Search

In this paper we discuss the application of observability as a measurement of observation quality for data assimilation in numerical weather prediction (NWP ). Observability is a measure of well-posedness of a dynamical system and provides a flexible framework to address questions in data assimilation. We review the concept of observability for differential equations and high dimensional numerical models. We discuss the relationship of observability to observation impact . We conclude with a discussion of various applications of observability to data assimilation including optimal sensor placement and data thinning .