Professional Documents
Culture Documents
Abstract
Factor analysis provides a quantitative interpretation for synthesizing and correlating data from airborne
spectrometric surveys. Factor analysis is applied on such data of Al-Awabed area, Northern Palmyrides. The seven
variables used in this research are: total radioactivity Ur1, eU, eTh, K%, eU/eTh, eU/K and eTh/K. The analysis and
interpretation show that a model of four factors (F1, F2, F3 and F4) is sufficient to represent them, where 94% of total
data variance is interpreted. Mapping of these four factors proved to be a powerful tool for a direct differentiation of
various rocks units, and a score lithological map of 11 radiometric units was established.
r 2004 Elsevier Ltd. All rights reserved.
0969-8043/$ - see front matter r 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.apradiso.2004.08.050
ARTICLE IN PRESS
650 J. Asfahani et al. / Applied Radiation and Isotopes 62 (2005) 649–661
Fig. 1. Simplified geological map of the Palmyrides Central Syria, showing Phosphatic deposits.
outcropping units. In fact, this technique is considered as a Upper Triassic to Neogene (Fig. 1); among these, the
powerful tool in providing a better description of the Soukhneh group is characterized by its significant
geophysical data, and an assessment of the underlying phosphorite deposits (Al-Maleh and Mouty, 1994). It is
factors controlling their variance. Statistical characteristics composed mainly of two rock types: calcareous and
of each radioelement (eU, eTh, K%), and their ratios (eU/ siliceous. The first type is dominated by limestone, marly
eTh, eU/K, and eTh/K) are also examined for determining limestone, limy marl and marl with characteristic limy
and outlining the significant radiometric anomalies concretionary structures of few centimeters up to 2 m in
occurring through the choice of a statistical threshold size. The siliceous rocks are composed of thin layered flint
value of two standard deviations plus the arithmetic mean. bands, lenses and nodules. The Soukhneh group is divided
into two lithological formations (Rmah and Swwaneh).
Fig. 2 shows a typical geologic column of phosphatic
2. Geological setting deposits in Central Syria (modified by Jubeli, 1998).
Phosphatic layers thicken in the Central Palmyrides and
The Palmyrides are located in Central Syria and are thin eastwards until they pinch out under marl Arak and
subdivided into northern and southern ranges separated Tantour formations. Phosphatic deposits in Syria can be
by an intermontane extensive basin filled with Neogene– classified in two types, A and B (Fig. 1). Type A presents
Quaternary deposits, i.e., Ad-Ddaw. Sedimentary forma- the Upper Cretaceous phosphates, while Type B presents
tions exposed in the Palmyrides range in age from the Paleogene (Lower Eocene) Syrian Desert phosphorites.
ARTICLE IN PRESS
J. Asfahani et al. / Applied Radiation and Isotopes 62 (2005) 649–661 651
The phosphorite deposits in the region are In addition, they are associated with primary and
attributed to enrichment processes involving secondary uranium mineralization. The former is
phosphorus and plankton particles which indicates associated with phosphate precipitation while the latter
a paleogeographic evolution related to Cenonian fills open cracks and pores, due to surface and subsur-
transgression and the subsidence of the Arabian face water percolation (Abbas, 1987). Fig. 3 shows the
Platform. geological map of the Al-Awabed area.
ARTICLE IN PRESS
652 J. Asfahani et al. / Applied Radiation and Isotopes 62 (2005) 649–661
3. The airborne radiometric survey Settings of the four gamma-ray energy windows are
listed in Table 1. Fig. 4 shows the regions surveyed and
The airborne radiometric survey was carried out using the resulting radiometric map.
a compact, lightweight, four windows gamma-ray The research area is located in the Northern
spectrometer (Scintrex GAD-6, Canada) with a detector Palmyrides to the northwest of T4 oil pump station,
of 12.5 L NaI(Tl) volume. The detector was maintained and is situated between 371300 0000 E and 371450 0000 E, and
at constant temperature and the gain was stabilized between 3414l0 3000 N and 341540 0000 N. Fig. 5 shows the
before recording each mission. The system calibration surveyed eight E–W oriented parallel flight lines, and the
took place at the calibration pads at the Dala airport in resulting radiometric map. Ur, eU, eTh and K% were
Sweden (Riso, 1987). An IGI Loran-C navigation measured at 5190 stationpoints, as were the coordinate
system was used to provide efficient flight path control. points by Loran-C.
It defines and records automatically all the waypoints
and fiducials of the flight tracks along the flight lines.
Details of the equipment and the survey operations can 4. Statistical analysis of the data
be found in (Riso, 1987). The area covered by the
airborne survey consists of the following three areas: Radioactive measurements acquired by the spectro-
metric gamma technique applied in the research area
The Syrian Desert (7189 line km at 4 km line spacing). were subjected to quantitative and qualitative statistical
Ar-Rassafeh Badyieh (2240 line km at 4 km line analysis in order to draw a valid conclusion regarding
spacing). the nature and significance of the distribution of the
The Northern Palmyrides range (1600 line km at 3 km radioelements in it. The applied statistical analysis
spacing). includes single variate and bivariate statistics. In
addition, some quantitative statistical measures were
In all cases, a constant terrain clearance of 30 m was determined, such as the threshold levels, frequently
maintained as a standard survey height and the aircraft defined as the mean plus two standard deviations, which
speed was 120 km/h. indicates the beginning of anomalous values.
