Integration of Geological Mapping and Remote Sensed Studies for the Discovery of Iron – Ore Mineralization in Mutomo – Ikutha Area, SE Kenya

This research integrates geology with remote sensing techniques to establish characteristic features that can be used to discover iron ore mineralization within the Neoproterozoic rocks of Mutomo – Ikutha area in south eastern Kenya. The association of hornblende gneiss and shearing as well as alteration processes near the mineralized regions appear to play an important role in the distribution and localization of the iron mineralization The methods used in this research include Image processing techniques applied on the digital subset ETM+ data that cover Mutomo – Ikutha area and geological field mapping. These techniques generated several products of enhanced satellite imagery, such as colour composite images, ratio images and principal component images. These techniques have been successfully used in the lithological discrimination of iron ore bearing sheared hornblende gneisses. The capabilities of remote sensing data to characterize the iron ore bearing gneisses, in addition to characterization and mapping the hydrothermal alteration zones helped in identification of iron mineralization regions. Extensive field geologic and geochemical investigations to the pronounced zones delineated by the image processing technique, led to discovery of four locations of high iron anomalies with some iron mineralization, mainly connected to the studied Neoproterozoic hornblende gneisses. Chemical studies were carried out using atomic absorption spectrophotometer (AAS) and X-Ray florescence, for some selected mineralized samples. Petrographic analysis and physical properties of the iron minerals were carried out as well. These investigations confirm the present iron mineral to be magnetite ore enriched with phosphates. The magnetite is found in close association with apatite chalcopyrite, quartz, and chlorite. Relics of corroded magnetite grains are occur along the Tiva river bed as well as along the road sides.


Introduction
The study area is located between longitudes 38º 4´E to 38º 20´E and latitudes 1º 48´S to 2º 8´S in South Kitui within the Kitui County occupying about 100 Km 2 ( Figure. 1). In this paper, we show how we have used remote sensing and field based investigation to come up with a structural and a mineral resource map. New geological structures discovered include shear zones and faults that control iron ore mineralization. Orientation analysis has revealed two shear zones associated with iron ore mineralization. The Tiva shear and Mutito shear trending in the NW-SE direction. Another important discovery is the location of Mineral deposits associated with alteration within the gneisses. Although previous researchers had done geological mapping and produced maps for Ikutha area, the structures, lithology and mineral resources discovered were missing. This research attempts to answers some of the questions for the gaps left by the previous researchers.
Image Processing techniques were applied in this work to define the main characteristic features of the gneisses rock-bearing iron mineralization. The enhanced thematic mapper plus (ETM+) digital data of Landsat 7 satellite was used to recognize and map the investigated area. The constructed colour composite images and application of principal component analyses of ETM+ data led to identify and characterize the iron ore -bearing gneisses in the investigated area.
Integration of Geological Mapping and Remote Sensed Studies for the Discovery of Iron -Ore Mineralization in Mutomo -Ikutha Area, SE Kenya

Geology of the Study Area
The Mozambique Belt is a major N-S trending metarmophic and lithotectonic domain that extends along the Eastern Coast of Africa, as well as in Saudi Arabia, Madagascar, India and Sri Lanka. It was first defined by Holmes [1], mostly on metamorphic and structural criteria. The Mozambique is a Neoproterozoic Belt (900 to 550) which was affected by the Pan African tectonothermal event about 650 Ma, as suggested by (Kennedy [2] ; Kazmin [3]; Kröner [4] ; Cartier [5]) who considered it as a major phase of division of the African continent linked to the development of mobile belts. The Pan African event encompasses a wide time span from 650 to 500 Million years. Mozambique Belt is considered as an orogen resulting from the oblique collision between two Gondwana fragments, [Muhongo [6] Burke and Derwey [7]; Gass [8]; Key et al. [9]; Kröner [4]). Stern [10] proposed its name to be changed to East African Orogen. The fact that the collision was oblique resulted in the development of large shear zones parallel to the direction of the orogeny after the main collision phase, owing to tectonic escape phenomena. The geology, and structures, tectonic the history of the Mozambique Belt in Kenya has been reviewed, by (Mathu and Tole [11] 1984;Key et al. [9]; Mathu [12]; Mathu et al.[13]; Mathu et al.[14]; Gaciri et al.[15]; Biyajima et al.[16]; Suwa et al.[17] ; Opiyo -Aketch and Nyambok, 1984;[18] Hackman et al. [19]; Mosley [20]; Nyamai [21], and [22][23][24][25][26][27].
An attempt has been made here to update the geology of Mutomo -Ikutha area. This is to fill the gap left by the previous geologists who missed out some of the geological details that have been noted during the current survey. The current survey has also managed to establish continuity of rock units from the area mapped by Saggerson [28] to the area mapped by Walsh [29]. The rocks of the Mozambique mobile belt are considered to be the metamorphosed equivalents of originally sedimentary rocks, the sedimentary pattern being retained. The repetition of certain beds, e.g. the meta-calcareous horizons in the area, the presence of originally carbonaceous limestone, and preserved currentbedding and stratiform succession are considered proof of sedimentary origin. Evidence of lateral variation is observed in this area. The pelitic gneisses grade into rocks which do not contain the higher-grade index mineral. An example of lateral variation is seen south of Mutomo area where the biotite gneisses grade imperceptibly into biotite-hornblende gneisses. Another lateral variation is noted in Ikutha area where biotite gneiss grades gradually into biotite hornblende gneiss.
An attempt has been made here to classify rock units of the area based on field relation, lithology and colour into groups and formation. These groups include; Ikutha and Mutomo groups. The formations and members of these groups have been subsequently discussed in the following sections.

