Recognition of a porphyry system using ASTER data in ١ Bideghan-Qom province ( central of Iran )

١١ The Bideghan area is located south of the Qom province (central of Iran). The most ١٢ impressive geological features in the studied area are the Eocene sequences which are ١٣ intruded by volcanic rocks with basic compositions. Advanced Space borne Thermal ١٤ Emission and Reflection Radiometer (ASTER) image processing have been used for ١٥ hydrothermal alteration mapping and lineaments identification in the investigated area. In this ١٦ research false color composite, band ratio, Principal Component Analysis (PCA), Least ١٧ Square Fit (LS-Fit) and Spectral Angel Mapping (SAM) techniques were applied on ASTER ١٨ data and argillic, phyllic, Iron oxide and propylitic alteration zones were separated. ١٩ Lineaments were identified by aid of false color composite, high pass filters and hill-shade ٢٠ DEM techniques. The results of this study demonstrate the usefulness of remote sensing ٢١ method and ASTER multi-spectral data for alteration and lineament mapping. Finally, the ٢٢ results were confirmed by field investigation. ٢٣


٢٥
Iran is located in the Alpine-Himalayan orogenic belt and Uromieh -Dokhtarmetalogenic belt ٢٦ is the most impotant zone in Iran which has high potentials for gold and copper as well as ٢٧ other base metal deposits.In Iran, satellite data such as TM, ETM+ and ASTER (Advanced ٢٨ 2 Spaceborne Thermal Emission and Reflection Radiometer) have been used by geologists for ١ exploration purposes (AsadiHaroni and Lavafan, 2007;Azizi et al. 2010).

٢
Intermediate to acidic igneous rocks from the late Cretaceous to Tertiary in the western ٣ ,northwestern and northern parts of Iran are important because the high value presence of ٤ copper and gold mineralization in these host rocks .These rocks have been found in the ٥ 1:100000 Kahak sheet .The studied area is located in this geological map as a part of Uromieh ٦ -Dokhtar belt.Therefore, high resolution remote sensing data and Geographic Information ٧ Systems (GIS) are important tools to map subtle anomalies associated with unknown gold and ٨ base metal deposits (Kujjo, 2010;Bell, 2011;Albanese, 2011).

٩
Because the alteration zones recognition are the most important keys for base metal deposit ١٠ exploration therefore using of methods such as remote sensing which help to separate these ١١ zones will be useful.

١٢
In this research, a spectral analysis was carried out on the ASTER satellite imagery data of the ١٣ investigated area to map spectral signatures associated with the hydrothermal alterations.

١٤
Least Square Fit (LS-Fit), Principal Components Analysis and band ratio methods were also ١٥ performed to achieve better accuracy along with spectral analysis.In addition, lineaments by ١٦ using of False Colour Composite (FCC) images, high pass filters and hill-shade DEM ١٧ techniques were extracted.

١٨
Finally the integration of alteration, lineaments and litho logical features can show the hopeful ١٩ areas for exploration.

٢٠
There are a lot of researches in different case studies with these methods which have very ٢١ good results, for example remote sensing studies in 1:100000 Soltanieh map (Feizi et al.

٧
The ASTER sensor is one of the multi-spectral sensors that has been installed on the TERRA ٨ satellite.This sensor can measure the reflection of the Earth's ground in three bands, that is, ٩ between the wavelengths of 0.52-0.86l m with a resolution of 15 m (visible and near-١٠ infrared: VNIR), six bands between the wavelengths of 1.6-2.43lm with a resolution of 30 m ١١ (SWIR), and five bands between the wavelengths of 8.125-11.65 lm with a resolution of 90 m ١٢ (thermal infrared: TIR) (Figure 2).(Rowan & Mars, 2003;Moghtaderi et al., 2007;Yousefifar ١٣ et al., 2011).Furthermore, the TERRA satellite has a back-looking telescope with a resolution ١٤ of 15 m in the VNIR that matches with the wavelength of the band 3 that is used to extract 3D

١٠
The studies on re-sampling of USGS standard curves on ASTER bands show that Al-OH ١١ minerals such as kaolinite, muscovite, montmorilonite and illite (major minerals for phyllic ١٢ and argillic alteration zones) have the most reflection in B4 on SWIR (Figure 3).Also, Mg-

١٣
OH minerals such as chlorite and epidote that are remarkable for propylitic alteration zones

٢٠
The Principal Component Analysis (PCA) is a multivariate statistical technique that selects ٢١ uncorrelated linear combinations (eigenvector loadings) of variables.Each successively ٢٢ extracted linear combination, or principal component (PC), has a smaller variance.The PCA ٢٣ is widely used for alteration mapping in metallogenic provinces (Myint, et al., 2005) and has

