About the book
Data assimilation (DA) is the process of updating model forecasts with information from the observations of complete or incomplete state variables. The objective is to develop an improved model that better represents the dynamical system. Conventional methods for assimilation include Kalman filters and variational approaches which allow for problems with uneven spatial and temporal data distribution and redundancy to be addressed in a way that models can ingest information. However, because these methods rely on Gaussian assumptions, performance is severely degraded when the prior facts are described in terms of complex distributions and based on unrealistic assumptions, particularly linearity, normality. Application of artificial neural networks (ANN) and machine learning (ML) can be exploited to forecast and to some extent, learn the dynamics of a model from its output and to obtain the initial condition precisely.
The book primarily addresses the student's needs in the field of data assimilation. It is intended to provide the reader with a comprehensive overview of the subject, from basic principles to advanced state-of-the-art in the interdisciplinary field of data assimilation. It is compiled in a pedagogical way to become a reference book for researchers interested in the application of advanced analytical tools, deep learning, ANN in data assimilation implementation, data analytics, machine learning, and decision support systems.