Study On High Resolution Retrieval Of Aerosol Optical Depth And Its Correlation With Land Surface Features And Thermal Environment

Aerosol optical depth (AOD) is not only an important factor of affecting quantitative retrieval of remote sensing, but also the key paramerter in earth-atmosphere system. It affects the balance of surface enegy payments with other surface parameters. Besides of these parameters, AOD is also closely related with landscape metrics and land surface temperature (LST). Study of the temporal-spatial characteristics of the correlation between AOD and these variables are valuable for reductiing air pollution and improving urban thermal environment, and is helpful for understanding the relation between LST and the atmosphere optical effects.In this paper, the optimal scale of AOD, as well as the relationships between AOD and land cover/land surface parameter/landscape metrics are studied. Dense dark vegetation algorithm and deep blue algorithm are employed to retrieve the AOD of Beijing and the surrounding area based on Landsat8 data and MODIS surface reflectance products. The result is vaildated referring to AERONET data. Based on the AOD retrieved by dense dark vegetation algorithm, the multispectral remote sensing image is corrected pix by pix. Finally, the land surface temperature (LST) is retrieved by improved mono-window algorithm and the relationship between AOD and LST is stuied by correction analysis. The main conclusions can be summarized as following.(1)The dense dark vegetation algorithm is more accurate than the deep bule algorithm on the whole. According to the theory of the spatial dependence, the optimal scale(240m) is achieved.(2)There is a significant correction between the distribution of AOD and road. With the distance to road increase, the AOD increases at first and then decrease.(3)The AOD varies with land types in the following order overall:forest land< grassland<cultivated land<water area<construction land. The realtionship between AOD and land surface parameters varies with the altitude. At lower altitudes, the correlation between AOD and NDVI, LSWI, UI is strong. NDVI has the greatest effect on AOD, LSWI takes the second place and UI is in the last. At higher altitudes, the correlation between AOD and altitude shows significant, but the correlation between AOD and surface parameters shows very weak. Besides, the effect of these parameters on AOD is uncertain.(4)The correlation between AOD and LPI, PLAND, AI is significant negative for vegetation, while significant positive for impervious areas.(5)There is a significant positive correlation between AOD and LST in terms of spatial relevance, while weak negative correlation between AOD and LST in terms of temporal relevance. The spatial positve correlation reflects the land cover types’ effect to AOD and LST, while the temporal negative correlation implies that the main effect of aerosols in the atmospheric radiation transfer is blocking the solar radiation and reducing the land surface temperature.