الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 329 |  |  |
| | | from | 329 | | |
Abstract
In the present study; remote sensing and geophysical radiometric data were used for detecting the alteration zones at the Wadi Saqia area supported by field work and lab spectral analysis. Accordingly, different remote sensing and airborne gamma-ray spectrometry techniques were applied. Maps showing the distribution and potentiality areas of alteration zones in the study area were produced, moreover an enhanced geologic map based on remote sensing data was created.
For achieving the objective of the study, different remote sensing data including; multispectral (ASTER), hyperspectral (Hyperion), hyperspectral reflectance measurements (spectral signatures), and Allos PALSAR DEM data were used. In accordance, geophysical airborne gamma-ray spectrometry data were used as well.
5.1. Preprocessing of the satellite data
Prior any interpretation or processing techniques be applied to remote sensing satellite data, different preprocessing procedures must be applied to remove the radiometric, atmospheric and geometric conflicts occurred during data capture, reception, storing and archiving.
5.1.1. Preprocessing of ASTER data
The preprocessing procedures applied to ASTER data are as the following:
5.1.1.1. Cross talk correction (CTC)
A cross talk signal scattering problem in the SWIR sensor of the ASTER data is common. The light incident on the detectors or recorders of bands 4 and 9 is reflected at the detector edge and the filter boundary then transported to the other detectors of SWIR bands by multiple reflections. ASTER images acquired after March, 2008 don’t include the SWIR subsystem bands. Both the ASTER images used in this study were corrected against cross talk errors.
5.1.1.2. Radiometric calibration from Digital numbers to spectral reflectance
The raw ASTER satellite data were converted from raw digital numbers to spectral reflectance using “radiometric calibration” tool of ENVI V.5.1 software. The result of this step is calibrated Top of Atmosphere (TOA) reflectance data including the atmospheric reactions.
5.1.1.3. Atmospheric correction
The TOA reflectance was converted to surface reflectance using the “FLAASH” module of ENVI V.5.1 to ensure pure surface reflectance for both the VNIR and SWIR ASTER data bands.
5.1.1.4. Spatial resolution merge
Using the “Gram-Schmidt spectral sharpening” tool of the ENVI V.5.1, the higher resolution data (VNIR 15m) was merged with the lower resolution data (SWIR 30m) to obtain higher spatial resolution for the SWIR data. Consequently, the SWIR bands were obtained with spatial resolution of 15m instead of 30m, the higher the spatial resolution, the higher the accuracy of the results.
5.1.1.5. Layer stacking
The “layer stacking” function in ENVI V.5.1 was applied to group the VNIR and SWIR 15 m datasets into one dataset containing bands (1, 2, 3, 4, 5, 6, 7, 8 and 9).
5.1.1.6. Mosaicking
Since our study area is acquired in two different ASTER images, then, those two ASTER images were mosaicked together into one dataset. For mosaicking, the “Georeferenced mosaic” function in ENVI V.5.1 was used.
5.1.1.7. Area of interest
Since the two ASTER images mosaic has been larger than the area of interest, as one ASTER scene covers (60 km * 60 km). So, the “subset” function was used to extract the study area.
5.1.2. Preprocessing of Hyperion data
In parallel to the ASTER data, Hyperion satellite images were used in the present study as well; the preprocessing procedures applied to the Hyperion data are as the following:
5.1.2.1. Bad band removal (BBR)
The Hyperion level 1R data contains 242 spectral bands; upon visual inspection of these 242 bands, only 155 bands were found informative. The other 78 bands were defined as bad bands, they were found either badly illuminated (black blank images) or extremely illuminated (white blank images) or having high noise contents hiding the image features due to spectral overlapping. These bands were removed using the “spectral subset” function in ENVI V.5.1 software.
5.1.2.2. Bad column removal (BCR)
Several bands of the 155 informative bands were found to contain vertical stripping. Using the “de-stripping” tool in ENVI 5.1 software, the stripping was removed.
5.1.2.3. Atmospheric correction
The atmospheric correction was applied to obtain surface reflectance data; it was carried out using the FLAASH atmospheric correction tool in ENVI software after the radiometric calibration process that was applied using the “EO-1 Hyperion tools” in the same software.
5.1.2.4. Geometric Projection (Georeferencing)
This process was performed to make sure that the satellite image is rectified to the actual coordinates on the ground. Level 1Gst image which is already geometrically registered was used as the reference base image to which the registration of the level 1R image was carried out. Using the “Georeferencing and registration” tools available in ENVI V. 5.1 and ArcMap v 10.5, control points were selected on both the reference registered and the unregistered images and the registration process was run.
