Statistics of atmospheric correlations

Abstract

Atmospheric correlation matrices were analyzed from the perspective of the random matrix theory. It was found that these matrices can be modeled as random matrices chosen from an appropriate RMT ensemble. The eigenvalue statistics exhibits short- and long-range RMT-type behavior. Most of the eigenmodes also follow the RMT-type eigenvector distribution. Few dominant eigenmodes that have physical significance deviate from RMT predictions.