Integrating remote sensing and geophysical data for mapping potential gold mineralization localities at Abu Marawat area, central Eastern Desert, Egypt

Abstract

Abu Marawat area in the Central Eastern Desert of Egypt is a very promising mineralization district located in the Golden Triangle area. The current study provides an integrated approach from multisource datasets including; remote sensing, airborne geophysical spectrometry and magnetic data supported by field studies and spectroscopic analyses for delineating potential mineralization localities. Several remote sensing techniques were adopted including; Band Ratios, Relative Band Depth, Mineralogical Indices, Spectral Angle Mapper, and Constrained Energy Minimization. These techniques showed that the alteration mineral assemblage is mainly, kaolinite, sericite, and iron oxides, with less abundant chlorite, epidote, and carbonates. In addition, the radiometry data were processed to map the localities with the highest possibility of potassic alteration abundance by integrating the potassium distribution, K/eTh ratio, and the F-parameter maps. The surface and subsurface linear structural features were also mapped using Digital Elevation Model (DEM) and aeromagnetic data, respectively. The surface linear structures were found exhibiting E-W and NE-SW trends, while, the subsurface structures showed dominant NW–SE trend. All the depicted fault trends match well with the local and regional geological and tectonic setting of the study area suggesting structural control on the mineralization in this area. Integration between the results obtained from both the remote sensing and the geophysical data was conducted by a GIS weighted overlay model. The obtained mineralization potentiality map highlights eight potential localities for mineralization. The accuracy of the adopted methodology was demonstrated through fieldwork and spectral analyses; several alteration indicators were observed, including quartz veins, iron oxides, kaolinite, malachite, montmorillonite, chlorite, talc, and sericite alteration indicator minerals. The adopted remote sensing-geophysical approach showed being very effective for mapping the hydrothermal gold-related alteration zones, and is recommended for other similar investigations.

Introduction

Hydrothermal alteration zones are highly important and have always attracted the attention of exploration geologists due to their high potential of containing economic minerals and ore deposits that are critical for various industries^1,2. They are often related to causative major geologic structures that act as the passage for the hydrothermal solutions which are the main player in the origination of the hydrothermal alteration zones^3,4,5. Consequently, the geologic structures are very important to be considered while investigating the hydrothermal alteration zones. The Eastern Desert of Egypt is known for containing a lot of mineral resources; alot of which are related to hydrothermal alteration. The hydrothermal-related alteration zones and the associated ore deposits and their related structural controls are of the most interesting remote sensing applications adopted especially for regional mineral exploration; this not only include the satellite imagery data but also, the remotely sensed geophysical aero-magnetic and aero-spectrometric data as well^{6,7,8,9,10,11,12,13,14,15,16}.

Abu Marawat area in the Central Eastern Desert (CED) of Egypt is considered an example of the hydrothermal alteration-related gold deposits. Abu Marawat is a Bedouin name which means the area with a lot of quartz veins; these quartz veins are the main host of the gold mineralization in Abu Marawat area. This area is located south of the Safaga-Qena highway between latitudes 26° 27′ 30′′ to 26° 33′ 30′′ N and longitudes 33° 35′ 30′′ to 33° 44′ 00′′ E (Fig. 1). It gained the focus of several authors for exploring potential mineralization occurrences being part of the well-marked golden triangle in the CED^{3,17,18,19,20,21}, with only a few studies addressed this area from remote sensing perspective^17,18,19. Of the most prominent studies conducted on the study area is Zoheir and Akawy³. This study investigated the nature and genesis of the gold-bearing quartz veins at Abu Marawat area and their relation to the shear zones and banded iron ores present at this area in light of the geologic and structural controls. It also provided supportive ore mineralogy and fluid inclusion investigations that helped identifying and distinguishing the auriferous quartz veins at Abu Marawat area and assigning them a hydrothermal origin.

Figure 1

Location and the main landmarks of Abu Marawat area of study. Created by ArcGIS Desktop v 10.7.1. https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview.

The previous remote sensing studies on the considered area have been mainly forwarded to mapping the hydrothermal alteration zones and discriminating the lithological units. Amer, et al.¹⁷, maintained analyzing the spectral signatures of the alteration minerals from the JPL and USGS spectral libraries to map the hydrothermal alteration zones utilizing the SID and SAM techniques on ASTER satellite data. Asran, et al.¹⁸, utilized ETM+ data to discriminate the exposed rock units at Abu Marawat area through adopting the band ratio technique which enabled them to differentiate among various basement lithological units. Hassan and Ramadan¹⁹ utilized band ratio technique on Landsat-8 data to discriminate the various exposed lithological units, while adopted the Minimum Noise Fraction (MNF) and Principal Component Analysis (PCA) applied to ASTER data to delineate the potentially mineralized alteration zones. Their study could spot two different types of alteration zones based on the remote sensing investigations and geochemical analyses.

In accordance, Abu Marawat area lakes detailed remote sensing investigation utilizing recent remote sensing techniques instead of the conventional techniques adopted by previous studies. Also, the study area lacks detailed subsurface structural investigations as there is no sufficient subsurface data available for the targeted area. Consequently, the current study comes with main objectives of mapping the potentially mineralized alteration zones localities at Abu Marawat area and delineating the effective subsurface structures controlling the mineralization at this area using integrated remote sensing-geophysics approach supported and validated by fieldwork and spectral analyses. This study integrates multiple mapping techniques applied to multi-source data to ensure maximum accuracy of the results^6,7 instead of relying on single dataset or technique. Moreover, the present study aims at presenting a tested and validated procedure for integrating remote sensing, GIS, gamma-ray spectrometry, magnetic, and field data to delineate potential mineralization occurrences within this area, which could be extended to be used for other areas with similar geological environments.

Geological setting

Abu Marawat area is located in the Central Eastern Desert (CED) of Egypt in a wrench tectonic province^5,22,23, it suffers high degree of deformation as it is located between Duwi and Atalla mega shear zones that bound the area from the east and west directions, respectively²⁴. The wrench and shear zones tectonic features that are dominating the CED comprising the study area, are characterized by strong NW–SE structural trend expressed in dissected by ENE deep-seated faults^25,26.