4.1. Single variate analysis mean ðx̄Þ; measures of dispersion of the data and
standard deviation (s), as shown in Table 2.
Single variate statistics were used to describe the
statistical characteristics of the distribution of each 4.2. Bivariate analysis
radioactive variable. These include range, arithmetical
Table 1 Correlation analysis has been applied as a bivariation
Specification of the counting windows statistics in order to examine the mutual relations and
strength of association between pairs of variables
Window Window setting Radioisotope through calculation of the linear Pearson product
designation (MeV) mainly detected moment correlation coefficient ‘‘r’’.
40
Potassium 1.38–1.56 K
Uranium 1.66–1.90 214
Bi 4.3. Factor analysis
208
Thorium 2.44–2.77 Tl
It is a multivariate statistical technique by which
Total-count 0.40–2.77
variables on a set of samples are linearly combined
Fig. 4. A: Areas surveyed by airborne gamma-ray spectrometry. B: Radiometric map resulting from spectrometric survey.
ARTICLE IN PRESS
654 J. Asfahani et al. / Applied Radiation and Isotopes 62 (2005) 649–661
Fig. 5. A: Flight lines of the spectrometric airborne survey in Al-Awabed area. B: Resulting total radioactivity (Ur) in Al-Awabed
area.
Table 2
Statistical characterestics of the 7 radioactive variables in Al-Awabed area
giving rise to new fundamental factors, which can be variable are determined for the Al-Awabed area
named and interpreted based on geological reasoning. (Table 2).
The number of variables is reduced to a minimum The maximum equivalent uranium in the studied area
number of independent variables, which will adequately is 22.33 ppm, with an average of 3.04 ppm and a
describe the data. standard deviation of 2.69 ppm.
This technique has been applied on airborne gamma- Review of the computed coefficient of variability (CV)
ray spectrometric data from a survey performed in of the radiometric variables (eU, eTh, and K%) showed
South Texas (Duval, 1976), in the Colorado quadrangle that uranium exhibits a relatively high value (88.57%),
(Wecksung, 1982), and in several prospected areas in when compared with those of thorium (37.84%) and
Egypt (Moustafa et al., 1990). potassium (38.86%). A higher coefficient of variability
Using this technique, a system of factors is obtained implies a lower degree of homogeneity. In other words,
through the transformation of the three radioelements the observed relative tendency of uranium toward
(eU, eTh, and K%), their ratios (eU/eTh, eU/K%, and eTh/ heterogeneity is interpreted to be attributed to the
K%), and the total radioactivity (Ur). These new factors relatively higher mobility of the uranium in comparison
are constrained to reproduce as much as possible the
total variance of the original data. Each original data
point gains factor scores, representing the affiliation of Table 5
the samples to the newly defined factors. The mapping The four rotated factors
of factor scores produces a set of new maps. Compara- Factor variable 1 2 3 4
tive study of these maps with a geological map serves as
a powerful tool in reinterpreting the data to provide Ur 0.926 0.091 0.340 0.082
direct differentiation of all rocks units on a lithological K% 0.024 0.478 0.824 0.011
score map. eU 0.986 0.039 0.0142 0.075
eTh 0.091 0.221 0.915 0.0748
eU/eTh 0.854 0.149 0.299 0.144
eU/K% 0.80 0.326 0.261 0.372
eTh/K% 0.05 0.982 0.0057 0.027
5. Results and discussion Eigenvalue 3.126 1.695 1.588 0.155
% 44.654 24.215 22.688 2.22
By applying a single variate analysis, the statistical Cum% 44.654 68.867 91.557 93.777
characteristics of the distribution of each radioactive
Table 3
Correlation matrix of the 7 radiometric variables in Al-Awabed area
Ur 1
K% 0.35 1
eU 0.92 0.03 1
eTh 0.21 0.65 0.11 1
eU/eTh 0.69 0.15 0.85 0.39 1
eU/K% 0.59 0.35 0.76 0.27 0.77 1
eTh/K% 0.14 0.47 0.09 0.23 0.19 0.29 1
Table 4
Unrotated factors
Variable F1 F2 F3 F4 F5 F6 F7
with Th and K, which show relatively higher stability Bivariate correlation analysis is also carried out, and
under the same conditions. the coefficient matrix between the seven data variables in
Asfahani (2002) adopted this concept as a critical the research area is shown in Table 3.
parameter distinguishing between phosphatic rocks and This correlation matrix shows a cluster of high
phosphatic sands in Khneifiss mine by interpreting positive correlation between three variables (Ur, eU
natural gamma-ray well logging measurements through and eU/eTh), High positive correlations have also been
a statistical approach. found between eU/eTh and eU/K (0.77), and between eU/
eTh and eU (0.85).
