172
IAN ROBINSON
surface. Their great benefit is that their view is not impeded by cloud and
very little attenuation occurs in the atmosphere, although water present as
large liquid drops in precipitation does attenuate the signal. However, the
emissivity, H, of the sea surface in the microwave part of the spectrum is less
than 0.5. H also depends on factors such as the temperature, the salinity and
the viewing incidence angle. This in turn means the brightness temperature
is also a function of the mean square slope and hence of the sea surface
roughness and wind speed. While this complicates the retrieval of SST from
microwave radiometry compared with infra-red methods, the corollary is
that microwave sensors can be used to measure the surface roughness,
rainfall or even salinity as well as SST.
It is possible to distinguish between the different contributions to the
brightness temperature of SST, surface roughness and salinity, as well as to
identify atmospheric contamination by liquid water, because each factor
differentially affects different microwave frequencies. For example SST
strongly affects wavebands between 6 and 11 GHz whereas the effects of
salinity are found only at frequencies below about 3 GHz. Surface
roughness effects influence frequencies at 10 GHz and above, and are also
polarisation specific.
Thus a multi-frequency and multi-polarisation
radiometer can, in principle, be used to measure SST, surface wind and
precipitation (see chapter 8 of Robinson (2004)). Each of these has potential
for use in ocean models, and coincident measurement of SST and winds has
potentially useful applications in the estimation of air-sea fluxes.
Despite a long series of microwave sensors flown for atmospheric remote
sensing, serious consideration of microwave measurements of SST from
space started only when a microwave radiometer having a 10.7 GHz channel
was flown on the Japanese-US Tropical Rainfall Mapping Mission. Called
the TRMM microwave imager (TMI) it has a spatial resolution of 0.5q
(about 50 km) and because it over-samples it is capable of mapping
mesoscale eddies quite effectively using a grid scale of 25 km. It lacks the
preferred SST waveband of 6.6 GHz, but its 10.7 GHz channel is sensitive to
SST in tropical water temperatures (Donlon et al., 2001), and its usefulness
for measuring the thermal signatures of tropical instability waves has already
been demonstrated by Chelton et al. (2000). It covers only latitudes lower
than 40º.
In 2002 the Japanese Advanced Microwave Scanning Radiometer
(AMSR-E) was launched into a near-polar orbit on the NASA Aqua satellite.
This sensor includes a channel at 6.6 GHz, which is effective over the full
range of sea temperatures, and has opened the way for routine, high quality,
global mapping of SST by microwave radiometry. AMSR-E is now
providing global cloud free SST to an accuracy of ~0.3 K derived from oversampled 76 km resolution data. The composite daily, weekly and monthly
SST products are supplied on a ¼q grid (Wentz and Meissner, 2000).
IAN ROBINSON
surface. Their great benefit is that their view is not impeded by cloud and
very little attenuation occurs in the atmosphere, although water present as
large liquid drops in precipitation does attenuate the signal. However, the
emissivity, H, of the sea surface in the microwave part of the spectrum is less
than 0.5. H also depends on factors such as the temperature, the salinity and
the viewing incidence angle. This in turn means the brightness temperature
is also a function of the mean square slope and hence of the sea surface
roughness and wind speed. While this complicates the retrieval of SST from
microwave radiometry compared with infra-red methods, the corollary is
that microwave sensors can be used to measure the surface roughness,
rainfall or even salinity as well as SST.
It is possible to distinguish between the different contributions to the
brightness temperature of SST, surface roughness and salinity, as well as to
identify atmospheric contamination by liquid water, because each factor
differentially affects different microwave frequencies. For example SST
strongly affects wavebands between 6 and 11 GHz whereas the effects of
salinity are found only at frequencies below about 3 GHz. Surface
roughness effects influence frequencies at 10 GHz and above, and are also
polarisation specific.
Thus a multi-frequency and multi-polarisation
radiometer can, in principle, be used to measure SST, surface wind and
precipitation (see chapter 8 of Robinson (2004)). Each of these has potential
for use in ocean models, and coincident measurement of SST and winds has
potentially useful applications in the estimation of air-sea fluxes.
Despite a long series of microwave sensors flown for atmospheric remote
sensing, serious consideration of microwave measurements of SST from
space started only when a microwave radiometer having a 10.7 GHz channel
was flown on the Japanese-US Tropical Rainfall Mapping Mission. Called
the TRMM microwave imager (TMI) it has a spatial resolution of 0.5q
(about 50 km) and because it over-samples it is capable of mapping
mesoscale eddies quite effectively using a grid scale of 25 km. It lacks the
preferred SST waveband of 6.6 GHz, but its 10.7 GHz channel is sensitive to
SST in tropical water temperatures (Donlon et al., 2001), and its usefulness
for measuring the thermal signatures of tropical instability waves has already
been demonstrated by Chelton et al. (2000). It covers only latitudes lower
than 40º.
In 2002 the Japanese Advanced Microwave Scanning Radiometer
(AMSR-E) was launched into a near-polar orbit on the NASA Aqua satellite.
This sensor includes a channel at 6.6 GHz, which is effective over the full
range of sea temperatures, and has opened the way for routine, high quality,
global mapping of SST by microwave radiometry. AMSR-E is now
providing global cloud free SST to an accuracy of ~0.3 K derived from oversampled 76 km resolution data. The composite daily, weekly and monthly
SST products are supplied on a ¼q grid (Wentz and Meissner, 2000).