In Part II, I will start by reviewing the many forms of loose coupling of chaotic systems that give rise to synchronized motion, to argue for the ubiquity of the phenomenon, as illustrated by predicted relationships between large-scale weather phenomena at distant points on the globe. The same phenomenon of synchronization of fluid-dynamical channel systems, coupled through only medium-scale Fourier components, also forms the basis for a new approach to data assimilation with an interpretation of one system as "truth," and the other as "model". It is argued that the sufficiency of coupling the medium-scale components for synchronizing the entire systems follows in part from the existence of inertial manifolds for the two systems separately.
I will conclude by suggesting that the synchronization-based approach to data assimilation is in harmony with a theory that human consciousness is a manifestation of brief periods of synchronized activity among widely spaced neurons. Related neural-synchronization models of auditory segmentation in the cocktail-party problem motivate potential applications to image segmentation and combinatorial optimization problems generally.