Integrated geophysical methods to constrain subsurface structures of Tulu Moye-Bora-Berecha axial volcanic complex, main Ethiopia rift: Implications for geothermal resources

Abstract

The Main Ethiopian Rift (MER) is a well-known continental rift whose axial sector is characterized by the occurrence of regularly spaced silicic caldera complexes and central stratovolcanoes, interspersed with large fields of fissural basalts, small mafic scoria cones and numerous young normal faults and fissures. The Tulu Moye-Bora-Berecha volcanic complex is found in the central portion of the MER and includes the Tulu Moye geothermal prospect area. A combination of gravity and magnetic methods was used to better constrain the subsurface volcanic stratigraphy and tectonic structures. Regional and residual anomaly maps were produced from the gravity data and the magnetic data were corrected to produce anomaly and enhanced maps. A complete Bouguer anomaly contour map was produced after the necessary reduction was applied to the gravity data. Due to the geomagnetic field's dipolar nature and since the study area is in the equatorial region (<15°), the reduced-to-equator (RTE) technique was used to minimize external effects and correct the data as if the body had been laid at the magnetic equator. The anomalous source's depth was calculated using a Euler depth solution and spectral analysis approach. Joint 2D forward models on three profiles were developed using GM-SYS of Oasis Montaj software. From the interpretation of the geophysical results, the following conclusions have been reached: (1) the crystalline basement is more raised around Salen ridge and west of it; (2) the main heat source of the geothermal system appears to be a central region of the studied area near Salen ridge and is estimated to be at 4–5 km depth, (3) under Gnaro obsidian dome, the basaltic and silicic volcanic horizon is thin, and the deep-seated regional fault serves as the main conduit for the passage of hot fluid to the surface and (4) the gravity and magnetic anomaly plots, regional-residual maps and enhanced data plots all indicate towards the existence of a major geological feature-a large caldera-comprising of Tulu Moye, Bora and Berecha volcanic centers.

1. Introduction

Geothermal energy is energy that is derived from the heat existing within the interior of the Earth. The heat is generated from the Earth interior due to stored heat within the Earth and process of nuclear decay; and has the potential to melt rocks at deep [1,2]. Because these molten rocks are lighter in density, they experience upward motion, resulting in thermal gradient-driven convection of the hot material to the near surface. The upward-moving molten rocks and the associated heat are confined in some sections of the Earth's upper crust, where they interact with groundwater to produce geothermal resources such as thermal springs, fumaroles, and hot grounds. In addition, geothermal energy systems have a minimal environmental effect, are not affected by climate change, and have the potential to be the most affordable source of sustainable thermal fuel for zero-emission, base-load direct usage, and power production [3]. The resource can be used to generate electrical energy as well as in space heating, farming and industry. Based on its temperature and fluid presence, the geothermal resource can be classified as semi-thermal, hyperthermal-wet, or hyprethermal-dry geothermal system [4].

Areas of geothermal potential are concentrated in regions of the Earth where the heat could rise to near surface due to crustal thinning and/or the presence of geological structures that could act as conduits for the rise of the heat. Areas where the crust has relatively thinned due to extensional tectonics and also where structures play a major role in bringing this heat to the surface are identified as locations of higher geothermal potential. One such area is the East African Rift System (EARS) [5,6].

The EARS, spans ∼6500 km and consists of the eastern and western branches. The geothermal resource base on the eastern branch, which creates the Ethiopian and Kenyan rifts, is by far the largest in Africa and among the largest in the world. The EARS is therefore considered to have tremendous geothermal potential, with the capacity to produce between 10,000 and 15,000 Megawatts of energy using current technology [7].

