Model Development for Passive Microwave Remote Sensing ...
581
Water vapour over sea
International programmes have been initiated to understand the dynamic ocean/atmosphere
system. The Tropical Ocean and Global Atmosphere (TOGA) Programme started in 1985
and focussed on the tropical ocean/atmosphere system and its effect on the climate at higher
latitudes. The goal of the World Ocean Circulation Experiment (WOCE) is to develop a
data set of unprecedented magnitude and quality for use in developing numerical models of
ocean circulation and physical processes and to couple them with atmospheric models. Passive
remote sensing with ESMR, SMMR, SSM/I and MIMR can provide global information about
sea surface temperature, sea surface wind speed and other ocean parameters (Goodberlet et
al., 1990).
For water vapour (WV) Lojou (1990) has proposed algorithms for SMMR, SSM/I and MIMR
(Scire Scappuzzo, 1992). Using the MIMR channels (see Table 21.1) WV can be expressed as
a function of the 18.7, 23.8 and 37 GHz channels as follows
WV[g/cm 2 ] = 8.45 + 3.71ln(280 - T H19 ) + 1.05ln(280 - T V1S ) - 6.52ln(280 - T H23 )
-2.69ln(280 - T V23 ) + 3.32ln(280 - T H37 ) - 0.4171n(280 - TV37)' (21.3)
This algorithm (Eq. 21.3) is valid only for TB < 280 K and assumes an emittance with an
averaged homogeneous distribution that can only be found when observing ocean surfaces.
After applying this formula to measured data the water vapour is obtained and its accuracy
proved by comparing the results with measurements taken from other sensors. Similar formulae
can be applied for other parameters like rain rate and liquid water content.
Most inversion algorithms that are used in an operational environment are based on formulations as simple as the one given above. These type of formulae are obtained through parameter
fitting to either existing measurements or to calculated values obtained from previously tested
direct models.
Wind speed and sea surface temperature
Brightness temperature as a function of ocean parameters is presented as an example for direct
algorithms. Three major physical properties influence the emission characteristics, namely the
dielectric constant, the surface geometry and the foam cover (or indirectly the wind speed).
The permittivity lOw of sea water is a function of frequency, salinity and water temperature and
can be calculated by the modified Debye equation (Ulaby et al., 1986). At frequencies higher
than 6 GHz (all SSM/I channels) the effect of salinity is negligible. The surface geometry is
regarded as a rough surface with partial foam cover depending on the wind speed W. Wilheit
(1979) has proposed an algorithm to model this behaviour where the reflectance fh,v, given
by the Fresnel reflections coefficients with fh,v = R~,v(f, e), is averaged over the surface slope
distribution of a wind roughened sea and it uses a Geometrical Optics (GO) approach of the
sea surface. A partial foam cover is taken into account for W > 7 mis, as it reduces the surface
reflectance. An empirical expression for the foam reduction is given as
_ { 00(1 - e-J1Jo)(W[m/s]- 7) for W;::: 7m/s
K Joam -
0
for W < 7m/s '
(21.4)
where a = 0.006 and fo = 7.5 GHz.
Finally the emittance esea is defined by the foam reduced averaged reflectance as
esea(f, e, w, SST) = 1 - [fh,v;ave(f, e, w, SST) (1 - KJoam(f, W))]
(21.5)
and has a typical behaviour shown in Fig. 21.6 as a function wind speed and in Fig. 21.7 as a
function of SST.
581
Water vapour over sea
International programmes have been initiated to understand the dynamic ocean/atmosphere
system. The Tropical Ocean and Global Atmosphere (TOGA) Programme started in 1985
and focussed on the tropical ocean/atmosphere system and its effect on the climate at higher
latitudes. The goal of the World Ocean Circulation Experiment (WOCE) is to develop a
data set of unprecedented magnitude and quality for use in developing numerical models of
ocean circulation and physical processes and to couple them with atmospheric models. Passive
remote sensing with ESMR, SMMR, SSM/I and MIMR can provide global information about
sea surface temperature, sea surface wind speed and other ocean parameters (Goodberlet et
al., 1990).
For water vapour (WV) Lojou (1990) has proposed algorithms for SMMR, SSM/I and MIMR
(Scire Scappuzzo, 1992). Using the MIMR channels (see Table 21.1) WV can be expressed as
a function of the 18.7, 23.8 and 37 GHz channels as follows
WV[g/cm 2 ] = 8.45 + 3.71ln(280 - T H19 ) + 1.05ln(280 - T V1S ) - 6.52ln(280 - T H23 )
-2.69ln(280 - T V23 ) + 3.32ln(280 - T H37 ) - 0.4171n(280 - TV37)' (21.3)
This algorithm (Eq. 21.3) is valid only for TB < 280 K and assumes an emittance with an
averaged homogeneous distribution that can only be found when observing ocean surfaces.
After applying this formula to measured data the water vapour is obtained and its accuracy
proved by comparing the results with measurements taken from other sensors. Similar formulae
can be applied for other parameters like rain rate and liquid water content.
Most inversion algorithms that are used in an operational environment are based on formulations as simple as the one given above. These type of formulae are obtained through parameter
fitting to either existing measurements or to calculated values obtained from previously tested
direct models.
Wind speed and sea surface temperature
Brightness temperature as a function of ocean parameters is presented as an example for direct
algorithms. Three major physical properties influence the emission characteristics, namely the
dielectric constant, the surface geometry and the foam cover (or indirectly the wind speed).
The permittivity lOw of sea water is a function of frequency, salinity and water temperature and
can be calculated by the modified Debye equation (Ulaby et al., 1986). At frequencies higher
than 6 GHz (all SSM/I channels) the effect of salinity is negligible. The surface geometry is
regarded as a rough surface with partial foam cover depending on the wind speed W. Wilheit
(1979) has proposed an algorithm to model this behaviour where the reflectance fh,v, given
by the Fresnel reflections coefficients with fh,v = R~,v(f, e), is averaged over the surface slope
distribution of a wind roughened sea and it uses a Geometrical Optics (GO) approach of the
sea surface. A partial foam cover is taken into account for W > 7 mis, as it reduces the surface
reflectance. An empirical expression for the foam reduction is given as
_ { 00(1 - e-J1Jo)(W[m/s]- 7) for W;::: 7m/s
K Joam -
0
for W < 7m/s '
(21.4)
where a = 0.006 and fo = 7.5 GHz.
Finally the emittance esea is defined by the foam reduced averaged reflectance as
esea(f, e, w, SST) = 1 - [fh,v;ave(f, e, w, SST) (1 - KJoam(f, W))]
(21.5)
and has a typical behaviour shown in Fig. 21.6 as a function wind speed and in Fig. 21.7 as a
function of SST.