Abu Marawat area forms a large belt of metavolcanic rocks and their related pyroclastics and tuffs. The metavolcanics vary compositionally from acidic through intermediate to basic volcanics including; rhyolite, dacite, andesite and basaltic andesite. These rocks are often thrusted by slices of serpentinites and their derivatives of talc carbonates and listwaenites east of the Abu Marawat gold mine²⁷ (Fig. 2). The metavolcanics are intruded by monzogranites in the eastern and northwestern parts of the investigated area. In the southwestern part, molasse-type Hammamat sediments of siltstone and mudstone composition are intercalated with the highly deformed metavolcanic rocks³. At the high peaks of the metavolcanic rocks to the east of Abu Marawat gold mine and west of Gabal Wa’era, the Banded Iron Formation (BIF) is abundant as alternating bands with the metavolcanic rocks²¹.

Figure 2

Geological map of Abu Marawat area, modified after²⁷. Created by ArcGIS Desktop v 10.7.1. https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview.

Materials and methods

To achieve the objective of this study, multiple datasets were considered, including (1) Level-1T ASTER image (VNIR and SWIR bands), acquired on October 7th, 2007, obtained from the NASA Earth Data Collection website, (2) ALOS PALSAR DEM, (3) airborne spectrometry and magnetic data; obtained as paper maps from the Egyptian Mineral Resources Authority (EMRA), (4) geological map obtained after the Egyptian Geological Survey and Mining Authority (EGSMA), and (5) field data (field observations, samples, and photographs).The proposed methodology is illustrated graphically as a flow chart (Fig. 3). The study used three main datasets including; remote sensing data (ASTER image and ALOS PALSAR DEM), airborne geophysical data (spectrometry and magnetic data), and field data (field observations and mineral/rock samples). The remote sensing techniques utilized in this study rely mainly on mathematical calculations and user-aided image classification procedures applied to ASTER satellite data. These techniques include Band Ratios (BRs), Relative Band Depth (RBD), Mineralogical Indices (MI), Spectral Angle Mapper (SAM) and Constrained Energy Minimization (CEM). Moreover, the potassium distribution map, the K/eTh map, and the F-parameter technique of the air-borne gamma-ray spectrometry were used to delineate the potassium- enriched localities as indication of potassic alteration zone^8,15,16,28. Automated extraction of linear structures²⁹ was implemented to map the surface weaknesses that are often regarded as potential pathways for mineralized hydrothermal fluids. In addition, the derivative maps of the aeromagnetic data were utilized to delineate the subsurface structures and were affirmed by the 3d Euler Deconvolution (ED) technique³⁰ to understand their relation to the origin of the alteration zones in the study area. The outcomes from the three main datasets were integrated together through weighted overlay analysis in a GIS model.

Figure 3

Methodological framework of the current research study.

Alteration minerals mapping using remote sensing data

Several remote sensing procedures (preprocessing and processing) were carried out on the ASTER imagery for mapping the mineralogical alterations in the targeted area.

Image preprocessing

• a. Cross-talk correction against signal scattering issues in the SWIR sensor.

• b. Atmospheric interference was eliminating using Dark Object Subtraction (DOS).

• c. The nine VNIR and SWIR 30m bands were stacked in one dataset.

• d. The study area was extracted (subset) from the whole ASTER scene.

Image processing

For alteration mapping; BR, RBD, MI, SAM and CEM techniques have been used with the integration of the extracted surface and subsurface linear structures. These different techniques were adopted mainly to take advantage of their different algorithms. By utilizing several mapping algorithms and integrating them, more reliable results can be obtained⁶. These techniques were used to map different alteration index minerals including; kaolinite, sericite, chlorite, alunite, calcite, and iron oxides.

BR, RBD, and MI techniques

The BR, RBD, and MI techniques are quite similar in that they depend on mathematical calculations applied to the spectral bands to isolate particular mineral information³¹. The BR technique is a direct application of the simple rationing process in which, the band corresponding to the maximum reflectance is divided by that of the maximum absorbance (minimum reflectance).

RBD technique is not as simple as the BR technique; where, the numerator in the RBD is the sum of the shoulder bands of the absorption feature and the denominator is the band at or nearest to the minimum of the absorption feature^31,32. MI technique is more complex than the BR and RBD techniques. Several authors have developed complex mathematical formulations to delineate particular mineral types. Among the most widely used mineralogical spectral indices are those provided by Ninomiya³³, which been widely used^32,34 and proven highly effective in detecting hydrothermally altered localities.

The BRs, RBDs, MIs of the indicator alteration minerals (Table 1) were applied to ASTER bands to extract and delineate their spatial abundance as indication for the alteration zones occurrence.

Table 1 ASTER ratios and RBD's for mineral features (after Refs.^33,35,36).

Supervised classification techniques (SAM and CEM)

SAM and CEM techniques were applied to the ASTER surface reflectance bands using the spectra of selected alteration minerals (alunite, calcite, chlorite, kaolinite, sericite, and iron oxides (hematite)) from the USGS minerals spectral library (Fig. 4.).

Figure 4

USGS spectral curves of the alteration index minerals mapped by SAM and CEM techniques, resampled to match the ASTER spectral bands. Created by ENVI software v. 5.3. https:// www.l3harrisgeospatial.com/software-technology/ENVI.

SAM compares the signatures of the provided end-members with the image pixels spectra. Pixels within the provided angle threshold were delineated as anomalous localities for each end-member. This technique yielded greyscale maps for the abundance of these index minerals illustrating the anomalous localities as the brightest pixels.

Similarly, CEM technique is another controlled classification procedure that maximizes the response of the desired target and suppresses the other objects considering them as unidentified background. Like SAM, CEM technique results in greyscale abundance maps illustrating the anomalous localities as bright pixels.