One of the most significant regions of the EARS is the Ethiopia Rift System (ERS), where heat energy from the Earth's subsurface escapes to the surface through volcanic eruptions (like Erta’Ale lava lake), hot springs, and natural vapor emanations (fumaroles) are found. From several geoscientific surveys like Geological Survey of Ethiopia (GSE) and Electroconsult [9,10] it was determined that 120 sites in the ERS have the potential for geothermal resources with high and low enthalpy [11,12]. Of these potential areas, 22 geothermal prospective regions have been identified to be of high-enthalpy geothermal resources [13]. The following seven areas are specifically identified for further detailed geothermal exploration: Corbetti, North Abaya, Tulu Moye, Dofan, Fentale, Aluto-Langano and Tendaho [8,14,15]. The Tulu Moye geothermal potential area, which is the focus of this study, is situated on the eastern edge of the CMER and is characterized by the Wonji Fault Belt fissural basaltic eruptions and significant extensional tectonics [[16], [17], [18]] (Fig. 1). Surface hydrothermal manifestations are mainly in the form of warm and hot steaming grounds with an extensive area of hydrothermal altered ground. The basaltic lava flows dating from the Pleistocene to more recent times and silicic centers are what make the region unique [19] (Fig. 2, Fig. 3).

Fig. 1.

DEM of the CMER region. Abbreviations: LA-Lake Hawasa, LSh-Lake Shala, LLa-Lake Langano, LAb-Lake Abijata, LK-Lake Koka, LZ-Lake Ziway, Alu-Aluto, G-Gedemsa, TM-Tulu Moye, B-Boset, Zi-Ziquala, Ch-Chilalo, Gr-Galema range, MKSH-Midre Kebd Structural High, YTVTL-Yere Tulu Wellel Volcano Tectonic Lineament, SDFZ-Silte Debre Zeyit Fault Zone. The square boundary indicates the Tulu Moye area.

Fig. 2.

Satellite image of the Tulu Moye geothermal prospect in the eastern rift margin of the Central Main Ethiopian Rift including the Gnaro-Salen-Jano's recent lava flow with mountains found in the region. Volcanic centers are represented by red triangles and caldera rims are represented by greenish blue lines with small triangles. The inset map shows Ethiopia and Horn of Africa countries.

Fig. 3.

Detailed map of volcanic complexes Tulu Moye-Bora-Berecha volcanic system. The silicic centers are labeled and greenish blue dots indicate small pumice cones, lava and domes. Blue squares (TG-1, TG-2, TG-3, TG-4, and TG-5) are shallow wells drilled in the region. MER faults are after Agostini et al. (2011) and shallow drill well sites after Ayele et al. (2002).

Throughout the last several decades, beginning in 1969, Ethiopia has been conducting an extensive geothermal exploration. UNDP [12] and Di Paola [20] supplied concise geologic explanations of hydrothermal sites in the lakes district as well as geologic studies of the various rock types and geothermal conditions in the region. The history of rifting and volcanism in the central sector of the MER, including Tulu Moye, was documented by Tadesse et al. [21] and WoldeGabriel et al. [22] using K/Ar dating. The northeast lake Ziway region, which includes a portion of Tulu Moye, was lithologically and structurally mapped by Admassu and Woku [25] and Korme [23] using data from Landsat TM. During the years 1985–1987, Electroconsult (ELC) collaborated with Geothermica Italiana and the Ethiopian Institute of Geological Surveys (EIGS) to conduct research and identify attractive geothermal areas in the Main Ethiopian Rift system [10]. Using the Magnetotelluric (MT) approach, Samrock et al. [24] conducted geophysical investigation at the Tulu Moye geothermal prospect and created the first comprehensive 3-D picture of a magmatic-hydrothermal system. Moreover, the Tulu Moye regions’ fault initiation and development, as well as their link with magmatism were examined using remote sensing data by Admassu and Worku [25].