Belt of SE Kenya
Mineralization in the study area is mainly controlled by the stratigraphy and structures. Stratigraphy of Mutomo -Ikutha area is largely classified on the basis of the lithology of the rocks (Table 1). According to the regulation of the international Subcomission on Stratigraphic Classification [30] of the international Union of the Geological Sciences (IUGS), the stratigraphy of Mutomo -Ikutha area is considered to be composed of lithostratigraphic rock units. In designing nomenclature of the rocks in the study area, the North American Stratigraphic code prepared by the North American Commission on Stratigraphic Nomenclature, and the rules for naming geological units in Norway [31], have been used. It involves a general description of the unit, origin of its name and where possible to its earlier work, areal distribution, including the best exposed areas and stratotype; the contact relations and thickness estimates; the division of the unit into sub-units and origin of the unit based on the available evidence. The stratigraphic descriptions of the study area starts with Ukamba super group, which is referred to as Ukamba group by Mosley [20].

Lithostratigraphy of Mutomo -Ikutha Area
The rocks of the Mozambique mobile belt in this area are classified lithostratigraphically as above.

Image Processing of Remotely Sensed Data
The remotely sensed raw data of the studied areas are included in Landsat-7 Enhanced Thematic Mapper Plus (ETM+) data scenes number 167/61 (Path/Row) covers Mutomo -Ikutha area. Subsets of the digital ETM+ imagery covering the studied areas were obtained. The image processing techniques were applied in this work using ENVI 4.8, ArcGIS 10 and PCI, GeoAnalyst software-package. Geometric correction has been done for the digital data of the ETM+ bands of the study areas. Raw digital satellite data usually includes geometric distortions due to sensor geometry, scanner, platform instabilities, earth rotation, earth curvature, etc. and it was necessary to correct and adapt them ([32], Lillesand et al [33], [34]). The georeferencing was carried out using ground control points selected from topographic sheets of scale 1:50,000 (175/1 of Ikutha, 164/3 of Mutomo, 164/4 of Mutha and 172/2 of Ithumba).The root mean square error (RMS) in the geometric processing was 0.54 [35]. The following parameters have been used in the registration procedures: UTM Projection, Zone 37S, and Arc 1960 Datum.
Image enhancement techniques were applied to the selected subset of the ETM+ data for the study area. The

42
Integration of Geological Mapping and Remote Sensed Studies for the Discovery of Iron -Ore Mineralization in Mutomo -Ikutha Area, SE Kenya achievement of the goal of the application of remote sensing techniques on the studied gneisses led to identify and characterize the studied iron ore bearing gneisses based on the different remote sensing processes of ETM+ data such as color composite image, principal component and band ratio images.

Colour Composite Images of ETM Data
Digital images are typically displayed as additive colour composites using the three primary colours, red, green and blue (RGB). Different spectral bands of ETM+ data have been selected and combined in RGB colour system to make colour composite images for the studied areas. These combinations were tried to select the best colour composite ETM image to be useful in extracting meaningful information about visual lithological discrimination of the studied hornblende gneiss rocks such as combination of ETM bands 7, 4 and 2 (Figure.5).

