Numerical Regularization for Atmospheric Inverse Problems

The retrieval problems arising in atmospheric remote sensing belong to the class of the - called discrete ill-posed problems. These problems are unstable under data perturbations, and can be solved by numerical regularization methods, in which the solution is stabilized by taking additional information into account. The goal of this research monograph is to present and analyze numerical algorithms for atmospheric retrieval. The book is aimed at physicists and engineers with some ba- ground in numerical linear algebra and matrix computations. Although there are many practical details in this book, for a robust and ef?cient implementation of all numerical algorithms, the reader should consult the literature cited. The data model adopted in our analysis is semi-stochastic. From a practical point of view, there are no signi?cant differences between a semi-stochastic and a determin- tic framework; the differences are relevant from a theoretical point of view, e.g., in the convergence and convergence rates analysis. After an introductory chapter providing the state of the art in passive atmospheric remote sensing, Chapter 2 introduces the concept of ill-posedness for linear discrete eq- tions. To illustrate the dif?culties associated with the solution of discrete ill-posed pr- lems, we consider the temperature retrieval by nadir sounding and analyze the solvability of the discrete equation by using the singular value decomposition of the forward model matrix.

Presents regularization methods for atmospheric retrieval, based on the authors work Focuses on computational aspects but also provides some theoretical results Surveys the state-of-the-art numerical methods for solving discrete ill-posed problems Anlayzes the existing numerical alorithms and discusses practical implementation issues Illustrates with examples from atmospheric remote sensing the variouis methods in action Includes supplementary material: sn.pub/extras

Autorentext

Adrian Doicu is Research Scientist at the Remote Sensing Technology Institute, DLR Deutsches Zentrum für Luft- und Raumfahrt e. V., Oberpfaffenhofen, Germany, prior to this he was Research Scientists at the Department of Chemical and Process Engineering, University of Bremen, Bremen, Germany. Thomas Wriedt is Head of the Particle Technology and Particle Characterization Department of the Institut für Werkstofftechnik, Bremen, Germany. Yuri Eremin is Research Professor at the Applied Mathematics and Computer Science Faculty, Moscow State University, Moscow, Russia.

Klappentext

The subject of this book is a hot topic with currently no monographic support. It is more advanced, specialized and mathematical than its competitors, and a comprehensive book on regularization techniques for atmospheric science is much needed for further development in this field. Written by brilliant mathematicians, this research monograph presents and analyzes numerical algorithms for atmospheric retrieval, pulling together all the relevant material in a consistent, very powerful manner.

The first chapter presents the typical retrieval problems encountered in atmospheric remote sensing. Chapter 2 introduces the concept of ill-posedness for linear discrete equations, illustrating the difficulties associated with the solution of the problems by considering a temperature retrieval test problem and analyzing the solvability of the discrete equation by using the singular value decomposition of the corresponding matrix. A detailed description of the Tikhonov regularization for linear problems is the subject of Chapter 3, in which the authors introduce a set of mathematical and graphical tools to characterize the regularized solution. The goal of Chapter 4 is to reveal the similitude between Tikhonov regularization and statistical inversion regarding the regularized solution representation, the error analysis, and the design of parameter choice methods. The following chapter briefly surveys some classical iterative regularization methods such as the Landweber iteration and semi-iterative methods, and then treats the regularization effect of the conjugate gradient method applied to the normal equations.

Having set the stage in the first part of the book, the remaining chapters dealing with nonlinear ill-posed problems. The authors introduce four test problems that are used throughout the rest of the book to illustrate the behaviour of the numerical algorithms and tools. These deal with the retrieval of ozone and BrO in the visible spectral region, and of CO and temperature in the infared spectral domain. Chapter 6 looks at the practical aspects of Tikhonov regularization for nonlinear problems, while Chapter 7 presents the relevant iterative regularization methods for nonlinear problems. The following chapter reviews the truncated and the regularized total least squares method for solving linear ill--posed problems, and include the similarity with the Tikhonov regularization. Chapter 9 brings the list of nonlinear methods to a close. It describes the Backus-Gilbert approach as a representative member of mollifier methods and finally, addresses the maximum entropy regularization. For the sake of completeness and in order to emphasize the mathematical techniques which are used in the classical regularization theory, five appendices at the end of the book present direct and iterative methods for solving linear and nonlinear ill-posed problems.

Inhalt

Chapter 1. Atmospheric remote sensing

Chapter 2. Ill-posedness of linear problems

Chapter 3. Tikhonov regularization for linear problems

Chapter 4. Statistical inversion theory

Chapter 5. Iterative regularization methods for linear problems

Chapter 6. Tikhonov regularization for nonlinear problems

Chapter 7. Iterative regularization methods for nonlinear problems

Chapter 8. Total least squares

Chapter 9. Two direct regularization methods

Appendix A. Analysis of continuous ill-posed problems

Appendix B. A general direct regularization method for linear problems

Appendix C. A general iterative regularization method for linear problems

Appendix D. A general direct regularization method for nonlinear problems

Appendix E. A general iterative regularization method for nonlinear problems