Mason, R.A., H.N. Shirer, R. Wells, G.S. Young, 2002

Improved description of fluxes by convective plumes within the marine atmospheric surface layer

J. Atmos. Sci., 59, 1337-1355

Abstract

Bursts in the instantaneous kinematic fluxes of temperature and horizontal momentum in a moderately convective marine atmospheric surface layer are studied by applying the VITA (variable interval time averaging) detection method to PCA (principal components analysis)-filtered data sets obtained from the FLIP moored vessel during the 1995 April/May Pacific Marine Boundary Layer (PMBL) experiment. For convective plumes, a well-defined dimensionless relationship is shown to exist between the vertical fluxes of temperature and horizontal momentum; this relationship cannot be easily deduced if PCA and VITA are not both applied.

PCA filters a data set using correlations within that data set instead of band-pass filtering it to retain energy in a predetermined range of scales; PCA thus respects all scales contributing to the phenomena retained in the data set. Subsequent use of cross-spectral techniques to group the PCA-filtered data set into coherent structure types leads to, amongst other types of coherent structures, PCA-derived plumes. The VITA method is applied to a filtered data set in order to identify updrafts (bursts) and downdrafts (sweeps) in the time series of correlated variables by searching the signal for events that satisfy user-specified criteria. With proper use of PCA, surface-layer plumes can be reassembled in a way that yields the same flux configurations for two different detecting variables. This robustness overcomes a common criticism directed at application of conditional sampling to a non-PCA-filtered data set because PCA removes small-scale noise that contributes to the total variance of the transported quantities but not greatly to the covariance between them.