According to Peacock et al. [26], geophysical technologies enhance drilling success and lower exploration and development expenses. The chance of drilling an unsuccessful geothermal well can be decreased by evaluating several geophysical data sets to determine the best drilling locations [27]. Gravity and magnetics, the two geophysical techniques utilized in this study, complement one another since they are sensitive to various physical variables [26]. In order to map geological features, magnetic methods are frequently used with gravity data in geothermal investigation. In geothermal exploration, magnetic measurements often focus on finding buried intrusive bodies and determining their depths as well as tracking specific dykes and faults [28]. On the other hand, gravity measurements are used to find denser and lateral geological barriers that potentially regulate the flow of hydrothermal fluids in geological formations [29].

In this study, a combination of geophysical methodologies was used to determine the subsurface geological architecture including the stratigraphy and thicknesses of the various lithologic units, major tectonic structures and geothermal characteristics of the research area (Fig. 3). Accordingly, regional and residual anomaly maps were created using the gravity and magnetic data. Euler and radially averaged power spectrum approach were used to calculate the depth to the geothermal resource. A joint 2D forward model for the study area were developed using the value determined from these methodologies and prior knowledge as a basis [30].

2. Geological setting

The Main Ethiopian Rift (MER) is an active intra-continental rift that lies between the Nubian and Somalian lithospheric plates and has an NNE-SSW to N–S trend [31,32]. It represents the northernmost sector of the East African Rift System (EARS) and separates the northwestern Ethiopian plateau to the west from the southeastern Ethiopian (Somali) plateau to the east; and opens at rates of ∼4–6 mm/yr [[33], [34], [35], [36]]. The MER is divided into three main segments: the North Main Ethiopia Rift (NMER), the Central Main Ethiopia Rift (CMER) and South Main Ethiopia Rift (SMER) [17,22,[37], [38], [39]].

The main boundary faults in the NMER region developed roughly 10–11 Ma and they have an average N50° trend [37]. The CMER is 80 km wide and 180 km long. It also stretches southward to Lake Hawassa area (or Goba-Bonga lineament). In the CMER, the rift boundary faults trend around N30°-35° and the age of faulting is considered to have occurred during 8.3–9.7 Ma [37] (Fig. 1). The SMER is 600–700 km long and ranges 60–300 km in width south of Lake Hawassa; and its southern end is bordered by the Kenya Rift in the Turkana depression [37,40]. Faults in the SMER show a dominant N–S to N20° trend and age of faulting is estimated to be around ∼18 Ma [37,[41], [42],43,44,45,46]. Beginning in early Pleistocene (about 1.6 Ma ago), the youngest faults (Wonji Fault Belt -WFB) began to form on the rift floor [18,[47], [48], [49], [50], [51], [52], [53]]. Despite the MER's general NE-SW trend, the WFB is characterized by active extension fractures and normal faults that have an NNE-SSW trend. The WFB system of faults are arranged in a right-stepping, en-echelon pattern [17,50,54,55]. All recent volcanic eruptions and seismic events are taking place on the WFB which has been divided into several discrete volcano-tectonic segments. Accordingly, the Tulu Moye- Bora-Berecha (TMBB) volcanic complex (also referred to as the Bora-Berecha-Tulu Moye, TMBB volcanic complex, Tadesse et al. [16]) lies on the so-called Aluto-Gedemsa magmatic segment [38,56].

2.1. Geology of the study area

The Tulu Moye-Bora-Berecha (TMBB) region is located close to the MER eastern boundary and is characterized by very steep, west-dipping normal faults with an average strike of N30°-N40°E. This rift sector is considered to be part of the CMER [22,37,58,58], while others put it as part of the NMER [59,60]. In this study, it is preferred to regard it as part of the CMER since its geological characteristics seem to be closer to the latter (Fig. 1). The TMBB sector is bordered to the north by the Gedemsa and Boku calderas (Fig. 2). The southern boundary is Lake Ziway which has been described as an accommodation zone with a dextral transverse displacement [16,18,22,57]. The eastern side is bordered by the rift margin and the off -axis Chilalo stratovolcano, which is situated on top of the southeast Ethiopian (Somali) plateau (Fig. 2, Fig. 3). The western boundary is not well-defined tectonically and the sector gradually passes through a wider flat topography covered by poorly welded pyroclastic deposits and lacustrine sediments. The rift sector in which TMBB lies thus consists of major boundary faults, young WFB system of faults, fault swarms and fissures, large caldera structures, young lava flows and domes and numerous scoria and pumice cones (Fig. 3).

