Atmospheric Carbon Dioxide （CO₂）Retrieval And Sensitivity Studies From Satellite Observations

Satellite remote sensing of atmospheric CO2is of great significance to many important scientific issues, such as investigations on sinks and sources of atmospheric CO2and global climate changes, due to the real-time and global monitoration of CO2concentratoins with high precisions. From the beginning of this century, researchers abroad started to work on this subject, and so far made some advances. However, in china, all the works are still in early stages at present. In this paper, we focused on the investigations of atmospheric troposphere CO2retrieval and sensitivity studies from satellite observation, which gathered experience for our developing technology for monitoring CO2concentratoins from satellite observation. Firstly, we made the quality analysis on the product of tropospheric CO2concentrations. Focused on the region of china, the works on the space-time distribution characteristics of CO2and its contributing factor have been discussed. Secondly, we resumed the theory of monitoring the concentration of atmosphere CO2with near-infrared remote sensing and the principle of retrieval. Based on the atmospheric radiative transfer principle, the error estimation and retrieval sensitivity of tropospheric CO2concentration products have been investigated and the tropospheric CO2concentration retrieval model has also been established. This is the realization of the tropospheric CO2concentration retrieval. Finally, the sensitivity studies were investigated on retrieval parameters to evaluate the estimate and resource of retrieval errors. The paper is mainly divided into the following sections:1. For the development of the hyperspectral technology and the needs for the atmospheric CO2monitor, the AIRS sensor selected as a research object, firstly we evaluated the sensor parameters and chose the CO2observation data, provided by the World Data Center, to compare worldwide CO2concentration products of AIRS for the analysis of the data quality. The data of six sits above the altitude of3000m was selected to investigate the time series and spatial error distribution, and the relationship between AIRS tropospheric CO2product data error and latitude and seasonal variation was obtained. The result of CO2concentration by AIRS satellite observations was in good agreement with that by station observation with error in the range of2ppmv. In the Northern Hemisphere with high density of human activity, the correlation coefficient is low; the difference is high, but in the sparsely populated southern hemisphere, a high correlation was obtained, which shows the great impact of human activity on the distribution of CO2concentration.2. Based on AIRS data of CO2concentration product, the temporal and spatial distribution characteristics of tropospheric CO2concentration in china were investigated. In areas of china, troposphere CO2distribution in middle area generally shows a north to south distribution, high-value areas are concentrated in the35°-45°N. Due to the seasonal meteorological characteristics, the middle troposphere CO2concentration showed significant seasonal variation, high in spring and low in autumn. Then, we investigated the influence of the NPP, precipitation and energy emissions on CO2spatial and temporal distribution. The results showed that the CO2concentration mainly was impacted by the delivery of the northern hemisphere atmosphere in china. In some area, the distribution and growth rate of the CO2concentration are lower than those in the north due to local rainfall erosion and NPP factor.3. Basic theory of infrared thermal radiation and adsorption property of the main atmosphere adsorbed gases were introduced. And, infrared thermal radiation transfer equation in the atmosphere, which is the calculation formula of satellite forward modelling method, was also induced. Based on CO2retrieval theory in the infrared bands and VPD algorithm, retrieval model of tropospheric CO2concentration was constructed with multi-source satellite data as input and small aircraft data as the validation data. We selected five sample points between0-30°N in the Pacific to process the inverse modeling experiments. The results show that its accuracy in the CO2concentration is less than1%, and meantime CO2concentration retrieval errors with a certain latitude and seasonal dependency.4. According to the retrieval theory and atmospheric radiative transfer model, the impact factors of CO2retrieval accuracy were investigted to make evidence on analysis of error resources. The ability of the channel for remote sensing of atmospheric CO2concentration in ARIS was also studied. Based on atmospheric molecular absorption spectral data, the line integral radiative transfer model was used to theoretically calculate the radiation values of infrared CO2concentration by satellite, and to investigate the dependence of different parameters of linear function and their impacts on sensitivity of CO2concentration. Meanwhile, the effect of other atmosphere factors of temperture, H2O and O3on the sensibility of CO2retrieval were also calculated to evaluate the precise and error resource of CO2retrieval, which attribute to the development of retrieval technique and improvement of retrieval accuracy. The results show that, in most cases, the temperature and O3data is the largest interference factors, and H2O is relatively a smaller influence factors especially in the high latitude winter. CO2concentration retrieval error caused by the temperature, H2O and O3was less than1%in the tropics and polar regions.