SURFACE FLUXES
231
found in WGASF (2000) and fluxes from the recent 40 year ECMWF
reanalysis (ERA-40) are beginning to appear. Different fields from various datasets have been corrected for known biases and merged (Large
and Yeager, 2004) to produce the fluxes presented in Section 6. Although sea-ice concentration has been well observed from satellites since
the late 1970s, other necessary ice data (e.g. thickness and surface temperature) are not available to compute companion air-ice and ice-ocean
flux climatologies.
2.
Ocean surface flux problems
The fundamental problem of ocean surface forcing is that it can not
be directly observed. In the case of the momentum flux between the atmosphere and a uniform land surface, it is possible to construct a ”drag
plate” (Bradley, 1968) and actually observe the wind stress over a small
area, at high frequency. This standard can then be transferred to less
direct, more easily implemented measurement techniques such as eddy
correlation (Haugen et al., 1971), which themselves can be used as a
transfer standard to even more indirect and easily implemented methods, such as inertial dissipation (Large and Pond, 1981). The problem
for GODAE is that there are no prospects for an ocean surface drag
plate. It is conceivable that detailed measurements of heat and moisture changes over a small patch of ground could yield direct estimates
of the heat and water fluxes that could serve as a basic truth standard.
Unfortunately, a similar approach in the ocean appears hopeless, largely
because of transports across the non-surface boundaries of the control
volume, and also because of the vertical penetration of solar radiation
and the presence of bubbles that are carried to depth after wave breaking.
At the end of the transfer of standard chains are the bulk aerodynamic
formulae (Roll, 1966), whose inputs are relatively easily measured mean
quantities such as atmospheric wind speed, U , potential temperature,
θ, and specific humidity, q, as well as sea surface temperature (SST).
But, without a direct standard there always remains an unquantified
uncertainty in estimates of ocean surface fluxes.
Since the nature of the sea surface dictates that flux measurements be
made at some distance, d of order 10 m above the sea surface, there is
another basic problem; namely, this flux is not equal to the surface flux.
To estimate the fractional error in wind stress, δτ = |τ −τ (d)|/τ , consider
the momentum equation for a mean near surface wind, U, aligned with
the x-coordinate axis. The dominant terms in horizontally homogeneous
flow are (Busch, 1977)
ρ ∂ t U = ∂ z τ (z) − ∂ x P o ,
(1)
231
found in WGASF (2000) and fluxes from the recent 40 year ECMWF
reanalysis (ERA-40) are beginning to appear. Different fields from various datasets have been corrected for known biases and merged (Large
and Yeager, 2004) to produce the fluxes presented in Section 6. Although sea-ice concentration has been well observed from satellites since
the late 1970s, other necessary ice data (e.g. thickness and surface temperature) are not available to compute companion air-ice and ice-ocean
flux climatologies.
2.
Ocean surface flux problems
The fundamental problem of ocean surface forcing is that it can not
be directly observed. In the case of the momentum flux between the atmosphere and a uniform land surface, it is possible to construct a ”drag
plate” (Bradley, 1968) and actually observe the wind stress over a small
area, at high frequency. This standard can then be transferred to less
direct, more easily implemented measurement techniques such as eddy
correlation (Haugen et al., 1971), which themselves can be used as a
transfer standard to even more indirect and easily implemented methods, such as inertial dissipation (Large and Pond, 1981). The problem
for GODAE is that there are no prospects for an ocean surface drag
plate. It is conceivable that detailed measurements of heat and moisture changes over a small patch of ground could yield direct estimates
of the heat and water fluxes that could serve as a basic truth standard.
Unfortunately, a similar approach in the ocean appears hopeless, largely
because of transports across the non-surface boundaries of the control
volume, and also because of the vertical penetration of solar radiation
and the presence of bubbles that are carried to depth after wave breaking.
At the end of the transfer of standard chains are the bulk aerodynamic
formulae (Roll, 1966), whose inputs are relatively easily measured mean
quantities such as atmospheric wind speed, U , potential temperature,
θ, and specific humidity, q, as well as sea surface temperature (SST).
But, without a direct standard there always remains an unquantified
uncertainty in estimates of ocean surface fluxes.
Since the nature of the sea surface dictates that flux measurements be
made at some distance, d of order 10 m above the sea surface, there is
another basic problem; namely, this flux is not equal to the surface flux.
To estimate the fractional error in wind stress, δτ = |τ −τ (d)|/τ , consider
the momentum equation for a mean near surface wind, U, aligned with
the x-coordinate axis. The dominant terms in horizontally homogeneous
flow are (Busch, 1977)
ρ ∂ t U = ∂ z τ (z) − ∂ x P o ,
(1)