The TMBB volcanic field constitutes a series of Late Quaternary volcanic edifices, namely Tulu Moye, Bora and Baricha centers. In addition, there are a number of smaller eruptive vents, cones, and domes across the field (Fig. 3). Tulu Moye is an edifice close to the eastern rift escarpment, Bora is located in the west of the volcanic field near Meki town and Berecha is located near the southern shore of Koka reservoir. In this volcanic field, the WFB controls the eruption of both mafic and silicic magmas, which have clearly erupted contemporaneously in the recent past [47,57]. Historical eruptive activity in the area is reported for the Giano lava flow immediately north of Tulu Moye (Fig. 4). There are two fresh-looking lava flows in Giano, possibly sourced from two different eruptions. According to Fontijn et al. [47] and Gouin [61], local elders account for the fissure eruption to early 1900s. On the other hand, Bizouard and Di Paola [62] suggest the same eruption occurred in the 18th century.

Fig. 4.

Geological sketch map of the study area showing lithology and distribution of surface hydrothermal alteration areas (the lithology is modified from Abebe et al., 1998; Boccaletti et al., 1999; Abebe et al., 2005 and Birhan, 2021).

From previous studies the TMBB volcanic complex [16,[23], [24], [25],47,[62], [63], [64], [65], [66], [67], [68]], the following general geologic evolution has emerged:

The Late Pleistocene ignimbrites, which have been dated to 1.58 $\pm$ 0.2 Ma [21,22], are overlain by volcanic products from the recent TMBB volcanic centers. Trachyte and rhyolite lava flows are found at the bottom of the TMBB subsurface geology, and these flows are overlain by weakly to moderately compacted layers of ignimbrites and unwelded pyroclastic deposits, as revealed from lithologic logs of three shallow wells TG-1, TG-2, and TG-3 [68,70] (Fig. 3). According to well TG 2, the rhyolite lavas grow thinner and the fine-grained ignimbrites thicken to a depth of 87.5 m. Another nearby shallow well (TG 4) drilled near a rhyolite dome reveals very thick (100 m) trachytic and rhyolitic lavas with different porphyritic to completely glassy obsidian covering an ignimbrite that resembles those found in the rift escarpment [70] (Fig. 3). A thick layer of porphyritic and scoriaceous basalt (90.5 m) is present at the base of the well TG 5, which is adjacent to the Tulu Moye volcano. According to Ayele et al. [70], the subsurface lithologies exhibit alteration that is probably connected with the area's ongoing hydrothermal activity.

The volcanic products exposed in the TMBB show that the volcanic field has undergone many eruptions and of variable styles until the present day. There is a strong suggestion for the existence of an earlier caldera, but there is no obvious surface expression of one. A potential caldera structure was previously interpreted from the accommodation faults trending NW-SE [23,25]. Products exposed at the surface and presumably all of the post-caldera age, include basaltic and rhyolitic lava flows, tephra falls and pyroclastic density current (PDC) deposits. To the NE of the Tulu Moye edifice, NNE-SSW faults have accommodated eruptions of both basaltic and silicic lava flows [47,67]. Tulu Moye has erupted comenditic compositions, and Bora-Berecha pantellerite compositions [47]. This has been interpreted as an indication that Tulu Moye and Bora-Berecha may have separate magmatic systems (Fig. 1, Fig. 4).

