(e.g. Bunker, 1976a) used bulk formulae and relied
on a ship’s human observer for the cloud reports.
The Frouin et al. method uses satellite-derived
cloud information instead (see also review by
Katsaros, 1990). Improved atmospheric temperature sounders such as the Atmospheric Infrared
Sounder (AIRS) scheduled for launch early in the
year 2000 might advance estimation of both of
these difficult surface fluxes.
3.4.4.4 Present capability
With better than 20 W m
92 accuracy, we are able
to discern the annual cycle and ENSO variations
of the largest variable components of the ocean
surface heat flux – the solar heating and the evaporative cooling. To close the net air–sea energy
budget, or to estimate the meridional heat transport in the ocean, we need an absolute accuracy of
the total flux of at least 10 W m
92 on a monthly
2°2° resolution. We cannot meet this stringent
requirement at present. More intense effort is
needed to estimate SH and SR contributions, even
if the state of the art leaves much to be desired.
Since these terms are typically much smaller than
SR and LH at low latitudes and in the tropics,
even large percentage errors, of 20–30%, give
relatively small absolute errors in the net budget
(exceptions are cold, clear high-pressure situations
in the subtropics). Estimates within a few W m
92
may be possible for the tropical oceans but we
should be prepared to assess accuracy based on the
region and the general pattern of the circulation.
3.4.5 Hydrologic forcing
The computation of evaporation has been
described in Section 3.4.4.1. To estimate ocean
surface hydrologic forcing, we also need the precipitation (P) at the surface. In the past, attempts
have been made to estimate P from space-borne
sensors at visible, infrared and microwave wavelengths (e.g. Arkin and Ardanuy, 1989; Wilheit
et al., 1991). There are visible and infrared sensors,
but no microwave sensors, on geostationary satellites. Microwave sensors such as SSM/I have the
advantage of being more directly related to rain but
are limited by the insufficient sampling of the diurnal cycle on polar orbiters. Two or three satellites
flying in tandem or a satellite in low-inclination
orbit may alleviate this problem. Production and
validation of space-based rain data have been
undertaken by the Global Precipitation Climatology Project (GPCP), and data at 2.5° and monthly
resolutions from 1986 to 1994 are available. Precipitation for a very long period of time has also
been derived from the operational Microwave
Sounding Unit (Spencer, 1993).
The launching of TRMM in November 1997,
with its low-inclination orbit (higher sampling
rate), opened up a new era of estimating both
surface rainfall and the precipitation profile
(Kummerow et al., 1998). There are a number of
instruments on TRMM, but only TMI and the
precipitation radar (PR) are directly relevant in
estimating the hydrologic forcing. TMI operates
at five frequencies and has a swathe width of
760 km. From these measured radiances, the integrated column precipitation over the ocean can be
estimated, among a number of parameters. The
spatial resolution varies with frequency, being
15 km at 10 GHz and 5 km at 85 GHz. PR sends
radar pulses at 13.8 GHz and measures the
backscatter, from which the three-dimensional rainfall distribution over both land and ocean can be
derived. The horizontal resolution is 4.3 km, but
the swathe is narrow at only 220 km.
Figure 3.4.5 (see Plate 3.4.5, p. 172) shows that
a large value of precipitation is found in the
Intertropical Convergence Zone and South Pacific
Convergence Zone all year round. Heavy precipitation is also obvious along the storm tracks in the
western North Pacific and North Atlantic during the
boreal summer. The strong and steady trade winds
and dry air cause strong evaporation in the subtropical oceans, except in the eastern part of the ocean
basin of the summer hemisphere, where coastal
upwelling produces low T s , which reduces E.
In addition to the extreme sparsity of in-situ
validation standards over the ocean, the problem of space-based precipitation estimates is
compounded by the short period of rainfall. There
is no comprehensive evaluation of precipitation
over the global ocean, but many validation experiments are being conducted, and we expect useful
accuracy estimates will be forthcoming in the near
future.
3.4.6 Future prospects
NASA will provide another scatterometer identical
to SeaWinds on QuikSCAT to fly on ADEOS-2,
scheduled for launch in November 2001. QuikSCAT
3.4 Air–Sea Fluxes from Satellite Data
179
Liu and Katsaros
on a ship’s human observer for the cloud reports.
