China
Abstract:
With the development of microelectronics and other technologies, low-cost optical sensors have been widely applied to monitor the concentrations of particular matter such as PM2.5. High-cost and high-precision reference instruments always adopt higher stability laser and are equipped with more precise environment control device and more refined air pump to sample the atmosphere. However, low-cost sensors are limited by size and cost, and there is no such control equipment like environment control device, so the data quality will be greatly reduced. This study quantitatively assessed the influence of the field environment (PM concentration, temperature, humidity) on the data quality of low-cost sensors without environment control device.
We evaluated five Plantower PMSA003 sensors deployed in Beijing, China, over 7 months (October 2019-June 2020). The sensors tracked PM2.5 concentrations, which were compared to the measurements at the national control monitoring station of the Ministry of Ecology and Environment (MEE) at the same location. The correlations of the data from the PMSA003 sensors and MEE reference monitors (R2=0.83~0.90) and among the five sensors (R2=0.91~0.98) indicated a high accuracy and intersensor correlation. However, the sensors tended to underestimate high PM2.5 concentrations. The relative bias reached -24.82% when the PM2.5 concentration was >250 µg/m3. Conversely, overestimation and high errors were observed during periods of high relative humidity (RH>60%). The relative bias reached 14.71% at RH >75%. The PMSA003 sensors performed poorly during sand and dust storms, especially for the ambient PM10 concentration measurements.
Overall, we found that PMSA003 is promising for dense network PM monitoring. However, there are still some problems with the data quality of uncorrected low-cost sensors. The influence of the environment needs to be considered when applying low-cost PM sensors to the field. This study revealed the biases and limitations in using it, the implications of these results will be beneficial in future corrections of low-cost sensor data and dense network observations.
