Estimation of atmospheric temperature using airborne microwave remote sensing

Kurzfassung

The atmosphere surrounding the Earth makes the planet habitable, therefore, the status of the atmosphere is of great importance for human activities and natural phenomena. As the arising atmospheric issues such as climate change, air pollution and ozone depletion have extracted extensive attraction worldwide, observing and modeling of atmospheric quantities becomes critical to our understanding of the environment. In order to investigate the atmosphere, remote sensing has been widely and successfully used to collect global and regional observations. In this thesis, the research focuses on the performance of one high-resolution airborne passive microwave spectrometer which is called MTP (Microwave Temperature Profiler). Our objective is to obtain a reliable vertically distributed atmospheric temperature profile from intensities that are measured by the instrument at certain frequencies and angles. To accomplish this purpose, an efficient tool TIRAMISU (Temperature InveRsion Algorithm for Microwave Sounding) is utilized for processing the MTP data and characterizing the atmospheric temperature. As for the forward simulation, a line-by-line radiative transfer model is an essential prerequisite for the analysis of data recorded by the spectrometer. As for the inversion, the solution to an ill-posed problem can be easily influenced by perturbations. To solve the ill-posed problem, means of numerical regularization can be applied by adding a penalty term onto the objective equation such that the new solution becomes less sensitive to perturbations. A direct method called Tiknonov Regularization (TR) and two indirect methods called Regularizing Levenberg-Marquardt (RLM) and Iteratively Regularized Gauss-Newton (IRGN) are used and their numerical performance are analyzed. The method IRGN and the matrix L1 are found to be the optimal setting for the equidistant vertical grid. The numerical analysis can help us to better understand pros and cons of these regularization methods and to investigate the measurement capabilities of MTP.