See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/320160445

Regosols

Chapter · January 2018

DOI: 10.1007/978-3-319-64392-2_16

CITATIONS READS

0 46

4 authors:

Yusuf Kurucu Mustafa Tolga Esetlili

Ege University Ege University
69 PUBLICATIONS   248 CITATIONS    49 PUBLICATIONS   134 CITATIONS   

SEE PROFILE SEE PROFILE

Erhan Akça Mehmet Ali Çullu

Adiyaman University Harran University
79 PUBLICATIONS   446 CITATIONS    28 PUBLICATIONS   296 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

TRACING ORGANIC AND INORGANIC POLLUTION SOURCES OF THE AGRICULTURAL CROPS AND WATER RESOURCES IN GÜZELHISAR BASIN OF AEGEAN REGION-TURKEY.
View project

Landscape Project View project

All content following this page was uploaded by Yusuf Kurucu on 28 September 2018.

The user has requested enhancement of the downloaded file.

Turk J Agric For
29 (2005) 19-28
© TÜB‹TAK

Using Advanced Spectral Analyses Techniques as Possible Means of

Identifying Clay Minerals

Ünal ALTINBAfi*, Yusuf KURUCU, Mustafa BOLCA

Ege University, Faculty of Agriculture, Soil Department, ‹zmir - TURKEY
A. H. El-NAHRY
National Authority for Remote Sensing and Space Sciences, Cairo - EGYPT

Received: 15.01.2004

Abstract: Spectral analyses, one of the most advanced remote sensing techniques, were used as a possible means of identifying the
mineralogy of the clay fractions that corresponded to the Küçük Menderes Plain, south of ‹zmir, Turkey. Different spectral processes
were used to execute the prospective spectral analyses. The processes include: i. the reflectance calibration of TM images belonging
to the studied area, ii. using minimum noise fraction (MNF) transformation and iii. creating the pixel purity index (PPI), which was
used to the most “spectrally pure”, extreme, pixel in multi-spectral images. Spectral analyses of the clay mineralogy of the studied
area were obtained by matching the unknown spectra of the purest pixels to pre-defined (library) spectra providing scores with
respect to the library spectra. Three methods, namely Spectral Feature Fitting (SFF), Spectral Angle Mapper (SAM) and Binary
Encoding (BE) were used to produce a score between 0 and 1, where the value of 1 equals a perfect match showing the exact mineral
type. We were able to identify 4 clay minerals i.e., vermiculite, kaolinite, montmorillonite and illite, recording different scores related
to their abundance in the soils. In order to check the validity and accuracy of the results obtained regarding the spectral signatures
of the minerals identified, soil samples taken from the same localities were subjected to X- ray analysis. As a result a good correlation
was found between the spectral signatures and the X- ray diffractions.

Key Words: Remote sensing, spectral analysis, X-ray diffraction, clay mineralogy.

‹leri Spektral Analiz Teknikleri Kullan›larak Kil Minerallerinin Belirlenebilirli¤i

Özet: ‹zmir ilinin güneyinde bulunan Küçük Menderes deltas› topraklar›nda kil mineral tiplerinin belirlenmesinde uzaktan alg›lama
tekniklerinden spektral analiz tekni¤i kullan›lm›flt›r. Bu amaçla çal›flma alan›na ait TM görüntülerinin yans›ma kalibrasyonu yap›lm›flt›r.
Minimum Noise Fraction (MNF) yöntemi ile görüntüdeki bozukluklar gerekli matematiksel algoritmalar kullan›larak azalt›lm›flt›r. Pixel
Purity Index (PPI) tekni¤i kullan›larak görüntünün piksel boyutundaki mineral tanecik yans›malar› belirlenmifltir. MNF ve PPI
teknikleri beraberce kullan›lm›fl ve 3 boyutlu görünüm yard›m›yla en iyi yans›may› veren piksellerin yerini bulmak, tan›mlamak ve
s›n›fland›rarak ayr›ml› mineraller için en iyi spektral yans›malar belirlenmifltir. Çal›flma alan›na ait bilinmeyen spektral yans›ma
analizleri minerallerin yans›ma özellikleri ile karfl›laflt›r›lm›flt›r. Spectral Feature Fitting (SFF), Spectral Angle Mapper (SAM) ve Binary
Encoding (BE) teknikleri kullan›larak spektral yans›malarla mineral yans›malar› aras›nda efllefltirilme yap›larak ayr›ml› mineral tipleri
belirlenmifltir. Bu yöntemlerde 0 ile 1 aras›nda de¤iflen say›lar kullan›lm›flt›r. 1 en uygun efllemeye karfl›l›k gelen spektral özellik olup,
bir mineralin spektral özelli¤ine tam olarak uyan yans›may› göstermektedir. Çal›flma alan› topraklar›nda bulunma yo¤unlu¤una gore
kil mineralleri vermikulit, kaolinit, illit ve montmorillonit’tir. Kullan›lan spektral yöntemlerde bulunan minerallere ait spektral
de¤erlerin do¤ruluk analizleri için spektral analizlerin yap›ld›¤› alanlardan toprak örnekleri al›nm›fl ve toprak örneklerinin X-ray
sonuçlar› ile spektral analiz sonuçlar› aras›nda pozitif iliflki belirlenmifltir.

