| At present our country is facing an unprecedented challenge in environmental issues, andthe air pollution problem has become an important factor affecting our country city and regionalin sustainable development and environmental diplomacy. It is particularly important to fullygrasp the regional distribution and change rule of atmospheric pollution for environmentmanagement and planning. China is vast area, but the general business of environmentmonitoring is often rely on ground monitoring routine. It has not yet formed an integratedmonitoring system. The master degree on the atmospheric environment quality of regional andglobal is still at a low level. Satellite remote sensing monitoring compared with the traditionalmethod of physical or chemical analysis and measurement which set up on the ground, has thecharacteristics of wide coverage, continuous, dynamic. It has the obvious advantages ofcontinuity, space and trend monitoring in the atmospheric environment quality changes. Butatmospheric environment application level is low in the domestic because of remote sensing dataand technical support conditions. Especially the atmospheric environmental satellite remotesensing monitoring of environmental protection is still in the initial stage.In this paper, we use Ji’nan city as an example and the environmental pollution in Ji’nan cityis very serious in recent years. Based on the MODIS data from2012to2013, we create a sampleof sunny image. We invert the Aerosol Optical Depth of Ji’nan region with the dark pixelalgorithm which currently has a mature system. At the same time, we get the instantaneousPM2.5concentration of all monitoring points in Ji’nan city at the corresponding period. Weanalyzed the characteristics of the overall pollution of Ji’nan PM2.5in the dimension of time anddimension of space based on statistical analysis principle. Finally we selects high quality parts inall the retrieval of aerosol optical depth and the instantaneous corresponding PM2.5concentration value of each monitoring point to do the correlation analysis and regression analysis. So we can get the selection of the best fitting model from all the different functionregression models.Based on the above research, we can draw the following conclusions.As the land surface types is varied, at the same time, most of1km scaled pixel in landsurface is mixed pixel, it is more difficult to accurately construct the pixel scaled surfacereflectance model so that accurately estimate reflection noise of surface. Aerosol scatteringinformation is weak especially for the reflective bright target, and the surface of high reflectivityestimation model often has error which could obscure the aerosol information. More maturealgorithm of land is only used for the surface type which has low reflectivity. Therefore in theMODIS load USA NASA specially designed shortwave infrared band(2.1μm) which has verysmall atmospheric effects, in order to get the band surface reflectance information, and then getthe surface noise algorithm of red and blue light band by the linear relationship with the red bandreflectance and blue band reflectance. The algorithm through continuous validation andimprovement by domestic scholars and foreign scholars has tended to be mature, and we call itthe dark pixel algorithm. Other satellite data which do not have similar bands need to beidentified false dark target when use dark pixel algorithm in the retrieval of aerosol opticaldepth.Therefore, to the dark pixel algorithm, MODIS data is the most appropriate SatelliteSounding data.The air quality of Ji’nan City is a more serious level than other cities. Strengthen regulationand scientific prediction has been become more and more important, and it is a large factor ofurban economy. The characteristics of PM2.5pollution is presented as follows, in the space ofthe trend, Ji’nan city pollution characteristics of PM2.5in spring, autumn, winter are decreasedgradually from north to South. On the other hand, it is decreasing from the center to both sides inprojection from east to west. But in summer, oppositely, it is characterized by decreasedgradually from the center to both in North and South projection, and in what direction isgradually descends from west to East trend.When the Correlation analysis is used in the establishment of Aerosol Optical Depthretrieval results and PM2.5concentration, we found the correlation coefficient is-0.603, at thesame time significance (bilateral) is0<0.01, so it is significant to certain. Therefore, setting upobservation and forecasting system of PM2.5by using MODIS data to retrieve Aerosol OpticalIV Depth is scientific and feasible.According to the observation data and the inversion results, establish the regression modelof Aerosol Optical Depth and PM2.5concentration value from linear function, logarithmicfunction, two times function, three times function, power function, exponential function and soon. By comparing fitting results, related parameters and test accuracy, we will find that a cubicequation function has the best fitting effect and R2is above0.5. This function has a goodexplanation. |
PDF Full Text Request