The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B7. Beijing 2008
240
the surface down to 100cm (limited by the fiber optic length).
2.4 Way of operation
Eighteen boreholes were dug according to the surface IS-based
map along the area. For every depth, the soil properties were
estimated using the NIRS models. Based on this information,
we generated a surface soil property map for every depth layer
for five soil properties (see below). In addition, the soil color
was estimated using a digitally based calculation of the VIS
region. Each drill hole was vertically described and classified
according to a certain soil order. At several locations, the soil
profile description of the drill hole (based on the POS method)
was compared to the tradition soil profile description using
open trenches.
2.5 Using the spectral-based model NIRS to evaluate soil
properties
One hundred and sixty soil samples, taken from Israeli soils,
were used to generate the spectral models based on the NIRS
approach. For that purpose, we used the surface reflectance and
its equivalent wet laboratory information. The soils were
analyzed in the laboratory for six soil properties that are
considered to be important for the soil survey mission and were
determined in the laboratory: Soil Moisture (SM), Organic
Matter (OM), Hygroscopic Moisture (HIG), Field Moisture
(FM), Soil Carbonate Content (SCC), Specific Surface Area
(SSA), and iron oxide content (Fed). The models were validated
in several drilling points against the (true) chemical properties
from the area.
2.6 Spatial processing of the soil drill information
Every depth for a given property’s layer, using all POS points,
was interpolated by Kriging manipulation to provide a spatial
layer view of the property in question. Five layers per property
were generated at 20-cm intervals (to a depth of 100cm), all
together, summing up to 25 layers. The layers were projected on
a DTM map of the area that was generated by using the GPS
measurements that were taken during the POS field
measurements.
2. RESULTS AND DISCUSSION
Figure 1 presents the proposed paper concept, where a new
spectral-based approach has replaced the traditional soil
mapping approach that used open trenches and laboratory work.
Figure 1: The concept of the optically based approach to map
soil against the traditional way (Acknowledgement is given to
Dr. J. Dematte for part of the illustration)
In general, the airphoto is replaced by an imaging spectrometer
view that provides not only a cognitive presentation of the area;
it also enables accurate classification of the soil surface based
on its spectral properties. This information is then used to select
the proper location for the drilling holes. In these points, the
POS approach is used and the entire soil profile is described
qualitatively and quantitatively in situ. An equivalent activity,
using the convention approach, is characterized by opening
trenches and sending the soils to the laboratory, which is a time-
and money-consuming process. Finally, soil maps (or profile
descriptions) are generated and a comparison between the two
methods is possible. Tables 1 and 2 present a detailed
description of four soil profiles, as obtained from the POS
approach, for four different soils selected to validate the
approach. Table 1 presents the traditional soil profile
descriptions obtained from trenches that were opened at nearby
positions and Table 2 presents the optical (POS) profile
descriptions
Horizon
Depth
(cm)
Description
Rhodoxeralf
A
0-40
Bare cover, small amount of dry weed, no carbonate
content, dry soil, color 2.5 YR 4/8 red, sandy texture.
Bt
40-55
Elovial horizon, accumulation of clay and iron oxides,
no carbonates, color 2.5 YR 3/6 dark red.
В
55-80
No carbonates, sandy soil, color 2.5 YR 4/6 red.
C
80-100
Sandy soil, brighter color, higher humidity, color 5 YR
4/6 yellowish red.
Haploxeralf
A
0-20
The surface is covered with wild vegetation, high
carbonate content, clay loam texture, color 10 YR 6/4
light brown.
В
20-70
High carbonate content, decreased organic matter,
sandy clay loam texture, a high concentration of
carbonate pebbles in the horizon, color 10 YR 6/4
light brown.
c
70-90
Sandy soil texture, decreased carbonate content, color
10 YR 8/8 reddish yellow.
Haploaquept
A
0-15
Very rich in organic matter, dark color, rich in
carbonates, loam sandy texture, color 10 YR 4/1 dark
gray.
A3
15-50
Very rich in organic matter, high soil moisture, loam
sandy texture,
color 10 YR 4/1 dark gray.
В
50-80
High water table and bad drainage in the soil increase
the moisture; the horizon contains evidence of its
original lake materials as shells. Drastic change in
color; color 10 YR 7/1 light gray.
C
80-100
Highly rich in carbonates, high moisture, contains free
water, very bright color 10 YR 8/3 very pale brown,
clay loam texture.
Chromoxerert
A
0-65
Rich in organic matter, high carbonate content, clay
texture, color 10 YR 4/3 brown.
AC
65-100
Rich in organic matter, high soil moisture, clay
texture, color 10 YR 3/3 dark brown.
Table 1: Traditional soil profile descriptions, as was done in the
field in nearby trenches
As seen, the two tables are in good agreement; however, when
using the POS approach, the information is more detailed.
Bearing in mind that the POS information was obtained
objectively in situ, we can conclude that the optically based
approach is a much more effective method for carrying out soil
survey missions than the traditional method. Since the POS
technology permits many measurements in the space domain,
interpolation maps may be generated in the field to provide a
spatial view if the soil entity is in question. In the current
study, we used an unsupervised technique applied to the IS
image, which resulted in 6 classes using the ISO Data classifier.