5.1.2.5. Mosaicking the two Hyperion images
The two Hyperion scenes; each having swath width of 7.75 km were mosaicked together into one dataset using the ENVI v 5.1 software.
5.1.2.6. Extracting the area of interest
The “subset” function was used to crop the area relative to our study and trim out the undesirable areas.
5.2. Processing techniques applied to the data
Different processing techniques were applied to the remote sensing data for producing an enhanced geological map based on the remote sensing data. On the other hand, different processing techniques were applied to both the remote sensing and radiometric data in order to map the alteration zones in the study area which is the main objective of our study.
5.2.1. Lithological mapping from remote sensing data
For lithological mapping, several remote sensing techniques were applied on ASTER level 1T data after being preprocessed. False color composites (FCC), Principal Component Analysis (PCA), and Color Ratio Composites (CRC) techniques were utilized. These techniques showed to be effective for lithological discrimination of the rock units in the study area.
5.2.1.1. False Color Composite (FCC) images.
FCC means viewing a colored image through selecting three bands for the three color guns (Red, Green, and Blue (RGB)). For the nine ASTER VNIR-SWIR bands, there are 84 possible combinations. For Hyperion data, there are 608685 possibilities. The OIF statistical calculation method was applied to rank all the 84 possibilities of ASTER data according to the data variance. The best triplet ASTER band combination was bands 1, 5, and 8 (rank=1) in RGB respectively. For Hyperion data, it is impossible to perform the OIF analysis due to the enormous number of possibilities. Another approach was used; the equivalent Hyperion bands to the selected optimal ASTER bands were blended together in a false color composite. The selected Hyperion bands for false color composite were 21, 201, and 217 in RGB, respectively. The false color composites obtained from both ASTER and Hyperion data provided good lithological discrimination of the exposed rocks in the study area.
5.2.1.2. Principal component analysis (PCA)
PCA technique was used to reduce the spectral redundancy and produce less correlated spectral bands. This technique was applied to both the ASTER and Hyperion data of the study area. Based on the analysis of the PCA Eigenvalues of both ASTER and Hyperion data, it was found that the first three principal component images contain almost 99% of the variance in both the data sets. Consequently, these first PCA images for both the data sets were visualized in RGB and showed perfect lithological discrimination.
5.2.1.3. Color Ratio Composites (CRC)
Ratio images are obtained by dividing the reflectance value of the pixels in one band by the value of the corresponding pixels in another band, and then the resulting values are stretched and plotted as a new ratio image. For ASTER and Hyperion data, equivalent color ratio composites to Abram’s Landsat TM ratio were used. The equivalent ASTER Abram ratio composite was (4/7, 3/4 and 2/1) in RGB, respectively. The equivalent Hyperion Abram ratio composite was (150/211, 37/150 and 31/21) in RGB, respectively. Both the color ratio composite images of ASTER and Hyperion data showed excellent lithological discrimination for the rock units in the study area.
Finally, the results of the FCC, PCA, and CRC techniques were integrated together, resulting in an enhanced geological map of the study area.
5.2.2. Alteration zones mapping from radiometric data.
Firstly, radiation distribution maps regarding the K (%), eTh (ppm), and eU (ppm) radio-elements were obtained and carefully examined. For alteration zone mapping from the gamma-ray spectrometry data, several techniques were used including; radio-element ratio maps, potassium point anomaly map, and F-parameter technique.
5.2.2.1. Radioelements concentration distribution maps:
The K (%) concentration distribution map showed that the potassium concentration over the study area ranges from " ~ " 0.1 to 3.0 %. The highest K concentrations are hosted in the Younger granites, Gneisses and Schists, Hammamat molasse sediments, Post Hammamat Felsites, undifferentiated Dukhan volcanics and in the sandstone of Taref Formation. Moderate concentrations were found dominated mostly by Volcanic and Metavolcanic rocks. The lowest concentrations were aligned with the Metagabbros, Ophiolitic Serpentine & Talc Carbonate rocks, older granites, and in the rest of the Phanerozoic sedimentary rocks.
Over the study area, eTh concentration varies from " ~ " 0.54 to " ~ " 10.0 ppm. Areas with the highest eTh (ppm) concentrations are highly correlated to the areas with the highest K (%) concentrations. Moderate concentrations were found present in Metavolcanic rocks and in undifferentiated sedimentary rocks. The lowest Th (ppm) concentrations were found hosted in the Metagabbros, Ophiolitic Serpentine & Talc Carbonate rocks, and older granites.
