Complexity in the atmosphere

Two complexity measures, mutual information and statistical complexity, were applied to two types of atmospheric turbulence data sets. One data set is time series measurements of ocean surface winds. The other is a time-height, Doppler lidar image of vertical winds in the atmosphere. High values of complexity are seen where they might be expected: in the time series data in wind regimes where statistical models of turbulence perform poorly and in the time-height Doppler lidar images at the top of the boundary layer, where the atmosphere is undergoing a transition from an unstable to a stable flow regime. Finite-state machines known as “epsilon machines”, were constructed to model the information flow in the data. The results are a step toward establishing new, unbiased, modeling frameworks for atmospheric turbulence that will make optimal use of the computational resources being applied to sensing, data management, and numerical modeling of the atmosphere