资源环境科技发展态势分析平台

Classifying aerosol types using spectral aerosol optical properties plays a key role in promoting the accuracy of remote sensing observations and expanding our knowledge of the internal relations between aerosol and climate. However, studies on verifying the correctness for aerosol type classification methods are particularly scarce. In this study, data from three month optical and chemical measurements of atmospheric aerosols in Harbin, Northeast China, were used to verify the optical classification methods by comparing their results with the actual aerosol types. Aerosol optical properties like aerosol optical depth (AOD), Angstrom Exponent (AE), and single scattering albedo (SSA) were obtained through a CE318 sun-photometer, and aerosol type was then identified based on four previously published aerosol classification schemes. To acquire the actual aerosol states, satellite observations from MODIS and CALIPSO was combined and then the morphology, chemical composition, and molecular structure of sample particles were examined by energy-dispersive X-ray spectroscopy (SEM-DES), X-Ray fluorescence spectrophotometer (XRF), and Fourier transform infrared spectroscopy (FTIR) measurements. Results show that aerosol types over urban Harbin varied a lot during the studying period and mixed aerosol was the dominant type. The optical classification methods mentioned in this study could identify aerosol types correctly in most cases, but sometimes made improper estimations. Besides, they are able to descript the distinct variation of aerosol type on seasonal and daily scale, but at present the result is still too rough to discriminate the detailed changes of actual aerosol state accurately. The combination of different classification methods or chemical measurements would help reducing the misjudgment of aerosol types. Our findings demonstrate that a more comprehensive way to identify changes in aerosol state is urgently needed to enhance our understanding of the impacts of aerosols on climate and remote sensing.