Computational Methods for Data Evaluation and Assimilation

摘要: Experimental Data Evaluation: Basic Concepts Uncertainties and Probabilities Moments, Means, Covariances Computation of Means Variances from Measurements Statistical Estimation Covariances, Confidence Intervals Assigning Prior Probability Distributions under Incomplete Information Evaluation Consistent with Independent Random Errors Systematic Discrepant Unrecognized Notes Remarks Optimization Methods for Large-Scale Assimilation Introduction Limited Memory Quasi-Newton (LMQN) Algorithms Unconstrained Minimization Truncated-Newton (T-N) Hessian in Nondifferentiable Minimization: Bundle Step-Size Search Trust Region Scaling Preconditioning Nonlinearly Constrained Global Principles 4D VAR Nudging (Newtonian Relaxation) Optimal Interpolation, Three-Dimensional Variational, Physical Space Analysis Error Covariance Matrices Framework Time-Dependent Four-Dimensional Variational (4D VAR) Numerical Experience Using the Shallow Water Equations Treatment Model Weather Prediction Models The Objective Cost Functional Gradient Adjoint Coding FFT Inverse Developing Programs Interpolations "On/Off" Processes Construction Background Characterization Incremental Algorithm Appendix A Frequently Encountered B Elements C Parameter Identification Bibliography Index