Principal Component Analysis of ETM+ Data
Different bands of multispectral data are often correlated and thus contain similar information i.e. have similar visual appearances. This correlation means that there is redundancy of information. The principal components (PC) transformation is used, to reduce this data redundancy, by compressing multispectral data sets and calculating a new coordinate system [36] . The application of the PC on the present data is to compress all of the information contained in an original n-channel (band) data set, into fewer number of channels or components, that could be displayed separately as single stretched PC-images, or as component in color composite PC-image [37]. The colour composite image of principal components PC1, PC2 and PC3 for the studied areas successfully distinguish and characterize the studied gneisses (Fig. 3).

Band Ratios of ETM+ Data
Addition, subtraction, multiplication and division of the pixel brightness from two bands of image data to form a new image are particularly simple transformations to apply in image processing technique. Multiplication seems not to be as useful as the others, band differences and ratios being most common. Differences can be used to highlight regions of change between two images of the same area. Ratios of different spectral bands from the same image are found useful in reducing the effect of topography and for enhancing subtle differences in the spectral reflectance characteristics for rocks and soils [32] . The use of the band ratio technique was applied to the satellite ETM+ Landsat 7 digital data for this study area.  The reflectance values in band 7 (2.08 to 2.35 µm) of TM data depend mainly on the hydroxyl content of the rocks. The ratio of band 5 to band 7 was used as a measure of the intensity of the hydroxyl absorption (2.2 to 2.4 µm region). This ratio was used because band 5 is not within the confines of the Fe -bearing aluminosilicate related or hydroxyl-related absorption features, whereas band 7 is within the hydroxyl absorption wave lengths [38]. The high value of band 5/7 ratio appears in light tone due to the high content of hydroxyl bearing minerals.
The spectral band ratio is one of the most common powerful techniques applied for mapping the minerals of alteration zones (clay, alunite and iron minerals) [36] . Recognition of hydrothermal altered rocks associated with mineral deposits was carried out using image processing techniques such as band ratio images in addition to colour ratio composite images. The spectral bands of ETM+ are well suited for recognizing assemblage of altered minerals. The hydrothermal minerals that were detected by the Landsat image processing of the data of selected areas could be classified into two groups: hydroxyl (clay minerals) and hydrated mineral (sericites) detected by band ratio 5/7 ( Figure 4) on one hand, and minerals containing iron (magnetite) detected by band ratio 3/1 on the other hand ( Figure 4). Applying these techniques led to the recognition of more than three zones of altered rocks within the studied gneisses along Tiva River.

Colour Composite ETM+ Band Ratio Images
Colour composite images have been constructed using combination of three ETM band ratios. Different ETM band ratio combinations were carried out to select the optimum colour composite image to use in the visual lithologic discrimination of the investigated area. Some of these combinations are illustrated in (Figure.

Lineaments Extraction
The automatic lineaments extraction from Landsat ETM+ panchromatic band of the study areas was carried out under the default parameters of GeoAnalyst-PCI package. These lineaments have been visually edited to extract only the structural lineaments. The obtained total structural lineaments in Ikutha area were 56. The rose diagram of these lineaments shows that their predominant trend is NNW (Fig.  8). The zones of high lineament intensities concentrated at the central and eastern part of River Tiva (Figure 7). The intersected lineaments, on the other hand, are found to be restricted and concentrated only in the southern part of the Ikutha town (Fig. 6).
It could be concluded that the alteration zones and/or mineralized zones in the investigated areas, delineated by band ratio images of the studied gneisses coincide more or less with high lineaments intensity and/or lineaments intersection or both together.