Anahtar Sözcükler: Uzaktan alg›lama, spektral analiz, X-ray difraktometresi, kil mineralojisi.

Introduction philosophy of remote sensing, which is based on

The current work focused on a small portion of the identifying objects or natural resources without physical
Küçük Menderes valley for the identification of the connection remotely by the use of different sensors. The
mineralogy of clay samples. The study was in line with the main objective of the current work was a trial to

* Correspondence to: altinbas@ziraat.ege.edu.tr

19
Using Advanced Spectral Analyses Techniques as Possible Means of Identifying Clay Minerals

recognize the dominant clay minerals of the Küçük Fieldwork

Menderes soils, by applying the highly advanced remote A reconnaissance survey was conducted in the
sensing techniques of spectral analyses. investigated area in order to gain an appreciation of the
Olsen et al. (2000) described research aimed at broad soil patterns according to the obtained spectra end
determining the feasibility of using reflectance members. GPS Garmin12XL (+-,5-8 m) was used in the
spectroscopy to identify and characterize the expansive field to recognize the accurate locations of the end
clays and clay-shales along the Colorado Front Range members for soil sampling. Four surface soil samples
Urban Corridor, both at the laboratory/field scale, and at representing the major soils in the area were collected for
the remote-sensing scale. X-ray diffraction analyses.
Chabrillat et al. (2001) established the correlation of Using USGS Spectral Library (Minerals)
clay mineralogy with swelling potential indices and the
expansion of existing correlations of laboratory The spectra that represent the USGS spectral library
reflectance spectra with clay mineralogy, examining also were measured on a custom-modified, computer-
direct correlations between reflectance data and swelling controlled Beckman spectrometer at the USGS Denver
potential. Spectroscopy Lab., U S A. Wavelength accuracy was in the
order of 0.0005 µm (0.5 nm) in the near-IR and 0.0002
µm (0.2 nm) in visible light (http://speclab.cr.usgs.gov/
Materials spectral-lib.html).
The Küçük Menderes valley occupies about 400 km2 X-Ray Diffraction Analyses
in the western part of Turkey. It is located between 37º
The clay fraction was separated after being pretreated
45´ and 38º 00´ N latitudes and between 27º 15´ and 28º
according to the procedure described by Jackson (1975).
30´ E longitudes. The valley represents the most
The clay samples were X-rayed after they had been Ca-
important agricultural area, the south part of the city
‹zmir. The valley is surrounded by highlands acting saturated at 25 °C, Ca-saturated and glycolated, and K-
together as catchments areas that supply the Küçük saturated at 25 °C. Apart from the above treatment,
Menderes river tributaries with rainwater and weathered sample no. 1 (Ca-saturated) was also heated to 300 °C
materials. The valley is separated from the other and 520 °C.
neighboring valleys by 2 mountainous chains, which X-ray diffraction patterns were obtained with the
define its natural boundaries. Near the Aegean coast the Philips P.W. (1060/100) X- ray diffractometer with Cu-α
Küçük Menderes penetrates a narrow valley, forming a radiation and an iron filter. Identification of the clay
delta near the city of Selçuk, where the meanders drain minerals was carried out following the guidelines
into the Aegean Sea. provided by Black (1965), MacEwan (1980), Carroll
(1970) and Weed (1977).
Methods
Digital image processing Results and Discussion
Image processing using ENVI 3.4 software includes Many obstacles and atmospheric and environmental
the following processes: conditions (i.e., weathering processes) are still hampering
the accurate identification of land resources due to their
a) Calibration of a Landsat 7 (taken in 2003) ETM
image to reflectance. Filtering by the adaptive filters to influence on the accuracy of measured spectra. Thus, the
reduce noise by smoothing while preserving sharp edges. tool of spectral analyses is not foolproof. Nevertheless, it
b) Stretching using Gaussian stretching methods using a is meant to be used as a starting point for identifying
mean DN of 127 with the data values of 3 standard materials in an image scene. When used properly, spectral
deviations set to 0 and 255. c) Geometric correction analysis tools in conjunction with a good spectral library
using image to image methods for adding ground control could provide excellent suggestions for identification of
points with 3 m average RMSE and nearest method use objects on the land surface. Spectral analyses were used
as a resampling method, and image projection using the to identify the clay minerals of the Küçük Menderes soils
Universal Transverse Mercator (UTM). using the following procedures.

