In the atmospheric surface layer, it is widely accepted that ejection
and sweep eddy motions, typically associated with coherent structures,
are responsible for much of the land-surface evaporation, sensible he
at, and momentum fluxes. The present study analyzes the ejection-sweep
properties using velocity and scalar fluctuation measurements over ta
ll natural grass and bare soil surfaces. It is shown that momentum eje
ctions and sweeps occur at equal frequencies (D-eject approximate to D
-sweep approximate to 0.29) irrespective of surface roughness length o
r atmospheric stability conditions. Also, their magnitudes are compara
ble to values reported from open channel velocity measurements (D-swee
p approximate to 0.33; D-eject approximate to 0.30). The scalar D-ejec
t is constant and similar in magnitude to the momentum D-eject(approxi
mate to 0.29) over both surfaces and for a wide range of atmospheric s
tability conditions, in contrast to the scalar D-sweep. The scalar swe
ep frequency is shown to depend on the scalar skewness for the dynamic
convective and free convective sublayers, but is identical to D-eject
for the dynamic sublayer. The threshold scalar skewness at which the
D-sweep dependence occurs is 0.25, in agreement with the accepted temp
erature skewness value at near-neutral conditions. In contrast to a pr
evious surface-layer experiment, this investigation demonstrates that
the third-order cumulant expansion method (GEM) reproduces the measure
d relative flux contribution of ejections and sweeps (Delta S-0) for m
omentum and scalars at both sites. Furthermore, a linkage between Delt
a S-0 and the scalar variance budget is derived via the third-order CE
M in analogy to momentum. It is shown that Delta S-0 can be related to
the flux divergence term acid that such a relationship can be estimat
ed from surface-layer similarity theory, and the three sublayer model
of Kader and Yaglom and proposed similarity functions.