Geography Reference
In-Depth Information
bottom type that do not appear to be directly related to
these properties, the water column cannot be ignored.
The optical characteristics of the water column deter-
mine how radiant energy is modified as it propagates
to and from the streambed. For applications focused on
the water conveyed within the channel rather than the
shape and composition of the channel boundary, optical
properties provide the critical link between radiance mea-
surements and concentrations of suspended sediment,
chlorophyll, or dissolved organic matter. In either case,
a useful approach to characterising the water column,
detailed by Bukata et al. (1995), is to consider the com-
posite, or bulk, inherent optical properties as the additive
consequence of the
specific inherent optical properties
of
the suspended and/or dissolved organic and/or inorganic
materials in the water column, as well as pure water itself.
For example, we can express the absorption coefficient
a
(
significant component must always be considered: pure
water. The absorption spectrumof pure water, denoted by
a
w
(
λ
) is relevant to many fields of study, and numerous
attempts have been made to measure this fundamen-
tal quantity with a high degree of accuracy. The most
widely used data on
a
w
(
) for visible and near-infrared
wavelengths were compiled by Pope and Fry (1997) and
Smith and Baker (1981) and are shown in Figure 3.7a.
Absorption of light by water is weakest in the blue and
green, and photons of these wavelengths can thus pen-
etrate considerable distances into a (pure) water body.
As wavelength increases into the red and especially the
near-infrared,
a
w
(
λ
) increases by an order of magnitude,
implying that photons in this portion of the spectrum
will not propagate through the water column with nearly
as much ease as their shorter-wavelength counterparts.
Importantly, the spectral shape of
a
w
(
λ
) dictates that red
and near-infrared bands are most useful for estimating
water depth in shallow stream channels with depths of
1m or less. The strong absorption by pure water over this
spectral range implies that small changes in depth will
correspond to relatively large changes in radiance. This
sensitivity allows for precise depth estimates in shallow
water, but strong absorption also leads to saturation of the
radiance signal at greater depths. Beyond a certain point,
an additional increase in depth produces a very small,
potentially undetectable, decrease in radiance because the
vast majority of photons have already been absorbed. To
map bathymetry across a broad range of depths, informa-
tion from across the visible spectrum is thus necessary.
Similarly, because the absorption spectrum of pure water
dictates which wavelengths are most likely to propagate to
the bed,
a
w
(
λ
λ
), a
bulk inherent optical property
of the water col-
umn as a whole, as the sum of the absorption coefficients,
denoted by
a
i
(
), associated with each of
n
components,
weighted by their concentrations
x
i
:
λ
n
a
(
λ
)
=
x
i
a
i
(
λ
)
(3.15)
i
=
1
Analogous expressions can be used to describe the scat-
tering and back-scattering coefficients as well. The
a
i
(
)
in Equation (3.15) represent the amount of absorption
attributable to specific components of the water column;
similarly,
b
i
(
λ
) describe scattering by specific compo-
nents. Specific inherent optical properties of this kind
are referred to as
optical cross-sections
and have units of
area per unit mass (i.e., m
2
g
−
1
). Multiplying an optical
cross-section by a concentration in gm
−
3
thus yields an
absorption or scattering coefficient with units of m
−
1
.
Legleiter et al. (2004, 2009) used optical cross-sections to
model the optical properties of a simple two-component
water column consisting of pure water and suspended
sediment and quantify the effects of suspended sediment
concentration on the water-leaving radiance. Conversely,
if optical cross-sections for the components of interest
are known and the bulk inherent optical properties of the
water column can be derived from remotely sensed data,
this information could be used to determine the concen-
tration of each component. This approach is described in
greater detail by Bukata et al. (1995), who focus on coastal
waters and lakes. In principle, similar methods could be
applied to fluvial environments, but in shallow streams
the influence of the bottom could complicate such efforts.
Whether the attribute of primary interest pertains to
the streambed or to the water column, one optically
λ
) influences which bands are useful for map-
ping bottom type. Scattering by pure water is strongest in
the blue and decreases withwavelength but is insignificant
relative to suspended sediment (Figure 3.7b).
The optical properties of suspended sediment, chloro-
phyll, and dissolved organic matter are best summarised
in terms of their optical cross-sections, shown in
Figure 3.7 (Bukata et al., 1995). For suspended sediment
in particular, data on optical cross-sections are sparse
and reflect natural variability in particle size, shape, and
lithology. In practice, one of a handful of existing optical
cross-sections is assumed to be representative of the
area of interest. Nevertheless, some general observations
can be made. Absorption by chlorophyll and dissolved
organic matter tends to be strongest at shorter wave-
lengths, decrease with increasing wavelength, and be of a
similar magnitude as absorption by suspended sediment.
Absorption by suspended sediment is greatest in the
λ
Search WWH ::
Custom Search