20
 Ü. ALTINBAfi, Y. KURUCU, M. BOLCA, A. H. EL-NAHRY

Display of Color Composite ETM Image Fraction (MNF) and Pixel Purity Index (PPI) routines.
1. A color composite ETM image has been filtered to ETM image calibration was been used with pre-launch
produce output images in which the brightness value at a gains and offsets calculated for Landsat sensors
given pixel is a function of some weighted average of the (Markham and Barker, 1986).
brightness of the surrounding pixels. Minimum Noise Fraction
2. Enhanced and stretched. The results obtained from MNF transformation is a method similar to principal
Gaussian stretching improved the visual display of the components. It was used to determine the inherent
spectra information as shown in Figure 1 dimension of the image data, to segregate noise in the
3. Geometrically corrected and displayed using a band data, and to reduce the computational requirements for
combination of 3 2 1 (RGB - true color) as shown in subsequent processing (Boardman and Kruse, 1994). The
Figure 2. MNF is used as a preparatory transformation to put most
of the essential components into just a few spectral bands
Calibration of the ETM Image and to order those bands from the most interesting (that
A reflectance calibration was required for Landsat can segregate noise perfectly) to the least interesting.
ETM data to compare image spectra with library Two cascaded principal components transformations
reflectance spectra and to run some Minimum Noise were implemented in the current work. The first

Figure 1. Gaussian-stretching results of TM image.

Figure 2. The true color composite 3 2 1 of the investigated area.

21
Using Advanced Spectral Analyses Techniques as Possible Means of Identifying Clay Minerals

transformation, based on an estimated noise covariance

matrix, decorrelates and rescales the noise in the data.
This first step results in transformed data in which the
noise has unit variance and no band-to-band correlations.
The second step is a standard principal components
transformation of the noise-whitened data. For the
purposes of further spectral processing, the inherent
dimension of the data is determined by the examination
of the final eigenvalues (of noise segregation) and the
associated image bands. The data space could be divided
into 2 parts: one part associated with large eigenvalues
and coherent eigenimages, and a complementary part
with near unity eigenvalues and noise dominated images. Figure 5. MNF band 6.
By using only the coherent portions, the noise was
separated from the data, thus improving spectral Pixel Purity Index
processing results. The decreasing eigenvalue with
The PPI function finds the most spectrally pure or
increasing MNF band as shown in the eigenvalue plot in
“extreme” pixels in multispectral and hyperspectral data
Figure 3 shows how noise is segregated in the higher
(Boardman and Kruse, 1994; Boardman et al., 1995).
number MNF bands and it was noted that there was a
The extreme pixels correspond to the materials with
decrease in spatial coherency with increasing MNF band
spectra that combine linearly to produce all of the spectra
number as shown in Figures 4 and 5.
in the image. The PPI was computed by using projections
of n-dimensional scatter plots to 2-D space and marking
the extreme pixels in each projection. The extreme pixels
in each projection were recorded and the total number of
times each pixel was marked as extreme was noted. The
Noise segregation

