230
WILLIAM B. LARGE
sampled Southern Ocean and ice covered polar seas. Another challenge
is proper specification of uncertainties. Also, the requirement for ”near
real time” fluxes precludes the use of some data, and makes this application very different from its cousin; Ocean Reanalysis. In both cases,
flux variability should be properly represented across all the important
ocean time scales; from the 12-hour polar inertial period, to the decadal,
but GODAE may place a higher premium on time scales comparable to
an assimilation cycle, such as the diurnal solar and three to seven day
synoptic cycles.
Of primary importance are the surface fluxes of heat, Q, freshwater,
F, and momentum, τ , with components τ λ in the zonal and τ φ in the
meridional directions. Nearly all aspects of this rather complicated subject are comprehensively covered in the final report of the WCRP/SCOR
Working Group on Air-Sea Fluxes (WGASF, 2000). With this report as
a solid base, it is possible here to focus on what practitioners of GODAE
should understand about the ocean surface fluxes they may utilize.
Perhaps the most well known constraint on surface fluxes is that the
global long term heat and freshwater fluxes into the ocean should both
be near zero. Observations of long-term changes in ocean heat content
suggest a heat flux of at most a few W/m 2 . For example, Levitus et
al. (2000) find that the temperature of the world’s oceans increased
from the 1950s to the 1990s at a rate equivalent to a surface heat flux
imbalance of only 0.3W/m 2 . Similarly, the global ocean salinity record
doesn’t support a significantly non-zero global long-term freshwater flux.
However, fluxes on shorter time and space scales are not known nearly so
well. This issue and other surface flux problems are outlined in Section
2. Sections 3 and 4 deal with the measurement and parameterization
of the turbulent flux components. Satellite techniques have been most
successful at estimating the radiative heating, precipitation and wind
stress, so only these products are discussed in Section 5.
Section 6 is concerned with what the various surface flux fields might
really look like. Any specified surface forcing will be incremented by
the data assimilation cycle. The resulting fluxes should at least satisfy
global constraints and should be compared to existing air-sea flux climatologies to see if there are unacceptably large discrepancies in seasonal
cycles, variability on inter-annual and longer time scales, and the mean.
A number of such climatologies are compared by Beranger et al. (1999);
including the SOC (Southhampton Oceanography Centre; Josey et al.
1998), COADS (University of Wisconsin-Milwaukee version of the Cooperative Ocean Atmosphere Data Set; Da Silva et al., 1994), NCEP
(NCEP/NCAR reanalysis fluxes; Kalnay et al., 1996 ) and ERA-15 (15
year ECMWF reanalysis fluxes, Gibson t al., 1997). Others can be
e
WILLIAM B. LARGE
sampled Southern Ocean and ice covered polar seas. Another challenge
is proper specification of uncertainties. Also, the requirement for ”near
real time” fluxes precludes the use of some data, and makes this application very different from its cousin; Ocean Reanalysis. In both cases,
flux variability should be properly represented across all the important
ocean time scales; from the 12-hour polar inertial period, to the decadal,
but GODAE may place a higher premium on time scales comparable to
an assimilation cycle, such as the diurnal solar and three to seven day
synoptic cycles.
Of primary importance are the surface fluxes of heat, Q, freshwater,
F, and momentum, τ , with components τ λ in the zonal and τ φ in the
meridional directions. Nearly all aspects of this rather complicated subject are comprehensively covered in the final report of the WCRP/SCOR
Working Group on Air-Sea Fluxes (WGASF, 2000). With this report as
a solid base, it is possible here to focus on what practitioners of GODAE
should understand about the ocean surface fluxes they may utilize.
Perhaps the most well known constraint on surface fluxes is that the
global long term heat and freshwater fluxes into the ocean should both
be near zero. Observations of long-term changes in ocean heat content
suggest a heat flux of at most a few W/m 2 . For example, Levitus et
al. (2000) find that the temperature of the world’s oceans increased
from the 1950s to the 1990s at a rate equivalent to a surface heat flux
imbalance of only 0.3W/m 2 . Similarly, the global ocean salinity record
doesn’t support a significantly non-zero global long-term freshwater flux.
However, fluxes on shorter time and space scales are not known nearly so
well. This issue and other surface flux problems are outlined in Section
2. Sections 3 and 4 deal with the measurement and parameterization
of the turbulent flux components. Satellite techniques have been most
successful at estimating the radiative heating, precipitation and wind
stress, so only these products are discussed in Section 5.
Section 6 is concerned with what the various surface flux fields might
really look like. Any specified surface forcing will be incremented by
the data assimilation cycle. The resulting fluxes should at least satisfy
global constraints and should be compared to existing air-sea flux climatologies to see if there are unacceptably large discrepancies in seasonal
cycles, variability on inter-annual and longer time scales, and the mean.
A number of such climatologies are compared by Beranger et al. (1999);
including the SOC (Southhampton Oceanography Centre; Josey et al.
1998), COADS (University of Wisconsin-Milwaukee version of the Cooperative Ocean Atmosphere Data Set; Da Silva et al., 1994), NCEP
(NCEP/NCAR reanalysis fluxes; Kalnay et al., 1996 ) and ERA-15 (15
year ECMWF reanalysis fluxes, Gibson t al., 1997). Others can be
e