٢٤
been applied in this study.An approach based on the examination of eigenvector loadings in ٢٥ each PC image is used for determining which image contains information related to the ٢٦ spectral signatures of specific target minerals.It is expected that the PC image that collects ٢٧ moderate to high eigenvector loadings for the diagnostic absorptive and reflective bands of the ٢٨ index mineral could be considered as the specific image for that mineral.If the loading of the ٢٩ absorptive band is negative in sign the target area will be enhanced in bright pixels, and if the ٣٠ loading of the reflective band is negative the area will be enhanced in dark pixels (Crosta and ٣١ Moore, 1989;Khaleghi, and Ranjbar, 2011).

٣٢
Table 1 shows the eigenvector loadings for bands 1, 2, 3 and 4. In this table at a ratio and ١ at a ratio should not be + or -both.These ratios must be (normal case) or (Inverse ٢ case).Inverse case has to correct.The analyses which have done on table 1 show in PC3,

٣
ratio is so it should study in inverse case.In PC4 ratio is and ratio is so PC4 ٤ have to study in both case.Base on this investigation PC3 shows good results in inverse case.

٥
Inverse of PC3 can show the areas with iron oxide.6).

٢٦
Distribution of iron oxide was created by using all three visible and near-infrared (VNIR)

Minimum Noise Fraction (MNF)
٣ The Minimum Noise Fraction (MNF) transformation was used to determine the inherent ٤ dimensionality of image data, segregate noise in the data and reduce the computational requirements for subsequent processing (Boardman et al. 1995;Green et al. 1988 iron oxide, argillic, phyllic and propylitic alterations (Figure 8).

٢١
The cause of using these methods is very good results in determination of alteration zones ٢٢ with these bands combination.For example in a case study in the north of this region (Feizi et ٢٥

٢٦
In this part for integration the data layers in GIS area, first of all, the shape files of all ٢٧ alteration zones which were carried out with different methods were drown.Then the layers

٢٨
were overlapped on each other.Afterwards the most overlapped zones were chosen and were

٢٩
controlled with field investigations.As the last step, the lineament map of studied area was ٣٠ integrated with the final alteration map in GIS area.(Figure 12).As the figure shows, there is ٣١ a very good adaptation between these two layers, especially in the east and a band with a NW-٣٢ 8 SE trend in the south of the area.There is also a good adaptation in north of the area

١
.Therefore, a circular band that begins from the northwest corner to the east, southeast and to

٢
the west has been recognized.

٤
After all these software analyses, field investigations were necessary.Figure 13 shows a full ٥ view of studied area.The control points were detected, and after the field studies, the ٦ correction of alteration zones were confirmed.Figure 14 shows the three check points for Iron ٧ oxide which were recognized with using remote sensing processes.These checking have ٨ confirmed the results of the RS methods.

٩
As Figure 15 shows, Sericitic Muscovites, Quartz and a few of Illite have been seen in the

١٠
Phllic alteration zones.The check field for Argillic alteration zones were confirmed by the ١١ results of RS methods (Figure 16).Presence of Chlorite and Epidote minerals in check field

١٢
processes confirmed the Propylitic alteration zones (Figure 17).The adaptation of Iron Oxide

١٣
and Argillic alteration zones is a very important guide for hydrothermal mineralization ١٤ deposits, especially sulfide minerals (Figure 18).

٢٧
The existence of these alterations is important key for exploration of this kind of deposits,

٢٨
especially, with remote sensing methods.and Phyllic zones.

٣
The porphyry copper deposits have a red to brown iron-stained crust called Gossan or leached ٤ capping which can be useful for exploration.(Sabins, 1999).begins from the northwest corner to the east, southeast and to the west.

١٥
There is probably a porphyry system, caused by locating of the propylitic alteration zone

٢٢ 2013 )
and Iron prospecting in 1:100000 Sfordi sheet(Sadeghi et al. 2013). is located between 50˚ 56ꞌ, 50˚ 59ꞌ longitude and 34˚ 17ꞌ, 34˚ 20ꞌ latitude in ٢٦ the south of the Qom province (Northeast of 1:100000 KAHAK Sheet).Based on 1:100000 ٢٧ geological map of Qom, the most impressive lithological features in the studied area are ٢٨ the volcanic rocks with basic combinations.These features are included; Andesite, ٢٩ Andesibasalt and mega porphytite andesite.There are tuffaceous sandstone and lime stones in ٣٠ the east and southeast of the investigated area.Hematitation, limonitation, silisitaion and ٣١ argillic alterations are seen in central parts that have been formed in north-south trend in the ١ east and northeast of the studied area.Volcanic rocks have specific conditions.For example, ٢ ferrous ore such as Oligist formed along fractures in a tectonic environment.(Agha Nabati ٣ ,2004).Based on 1:100000 geological map of Kahak, three NE-SW faults have been seen.٤Accordingto study results conducted around the faults, alteration and ferrous fluid have been ٥ observed (Figure1).