Combining the remote sensing mapping outputs

The outputs of alteration mapping using the various adopted remote sensing techniques were combined through a weighted overlay GIS model with equal weights assigned to all the implemented layers. The combined data included; the alteration maps obtained by BR, RBD, MI, SAM and CEM techniques with 20% weight/each.

Acquisition and preprocessing of the airborne geophysical data

The airborne geophysical gamma ray spectrometry and magnetic data used in the present study are part of the airborne survey conducted by Aero-Service Division in 1983 within a mission covered a large portion of the Eastern Desert of Egypt aiming at assisting in mineral, petroleum and groundwater exploration. The survey lines were flown in a NE-SW direction at 1.0 km spacing, the tie lines were spaced 10 km in a perpendicular NW–SE direction, and the ground terrain clearance was 120 m above the ground level³⁷.

A Varian (V-85) proton precession magnetometer sensor and a high-sensitivity 256-channel gamma-ray spectrometer (model Hisens AGRS 3000F system) were used for data collection³⁷. Aero-Service Company performed the airborne gamma-ray spectrometric data processing according to the recommendations of IAEA³⁸.

The available data were corrected from the Compton Effect, flight effective height and background radiation removal from the aircraft, atmospheric radon gas and cosmic rays³⁷.

Alteration mapping using the radiometry data

Digitizing the radiometry data

Three paper maps (hard copies) of the radiometry data covering the targeted area corresponding to the K (%), eTh (ppm), and eU (ppm) radioelements were obtained from EMRA. These sheets were scanned, georeferenced (datum/coordinate system: WGS 1984/UTM zone 36 N), digitized and codded using ArcMap software. As a result, codded vector contour maps were obtained for the three radioelements. These line vector maps were then converted to point data using the “feature-to-point” tool.

Gridded maps were generated from these point data using the “Geosoft Oasis Montaj software v. 8.4, https://www.seequent.com/products-solutions/geosoft-oasis-montaj”, which was used throughout the whole processing procedure and exporting the layouts of the radiometry data.

Potassic alteration delineation

Several procedures have been applied to the digital radiometry data including; (1) radioelement concentration distribution maps, (2) K/eTh ratio map showing areas of potassium enrichment at the expense of thorium; this is a widely accepted technique for mapping the potassic alteration zone^8,15, and (3) the F-parameter of Efimov,³⁹ that provides a good indication of the areas with potential alteration abundance, it has been successful for mapping the potassic alterations associated with gold and massive sulfides-related mineralization^{8,40,41,42,43}. F-parameter can be calculated by applying one of the different equivalent formulas [K*(U/eTh) or K/(eTh/U) or U/(eTh/K)].

Combining the outputs of the spectrometry data

The outputs of the spectrometry data were combined producing one final output highlighting the distribution and abundance of potassium enriched localities (potassic alteration) across the investigated area. The combined radiometry data included; the K (%) abundance, K/eTh ratio, and the F-parameter layers/maps, which were integrated together through a WOM with equal weights for each map/layer (~ 33%).

Mapping the surface and subsurface linear structures

Surface lineaments extraction using DEM data

In deserts, a lineament can represent fault, crack, or other weakness spots^29,44. Lineaments may be related to zones of highly fractured rocks, and thus can be regarded as remarkable ascending passes for the hydrothermal fluids⁴⁴.

Satellite data have been widely used as a vital source for depicting the lineaments. For example, Mbianya et al.⁴ extracted the lineaments automatically from satellite data and integrated them with alteration data to map the mineralized districts in the Ketté goldfield area in Eastern Cameroon. Wahbi et al.⁴⁵ utilized satellite data for extracting the linear features automatically using spatial analysis techniques.

In the present study, a hill shade or shaded relief image of the study area created from the ALOS PALSAR 12.5 m DEM was used for lineament extraction. Extracting lineaments from shaded relief images depends on the idea that when a light source illuminates landscapes with different azimuth directions and multiple elevations, it creates shadows that depend on the azimuth and elevation of each of these illuminated landscapes. This process makes the lineaments more traceable⁴⁶. The delineation or depiction of any lineament is enhanced by creating these shadowed images that clearly illustrate the topographic variations. The automatically extracted lineaments in the present study were inspected manually to remove the non-geologic lineaments.

Subsurface structural mapping using the magnetic data

The Geophysical magnetic surveys (airborne and ground based surveys) have been widely used to delineate shallow and deep subsurface structural features, especially for mineral exploration purposes^47,48. In this study, airborne total magnetic intensity (TMI) data of the study area were processed to extract the subsurface structural linear features affecting the alteration zones in the study area.

Reduction to the magnetic pole

The TMI data that were acquired during an airborne survey in 1983 were reduced to the magnetic pole (RTP), to ensure all the magnetic anomalies are correctly overlain on their sources. The RTP map was generated through applying the RTP filter to the TMI data by the means of a Fast Fourier Transformation (FFT) mathematical operation as the following:

$${F}_{r}\left(U, V\right)=\frac{{F}_{v}\left(U, V\right)}{(sin I+{i\mathit{cos}I\mathit{cos}(D-arc\text{tan}\left[\frac{U}{V}\right]))}^{2}}(1)$$

where; ${F}_{r}\left(U, V\right)$ is the FFT of the observed or recorded magnetic data from TMI to RTP, ${F}_{v}\left(U, V\right)$ is the FFT of the vertical magnetic field, $\left(U, V\right)$ represents the wavenumber of (x, y) directions, respectively, whereas, I and D represent the inclination and declination parameters of the magnetic field, respectively.

Inclination and declination values of 32.8 North and 1.9 East respectively, were used to convert the TMI data to RTP. These values were taken after the surveying project parameters that were recorded during the data acquisition in 1983.

Vertical derivatives

The first vertical derivative (FVD) filter is widely used for mapping subsurface structural features^48,49. FVD filter functions to equalize the response of small and large amplitudes allowing more enhancement of representing the linear structural features. It removes the long wave forms originating from causative magnetic bodies, thus enhancing the response of local anomalies and attenuating regional anomalies. In the present study, FVD was applied to the RTP data to highlight the local subsurface structural features.