The Frouin et al. method uses satellite-derived
cloud information instead (see also review by
Katsaros, 1990). Improved atmospheric temperature sounders such as the Atmospheric Infrared
Sounder (AIRS) scheduled for launch early in the
year 2000 might advance estimation of both of
these difficult surface fluxes.
3.4.4.4 Present capability
With better than 20 W m
92 accuracy, we are able
to discern the annual cycle and ENSO variations
of the largest variable components of the ocean
surface heat flux – the solar heating and the evaporative cooling. To close the net air–sea energy
budget, or to estimate the meridional heat transport in the ocean, we need an absolute accuracy of
the total flux of at least 10 W m
92 on a monthly
2°2° resolution. We cannot meet this stringent
requirement at present. More intense effort is
needed to estimate SH and SR contributions, even
if the state of the art leaves much to be desired.
Since these terms are typically much smaller than
SR and LH at low latitudes and in the tropics,
even large percentage errors, of 20–30%, give
relatively small absolute errors in the net budget
(exceptions are cold, clear high-pressure situations
in the subtropics). Estimates within a few W m
92
may be possible for the tropical oceans but we
should be prepared to assess accuracy based on the
region and the general pattern of the circulation.
3.4.5 Hydrologic forcing
The computation of evaporation has been
described in Section 3.4.4.1. To estimate ocean
surface hydrologic forcing, we also need the precipitation (P) at the surface. In the past, attempts
have been made to estimate P from space-borne
sensors at visible, infrared and microwave wavelengths (e.g. Arkin and Ardanuy, 1989; Wilheit
et al., 1991). There are visible and infrared sensors,
but no microwave sensors, on geostationary satellites. Microwave sensors such as SSM/I have the
advantage of being more directly related to rain but
are limited by the insufficient sampling of the diurnal cycle on polar orbiters. Two or three satellites
flying in tandem or a satellite in low-inclination
orbit may alleviate this problem. Production and
validation of space-based rain data have been
undertaken by the Global Precipitation Climatology Project (GPCP), and data at 2.5° and monthly
resolutions from 1986 to 1994 are available. Precipitation for a very long period of time has also
been derived from the operational Microwave
Sounding Unit (Spencer, 1993).
The launching of TRMM in November 1997,
with its low-inclination orbit (higher sampling
rate), opened up a new era of estimating both
surface rainfall and the precipitation profile
(Kummerow et al., 1998). There are a number of
instruments on TRMM, but only TMI and the
precipitation radar (PR) are directly relevant in
estimating the hydrologic forcing. TMI operates
at five frequencies and has a swathe width of
760 km. From these measured radiances, the integrated column precipitation over the ocean can be
estimated, among a number of parameters. The
spatial resolution varies with frequency, being
15 km at 10 GHz and 5 km at 85 GHz. PR sends
radar pulses at 13.8 GHz and measures the
backscatter, from which the three-dimensional rainfall distribution over both land and ocean can be
derived. The horizontal resolution is 4.3 km, but
the swathe is narrow at only 220 km.
Figure 3.4.5 (see Plate 3.4.5, p. 172) shows that
a large value of precipitation is found in the
Intertropical Convergence Zone and South Pacific
Convergence Zone all year round. Heavy precipitation is also obvious along the storm tracks in the
western North Pacific and North Atlantic during the
boreal summer. The strong and steady trade winds
and dry air cause strong evaporation in the subtropical oceans, except in the eastern part of the ocean
basin of the summer hemisphere, where coastal
upwelling produces low T s , which reduces E.
In addition to the extreme sparsity of in-situ
validation standards over the ocean, the problem of space-based precipitation estimates is
compounded by the short period of rainfall. There
is no comprehensive evaluation of precipitation
over the global ocean, but many validation experiments are being conducted, and we expect useful
accuracy estimates will be forthcoming in the near
future.
3.4.6 Future prospects
NASA will provide another scatterometer identical
to SeaWinds on QuikSCAT to fly on ADEOS-2,
scheduled for launch in November 2001. QuikSCAT
3.4 Air–Sea Fluxes from Satellite Data
179
Liu and Katsaros