flectance of
; (Top), but
eometry are
in forward
h as 45% of
signal lacks directionality (Fig. 4, top). The polarization rate is also small irrespective of the
viewing geometry, with maximum values of about 5% (Fig. 4, bottom). This suggests that it may
not be appropriate to work with the polarized component of reflectance to enhance the relative
contribution of the water body. As the ocean signal is weak and almost unpolarized, one may not
want to deal with a noisy polarized signal, even though there are backscattering geometries for
which the atmospheric and surface effects on the top-of-atmosphere polarized reflectance are
practically eliminated (see Fig. 3, bottom).
The polar diagrams in Fig. 5 show the relative contribution to total and unpolarized reflectance
of photons that have interacted with the water body. The characteristics of the ocean, atmosphere,
and surface are the same as those of Figs. 3 and 4. Working with total reflectance (Fig. 5, top), the
contribution of the water body does not exceed 20%. The highest values are obtained outside the
glitter region for viewing zenith angles of less than 60°. For those geometries, the aerosol phase
function is small. Apart from the glitter region, where values are obviously small, the minimum
values are found at high viewing zenith angles, where atmospheric scattering dominates. Working
with the unpolarized component of reflectance, by contrast (Fig. 5, bottom), the contribution of the
water body reaches 45%, which represents a gain in useful signal of a factor of 2.2. The best
viewing geometries are adjacent to the glitter region, at viewing zenith angles of about 50°
(scattering angles of 100° or so). At these angles, polarization by air molecules and aerosols
(combined) is maximum. As the relative azimuth angle decreases from 150°, the scattering angle
decreases, making generally the air molecules and aerosols less efficient at polarizing sunlight. The
result is a decreased contribution of the water boby (Fig. 5, bottom).
4-DISCUSSION
As shown above, using the unpolarized component of reflectance instead of total reflectance
enhances, for some viewing geometries, the relative contribution of the water body to the signal
measured at the top of the atmosphere. The gain obtained at 450 nm (a factor of 2) is substantial,
making it easier to correct the top-of-atmosphere signal for atmospheric and surface effects (45%
instead of 20% of the measured signal may now originate from the water body) and, therefore, to
retrieve the signal backscattered by the water body. In fact, even in backscattering, at high viewing
zenith angles, or in the glitter region, the unpolarized reflectance at the top of the atmosphere is
generally less influenced by atmospheric and surface effects than the total reflectance (see Fig. 5).
The results, though preliminary, strongly suggest that polarization is a useful property of light to
improve ocean color remote sensing from space.
Since aerosols polarize incident sunlight less than molecules, the enhancement obtained using
unpolarized reflectance is reduced as aerosol optical thickness is higher, all the more as multiple
scattering, being mor efficient, further contributes to reduce polarization rate. When the aerosol
optical thickness is 1, for instance, there is practically no gain in the relative contribution of the
water body. Such conditions, however, are extreme, and not suitable for ocean color remote
sensing.
Another point should be mentioned about aerosols. Mie calculations show distinct polarization
signatures depending on type. The favorable viewing geometries, therefore, may be type-
dependent. In general, however, the maximum polarization rates are found between scattering
angles of 80° and 150°, which determines somewhat the potentially favorable geometries. In fact,
simulations with “water-soluble“ aerosols (not shown here) have revealed significant changes in
the diagram of Fig. 5, bottom (e.g., a narrower range of favorable viewing geometries), but only a
slight displacement of the location of the maximum values toward lower vewing zenith angles
(higher scattering angles).
Depending on the surface conditions (e.g., wind speed), the glitter pattern may extend over a
wider range of viewing angles, displacing the favorable geometries further off the principal plane.
As solar zenith angle increases, those geometries become closer to nadir.
Using unpolarized reflectance, one may wonder about the sensitivity of the unpolarized signal
backscattered by the water body (obtained after atmospheric corrections) to water composition
and, in particular, phytoplankton pigment concentration. In the favorable geometry regions of Fig.
5, bottom, the water body exhibits polarization rates of about 20% with only 0.1 mgrrr^ of
pgments, the values obviously decreasing with increasing pigment concentration. These relatively
low polarization rates suggest a sensitivity of the unpolarized signal to pigment concentration
similar to that of the total signal, but slightly reduced since the polarization rate decreases when
Pigment concentration increases.
673