Subsurface lineament extraction

The enhanced linear structures on the FVD map were ultimately input into the workflow of “center for exploration target (CET-grid analysis)” in the “Geosoft Oasis Montaj software v. 8.4, https://www.seequent.com/products-solutions/geosoft-oasis-montaj”,. This workflow comprises a set of consecutively applied tools including; (1) texture enhancement; which maximizes the response of the edges using the standard deviation of the data, (2) phase symmetry, which acts to detect the linear discontinuities in the data, (3) skeletonization, which means line thinning and it allows to eliminate any features which is less than a user provided threshold, and (4) vectorization, which is the final step of lineaments extraction and the output is a vector map of the depicted subsurface linear structures.

Euler deconvolution (ED)

The extracted structural features were further confirmed and attributed to their causative sources and their corresponding relative depths using the widely acknowledged 3D Euler Deconvolution^30,50,51,52. The ED method uses x, y and z derivative to determine location and depth for various idealized targets (spheres, cylinders, thin dykes, pipes and contacts), each characterized by a specific Structural Index (SI) assigned to each of these targets³⁰. The SI is a measure of the fall-off rate of the filed with distance from the source. It plays an important role in identifying the sources based on their geometry; SI values normally vary from 0 to 3. Possible source bodies and their most suitable structural indices were given by³⁰ (Table 2). The structural index (SI=0) was used in the current investigation to delineate the magnetic sources related to geological faults and contacts as they are most likely to contribute in hydrothermal solutions migrations and thus could be regarded as a controlling factor for mineralization occurrence.

Table 2 Euler Deconvolution Structural Indices for various geological models (after Reid, et al.,³⁰).

Integrating remote sensing and radiometry data

The obtained data of the alteration minerals abundance/potentiality obtained from the remote sensing and the radiometry data were integrated with the lineaments density of the investigated area through GIS Weighted Overlay Modeling (WOM).

Two scenarios were applied; the first scenario was through integrating remote sensing alteration outputs (RS) with the gamma-ray spectrometry (GS) alteration outputs only without implementing the structural element (lineaments density), while the second scenario, maintained the implementation of the lineaments density maps along with the remote sensing and spectrometry alteration outputs.

In the first scenario; the RS and GS maps were integrated together through a WOM with assigning equal weights (50%) to these maps. As a result, a final weighted RS-GS alteration map was obtained.

In the second scenario; the surface and subsurface structural factors were implemented, as the RS and GS maps were integrated along with the lineaments’ density maps through a WOM with assigning equal weights of (35%) to each of the RS and GS alteration maps, and 15% weight for each of the surface and subsurface lineaments layers. Consequently, a final mineralization potentiality/prospectivity map was obtained suggesting the most potential localities for the occurrence of hydrothermal alteration-related mineralization (mainly gold).

Validation of the results by fieldwork and spectral analyses

The final output from integrating the remote sensing, radiometry, and magnetic data was validated through a field study, during which several alteration indicators were observed in the mapped highly anomalous localities. Samples were collected form the suggested highly potential mineralization localities. Spectral signatures were measured at the spectral signatures lab of the National Authority for Remote Sensing and Space Sciences (NARSS). The spectral signatures of the collected samples were measured for the Visible (VIS), Near Infra-Red (NIR), and Short Wave Infra-Red (SWIR) regions of the electromagnetic spectrum in wavelength range from 350 to 2500 nanometer (nm) with interval of 1 nm using handheld spectroradiometer named “ASD TerraSpec Halo mineral identifier”. This device allows collecting spectral signatures for a small footprint of the target sample via direct contact and a simple pull of trigger. Several measurements have been recorded for each sample; typically 3 to 5 measurements depending on the degree of homogeneity and to ensure maximum representation for all the different parts of the sample. Spectral analyses were carried out through comparing the measured spectra of the collected samples against the published minerals spectral by the United States Geological Survey (USGS) which is embedded in ENVI image processing software.

Results

Remote sensing mapping results

BR, RBD, and MI

The results of BR, RBD, and MI techniques were obtained as greyscale image for each of the targeted alteration indicator minerals (kaolinite, chlorite, sericite, alunite, calcite/carbonates, and iron oxides) in which the target is shown in the brightest pixels (white color). Density slicing was applied to delineate these bright pixels representing the optimum abundance localities of the targeted minerals. For ease of presentation, the delineated abundances of all the mapped key alteration minerals for each technique were independently combined (represented by red color on all the maps) as indication of the alteration zones distribution.

Alteration zones (cumulative abundance of the mapped key indicator minerals) obtained using BR technique (Fig. 5a) are found mapped mainly in the metavolcanic and monzogranite rock units. They are found associated with the localities of Abu Marawat gold mine, and east of the hammamat rocks belt in the southwestern part of the studied area. Moreover, the mapped alteration zones within monzogranite rocks are found corresponding to the localities of Gabal Wa’era area and also to the monzogranites in the northwestern corner of the study area. RBD map (Fig. 5b) suggests the existence of alteration zones mainly in sporadic localities within the metavolcanic rocks including the locality of Abu Marawat gold mine. Unlike the BR technique, RBD did not map alteration zones within the localities dominated by monzogranite rocks. Like in the RBD technique, the MI technique suggested that the alteration zones are linked to the metavolcanic rocks only. MI map (Fig. 5c) shows that the alteration zones are found in two specific localities; at Abu Marawat gold mining locality and in the southwestern parts of the investigated area.

Figure 5

Cumulative abundances of the mapped indicator alteration minerals by; (a) BR, (b) RBD, and (c) MI methods. Map created by ArcGIS Desktop v 10.7.1. https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview. Image processing performed by ENVI software v. 5.3. https:// www.l3harrisgeospatial.com/software-technology/ENVI.