output is an image (the PPI image) in which the digital

number (DN) of each pixel in the image corresponds to
the number of times that pixel was recorded as extreme.
Thus, bright pixels in the image showed the spatial
location of spectral endmembers. Image thresholding was
1 2 3 4 5 6 used to select several thousand pixels for further analysis,
Band No. thus significantly reducing the number of pixels to be
Figure 3. Noise segregation plot. examined (Figure 6).
The PPI, as shown in Figure 7, indicated the total
extreme pixels for the studied area, whereas 5212 pixels
were recorded throughout iteration no. equal to 4000
times for a pixels threshold of 3.
Chabrillat et al. (2002) used AVIRIS and HyMap
images acquired recently with a high signal-to-noise ratio
(SNR) to detect clays. The results showed the extent to
which laboratory spectra of swelling soils field samples
could be used to detect and discriminate different clays,
smectite, illite and kaolinite, related to variable swelling
potential.
Goetz et al. (2001) and Olsen et al. 2000 used near-
Figure 4. MNF band 1. infrared reflectance spectroscopy to discriminate among

22
 Ü. ALTINBAfi, Y. KURUCU, M. BOLCA, A. H. EL-NAHRY

Pixel Purity Index (PPI) Select Data Discard Low MNFs

Std. Deviation Threshold

Process Data
Maximize Iterations

Evaluate PPI Results

Display and Histograms

Figure 6. Flow chart of the PPI procedures.

Pixel Purfty Index of soile

5200
PPI Curve

5100
Total pixels

5000

4800

4800

1000 2000 3000 4000

Iteration no.

Figure 7. PPI index. Figure 8. PPI map.

pure smectite and mixed-layer I/S samples, based on

characteristic absorption bands in the 1900–2400 nm
spectral region.
The following single band images, as seen in Figures
8 and 9, represent the PPI, where the extreme (purest)
pixels are white in Figure 8. It is noted that the extreme
pixels occupied the region of interest, as shown in Figure
9.
n-Dimensional Visualization and Extracted
Endmember Spectra
The n-D visualization was used in conjunction with the
MNF and PPI tools to locate, identify and cluster the Figure 9. Region of interest.
purest pixels and the most extreme spectral responses in
a data set. If spectral signatures are recorded properly
et al.,1995). The coordinates of the points in n-space
and the curve shape is accurate they could be used for
consists of “n” values that are simply the spectral radiance
remote sensing applications (Salisbury et al., 1991).
or reflectance values in each band for a given pixel. The
Spectra can be thought of as points in a dimensional distributions of these points in n-space were used to
scatter plot, where n is the number of bands (Boardman estimate the number of spectral endmembers (4

23
Using Advanced Spectral Analyses Techniques as Possible Means of Identifying Clay Minerals

highlighted segmentations), as shown in Figure 10,

producing 4 pure spectral signatures, which were
extracted and plotted in an n-D visualizer plot

reflectance % (Offset for clarity)

representing the selected endmembers, as shown in
Figure 11.
n-D class1

n-D class3

n-D class2

0.5 1.0 1.5 2.0 2.5

Wavelength

Figure 11. Classes of the selected spectra.

De Jong’s (1992) correspondence analysis of the

spectral characteristics of soils revealed lime, clay, iron
Figure 10. 3-D visualization. and organic matter as the important variables, and pH
and the absorption features of Illite are generally broader
Matching Unknown Spectra to Library Spectra and less well defined compared with those of muscovite.
Spectral analyses and consequently clay mineral Nevertheless, the illite, muscovite and montmorillonite
identification could be obtained by matching the unknown spectra have similar absorption bands. Illite {(K,H,O)
spectra extracted from the 3-D visualizer to pre-defined (AI,Mg,Fe), (Si,AI),O,,((OH),,H,O)} shows broad water
(library) spectra, providing scores with respect to the absorption features near 1.4 and 1.9 ym, and additional
library spectra. Three weighting methods, i.e. Spectral Al-hydroxyl features at 2.2, 2.3 and 2.4 pm. Illite and
Feature Fitting (SFF), Spectral Angle Mapper (SAM) muscovite have absorption bands near 2.35 and 2.45 pm,
and/or Binary Encoding (BE), were used to identify that are lacking in the montmorillonite spectrum.
mineral type, producing a score between 0 and 1, where The output of the spectral analysis is a ranked score
1 equals a perfect match. As is known, some minerals are or weighted score for each of the materials in the input
similar in one wavelength range, yet very different in spectral library, as shown in Table 1. The highest score
another. For the best results, a wavelength range that indicates the closest match and shows higher confidence
contains the diagnostic absorption features was used to in the spectral similarity, where illite/smectite and
distinguish among the minerals. kaolinite, scored high values of 1.0 and 0.944,

Table 1. Weighting methods and mineral type/score of the extracted spectra.

