252
WILLIAM B. LARGE
5.1
Wind stress
Satellite wind products have been derived from both passive microwave
radiometers and active radars (altimeters and scatterometers), but the
principal source of wind direction data is scatterometers. In situ buoy
measurements of U N have been empirically related to scatterometer
backscatter, σ o , such that the accuracy of scatterometer wind speed,
U S , can be rather good; -.3m/s bias and 1.3 rms difference (Freilich and
Dunbar, 1999). Taking U S as an estimate of U N gives the stress directly
from (31) as
u
∗2 = C DN (U S ) U
2
S .
(53)
But is this the best calculation? From section 4 it can be argued that
any drag coefficient formulation is a representation of ”average” wave
conditions. There are recent formulations which purport to account for
at least some wave effects, given local wave conditions, which we denote
as C DN (waves). Figure 4 suggests that at the same wind speed, rougher
seas produce greater stress that a C DN (waves) formulation should be
able to capture, at least in large measure. However, the observations of
Li et al. (1987) suggest that the rougher seas themselves produce more
backscatter and a high satellite wind , so that in order not to account
for the rough surface twice, C DN as in (34) should be used in (53), as
given.
5.2
Radiation
The in situ measurement of radiation both incident on and reflected
from the ocean surface is essentially a direct measurement. The solar
and longwave components are measured by a pyranometer and a pyrgeometer, respectively. The two instruments are similar, consisting of a
blackened thermopile covered by a hemispherical dome whose transmissivity matches the desired measurement. In ideal conditions pyranometer accuracy can be excellent ( 2 %), but variations in the temperature of
the dome can be problematic. At sea conditions are far from ideal, with
platform motion a serious problem unless mitigated by use of a gimbal.
Other sources of error are calibration error, contamination of the dome,
so all together measurement accuracy is unlikely to be no better than
5%. Longwave radiation is naturally emitted near typical pyrgeometer
temperatures, so corrections need to be applied for dome temperature
and emissivity to transmissivity ratio, and for the temperature of the
instrument housing. Although the thermopile is shielded by a shortwave
filter, the solar radiation can be so large that even a small amount of
leakage can be serious. Again measurement errors of at least 5% are
expected.
WILLIAM B. LARGE
5.1
Wind stress
Satellite wind products have been derived from both passive microwave
radiometers and active radars (altimeters and scatterometers), but the
principal source of wind direction data is scatterometers. In situ buoy
measurements of U N have been empirically related to scatterometer
backscatter, σ o , such that the accuracy of scatterometer wind speed,
U S , can be rather good; -.3m/s bias and 1.3 rms difference (Freilich and
Dunbar, 1999). Taking U S as an estimate of U N gives the stress directly
from (31) as
u
∗2 = C DN (U S ) U
2
S .
(53)
But is this the best calculation? From section 4 it can be argued that
any drag coefficient formulation is a representation of ”average” wave
conditions. There are recent formulations which purport to account for
at least some wave effects, given local wave conditions, which we denote
as C DN (waves). Figure 4 suggests that at the same wind speed, rougher
seas produce greater stress that a C DN (waves) formulation should be
able to capture, at least in large measure. However, the observations of
Li et al. (1987) suggest that the rougher seas themselves produce more
backscatter and a high satellite wind , so that in order not to account
for the rough surface twice, C DN as in (34) should be used in (53), as
given.
5.2
Radiation
The in situ measurement of radiation both incident on and reflected
from the ocean surface is essentially a direct measurement. The solar
and longwave components are measured by a pyranometer and a pyrgeometer, respectively. The two instruments are similar, consisting of a
blackened thermopile covered by a hemispherical dome whose transmissivity matches the desired measurement. In ideal conditions pyranometer accuracy can be excellent ( 2 %), but variations in the temperature of
the dome can be problematic. At sea conditions are far from ideal, with
platform motion a serious problem unless mitigated by use of a gimbal.
Other sources of error are calibration error, contamination of the dome,
so all together measurement accuracy is unlikely to be no better than
5%. Longwave radiation is naturally emitted near typical pyrgeometer
temperatures, so corrections need to be applied for dome temperature
and emissivity to transmissivity ratio, and for the temperature of the
instrument housing. Although the thermopile is shielded by a shortwave
filter, the solar radiation can be so large that even a small amount of
leakage can be serious. Again measurement errors of at least 5% are
expected.