CEM and SAM

Similar to the BR, RBD, and MI alteration maps, the alteration distribution maps obtained by both CEM and SAM techniques represent the cumulative spatial abundance of all the six key indicator alteration minerals altogether. CEM alteration map (Fig. 6a) shows that the mineralogical alterations are abundant mainly in the metavolcanic rocks especially, at Abu Marawat gold mine and at the southwestern localities of the investigated area. Other alteration zones are mapped within monzogranite rocks especially at the localities of Gabal Wa’era and in the northwestern corner of the considered area. SAM map (Fig. 6b) shows that the mineral alterations are linked to the spatial distribution of the metavolcanic rocks, specifically at Abu Marawat gold mine locality southward to the southern limit of the research area.

Figure 6

Cumulative abundances of the mapped indicator alteration minerals by; (a) CEM, and (b) SAM methods. Maps created by ArcGIS Desktop v 10.7.1. https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview. Image processing performed by ENVI software v. 5.3. https:// www.l3harrisgeospatial.com/software-technology/ENVI.

Gamma-ray spectrometry results

Radioelement concentration distribution maps

The radioelement abundance maps of K (%), eU (ppm) and eTh (ppm) (Fig. 7a–c) show the levels of radiation and the distribution of concentrations for each of the three radio-variables across the investigated area. The ranges of radiation concentration values for potassium, equivalent uranium and equivalent thorium were found 0.16–2.54 (%), 0.01–3.31 (ppm) and 0.01–7.26 (ppm), respectively. The most important radioelement for alteration zones mapping is potassium (Fig. 7a). The elevated potassium values (orange, red, and pink colors) represent high enrichment in potassium and infer potential potassic zones of alteration⁵³. The potassium enriched zones are found spatially correlated with the localities of the granitic and acidic volcanic/metavolcanic rocks. This may be logical because the minerals that formed these rocks are potassium-rich, especially the feldspar-group minerals. The potassium-enrichment within the investigated area is considered as anomaly; as it is compared to the published average value of 1.5% K over the whole United States^54,55, it is found that the acidic rock units (granitoids and acidic volcanic metavolcanic rocks) in the study area are having potassium % values up to 2.54% K which exceed the published average. These anomalous values support the high potential for alteration zones abundance.

Figure 7

Radio-elements concentrations over the investigated area; (a) K (%), (b) eU (ppm), and (c) eTh (ppm). (All values are multiplied by 10). Aeromagnetic data are processed and maps are produced by Geosoft Oasis Montaj software v. 8.4. https://www.seequent.com/products-solutions/geosoft-oasis-montaj.

K/ eTh ratio map

The potassic alteration zone can be delineated by the potassium to thorium ratio (K/eTh) (Fig. 8). The potassium-to-thorium ratio is widely accepted for delineating the potentiality of hydrothermal alteration zones because thorium is relatively stable, unlike potassium that is highly mobile and thus can reflect the effect of hydrothermal activity when proportioned to the relatively stable thorium. Consequently, the high values promote the corresponding localities to be the most favorable for potassic alteration abundance. In the present study, high values are found corresponding to the northeastern, central and southwestern parts of the studied area (Fig. 8).

Figure 8

K/eTh ratio map highlighting the potassium-enriched localities with high values. Aeromagnetic data are processed and maps are produced by Geosoft Oasis Montaj software v. 8.4. https://www.seequent.com/products-solutions/geosoft-oasis-montaj.

The F-parameter

The F-parameter technique was applied to detect the localities with elevated potassium concentrations. Typically, for unaltered rocks, F- value is often lower than 1.3, but, for altered rocks, it can reach 5 and may exceed the value of 10⁴⁰.

F-parameter technique can be carried out via three formulas {K*(U/Th) or K/(Th/U) or U/(Th/K)]. Actually, these formulas are just the same, they are mathematical equivalents. F-parameter maximizes the response of potassium and equivalent uranium, while suppresses the equivalent thorium to delineate the localities with relatively higher concentrations of K and eU corresponding to low thorium values. Maximizing the K and eU responses is highly important because thorium is relatively stable and doesn’t reflect the hydrothermal activity.

In Abu Marawat area there is an obvious contrast in F-parameter anomalous values (Fig. 9) that we can delineate between the localities with high f-values. They were found corresponding to parts of the metavolcanic rocks, hammamat sediments, and granitic rocks (orange to pink zones with F- values more > 2 to 10.2). High F-parameter anomaly values (typically above 2) can be linked to high deformation and thus, high alteration potentiality.

Figure 9

F-parameter map of the study area. Aeromagnetic data are processed and maps are produced by Geosoft Oasis Montaj software v. 8.4. https://www.seequent.com/products-solutions/geosoft-oasis-montaj.

Surface and subsurface linear structures

Surface lineaments

The extracted lineaments were subsequently transformed into a lineaments density map (Fig. 10) with five major classes of lineaments density all over the whole study area. Higher lineaments densities are found corresponding to the localities of Gabal Wa’era, and Gabal Abu Marawat. Lower lineaments densities are found corresponding mainly to the western part of the studied area. The detected lineaments show two major trends in the E-W and NE-SW directions (Fig. 10) suggesting relation to the deep-seated faults²⁶. Other minor fault trends are in the NW–SE direction which is consistent to the major Red Sea trend²⁵.

Figure 10

The upper map shows the automatically extracted lineaments, while, the lower map shows the density of lineaments over the investigated area. Created using and ArcGIS Desktop software v 10.7.1. https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview.

Subsurface linear structures

TMI and RTP maps

The acquired TMI data (Fig. 11a) was converted to RTP (Fig. 11b), so that the magnetic anomalies are truly positioned over their sources. By comparing both the TMI and RTP maps, it is obvious that the anomalies in the RTP map (Fig. 11b) are shifted mostly towards north; this is attained after the correction of the inclination and declination of the magnetic field by 32.8 North and 1.9 East respectively.

Figure 11

(a) TMI map of the study area, (b) RTP map of the study area, (c) the subsurface linear structures overlain on the FVD map of the study area, (d) the subsurface linear structures (upper) and the map of relative proximity to these structures (lower) as obtained from the FVD map with the Euler solutions of SI = 0 overlain, and (e) Rose diagram for the delineated subsurface structural features. Maps are created by Geosoft Oasis Montaj software v. 8.4. (https://www.seequent.com/products-solutions/geosoft-oasis-montaj) and ArcGIS Desktop v 10.7.1 software (https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview).