Seismic [65,66], InSAR [71,72] and Magnetotelluric surveys [25] reveal that Bora and Tulu Moye are still active. InSAR observations between 2004 and 2011 by Biggs et al. (2011) show long-term (∼30 months) active ground deformation. Two pulses of uplift (+5.3 cm) with intervening subsidence in 2004 and 2008–2010 were identified in the region close to Bora and Tulu Moye. A shallow (<2.5 km) penny-shaped model was suggested for the deformation source [71]. More recently, in early 2016 a new episode of uplift (+5.8) was detected in the region centered between the three major edifices, and the uplift decreased rapidly with an average rate of 1.9 cm/yr for the period 2017–2020 [72].

The seismicity catalog developed by Greenfield et al. [65,66] highlights three significant earthquake clusters around Tulu Moye, Bora and the region between the two volcanic centers. The sources of the earthquakes are predominantly located shallower than sea level. The Magnetotelluric survey shows two main electrically conductive zones ponded at different depths, connected by an oblique westward-dipping conduit-like zone [24]. The deepest conductive zone is located below 14 km depth, underlying the entire TMBB region and has been interpreted as partial melt stored in the lower crust. The shallow conductive zone is located in the upper crust (4–5 km depth) underneath Tulu Moye. This zone is thought to represent the remnants of the melt that fed the last historical eruption and to be the main driver of the currently active subsurface hydrothermal circulation around Tulu Moye.

2.2. Geothermal manifestations

One of the MER's promising geothermal prospects is the Tulu-Moye geothermal area. Significant hydrothermal activity is present in the north and northwest portion of the central sector of the research area. In the region, different geothermal manifestations have been observed such as hot steam vents and fumaroles. During data acquisition phase, geothermal steam vents and thermally altered grounds were observed at different locations (Fig. 2, Fig. 4). The fumarole is being used by the local community for steam baths and for healing purposes. The comparatively high altitude has an impact on the hydrothermal manifestations in the potential region, and there are no hot or cold springs present in the area. This is possibly because the geothermal prospect area is located at a higher altitude (1800–2200 m a.m.s.l.) and has dense deep faults and fractures that may have controlled the groundwater movement. As a result, weak fumaroles, active steaming and altered grounds are the predominant hydrothermal manifestations.

3. Methodology

The Tulu Moye geothermal prospect, located between Lake Ziway (in the southwest) and Lake Koka (in the northeast) in the Central Main Ethiopia Rift, was studied in the current research using ground-based magnetic and satellite gravity data sets to understand the subsurface geological configuration. The data acquisition and processing methods are documented as follows.

3.1. Geophysical data

A total of 32,803 gravity data were adopted from International Gravimetric Bureau (BGI). The principal author and Tulu Moye Geothermal Organization (TMGO) jointly collected more than 92,000 magnetic data across a 513 km² area. A GSM-19T proton precession magnetometer with a precision of 0.01 nT was used to gather the data. Line-to-line spacing ranged between 2 and 4 km, and the data points were spaced out by 3 m (approximately 1s; depending on accessibility of the survey traverse). The magnetic data was mostly concentrated on the WFB and the transverse faults of the TMBB region.

3.1.1. Gravity data

The satellite gravity data were downloaded from the International Gravimetric Bureau (BGI) (via an available website, http://ddfe.curtin.edu.au/gravitymodels/GGMplus/). The grid resolution is 0.002deg (7.2 arc seconds) with the grid equally spaced in terms of geodetic (GRS 80) latitude and longitude. The complete Bouguer gravity anomaly map was created after taking the necessary reduction techniques (Theoretical gravity, Terrain correction, Free air correction) on observed gravity data. The study region has an elevation range of 1599.88 m–2240.92 m. Using the kriging gridding technique, the Geosoft Oasis Montaj software [73] displays the estimated Bouguer anomaly values as a contour map.

The Bouguer anomalies come from the interplay of shallow (high frequency) and deep (low frequency) causal sources, each of which needs to be broken down into its component elements. The 500 m upward continued Complete Bouguer Anomaly (CBA) grided map was processed using the Gaussian regional/residual filtering approach utilizing the interactive radially averaged power spectrum. The Geosoft MAGMAP module was used to do this. The residual gravity anomaly was calculated by deducting the regional gravity anomaly values from the Bouguer anomaly values. Using the residual complete Bouguer anomaly of the gravity data, Euler depth estimate and 2D forward modeling were carried out.