The above matrix is used to obtain the unrotated
Table 6 loading matrix of Table 4. The unrotated factors are
Factor score coefficients
difficult to be interpreted, and therefore it is necessary to
Variable F1 F2 F3 F4 rotate them into another form, which is equivalent to the
original unrotated matrix, but represents factor con-
Ur 0.509 0.00124 0.796 0.119 struction. This can be achieved by using the varimax
K% 0.0576 0.128 0.382 0.465 method (Comery, 1973), which allows a reduction from
eU 0.341 0.069 0.665 1.452 the data system of seven dimensional factors into four
eTh 0.661 0.102 0.376 0.113 principal factors (F1, F2, F3 and F4) without losing
eU/eTh 0.0622 0.0553 0.0745 0.221
significant information.
eU/K% 0.155 0.044 0.025 1.67
eTh/K% 0.0038 0.883 0.125 0.371
The four rotated factors are quite interpretable and
represent 94% of the total system information, which is
sufficient to interpret the variable data as shown in Fig. 10 shows the correlation of the chosen lithofacies
Table 5. cross-sections according to the profile P1–P2 of a
A high eigen value indicates the importance of each NE–SW direction. The thickness of the phosphatic layer
factor for data representation. Factor 1 has a value of varies between 20 and 240 cm. Radioactive measure-
3.126, whereas the other factors have lower values, ments have been carried out along cross-sections by the
which indicates that the first factor is the most important French SPP2 Saphymo. Phosphatic layers radioactivity
for representing the variation in the measurements. varies between 120 cps to more than 450 cps reflecting
The factor score coefficients of the four factors, shown P2O5 content and U concentration, since a positive
in Table 6, allow the construction of four standards relationship between uranium concentrations and phos-
factor score maps (Figs. 6–9). phate content have been established in most world
F1 explains 44.65% of the total variance, and has high phosphate deposits (Afteh, 1967; Gavshin et al., 1974;
loading values of 0.926, 0.986, 0.854, and 0.8 for the Altschuler, 1980; Abbas, 1987; Asfahani and Kamarji,
variables of total radioactivity (Ur), eU, eU/eTh, and eU/ 1996; Jubeli, 1998; Asfahani, 1999, 2002; Asfahani and
K%, respectively (Table 5). This factor is therefore Abdul-hadi, 2001).
composed of these four variables, and is directly related The second factor (F2) explains 24.21% of the
to the U presence in the phosphate deposits outcropped variability of the geophysical data. It is relatively highly
in the Al-Awabed area. Therefore it might be termed loaded for K% variable (0.478), and inversely highly
uranium phosphate factor (U–P factor). loaded for the ratio of eTh/K% (0.982). This factor is
More than 43 detailed lithofacies cross-sections have related to K presence and can distinguish between
been established in the research area (Fig. 3) eight of different lithostratigraphic units according to their
which were selected by F1. alkalinity, therefore it can be termed a limestone factor
ARTICLE IN PRESS
658 J. Asfahani et al. / Applied Radiation and Isotopes 62 (2005) 649–661
due to its association with chalky limestone intercalated with accumulated erosional products of phosphate rocks
with siliceous and marly clay. The N–S deep drainage in wadis and lowlands.
network’s effect is incorporated in such facies with the Eleven interpreted lithofacies units are determined
factor (F2) map, Fig. 7. This factor seems to be divided through the comparison and matching of the three
into four variable levels, depending on carbonate and mentioned score maps with the geological map of the
marly clay contents in the lithostratigraphic units. study area, as shown in Fig. 11. Table 7 shows the
The third factor (F3) explains 22.69% of the data ranges of the standard factor scores characterizing the
variability, and is inversely highly loaded with K% and outlining rock units of Al-Awabed area.
eTh, where the loading values are 0.824 and 0.915, The investigation of the standard factor score map of
respectively. This factor reflects tectonic effects, espe- Fl, (Fig. 6) indicates clearly the outlining of the
cially in the zones of hard rock exposures that are following phosphatic highly radioactive rocks:
fractured and faulted in NW and SE of the studied area.
Therefore, this factor can be termed fracturing factor. P1: Detrital phosphate sand with fish bones and flint
According to this factor, connected with hard rocks fragments.
like limestone, dolomitic limestone, three classified P2: Soft phosphate rocks with siliceous and calcareous
levels, were set. nodules.
The fourth factor (F4) explains only 2.22% of the P3: Phosphate rocks with siliceous and calcareous
data variability, and is inversely loaded with the ratio of nodules.
eU/K% (0.372). At this stage of research it is difficult P4: Marly phosphate, interbeded with siliceous and
to be precisely determined, though it may be connected calcareous beds.
ARTICLE IN PRESS
J. Asfahani et al. / Applied Radiation and Isotopes 62 (2005) 649–661 659
Fig. 10. Lithofacies cross-section, showing lateral variations of phosphate beds thickness and radioactivity, Al-Awabed area.
Table 7
Standard factor score characterizing the rock units of Al-Awabed area