| 英文摘要 |
The goal of this project is to audit instrument performance and validate data quality of the North Aerosol Supersite (Supersite) at the Taiwan Environmental Protection Administration (TEPA). To achieve this goal, work such as twice site inspections in every month, detailed calibration on instruments in every 6 months, and individual quality assurance meetings held by TEPA and this project in every month, respectively, were conducted. This year, the procedure of site inspection was renewed. The data acquisition rates for all instruments were above 93% except for 85% from carbon analyzer. For assessing performance of automated instruments on monitoring aerosol mass and compositions, this project conducted collocated comparisons between automated instruments and manual collection. The results showed that the variations of PM2.5 mass were quite consistent. PM2.5 carbon measurements tended to be lower. PM2.5 sulfate was with good agreement but was deviated in concentration. The agreement between automated monitoring and manual analysis for PM2.5 nitrate was greatly improved after compensating its volatile fraction. Owing to unstable conversion efficiency, out of production, and lacking technical support from the manufacturing company, the replacement of the latter two instruments with an appropriated system is imminent. Among the instruments in operation, a deviation of laser alignment was found for laser aerosol size spectrometer and is suggested to return to the manufacturing company for repairing. This year, the correlations between each paired measuring items were R2 =0.79 for PM10 and PM2.5 mass concentration, R2 =0.81 for PM2.5 mass concentration and total light-scattering coefficient, R2 =0.89 for PM2.5 mass concentration and backscattering light-scattering coefficient, R2 =0.76 for PM2.5 mass concentration and sulfate, R2 =0.46 for PM2.5 mass concentration and nitrate, and R2 =0.91 for black carbon (at 660 nm) and optical elemental carbon, respectively. All are above middle level in linear regression correlation.
|