١٥information.٤
Meanwhile, the sensor swath of ASTER is 60 km(Yamaguchi et al., 2001;    ١٦Rowan et al., 2006).According to extreme variations of spectral reflectance curves of ١٧ minerals in the SWIR region and high spectral resolution of the ASTER sensor, the sensor ١٨ identifies different rocks and minerals on the Earth's surface effectively.Considering the ١٩ differences of the sensor resolution capability between ETM + and ASTER sensors, usually ٢٠ ETM + images are used for the lineaments and ASTER images are used to identify minerals ٢١ and alterations of the Earth's surface.The ASTER is an advanced optical sensor comprised of ٢٢ 14 spectral channels ranging from the visible to thermal infrared region.It will provide ٢٣ scientific and also practical data regarding various field related studies of the earth (Watanabe ٢٤ and Matsuo, 2003).Various factors affect the signal measured at the sensor such as the drift of ٢٥ the sensor radiometric calibration, atmospheric and topographical effects.For accurate ٢٦ analysis, all of these corrections are necessary for remote sensing imagery.To this end, at the ٢٧ beginning of the path, data set in hierarchical data format (HDF) was used for this research ٢٨ and radiance correction such as wavelength, dark subtract and log residual by ENVI4.4 ٢٩ software which was essential for multispectral images were implemented.ASTER bands have ٣٠ good sensitivity for alteration minerals.For example VNIR band is good for Iron oxide and ٣١ SWIR is good for argillic alteration in band 4,propellitic alteration in band 6 and ٣٢ phyllicalteration in band 4, 5 or 8 usually.Also TIR include thermal bands for silica ١ Many image analysis and processing techniques can be used to interpret the remote sensing ٥ spectral data.In this research, False Color Composite (FCC), Minimum Noise Fraction ٦ (MNF), Principal Component Analysis (PCA), Least Square Fit (LS-Fit) and Spectral Angle ٧ Mapping (SAM) methods were used on the ASTER data for the discrimination of alteration ٨ zones.

١٤
have the most reflection in Band 5 and 6 on SWIR.(Yetkin et al. 2004.Figure 4) ١٥ Therefore, phyllic and argillic alteration zones are shown reddish to pink and propylitic ١٦ alteration zones are shown green in 468 False Color Composite (FCC) on SWIR.(Beiranvnd ١٧ Pour et al. 2011).(Figure 5) ١٨ ١٩ Squares Fitting (LS-Fit) ٢٠The technique assumes that the bands used as input values are behaving as the variables of a ٢١ linear expression and the 'y' value of the equation, namely the predicted band information, ٢٢ gives us a calculated output value.This predicted band is what the band should be according ٢٣ to the linear equation.The minerals which are sensitive to a specific band are then ٢٤ differentiated from the features which are reflective to the other bands as well by simply ٢٥ taking the difference between the predicted values and the original values (Yetkin et al. 2004).

