You are using an outdated browser that does not fully support the intranda viewer.
As a result, some pages may not be displayed correctly.

We recommend you use one of the following browsers:

Full text

Remote sensing for resources development and environmental management
Damen, M. C. J.

itification of
ea by the use
Season. CSIR
ng of winter
Advances in
m analysis of
c Engineering
matrix theo-
emote mapping
in of the pro-
8th Interna-
' Environment,
estigation of
note Sensing.
D.W. Deering
in the great
nposium, NASA
F.E. Guertin,
ladian LANDSAT
natible tape,
search Report
ntures derived
Symposium on Remote Sensing for Resources Development and Environmental Management / Enschede / August 1986
Theoretic reflection modelling of soil surface properties
B.P.J.van den Bergh & B.A.M.Bouman
Agricultural University, Wageningen, Netherlands
ABSTRACT: For a theoretical approach to surface reflection modelling,soil surface properties are divided on two
levels of distinct influence on reflection. On the first level, an attempt is made to formulate the influence
of particle size, mineralogical composition and moisture content and reflection. The developed equations are ba
sed on the Lambert attenuation law of radiation in diffusing media.
A natural soil surface is a heterogeneous combination
of various elements of different composition, size,
shape and spatial distribution. The combination of the
se elements can be described in terms of soil proper
ties such as textural class, mineralogical composition,
organic matter and moisture content. Further, a soil
surface has a certain roughness. There may be the
presence of aggregates, rills, crusts and/or the ef
fects of human influence like ploughing or harrowing.
Soil surfaces may also contain stones, boulders, plant
debris and other materials. Plant life in various
forms between algae and trees may be present, while
animal activity can have an important impact on the
structure of the surface. Finally, this whole com
plex of surface elements has a certain spatial orien
tation viz: slope and exposition.
For a theoretical approach of reflection from a natu
ral land surface, a division in two levels of surface
properties is suggested. The first will be called
"the intrinsic soil surface" and the second the "bare
land surface" (the presence of vegetation will not be
considered in this study).
Modelling reflectance from bare land surfaces
The intrinsic soil surface is defined in terms of
optical parameters:"a heterogeneous combination of
reflective and absorptive elements in a solid, liquid
or gaseous state, arranged in a specific spatial dis
tribution and orientation. The dimension of these ele
ments is such that the reflective process at this
intrinsic level is governed by multiple reflections
and re-reflections from these elements. The reflec
tion from this intrinsic surface is thus the result
of multiple internal reflections (mutually influen
cing each other) and hence the theory of diffuse re
flection should serve as a basis for further theore
tical model building. On this level grain size and
grain size distribution, shape and frosting of the
grains, mineralogy, content and nature of organic mat
ter and moisture content should be included as para
meters in a model.
The bare land surface is defined as a combination of
intrinsic soil and non-soil surface segments, arran
ged in acertain spatial distribution and orientation.
Here, following Cooper and Smith, 1985, only varia
tions in reflection arising from macroscopic features
of a soil (large enough that diffraction by the irre
gularities of the surface may be neglected) are consi
dered. A bare soil surface has a certain roughness of
many geometrically different intrinsic soil surface
segments. Also, non-soil elements like stones and boul
ders are included here, as well as slope and exposition
of the surface as a whole.
Above division in two surface levels is based on the
dimensions of the surface elements relative to the re
flective processes invoved. By this division, a hier
archical ordening in the reflection influencing surfa
ce properties is achieved. Modelling of the bare land
surface reflection is super imposed on modelling of the
intrinsic soil surface reflection. The approach is gra
dual in which, started from the skeleton of the soil,
different parameters are introduced in the model. In
this paper, attention will be paid to modelling at the
intrinsic soil surface level only; more specifically
to the parameters particle size, mineralogy and moistu
re content.
At this level, the equation of Lambert-Beer serves as
a basis for further elaboration. This equation descri
bes reflectance r from diffusing media in the optical
r = i/l = exp(-kd) eq.l
and hence:
Ln(r) = -kd eq.2
in which: r = reflectance
I = intensity of incident radiation (W/sr/u)
I = intensity of reflected radiation (W/sr/ju)
k = coefficient of absorption
d = mean penetrated layer thickness
The coefficient of absorption k is a characteristic of
the absorbing components of the medium, while mean pe
netrated layer thickness d depends on the geometrical
structure and texture of the medium. Both properties
are a function of the wavelength of incident radiation.
By calculating Ln(r), whereby r is measured for soil
samples of different variables, attempts can be made to
relate the coefficients k and d to these variables.
Since only limited laboratory measurements of reflec
tance could be performed, use has been made of measu
rements presented in literature. Own laboratory mea
surements have been carried out with a NIWARS-spectro-
fotometer, measuring bidirectional reflectance in the
range of 0.3 to 2.4 ^om. (Bunnik, 1978).
The influence of particle size
The mean penetrated layer thickness was studied in re
lation to particle diameter of sorted soil samples. In
1965, Bowers and Hanks (B&H) published the results of
a study to the effect of particle size on reflectance
at different wavelengths in the optical range. These
reflectance measurements of samples of kaolinite and
bentonite clays are used here for further investiga
tion, thereby introducing the following two hypotheses
1) d is dependent on wavelength and on particle size
2) k is dependent on wavelength but not on particle