Tijana Janjic Pfander | University of Munich | Germany
Data assimilation combines information in heterogeneous observations and a predictive model to learn about and predict phenomena of interest. It plays a key role in geosciences, astrophysics, finance, neuroscience and engineering, and utilises a vast range of mathematical and statistical techniques.
In this tutorial, we will introduce you to the basics of methods used in many meteorological offices world wide to initialize high resolution numerical models of the atmosphere that predict weather. We will focus on innovative algorithms that are required to deal with big data and big optimization problems in meteorology. We will show you how specifics of geoscience phenomena as for example governing physical laws, and noisy, non-uniform in space and time observations from many different sources or rare interesting events, are taken into account in this process.