Image Criteria of the Study Areas and the New Discovered Iron Mineralized Zones
From the previously studied remotely sensed data, it is worth mentioning that there are many criteria of the alteration zones that could be used as a guide for the exploration of Iron mineralization in the gneiss in the Mozambique mobile belt of Kenya. The most important criteria used as a tool to increase the potentiality of the iron mineralization was to distinguish high and low reflectance portions in the study area. Moreover, construction of the alteration map of the studied interpreted from the image processing helped in the discovery of the new zones of mineralization (Figure 9 and 10) and could be used in the exploration for iron mineralization in the future.
The methods applied in this work for the exploration of hydrothermal iron ore mineralization using satellite image processing, are based on the surface spectral features characteristic of this type of hydrothermal iron mineralization, which, will allow favourable areas for exploration. The occurrence of hydrothermal altered rock ( Figure 9) is one of the main features that can be utilized to determine the localization of iron mineralization. By comparing the intensity of the extracted structural lineaments, and their intersections to the altered zones, depicted from remote sensing processing technique, it could be recognized that the alteration zones, more or less coincide with the high frequency and intersected lineaments.
The distribution of hydrothermal alteration within the processed Landsat images become the key to locate the main outflows of a hydrothermal system, which led, after the field checking, to recognize high Fe 3 O 4 (Table 2) anomalies and mineralized zones within the study area. Therefore, many locations are delineated in this work as altered zones, whether by clay minerals and sericite. Iron oxides are frequently observed in the outcrops of hydrothermally altered gneisses. Therefore their identification is a useful key to define areas related to deeply mineralized zones ( Figure  10

Field Proving with Laboratory Investigations
Extensive geological, field proving and geochemical investigations were carried out for the alteration zones ( Figure 10), delineated in the studied gneisses by the Landsat image processing. During the field works, the geochemical investigations revealed that the Fe 2 O 3 is relatively high in the sheared gneisses as well as altered gneisses. The three locations discovered as high geochemical anomalies with some iron mineralization, are mainly connected to the hornblende gneisses and hornblende biotite gneiss of Tiva group.
These locations include Timboni, Kitambasye and Mutuluni in Ikutha area The mineralized zones of Tiva gneisses are found to be mainly connected or hosted by the hornblende gneisses that follow the direction of the strong and nearly vertical shear zone striking N320 o W-S140 o E in the western part of Ikutha town.
The Fe 2 O 3 anomaly at the northern part of Timboni, is mainly connected and hosted by highly altered and weathered gneisses and structurally controlled by NNW-SSE shear zone ( Figure 10). One of the striking feature of this anomaly is its association with high content sericite unlike the anomalies of Timboni area which are associated with apatite.

Summary and Conclusions
This study aimed to use the capability of remote sensing techniques integrated with geological investigations to characterize the iron bearing gneisses and to detect hydrothermal alteration zones, and their associated structures suitable for iron mineralization. The study area including the most promising iron bearing gneisses was selected in Mutomo -Ikutha area, SE Kenya. The achievement of the goal of the application of remote sensing techniques on the studied gneisses led to identify and characterize the studied iron bearing gneisses based on different remote sensing processes of ETM+ data, such as colour composite image, principal component, and colour ratio composite images.
Recognition of hydrothermal altered rocks associated with mineral deposits, was carried out using image processing technique such as band ratio images, in addition to colour ratio composite images. The spectral bands of ETM+ are well suited for recognizing assemblage of altered minerals. The hydrothermal minerals that were detected by the Landsat image processing of the data of selected areas could be classified into two groups: Hydroxyl (sericite minerals) and hydrated mineral/ detected by band ratio 5/7 on one hand, and minerals containing iron (magnetite) detected by band ratio 3/1 on the other hand. Applying these techniques led to recognize more extended zone of altered rocks within the studied gneisses. The association of hornblende gneisses, shearing and faulting along river Tiva in Ikutha area appear to play an important role in the distribution and localization of the mineralization.
Image processing techniques are applied to the digital subset ETM+ data of the studied areas. These techniques generated several products of enhanced satellite imagery such as colour composite images, ratio images and principal components. These techniques have been successfully used in the lithological discrimination of iron bearing gneisses.
The capabilities of remote sensing data to characterize the iron bearing gneisses in addition to characterization and mapping the hydrothermal alteration zones usually help in localization of iron mineralization.
Extensive field geologic and geochemical investigations to the zones delineated by the image processing technique led to the discovery of four locations of high Fe 2 O 3 anomalies (Table 2) with some Iron mineralization, mainly connected to the studied gneisses. These location include Timboni, Muangeni, Kitambasye and Mutulini (East of Timboni). This work showed the important criteria that can be considered as tools to increase the potentiality of the iron discovery in Mutomo -Ikutha area. Moreover, construction of the alteration map of the different studied gneisses, interpreted from the image processing can be applied in the discovery of new zones of mineralization. This technique can be used in exploration for iron mineralization in the future elsewhere in the Mozambique mobile belt having similar conditions.