The study of atmospheric aerosols lays the foundation of global climate change, air quality, and public health research. Different from those in the Chinese mainland, the special atmospheric circulation and local emission sources in Taiwan, china from its seagirt terrain have led to the difference in aerosols’ characteristics. Lack of information on aerosol temporal and spatial characteristics might cause errors in the retrieval of satellite aerosol parameters.On the basis of the historical data of several representative Aerosol Robotic NETwork (AERONET) sites in Taiwan, this study explored the temporal and spatial variation characteristics and differences of aerosol parameters and types of typical sites in Taiwan, China. First, the change trend of aerosol optical parameters was analyzed. The observation sample points of AERONET were divided into six categories, namely, maritime, continental, desert dust, sub-continental, urban industry, and biomass burning aerosols, using the graphical classification method. The differences in aerosol types at different sites and the effects of wind direction and speed on Aerosol Optical Depth (AOD) and aerosol types were explored, and they were comparing with the aerosol optical parameters of Beijing. Second, the MODIS data were validated against the AERONET data.The annual average of AOD at each station is decreasing annually, which suggests the highest seasonal variation in spring (0.5257) and the diurnal variation of bimodal structure. The dominant aerosol type is urban industrial, and only Lulin station is maritime type. The northeast wind prevails in Taiwan, China, and the AOD is lower and the maritime aerosol type occupies larger proportion when the wind speed is higher. Conversely, urban industrial aerosols dominate. The average values of Ångström Exponent, Single Scattering Albedo, Refractive Index-Imaginary Part, and Asymmetry Factor are 1.3283, 0.9564, 0.0054, and 0.7292, respectively. Compared with those in the Beijing site (39.9768°N, 116.3813°E), the annual average of AOD, seasonal variation, and dominant aerosol types in the “Central University” shows a dramatic difference. For the remote retrieval products, MODIS AOD has higher verification accuracy at different sites, and only the Lulin site is slightly lower (R²=0.5925). As for the verification results of different aerosol types, urban industry (R²=0.7238), biomass burning (R²=0.6161), and sub-continental (R²=0.5116) have higher accuracy, while maritime (R²=0.1585) and continental (R²=0.1111) have significantly lower accuracy.The types of aerosols in Taiwan, China show differences in temporal and spatial characteristics. The proportion of sub-continental types in the southwest coastal sites increases in autumn and winter, while the continental types in the northwest coast increase. Refining the characteristic changes of aerosol parameters plays an important role in guiding the aerosol satellite retrieval algorithm for island regions with distinct features of circulation.

atmospheric remote sensing;aerosol;AERONET;Aerosol Optical Properties;temporal and spatial analysis