Uranium concentrations over the study area were found ranging from " ~ " 0.17 to " ~ " 5.0 ppm. The highest eU (ppm) concentrations were found associated mainly with the Younger granites, Gneisses and Schists, Hammamat molasse sediments, Post Hammamat Felsites, Dukhan volcanics, and Taref formation. Moderate concentrations were mapped in Metavolcanic rocks, and in the older granites. The lowest eU (ppm) concentrations were found in the Metagabbros, and in the Ophiolitic Serpentines.
5.2.2.2. Radioelements Ratio maps:
In the present study, the radioelement ratio maps of K/eTh and K/eU were used to delineate areas in which potassium is relatively higher than eTh and eU, respectively. Areas with relatively high K anomalies are suggested to indicate the mineralized zones related to hydrothermal alteration. The K/eTh ratio is regarded the best indicator for areas of potassium enrichment related to hydrothermal alteration. In most rocks, the potassium to thorium ratio is nearly constant, typically varying from 0.17 to 0.2 (K/eTh). Anomalous high areas in the K/eTh ratio with values exceeding this constant range are regarded promising for alteration and mineralization. In the study area, the K/eTh ratio values were found exceeding the constant range reaching more than 0.45. The enriched areas were found scattered all over the study area associated mostly with; Younger Granites, Gneisses, Schists, Post Hammamat Felsites, Dukhan volcanics, and locally in the Volcanic and Metavolcanic rocks. The K/eU ratio map similarly, indicates areas of potassium enrichment relative to Uranium. Areas with high values of the K/eU ratio are highly correlated to that of the K/eTh ratio.
5.2.2.3. K point anomaly map:
Favorable areas for Potassium enrichment were delineated based on the statistical treatment of the three given spectrometric data of K (%), K/eTh, and K/eU. The statistical analyses, including; maximum (max.), minimum (min.), arithmetic mean (X), standard deviation (σ), and the thresholds for anomalous zones with different degrees of accuracy [(X + σ), (X + 2σ), and (X + 3σ)] were performed for each individual rock unit. The coefficient of variability (C.V.) = (σ/X)*100, was also calculated to reveal the degree of normality and homogeneity of the radio-elements distributions over certain different rock units.
The threshold of anomalous values was considered as the value equaling three standard deviations above the calculated arithmetic mean value (X + 3σ). If the maximum value is greater than the (X + 3σ) threshold, then (X + 2σ) or (X + σ) threshold was selected instead. A map was produced showing the potassium point anomalies exceeding the defined thresholds of the three radio-variables over the study area.
Areas in which all the three variables are coincident or at least very close to each other were regarded anomalous for potassium enrichment. On this basis; fourteen statistically significant anomalous areas were delineated and distinguished as Potassium enrichment zones and potential for alteration zones and mineralization.
5.2.2.4. F-parameter:
The Efimov F-parameter [K*(eU/eTh)] describes the abundance of K related to the eU/eTh ratio. Anomalous high F-parameter values are good indicators for altered rocks. It was found that the fourteen delineated areas from the K point anomaly map are remarkably coincident with areas of high F-values; this supports both the results.
5.2.3. Alteration zones mapping by remote sensing data
Remote sensing data and field work were integrated together for mapping the alteration zones in the study area. The approached methodology was as the following steps; 1) field work for collecting representative samples from the alteration zones in the study area, 2) measuring the spectral signatures of the collected samples, 3) spectral analysis of the measured spectral signatures for the identification of the abundant alteration minerals (end-members), 4) mapping the spatial distribution of the identified endmembers using ASTER and Hyperion data, 5) weighted overlay analysis to obtain the alteration potentiality maps for ASTER and Hyperion data, 6) lineaments mapping, 7) integrating alteration and lineaments maps to define the potential areas for alteration zones and mineralization, and 8) validating the results through comparing the mapping results with the locations of the already existing mineralization occurrences and with the locations of the field observations.
5.2.3.1. Field work and Sampling
A field survey was conducted during which about 100 samples were collected from the different exposed rock units and alteration zones in the study area. Different samples were collected from the already known mineralization occurrences such as, Semna gold mine. Other samples were collected from the wall rock alterations observed during the field study. There observed many alteration indications during the field study including; quartz veins/veinlets, clay minerals, Epidote, malachite, and iron oxides, samples from the different alteration types were collected for further spectral analysis.
5.2.3.2. Obtaining spectral signatures
For the present study; the spectral signatures of the collected rock/mineral samples were measured in wavelength range from 350 to 2500 nanometer using handheld spectroradiometer named “ASD TerraSpec Halo mineral identifier”. These measurements were carried out under laboratory conditions at Egypt’s National Authority for Remote Sensing and Space Sciences (NARSS). The spectral signatures of 15 samples were selected representing various alteration minerals in the study area.