Weighting Method (Score 0-1.0)

SAM SFF BE
Spectra class
Mineral type Score Mineral type Score Mineral type Score

1 - 0 Kaolinite 0.944 Kaolinite 0.833

2 - 0 Vermiculite 0.833 Vermiculite 0.667
3 - 0 Illite/Smectite 1.000 Illite/Smectite 0.883
4 - 0 Montmorillonite 0.667 Montmorillonite 0.500

24
 Ü. ALTINBAfi, Y. KURUCU, M. BOLCA, A. H. EL-NAHRY

respectively, while vermiculite and montmorillonite dominant clay minerals, followed by vermiculite and
scored 0.833 and 0.667, respectively, using SFF montmorillonite and small amounts of other minerals
weighting. The same clay minerals recorded scores of (Figures 12-16).
0.833, 0.833, 0.667 and 0.500, respectively, using BE Mermut et al. (1997) found that the soil loss from a
weighting. On the other hand, the SAM did not recognize soil dominated by smectite was high. The splash and wash
any kind of clay minerals (zero score). erosion in 80 mm of rain were 23 and 2.1 Mg ha-1,
According to Alt›nbafl (1982), using X-ray analysis, respectively, in a loamy soil in which smectite, mica and
illite and kaolinite were the dominant clay minerals in the vermiculite were the dominant clays, and 7.3 and 0.91
acidic brown forest soils (Typic Dystrustrepts), followed Mg ha-1 respectively, in a silt loam soil in which
by vermiculite and montmorillonite. Basically, the clay vermiculite, mica and kaolinite were dominant.
minerals found in these soil groups are products of the Illite
transformation and decomposition of biotite, muscovite
and feldspars. In order to check the validity and accuracy Illite as a 2:1 clay mineral was recorded in the X-ray
of our results concerning spectral signatures and to diffractogram at 10.04 Å. Illite is a widespread mineral in
define perfectly, the existing clay minerals, X-ray Kücük Menderes soils (Alt›nbafl, 1982). It may be formed
diffraction analysis was the obvious choice. The results by the alteration of mica minerals. Illite was found as the
obtained from X-ray diffraction indicated that the clay dominant clay mineral by X-ray analysis in the acidic
fraction contained mainly illite and kaolinite as the brown forest soils, which were developed on mica schist

1.0 montmorillonite 5spc

1.0 kaolinite 1spc
n_D class 1 n_D class 4

0.8
0.8
Reflectance %
Reflectance %

0.6
0.6

0.4
0.4

0.2
0.2

0.0
0.0
0.5 1.0 1.5 2.0 2.5
0.5 1.0 1.5 2.0 2.5
Wavelength µm
Wavelength µm

1.0 vermiculite 1 spc 1.0 illite 1. spc illite

n_D class 2 n_D class 3

0.8 0.8
Reflectance %

Reflectance %

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0
0.5 1.0 1.5 2.0 2.5 0.5 1.0 1.5 2.0 2.5
Wavelength µm Wavelength µm

Figure 12. ID of kaolinite, illite, montmorillonite and vermiculite.

25
Using Advanced Spectral Analyses Techniques as Possible Means of Identifying Clay Minerals

7.13Å

15.00Å
7.15Å
10.00Å
7.13Å

9.92Å
14.96Å

9.92Å 9.92Å
19.65Å

14.01Å
9.90Å

Ca.+Glyc.25°C
Ca.25°C
Ca.25°C K.25°C
Ca.Glyc.25°C
K.25°C
θ 14 13 12 11 10 9 8 7 6 5 4 3
θ 14 13 12 11 10 9 8 7 6 5 4 3

Figure 15. X-ray diffraction of the clay fractions-sample no. 3 (0-5 cm)
Figure 13. X-ray diffraction of the clay fractions-sample no.1 (0-4 cm) with Ca++, K+, Ca+++Glyc at 25 °C.
with Ca++, K+, Ca+++Glyc. at 25 °C.
7.15Å

