Cloud-Aerosol Satellite Borne Lidar Data Retrieval Algorithms’ Preliminary Study

In the earth radiation budget, aerosols and clouds occupy a very importantposition. First, they influence climate and environment through changing the radiationbudget, at the same time, the aerosol particles which suspend in atmospheric will beinhaled into the lungs and then affects human’s health. At present, there are a varietyof aerosol detection methods. Most of the methods are ground-based systems.Although Ground-based network can obtain the spatial distribution of aerosols andclouds in certain regions, but for the complexity of aerosols distribution and thegeographical constraints, ground-based network can not obtain the global aerosolspatial distribution.Space-based lidar can be achieved on the global context of clouds and aerosolsfor a long time continuous detection, accurately identify the aerosols’ and clouds’ layerheight, determine cloud constituents and aerosol type. Based on these reasons, we canstudy their influence on global climate change effectively. Using the Level1B data ofNASA CALIPSO system which is focuses on the feature detection and layerproperties algorithms and scene classification algorithm of the space-based lidar, thispaper introduces the extinction retrieval algorithms briefly. At last, according to thedata processing method, this paper designs the corresponding data processingsoftware.First this paper introduces the feature detection and layer properties algorithms indetail, including the single profile scanning algorithms and Selective IteratedBoundary Location Algorithms. Using the feature detection methods, this paper dealwith the data from September to October in2011of Shandong province, some resultsare given in this paper. Then, using the SCA, this paper classifies the layers and givesthe results. It introduces the method to determine the lidar ratio briefly, retrieval theextinction coefficient using the fernald method. At last, using in Visual studio2008environment to call Matlab method, this paper designs the data processing software, which can effectively complete the profile mean, feature detection and featureclassification.