The RTP map of the study area shows that the magnetic susceptibility values range from 42,185.6 to 42,730.6 nT, while the magnetic intensity datum is 42,425. This means that upon applying the International Geomagnetic Field Reference (IGRF-42425) correction, the corrected RTP magnetic intensity values are found ranging from − 225.8 to 236.5 nT. The positive magnetic anomalies in the RTP map (pink color) are found mostly related to the rocks enriched in iron and mafic and ultramafic minerals; serpentinites, talc carbonates, meta-basalts, and parts of the meta-andesite rocks. Strong negative magnetic anomalies (blue and green colors) are found related mainly to the localities where the sedimentary hammamat rocks, parts of the acidic meta-volcanic rocks (meta-rhyolites and the intensely altered acidic volcanic rocks), and alluvial wadi deposits are dominating. Moderate and weak negative anomalies (yellow and red colors) are found associated with parts of the acidic meta-volcanic rocks. Abu Marawat gold mining area is found coincident within strong negative anomaly being dominated by intensely altered acidic volcanic/metavolcanic rocks.

Subsurface lineaments from FVD and 3D Euler

The FVD map (Fig. 11c) was obtained by applying the first vertical derivative filter on the RTP map of the study area. Since FVD method is sensitive to the near-surface magnetic anomalies, it enhanced the contrast between the responses of different magnetic intensities allowing for the shallow subsurface structures to stand out sharp. A subsurface lineaments proximity map (Fig. 11d) was produced showing the relative proximity to the lineaments over the study area. It is found that Abu Marawat mining area is located within a locality which falls within the most proximal zone to lineaments (very high proximity). This supports the firm relationship between the alteration zones and the corresponding effective subsurface structures.

The Euler 3D technique with structural index (SI = 0) enabled to obtain the solutions corresponding to the fault or contact type magnetic anomalies³⁰ at the studied area. The ED result is overlain on the subsurface structures obtained using the FVD technique (Fig. 11c); it shows that several subsurface anomaly sources were detected at various depths reaching maximum depth of about 541 m; these anomalies are found highly consistent with the delineated structures by the FVD method with similar main NW–SE trend (Fig. 11e) suggesting authentic relation to the major Red Sea fault trend²⁵. At Abu Marawat gold mine area, the Euler solutions could delineate relatively shallow sources at depths from 100 to 200 m (green circles).

Integration of the data through GIS modeling

Combining the remote sensing (RS) mapping results

The alteration zones abundance maps obtained from the adopted five RS techniques (Fig. 5 a, b & c, and Fig. 6 a & b) were combined together using a weighted overlay analysis model with equal weights (20%) assigned to all the implemented maps. The resulting map (Fig. 12a) shows that the alteration zones are most favorably abundant in the localities of the metavolcanic rocks mainly in the western half of the investigated area including the locality of Abu Marawat gold mine.

Figure 12

(a) Alteration zones abundance map obtained from combining the BR, RBD, MI, CEM, and SAM RS techniques, (b) Potassium enrichment map showing the most favorable localities for the abundance of potassic alteration zone, and (c) Integrated RS-GS alteration abundance map. Created by ArcGIS Desktop v 10.7.1. https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview.

Combining the gamma-ray spectrometry (GS) mapping results

Three radioelement maps were combined together with equal weights (~ 33%) to highlight the potential localities for potassium enrichment. These maps included; the K (%) abundance map (Fig. 7a), the K/eTh ratio map (Fig. 8), and the F-Parameter map (Fig. 9).

This integration yielded a potassium enrichment map (Fig. 12b) indicating the possible existence of the potassic alteration zones, the potassium-enriched localities are found related to multiple sporadic localities across the investigated area including the locality of Abu Marawat gold mine.

Integrating the RS and GS alteration maps (the first scenario)

The combined RS alteration map (Fig. 12a) was integrated with the combined GS alteration map (Fig. 12b) through a GIS WOM with assigning equal weights (50%). This integration combined both the RS and GS data and resulted in a more reliable alteration abundance map with five alteration potentiality classes (Fig. 12c).

As per the used criteria, the highly potential localities for alteration are found to be spatially related to the geographic locations of Abu Marawat mine westward comprising almost all the western part of the research area, and southward reaching the locality east of the Hammamt metasediments belt. While, the locality of Gabal Wa’era was the only mapped locality within the eastern part of the study area.

Integrating the alteration and lineaments density maps (the second scenario)

The results for alteration abundance distribution obtained from both the remote sensing and radiometry data (Fig. 12a and b) in addition to the surface and subsurface lineaments densities data (Figs. 10 and 11d) obtained from the ALOS PALSAR DEM and aeromagnetic data, were integrated together using a Weighted Overlay Model (WOM). Equal weights were assigned to the surface and subsurface lineaments layers (15% each), and equal weights were assigned to the RS and GS alteration potentiality layers (35% each). This integration resulted in a final hydrothermal mineralization prospectivity map (Fig. 13). Eight highly potential localities for mineralization are delineated including the locality of Abu Marawat gold mine. It is found that the eight suggested localities are structurally controlled and in terms of the hosting rock units as well; all the localities are related to areas of high lineaments density. Moreover, four localities are found corresponding to the localities where the highly altered acidic volcanic/metavolcanic rocks are present, the other four localities are found spatially localized around Gabal Wa’era granitic intrusion suggesting to be formed during the intrusion event as a result of the accompanying hydrothermal solutions. The high robustness of the result is confirmed by the already known mineralization localities in the study area, as, Abu Marawat gold mine was mapped within the most potential mineralization classes (very high and high) in the resulting map.

Figure 13

Mineralization potentiality/prospectivity map with the highly mineralization potential localities (rectangles) and the field observation stations and sampling locations (black circles) plotted. Map created by ArcGIS Desktop v 10.7.1. https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview.