9.90Å
10.00Å
7.13Å
7.13Å

9.81Å
10.04Å

9.92Å

13.79Å

K.25°C
K.25°C
14.01Å

Ca.25°C
Ca.+Glyc.25°C Ca.+Glyc.25°C
Ca.25°C

θ 14 13 12 11 10 9 8 7 6 5 4 3
θ 14 13 12 11 10 9 8 7 6 5 4 3

Figure 14. X-ray diffraction of the clay fractions-sample no. 2 (0-6 cm)
Figure 16. X-ray diffraction of the clay fractions-sample no. 4 (0-7 cm)
with Ca++, K+, Ca+++Glyc at 25 °C.
with Ca++, K+, Ca+++Glyc at 25 °C.

26
 Ü. ALTINBAfi, Y. KURUCU, M. BOLCA, A. H. EL-NAHRY

parent material similar to that of Kücük Menderes formation and its formation is favored by alkalies and
(Alt›nbafl, 1982). alkaline earth’s enrichment of the pedo-environments.
Kaolinite Montmorillonite is inherited from parent materials prior
to sedimentation (Alt›nbafl, 1982).
Kaolinite is a 1:1 clay mineral, the presence of which
was shown by the strong peak at 7.13 Å (25 °C with K-
saturation). The peak disappeared only after heating to Conclusions
520 °C. The presence of kaolinite confirms the 1. The MNF method put most of the information into
hydromorphic condition of the soils and its inheritance a few spectral bands to reduce the volume of data and to
from parent materials. Kaolinite may be formed by the segregate the noise.
weathering of K-feldspars and Na-feldspars from
magmatic and metamorphic rocks or by a hydrothermal 2. The PPI is a means of finding the most “spectrally
attack of acid solutions on feldspars and micas. Kaolinite pure” pixel. The output is an image in which the DN of
is common in Kücük Menderes soils. Kaolinite also may be each pixel in the image corresponds to the number of
formed by the silicification of hydrargillite by silicic acid times that pixel was recorded with iteration running as
solutions (Alt›nbafl, 1982). extreme, thus significantly reducing the number of pixels.
Both the MNF and PPI operations effectively reduce
Vermiculite
the data volume to be analyzed interactively. The PPI
Vermiculite, a 2:1 clay enriched with Mg, was image is used as an input for n-dimensional scatter
identified by the presence of the 14.96 Å peak (Ca- plotting that allows real time rotation in n-dimensions.
saturated and glycolated), which shifted to 10.04 Å after The n-D visualizer for image clustering was performed to
Ca- saturation and heating to 520 °C. The presence of create classes (endmembers) by clustering the purest
this mineral is explained on the premise that Mg-affected pixels in the data set. Animation of the scatter-plots of
conditions stimulate its formation either through bands was used to select the endmembers. The results
digenesis or through neogenesis. This mineral is not as show that there are 4 classes can be distinguished by
widespread as illite in these soils. It may be formed by grouping pixels. After the classes were defined by
hydrothermal action on biotite in a magnesium–rich clustering, the selected classes were exported as regions
environment. After erosion, the mineral is found in the of interest and matched with the spectral library,
clay fraction of fluvial sediments. Vermiculite was found resulting in 4 classes representing different types of clay
in smaller amounts compared with illite and kaolinite in minerals of kaolinite, montmorillonite, vermiculite and
the acidic brown forest soils that developed on mica schist illite.
on the highlands that surround the study area (Alt›nbafl,
3. N-Dimensional visualization for image clustering
1982).
using scatter plotting animation was performed.
Montmorillonite
4. Using the spectra extracted from the ETM image
Montmorillonite is a 2:1 clay mineral formed as a with the aid of hyperspectral tools (MNF, PPI and N-
result of the hydrothermal alteration of volcanic ashes. dimensional visualization) clay mineral type on the soil
The extreme thinness and flexibility of the flake-shaped surface can be identified.
particles account for the plasticity of this mineral
Recommendations
(Alt›nbafl, 1982).
The success of this effort implies that spectral
Montmorillonite reflected at 19.65 Å. [The mineral of
signatures can be used broadly and economically for
this group consists of unit layers formed by 1 Al (Mg, Fe,
identifying clay minerals. Such an effort should
Zn, Cr, Li) – OH octahedral sheet and with 2 Si(Al, Fe)-O
concentrate on automating image analysis to permit the
tetrahedral sheets]. The pattern indicates basal reflection
analysis of large volumes of data in a short time. When
at about 14.96 Å for Ca-saturated samples, which implemented, the spectral signatures approach can be
expanded to 17.65 Å upon glycerol salvation with a used to supplement (and in some cases possibly even
second basal reflection. The peak collapsed to 9.92 Å replace) the X-ray analysis of clay minerals and potentially
upon Ca-saturation and heating for 4 h at 520 °C. This bring new information to the spectra used in a variety of
mineral reflects the contribution of water to soil soil subdisciplines.