٣٠(propylitic٤
Mapper (SAM) is a classification technique that permits rapid mapping by ٢٦ calculating the spectral similarity between the image spectrums and reference reflectance ٢٧ spectra.SAM measures the spectral similarity by calculating the angle between the two ٢٨ spectra and treating them as vectors in n-dimensional space (Kruse et al. 1993), (Beiranvnd et ٢٩ al. 2011).The image spectra were compared with USGS Digital Spectral Library (Minerals) Malekzadeh et al. 2009).Figure 8 shows selected minerals spectral library plots that related ٣١ to argillic, phyllic and propylitic alterations.Three mineral spectral representatives of the ٣٢ argillic zone include kaolinite, dickite and halloysite.Two minerals spectral representatives of ١ the phyllic alteration consist of illite, muscovite, and epidote.Chlorite representative of ٢ Lineament extraction in this study was carried out by manual method.In manual extraction ٥ method, the lineaments are extracted from a satellite image by using visual interpretation.In ٦ this research Shaded relief method and median filter were used for fault identification with ٧ using DEM image which was prepared by Geological survey of Iran.The lineaments usually ٨ appear as straight lines or "edges" on the satellite images which in all cases were contributed ٩ to the tonal differences within the surface material.The knowledge and the experience of the ١٠ user is the key point in the identification of the lineaments particularly to connect broken ١١ segments into a longer lineament (Sarp, 2005).١٢ False color images are produced for manual lineament extraction because they increase the ١٣ interpretability of the data.Different combinations of three bands are examined and the best ١٤ visual quality is obtained with a false color image utilizing three 7, 4, and 2 (in blue, green ١٥ and red respectively).This false color combination made it easier to identify linear patterns of ١٦ vegetation, geologic formation boundaries, river channels and geological weakness zones.١٧ Moreover, filtering operations are used to emphasize or deemphasize spatial frequency in the ١٨ image.The filtering operation will sharpen the boundary that exists between adjacent units.١٩ Furthermore, standard GIS techniques have been carried out to help in the evaluation of the ٢٠ lineaments detected.Digital Elevation Model (DEM) has the advantage of representing the ٢١ vertical extension of the earth's surface by assigning height values for every pixel (Papadaki et ٢٢ al. 2011).The Hill-shade DEM technique is also effective in creating images that enhance ٢٣ geomorphologic features (Weldemariam, 2009).Therefore, Hill-shades DEM with different ٢٤ azimuth direction and sun angle were used in this study (Figure11).
world's copper is mined from porphyry deposits which occur in a different ١٩ geologic environment.Hydrothermal alteration is also common at porphyry deposits and may ٢٠ be recognized by the same methods that were developed in remote sensing.model of hydrothermal alteration of porphyry copper deposits that was ٢٣ developed by Lowell and Guilbert _1970, Phyllic zone is the most intense alteration which ٢٤ occurs in the core of the porphyry body and contains Quartz, Sericite and Pyrite.Other ٢٥ alteration zones are Argillicand Propylitic.The most characteristic minerals inArgillic zone are ٢٦ Quartz, kaolinite and montmorillonite.Propyliticzone containsEpidote, calcite, and chlorite.
includes disseminated grains of Chalcopyrite, Molybdenite, Pyrite and other ١ metal Sulfides.The most of this mineralization occurs near the boundary between the Potassic

٦
Because the porphyry system in the research area was recognized by using remote sensing ٧ methods, field studies were necessary.These operations were successful and copper minerals ٨ indexes were evident.As Figure18shows, the copper hydro carbonates like Malachite and ٩ Azurite covered the surface of the rocks These minerals were recognized in the central parts ١٠ of propylitic and argillic rings.
dominant strikes: NE -SW, N-S, NW -SE were recognized in the studied ١٣ area.The result of integration between alteration and lineaments indicate a circle band that١٤

١٦
around the argillic alteration zone, especially in the central part of the area.The overlapping ١٧ between argillic alteration and iron oxide zones indicate the presence of sulfide deposits.The ١٨ phyllic alteration zone exists in the middle of the band, especially where the intrusive bodies ١٩ are.
Table 2 shows the eigenvector loadings ٩investigation PC3 shows good results in normal case.According to the results, the image ١٠ related to PC3 shows the argillic alteration.Table 3 shows the eigenvector loadings for bands ١١ 1, 3, 5 and 6.In PC2, ratio is so it should study in inverse case and In PC3, and ratios ١٢ are so they should study in normal case.In PC4, ratio is so it should study in inverse ١٣ case.Base on this investigation PC4 shows good results in inverse case.Inverse of PC4 can ١٤ show the areas with phyllic alteration.Table 4 shows the eigenvector loadings for bands 2, 5, ١٥ 8 and 9.In PC2, ratio is so it should study in normal case.In PC3, , ratios are so they ١٦ should study in normal case.In PC4, ratio is so it should study in normal case.Base on this ١٧ investigation PC4 shows good results in normal case.PC4 can show the areas with propylitic ١٨ alteration (Figure In his research, both methods were used for controlling the results with ١٠ comparing them.MNF involves two steps.In the first step, which is also called noise (Beiranvnd Pour et al. 2011) Pour et al. 2011)s for noise covariance matrix are calculated.This step ١٢ decorrelates and rescales the noise in the data.In the second step, principal components are ١٣ derived from the noise whitened data.The data can then be divided into two parts.One part ١٤ associated with large Eigen values and the other part with near unity Eigen values and noise ١٥ dominated images.Using data with large Eigen values separates the noise from the data, and١٦improves spectral results(Green et al. 1988;Beiranvnd Pour et al. 2011).MNF analysis can ١٧ identify the locations of spectral signature anomalies.This process is of interest to exploration ١٨ geologist because spectral anomalies are often indicative of alterations due to hydrothermal ١٩ mineralization(Beiranvnd Pour et al. 2011).MNF bands 2, 6, 4 (inverse) and 5 were used for ٢٠

Table 2 .
The result of PCA for enhancing argillic zone ١