5.2.3.3. Spectral analysis for end-members identification
The lab spectra were compared against the United States Geological Survey (USGS) mineral spectral library embedded in the ENVI V.5.1 software. The comparison was carried out using the “spectral analyst” tool of ENVI V.5.1. Spectral Angle Mapper (SAM) and Spectral Feature Fitting (SFF) algorithms were used for recognizing the closest matches to the lab spectra. The results of spectral analysis showed the spectral characteristics and alteration minerals dominating the alteration zones in the study area.
The identified alteration minerals through spectral analysis were; Montmorillonite, Muscovite/Sericite, Chlorite/Clinochlore, Epidote, Kaolinite/Smectite, and Iron oxides including, Hematite, Goethite, and Limonite. The spectral characteristics of these identified alteration minerals were illustrated in terms of the wavelength positions of the characterizing absorption features for each alteration mineral.
5.2.3.4. Alteration mapping using ASTER and Hyperion
Four alteration mapping techniques were used to map the abundance of the identified alteration minerals from the spectral analysis. These four mapping techniques are; Spectral Angle Mapper (SAM), Spectral Feature Fitting (SFF), Constrained Energy Minimization (CEM), and Mixture Tuned Matched Filtering (MTMF), they were applied using ENVI V. 5.1 software.
Each technique was supplied with the spectral signatures of the identified alteration minerals along with the ASTER and Hyperion data. Alteration minerals distribution maps were obtained; two maps from each technique; one representing mapping using ASTER and the other using Hyperion data.
5.2.3.5. Automatic Lineament Extraction
Lineaments are potential indicators of possible faults, cracks, joints or existence of other tectonic features; they can represent zones of deformation and fracturization which implies that they are zones of higher secondary porosity so becoming significant channel-ways for migration of fluids.
In the present study, multiple hill shade images with multi-illumination azimuths (0, 45, 90, 135, 180, 225, 270, and 315) were generated from the 12.5m Allos PAlSAR DEM data using the ArcMap V. 10.5 sotware. The obtained eight shaded relief images of the study area were overlain together producing one combined hill shade image that was used for automatic lineament extraction. The lineaments were obtained automatically using the Line Module in the PCI Geomatics software. A lineament density map was produced showing the frequencies of lineaments per unit area.
5.2.3.6. Weighted overlay analysis
Weighted Overlay analysis is a spatial analysis technique of the Geographic Information systems (GIS) that depends on a number of thematic layers mainly for suitability/potentiality analysis. The layers of all the alteration mineral were assigned the weight of 14 % except of Epidote and Limonite; each was assigned the weight of 8%. Consequently, four weighted alteration abundance maps were obtained for each technique individually from both ASTER and Hyperion. The weighted overlay analysis was applied once more for the weighted alteration maps of the individual techniques to produce weighted alteration maps integrated from all the four techniques. Consequently; one integrated alteration abundance map from the alteration layers of the four techniques was obtained for ASTER data and another map for Hyperion data.
5.2.3.7. Integration of alteration and lineaments Data
The mapped alteration minerals were integrated with the automatically extracted lineaments to delineate the most potential areas for the existence of alteration zones and mineralization as well. Coincidences of areas that have been mapped containing alteration minerals on areas with high lineaments density indicate high probability of alteration zones and mineralization.
Hence, the obtained final lineaments density map was integrated with the obtained final integrated alteration minerals abundance maps from both ASTER and Hyperion data resulting in the final alteration zones potentiality maps.
5.2.3.8. Validation of the alteration mapping results
For validating the results of alteration mapping, locations of the observed alteration areas during the field study including the location of Semna gold mine were plotted on both the alteration mapping results obtained from ASTER and Hyperion remote sensing data.
ASTER mapping results mapped the location of Semna gold mine as high potential for alteration and mineralization, this validates the ASTER result. Moreover, the validation checking results on ASTER data show that 12 locations out of 15 validation points were found located on areas having moderate, high, very high potentiality for mineralization.
From, the mapping results of Hyperion data, it is shown that all validation points were mapped successfully in the moderate and high potentiality with not only once in the low potentiality defined areas.
Consequently, the accuracy of Hyperion data is relatively higher than that of ASTER data. This is logic and mainly attributed to the great difference in spectral resolution between both the datasets that makes Hyperion more efficient in material identification through spectral information due to the high number of bands covering the VNIR and SWIR portions of the EMR.
5.3. Final Lithological and alteration map
At last, an integrated lithological and alteration map for the study area was introduced comprising all the results came out of this study regarding lithological and alteration zones mapping.
Finally, the adopted methodology and techniques applied to both remote sensing and airborne gamma-ray spectrometry data are recommended to be followed for lithological and alteration zones mapping in similar areas in the ANS or even across the globe.