Fieldwork and spectral analyses for validation

A field survey was carried out to validate the results and ensure they are matching the ground alteration evidences observed in the field. The sampling locations are overlain on the final mineralization potentiality map (Fig. 13). The sampling strategy was mainly based on the available accessibility options to the study area. The fieldwork and sampling was conducted by traversing the study area trying to reach all the 8 mapped localities through the existing several Wadis that dissect the rock units of the study area. It could be noticed from the distribution of the sampling localities that the survey was able to reach out to four of the suggested localities while failed to access the other four localities, so we tried to approach these localities as much as possible. During the survey, different field observations regarding the presence of alteration zones were evidenced. These evidences include; quartz veins/veinlets, iron oxides and malachite staining, carbonatization, kaolinitization and chloritaization evidences (Fig. 14a–f).

Figure 14

(a) Highly sheared metavolcanic rocks near Abu Marawat gold mine; (b) Quartz vein in the metavolcanic rocks around Gabal Wa’era; (c) High staining of Hematite in the metavolcanic rocks south of Abu Marawat gold mine; (d) Malachite (M) observed north of Abu Marawat gold mine; (e) Serpentinite (Sp) altered to talc carbonates (TC); (f) Kaolinite and iron oxides alterations in the metavolcanic rocks at the southern end of Wadi Abu Marawat within the study area.

In addition, 50 spectral signatures were measured for the collected samples. The spectral analyses revealed alteration index minerals in 15 samples. Montmorillonite with characteristic absorption features at 2.21 and 2.36 µm was identified in 4 samples; M21, M22, M23, and M32 (Fig. 15a). Chlorite with distinguishing absorption features at 2.26 and 2.34 µm was found in 4 samples; M14, M28, M35, and M36 (Fig. 15b). Muscovite with three prominent absorption features at 2.20, 2.35, and 2.44 µm was identified in 3 samples; M2, M10, and M16 (Fig. 15c). Kaolinite with well-marked absorption features at 2.21 and 2.24 µm was found existing in 2 samples; M42, and M50 (Fig. 15d). Talc (Fig. 15e) with three characteristic absorption features at 2.32, 2.39, and 2.47 µm and Hematite (Fig. 15f) with the wide absorption feature at central wavelength of 0.87 µm, each was observed once; in M37, and M8, respectively.

Figure 15

The spectral profiles of the measured spectra against their corresponding matchs from the USGS minerals spectral library; (a) Montmorillonite, (b) Chlorite, (c) Muscovite, (d) Kaolinite/Smectite, (e) Talc, and (f) Hematite.

For further validation, the location of Abu Marawat gold mine was plotted on the final resulting mineralization potentiality map; there found a complete coincidence of the site of Abu Marawat gold mine over the localities mapped as very high potential for mineralization. Consequently, the other suggested seven localities are of high potential for mineralization abundance.

Discussion

Abu Marawat gold deposit is considered a mesothermal type formed as a result of interaction between iron-rich host rocks (the metavolcanics) and structurally controlled hydrothermal fluids. It is present as a system of massive and sheared milky quartz veins cutting within island arc metavolcanic rocks and the related Banded Iron Formation (BIF) at the peaks of Gabal Abu Marawat³. The structural controls on gold mineralization and alteration zones occurrence maintained in the current study is consistent with the findings of the Zoheir and Akawy³; since the locality of Abu Marawat gold mine has been found coincident on a subsurface fault trending in NW–SE direction at depths of ranges between 100 and 200 m and grading northward to shallower depths of less than 50 m as depicted from the aeromagnetic FVD structures and ED solutions (Fig. 11c). The Major structural controls on gold mineralization can be attributed to NW–SE, E-W, and NE-SW trending faults, cracks, and joints. The NW–SE trends are attributed to subsurface faults delineated using the aeromagnetic data is largely attributed to the parent Red Sea major fault. On the other hand the E-W and the NE-SW trends are automatically from the ALOS PALSAR DEM data. Such structural trends are linked to variably-sized quartz veins and other shear-related fractures, cracks and joints^3,19. These automatically obtained surface lineaments mainly express the linear structural features that are attributed to geological faults, fractures, or joints. That’s because these lineaments were checked manually to ensure removal of the non-geologic artificial or manmade linear features such as road and buildings footprints. It is deduced that the vast majority of the delineated lineaments are due to geologic structures because neither asphaltic/highway roads nor urban communities or other buildings are present within the investigated study area. Only a few lineaments were found attributed to some desert tracks/routs and were removed manually.

In the current study, the aim of utilizing different datasets, and algorithms/techniques in the current study was to ensure maximum integrity of the results, especially when multiple datasets/techniques yield similar results. Each of the utilized five remote sensing techniques enhanced certain types of the alteration indicator minerals while were not able to delineate other alteration minerals within the studied area. BR and RBD techniques successfully mapped significant abundance of kaolinite, sericite, and iron oxides; Abu Marawat mining area was found dominated by kaolinite and iron oxides with less abundant sericite. The MI technique besides being able to map the abundance of the kaolinite, sericite, and iron oxides, it was also able to delineate the existence of alunite and carbonate minerals at the Abu Marawat mine locality and in other sporadic localities within the investigated area as well. SAM and CEM techniques highlighted the abundance of kaolinite, sericite, and iron oxides at Abu Marawat mine and at the other delineated localities as well confirming the results obtained by the other mathematical-based techniques. Cumulative abundance of the mapped alteration indicator minerals was considered to represent the spatial abundance of the potential alteration zones. The BR result (Fig. 5a) is highly comparable to that of the CEM technique (Fig. 6a). Both of these techniques mapped alteration zones within the monzogranite rocks in addition to the metavolcanic rocks mainly at Abu Marawat mine southward reaching the southern limit of the considered area. The RBD (Fig. 5b), MI (Fig. 5c), and SAM (Fig. 6b) results are highly comparable to each other. The alteration zones mapped by these techniques are found –unlike BR and CEM- confined to the metavolcanic rocks with minor or no alteration zones recorded in the monzogranite rocks. All in all, the results of the five adopted remote sensing techniques (Figs. 5a–c and 6a,b) show that the preferred alteration mineral assemblage is mostly; kaolinite, sericite, iron oxides, and less abundant chlorite, carbonates, and epidote as agreed by all the mapping techniques. All the remote sensing alteration mapping techniques also showed acceptable accuracy verified by mapping the locality of Abu Marawat Gold mine as a potential alteration zone locality.

