| Atmospheric aerosols refer to solid and liquid particles suspending in the atmosphere.The aerodynamic diameter of particles is between 0.001 and 100 um,and they constitute a stable mixed system.Although aerosols are less abundant in the atmosphere,they have more significant influence on atmospheric visibility,solar scattering and radiation,and atmospheric temperature.Because of the small size and large superficial area of aerosol particles,they can produce a variety of chemical reactions,which has a variety of effect on the atmospheric environment as well as the health of human beings and other living creatures.With the rapid development of economy and cities,most areas of our country have different degrees of environmental pollution.And the haze weather is more frequent,by which people's daily travel and life have been affected,and health has also been seriously threatened.Although the ground monitoring stations can detect aerosol particles,the number of monitoring stations is limited,which can not reflect the spatial and temporal changes of aerosols in various specific places.MODIS has its own unique advantages in monitoring aerosol data,for it can provide more abundant and detailed aerosol data.Moreover,MODIS data acquisition is more convenient,which makes up for the insufficient data of ground-based stations.Aerosol optical depth(AOD)is one of the most important parameters of aerosol.It is defined as the integration of extinction coefficient in vertical direction,which describes the effect of aerosol on the weakening of solar light.Aerosol optical depth(AOD)is often used in the study of climate effect and atmospheric pollutants for it affects the propagation of solar radiation through the reflection,scattering and absorption by particulate matter,which directly or indirectly affects the atmospheric climate effect.However,due to the cloudy and rainy weather,remote sensing satellite sensors can not obtain effective AOD data,which leads to a lot of missing AOD data.This thesis utilizes DT AOD,DB AOD of the Terra satellite and DT AOD,DB AOD of the Aqua satellite which are acquired by MODIS remote sensor.Firstly,linear regression between DT and DB is used to fill in each other,and inverse variance weighting method is used to fuse the data.The fused Terra AOD and Aqua AOD are obtained and verified by matching with ground monitoring AERONET AOD data.On the basis of the previous step,the linear regression between Terra AOD and Aqua AOD is used to fill in each other,and the average value is obtained for data fusion.And the fused MODIS AOD is obtained and verified by matching with ground monitoring AERONET AOD data.On the basis of data fusion,90%of the sample points are used for ordinary Kriging interpolation,and the spatial correlation interpolation is carried out within the range of each AOD data.Finally,the remaining10%of the sample points are used for verification.The result shows that the correlation coefficient R~2 between MOIDS AOD data and AERONET AOD data reaches 0.92,RMSE is 0.20,and the average deviation is0.10,which indicate that the precision of data fusion is very high.Compared with the DB AOD data of the Aqua satellite with the highest original coverage,the average effective days per grid after data fusion increased from 199 days to 237 days in 2016,and the average daily coverage rate after data fusion increased from 54.46%to64.79%.After data fusion,MODIS AOD uses ordinary Kriging method to interpolate,and uses leave-one method to verify the accuracy.The correlation coefficient R~2between predicted value and test value is 0.99,RMSE is 0.08,the mean absolute error is 3.85%,and the average relative error is 11.33%.After interpolation,the effective days of each grid increases to 332,and the effective coverage of each grid increases to90.65%.The result shows that the linear regression data fusion method between DT and DB,Terra and Aqua satisfies the accuracy requirement;the ordinary Kriging interpolation method also achieves the accuracy verification;the coverage of AOD is greatly improved by using various AOD data fusion and interpolation methods,which provide technical support and data basis for future AOD and other pollution estimation studies using AOD. |
PDF Full Text Request