The motivation for the reanalysis project (described in the chapter entitled "Pilot Reanalysis") was to demonstrate that valuable information regarding the general circulation of the atmosphere is gained by utilizing more modern models and data assimilation systems. We have applied extensive diagnostics to the output of the new data assimilation and compared it to the contemporaneous (ECMWF and NMC) analyses, with some results appearing in Paolino, et al. (1994). Two aspects of the general circulation which appear to be quite different in the new assimilation relate to the divergent circulation and the structure of the transient moisture fields. These were studied by using a comprehensive suite of energetics programs in spectral (spherical harmonic) space, as originally proposed by Straus and Huntley (1994).
The divergent component of the kinetic energy, shown as a function of global wavenumber n and level is shown in panels (a) (for the reanalysis) and (b) (from ECMWF) of "Figure 1" for the winter period Dec. 1982 - Feb. 1983. The higher intensity of this component of kinetic energy is apparent at both lower and upper levels, and leads to an increase in the vertically integrated measure by nearly a factor of two. In contrast, the rotational kinetic energy is nearly unchanged.
While we suspect, but can not prove, that the newer estimate of divergent kinetic energy in the reanalysis is more realistic, we are fairly certain that the newer estimate of moisture variance is more realistic. Both the moisture variance and transient meridional flux of moisture (not shown) peak at 1000 hPa in the reanalysis, in contrast to the ECMWF analyses in which they peak at 850 hPa. A quite detailed examination of tropical and subtropical Pacific stations has indicated that the ECMWF time series of moisture are simply much too steady in this region. (In fact ECMWF analyses for later years also show maxima at the lowest levels.) Panels (c) and (d) of Figure 1 show the moisture variance as a function of level and global wavenumber n. Besides clearly indicating the difference in the vertical structure of the variance, the panels show a larger moisture variance at high wavenumbers in the new results.
An aspect of the general circulation which has not received much emphasis previously is the non- linear spectral dynamics expressed in the framework of quasi-geostrophic turbulence. We expect that potential enstrophy will cascade towards larger scales, and that this cascade will co-exist with a higher wavenumber cascade of total energy towards smaller scales. While diagnostics to capture these cascades are under development, we have calculated the purely two-dimensional counterparts of these, in which ordinary transient enstrophy and kinetic energy are considered. Panels (e) and (f) of Figure 1, show the change in enstrophy due to non-linear processes, (i.e., the self-interactions of transients). The greater transfer of enstrophy from smaller to larger scales in the reanalysis attests to a more vigorous estimate of the general circulation.
Personnel: Huntley, Paolino, Q. Yang, Straus
Summary prepared by: Straus