Similarly, the adopted three spectrometry techniques; K (%), K/eTh, and F-parameter (Figs. 7a, 8, and 9), maintained maximizing the response of potassium as it is one of the best indicators on the alteration abundance. Highly-enriched potassium localities in the three techniques, were found corresponding to the spatial localities of monzogranites, acidic metavolcanics including Abu Marawat mine, and the acidic to intermediate metavolcanics near the hammamat metasediments. These localities show high coincidence with the localities of alteration zones mapped by the different adopted remote sensing techniques, which confirms the results of both the RS and GS techniques.

The GIS Weighted Overlay Model (WOM) utilized in the present study allowed integrating all the results obtained from the adopted multisource data utilizing several techniques. This WOM is very depends on the prior expert knowledge of the relative importance of the different implemented data in the spatial analysis process. It allows overlying multiple layers of different data and conducting correlative spatial analyses aiming at maximizing the accuracy of the final product depending on information from all the input criteria. However, this technique is very sensitive to the weights assigned to each layer as these weights significantly affects the final output, consequently should be used wisely. In this investigation, the remote sensing results were integrated through a GIS WOM with equal weights assigned to all the five techniques. Consequently, one integrated remote sensing alteration zones abundance map (Fig. 12a) was obtained. Similarly, the three adopted gamma-ray spectrometry techniques were integrated yielding one integrated GS alteration map (Fig. 12b). Additional two GIS models were applied; the first was to integrate the RS and GS alteration data, while, the second included the lineaments density maps from DEM and magnetic data in addition to the RS and GS alteration maps.

The RS and GS alteration maps were integrated as a first scenario without implementing the lineaments density data. Equal weights (50%) were assigned to both the layers. Consequently, a reliable alteration abundance map was obtained (Fig. 12c) with ranked five classes of alteration favorability/potentiality. In the second scenario, the lineaments density was implemented as an effective factor in the formation of alteration zones. A cumulative weight of 70% for the two alteration layers (RS and GS) by 35% each, and a total weight of 30% for the surface and subsurface lineaments density layers (15% each) were assigned. These weights reflect the relative importance of the implemented layers; they are in complete agreement with the weights suggested by Ghoneim, et al.,⁹. The result of this integration highlights the most potential localities for mineralization (Fig. 13), it was validated through the coincidence of the old mine in the area within the highest potential mineralization (very high and high potentiality) classes.

By comparing the results of both the two scenarios; it can be inferred the high importance of implementing the structural factor (surface and subsurface linear structures). It was functioned to increase the potentiality of specific localities (with high structural density/complexity) while, decreasing the potentiality of other areas (with low structural density/complexity). As a result, the final map is reshaped and the false positives (areas mapped containing alteration zones but with localities with low structural density) are reduced. Almost all the western parts of the investigated area is mapped in the first scenario (Fig. 12c) as highly potential for alteration zones existence, while in the second scenario (lineaments factor added) it is found that this part is mostly of low and very low structural density and thus its potentiality for mineralization abundance was reduced significantly (Fig. 13). The second scenario is more reliable being supported by the structural factor (surface and subsurface linear structural features) which is crucial for hydrothermal alteration and mineralization deposition. Eight localities – Including Abu Marawat mine—were delineated as the highest potential for mineralization occurrence based on the coincidence of alteration zones and structural (surface and subsurface) density data. The fieldwork and spectral analysis were very helpful for validating the outcomes of the present study. However, these measures are just for preliminary validation and are not reliable for supporting a final decision of exploitation. The delineated localities need more detailed investigations; intense sampling, chemical analyses, ground geophysical investigations, and core analyses are recommended to be conducted prior initiating any exploitation activity.

Conclusions

The main conclusions emerging from the study can be expressed as the following few bullet points:

• The present study provided a validated successful integration of remote sensing ASTER and DEM data together with the airborne geophysical gamma-ray spectrometry and magnetic data through GIS modeling for delineating the mineralized alteration zones. Based on the adopted approach Seven localities other than Abu Marawat mine locality were delineated as highly potential for mineralization occurrence, these localities need more detailed investigations to evaluate their economic feasibility.

• Utilizing multiple remote sensing mapping techniques with different algorithms in an integrated way strengthens the mapping results and gives rise to a more reliable output than that of individual technique.

• The gold mineralization in the studied area is structurally controlled as inferred from the aeromagnetic and the DEM data. Consequently, implementing the structural factor (surface and subsurface structures) is critical for alteration zones mapping and can significantly affect the mapping results. The delineated surface and subsurface structures match the local and regional tectonic setting of the study area and the CED of Egypt. These structures were found also highly correlated with the spatial location of Abu Marawat gold mine which suggests a structural control on the gold mineralization at this area.

• The GIS weighted overlay model has been effective for integrating the RS, GS, and the structural data in the current study and produced a reliable mineralization prospectivity map.

• The success in mapping the locality of the existing Abu Marawat gold mine validates the result and gives a very high confidence level of the other delineated seven localities; it also indicates high efficiency, credibility and reliability of the adopted methodology as well.

• The fieldwork and the spectroscopic analyses proved the existence of various index alteration remarks including; quartz veins, iron oxides and malachite staining, carbonate and clay minerals (kaolinite and montmorillonite), chlorite, talc, and muscovite (sericite).

• Finally, the approached methodology is proved highly effective in delineating the mineralization-loaded alteration zones, thus it is recommended to be followed for similar investigations in other regions of the Arabian-Nubian Shield and around the world.