27
Using Advanced Spectral Analyses Techniques as Possible Means of Identifying Clay Minerals

References
Alt›nbafl Ü. 1982. Potassium minerals in acidic brown forest soils Goetz, A.F.H., S. Chabrillat and Z. Lu, (2001). Field reflectance
developed on micaschist in Bozda¤ (Brown Mountains) and their spectrometry for detection of swelling clays at construction sites.
sources and their effects on the soil potassium reserves. Ege Univ. Field Analytical Chemistry and Technology, 5: 143–155.
Faculty of Agriculture Press. No. 433. ‹zmir, Turkey. Jackson, M.L. 1975. Soil chemical analyses advanced course, Pub. by
Black, C.A. 1965. Methods of soil analyses Part 1. Physical and author, Madison, Wisconsin.
mineralogical properties American Society Agronomy, Inc. Pub. MacEwan, D.M.C. and Wilson, M.J. (1980) Interlayer and intercalation
Madison,Wisconsin U.S.A. complexes of clay minerals. In G.W. Brindley and G. Brown, Eds..,
Boardman J. W. and F. A. Kruse, 1994. Automated spectral analysis: A Crystal Structures of Clay Minerals and their X-ray Identification,
geologic example using AVIRIS data, north Grapevine Mountains, chapter 3, Monograph 5 of Mineralogical Society, London.
Nevada: in Proceedings, Tenth Thematic Conference on Geologic Markham, B.L. and J.L. Barker, 1986, Landsat MSS and TM Post-
Remote Sensing, Environmental Research Institute of Michigan, Calibration Dynamic Ranges, Exoatmospheric Reflectances and
Ann Arbor, MI, p.I-407-I-418. At-Satellite Temperatures, EOSAT Landsat Technical Notes,
Boardman J.W., F.A. Kruse and R.O. Green, 1995. Mapping target August 1986, p. 3-8, U S A.
signatures via partial unmixing of AVIRIS data: in summaries, Mermut, A.R., S.H. Luk, M.J.M. Romkens and J.W.A. Poesen, 1997.
Fifth JPL Airborne Earth Science Workshop, JPL Publication 95- Soil loss by splash and wash during rainfall from two loess soils.
1, v.1p.23-26. Geoderma 75: 203–214.
Carroll, D., 1970. Clay minerals. A guide to their X-ray identification. Olsen, H.W., L. Krosley, K. Nelson, S. Chabrillat, , A.F.H. Goetz and
Geological Society of America. Spec. page 126, Boulder, Noe, D.C. (2000). Mineralogy-swelling potential relationships for
Colorado, U.S.A. expansive shales. In C.D. Shackelford, S.L. Houston and N.-Y.
Chabrillat, S., A.F.H Goetz, H.W. Olsen,, L. Krosley and D.C. Noe, Chang (Eds.), Advances in unsaturated geotechnics (Geotechnical
(2002). Near-infrared spectra of expansive clay soils and their Special Publication 99, pp. 361– 378). Reston, VA: ASCE.
relationships with mineralogy and swelling potential. Clays and Salisbury, J.W., L.S. Walter, N. Vergo and D.M. D'Aria, 1991. Infrared
Clay Minerals (in revision). (2.1-25 micrometers) Spectra of Minerals” Johns Hopkins
University Press, Baltimore.
De Jong, S.M. 1992. The analysis of spectroscopical data to map soil
types and soil crusts of Mediterranean eroded soils. Soil Weed, S.B., 1977. Minerals in soil environment. Soil Science Society of
Technology 5: p 199-211. America Journal, 56: September-October 1977: 1640-1645.

28

View publication stats