98
W.T. Liu et al.
6.3 Characterization of Stress Versus Wind Distribution
6.3.1 Dependence on Surface Temperature and Current
When the first Quick Scatterometer (QuikSCAT) data came back in 1999, the science team was surprised to see that the scatterometer signal in the equatorial Pacific
propagates westward with the ocean temperature front of the tropical instability
waves, in the area where we expected to see steady trade winds (e.g. Liu et al.,
2000; Chelton et al., 2001). Such coincident propagation was previously observed
by Xie et al. (1998) in European Research Satellite (ERS) data. Since then, the
spatial coherence between scatterometer measurements and sea surface temperature (T s ) has been observed over many locations and under various atmospheric
conditions, e.g. over the Kuroshio Extension (e.g. Nonaka and Xie, 2003), over the
circumpolar current (e.g. O’Neill et al., 2003), Indian Ocean (Vecchi et al., 2004), in
the East China Sea during winter cold air outbreak (Xie et al., 2002), over the Gulf
Stream Ring (Park and Cornillon, 2002), and over typhoon wake (Lin et al., 2003).
Following traditional paths to study atmospheric boundary layer processes, many
scientists were quick to postulate explanations of the wind and T s correlation, based
on boundary layer height change, pressure gradient force, secondary flow, cloud
entrainment, and organized convection, but none of these is generally applicable to
the ubiquitous correlation, as pointed out in the review by Small et al. (2008).
The first explanation by oceanographers, when they saw the results of Liu et al.
(2007) showing that QuikSCAT measurements deviate from the mean winds with
rotation in opposite direction to the underlying surface current of the Agulhas
Extension meanders, was that either the drifter (current) or the scatterometer measurements were erroneous. Their reasoning was that the strong current meanders
should impart its rotation on the prevailing westerly wind through drag, and the
wind anomalies should show the same rotation as the current. Misinterpreting scatterometer stress as wind is the cause of confusion. Stress must be spatially coherent
with T s and ocean current, which create buoyancy and wind shear. As pointed out
by Liu et al. (2007) and Liu and Xie (2008), at small turbulence scales at the
surface, factors that affect atmospheric boundary layer dynamics (wind), such as
Coriolis force, pressure gradient force, baroclinicity, cloud entrainment, etc., are
not important. That is why the spatial coherence is ubiquitous, under all kinds of
atmospheric circulations. Stress is the vector difference between wind and current. For a uniform wind blowing over a rotating current, the vector differences
will have opposite rotation to the current (Park et al., 2006). The ocean signals, of
course, will affect winds aloft through stress. The dynamic factors will then become
important.
Liu et al. (2007) observed T s signatures in cloud and atmospheric temperature
high in the atmosphere over the Agulhas Extension. Even stronger penetrating signals have been found, not only in temperature profiles but also in precipitation
profiles over the Kuroshio Extension. Present numerical models of atmospheric circulation in the mid-latitudes do not propagate the surface stress signal vertically
much beyond the atmospheric boundary layer. The observations posted a challenge
W.T. Liu et al.
6.3 Characterization of Stress Versus Wind Distribution
6.3.1 Dependence on Surface Temperature and Current
When the first Quick Scatterometer (QuikSCAT) data came back in 1999, the science team was surprised to see that the scatterometer signal in the equatorial Pacific
propagates westward with the ocean temperature front of the tropical instability
waves, in the area where we expected to see steady trade winds (e.g. Liu et al.,
2000; Chelton et al., 2001). Such coincident propagation was previously observed
by Xie et al. (1998) in European Research Satellite (ERS) data. Since then, the
spatial coherence between scatterometer measurements and sea surface temperature (T s ) has been observed over many locations and under various atmospheric
conditions, e.g. over the Kuroshio Extension (e.g. Nonaka and Xie, 2003), over the
circumpolar current (e.g. O’Neill et al., 2003), Indian Ocean (Vecchi et al., 2004), in
the East China Sea during winter cold air outbreak (Xie et al., 2002), over the Gulf
Stream Ring (Park and Cornillon, 2002), and over typhoon wake (Lin et al., 2003).
Following traditional paths to study atmospheric boundary layer processes, many
scientists were quick to postulate explanations of the wind and T s correlation, based
on boundary layer height change, pressure gradient force, secondary flow, cloud
entrainment, and organized convection, but none of these is generally applicable to
the ubiquitous correlation, as pointed out in the review by Small et al. (2008).
The first explanation by oceanographers, when they saw the results of Liu et al.
(2007) showing that QuikSCAT measurements deviate from the mean winds with
rotation in opposite direction to the underlying surface current of the Agulhas
Extension meanders, was that either the drifter (current) or the scatterometer measurements were erroneous. Their reasoning was that the strong current meanders
should impart its rotation on the prevailing westerly wind through drag, and the
wind anomalies should show the same rotation as the current. Misinterpreting scatterometer stress as wind is the cause of confusion. Stress must be spatially coherent
with T s and ocean current, which create buoyancy and wind shear. As pointed out
by Liu et al. (2007) and Liu and Xie (2008), at small turbulence scales at the
surface, factors that affect atmospheric boundary layer dynamics (wind), such as
Coriolis force, pressure gradient force, baroclinicity, cloud entrainment, etc., are
not important. That is why the spatial coherence is ubiquitous, under all kinds of
atmospheric circulations. Stress is the vector difference between wind and current. For a uniform wind blowing over a rotating current, the vector differences
will have opposite rotation to the current (Park et al., 2006). The ocean signals, of
course, will affect winds aloft through stress. The dynamic factors will then become
important.
Liu et al. (2007) observed T s signatures in cloud and atmospheric temperature
high in the atmosphere over the Agulhas Extension. Even stronger penetrating signals have been found, not only in temperature profiles but also in precipitation
profiles over the Kuroshio Extension. Present numerical models of atmospheric circulation in the mid-latitudes do not propagate the surface stress signal vertically
much beyond the atmospheric boundary layer. The observations posted a challenge