Proper Orthogonal Decomposition of Mesoscale Vertical Velocity in the Convective Boundary Layer

The proper orthogonal decomposition technique is used to analyze wind-profiler observations from the surface to the top of the convective boundary layer. The mesoscale thermal structures are identified by decomposing the vertical velocity measurements into a sequence of characteristic modes with random coefficients. Results show that the first two dominant modes contribute over 85% to the average kinetic energy. These most energetic modes also show mixed-layer similarity, which indicates that the non-local static instability plays a major role in determining the structure of the most energetic modes. Reconstruction of the vertical wind profiles by the first two dominant modes shows that they represent the most significant thermal structures. The probability distributions of the random coefficients related to these first two dominant modes are also analyzed and found to be Gaussian.