3.1.2. Magnetic data

Magnetic methods are widely used in geothermal exploration and focus on mapping changes in magnetism related to the distribution of magnetic minerals. To determine the total magnetic field (TMF), the obtained ground based magnetic data were corrected for diurnal and the International Geomagnetic Reference Field (IGRF). The difference between the total magnetic field determined by the IGRF and the measured magnetic value was then used to calculate the magnetic anomaly. Using the kriging gridding method in Oasis Montaj, magnetic anomaly results are then displayed as a contour map. The geomagnetic field is bipolar, therefore any inclination and declination away from the magnetic pole cause asymmetrical Earth magnetic field [74]. To solve this issue, researchers have adopted the reduce-to-pole (RTP), differential-reduced-to-pole (DRTP) [74,75], or the reduce-to-equator (RTE) transformation. The RTE technique for handling magnetic data, however, is preferred over lower latitude regions compared to the other methods [74,76]. Hence, the average inclination and declination values of the research area were used as 1.24° and 2.16°, respectively, to further process the total field magnetic anomaly data using the RTE transformation in Geosoft software. The RTE magnetic anomaly was separated into regional and residual anomaly grids, and the analytic signal grid was found using the Oasis Montaj program (Fig. 7D–F).

Fig. 7.

Magnetic anomaly maps of the study area: A) Total magnetic field (TMF); B) Magnetic anomaly; C) Reduced to Equator (RTE) of magnetic anomaly; D) Analytic signal (AS); E) Residual Magnetic anomaly and F) Regional magnetic anomaly. Black Lines indicate profile lines along A-A′, B–B′, and C–C′.

3.2. Depth estimation techniques

Regarding the geophysical interpretation of subsurface structures, the depth of the source of magnetic and gravitational anomalies is useful and significant. Two depth estimation approaches, i.e. 3D Euler deconvolution and radially averaged power spectrum analysis were used to comprehend the subsurface structure of the area.

3.2.1. 3D Euler depth

To solve the Euler equations and determine the depth and location of the target area, Euler deconvolution is a procedure of inversion by least squares using anomalous values of the magnetic field and gravitational density [[77], [78], [79]]. Various researchers have claimed that Euler deconvolution techniques can be used to obtain target positions for magnetic and gravitational anomalies [78,[80], [81], [82]]. To determine the target depth and its position, one primarily employs the 3D Euler deconvolution technique [[79], [80], [81],83] on gravity anomaly and RTE of the magnetic anomaly and the same techniques were adopted in this study.

The gradients along three orthogonal (dx, dy and dz) axes were estimated using the Fast Fourier Transform (FFT) from the grided gravity and RTE magnetic anomaly data using the MAGMAP application, an extension of Oasis Montaj 8.4. The estimated Euler depth solutions for the gravity source are given in (Fig. 8B). These solutions were estimated using a structural index (SI) of 1, a window size of 10, and a 15% depth tolerance. The estimated Euler solutions for the magnetic data were also calculated using the structural index of 2 (dike body), a window size of 10, and a 15% depth tolerance, as shown in Fig. 8A. Based on previous geological knowledge, the complete Bougier gravity anomaly and magnetic anomaly data were employed to define the lateral and vertical contact of the observed anomalous zones.

Fig. 8.

Euler depth solutions of A) RTE magnetic anomaly data using structural index = 2 (ranges 618.0–1602 m) overlayed on RTE magnetic anomaly map and B) gravity residual data sets using structural index = 1 (ranges 263.4–1747.2 m). The maps are accompanied by a color-coded bar that shows the depth of each source.

3.2.2. 2D Euler deconvolution

The homogeneity equation of Thompson [84], referred also by Usman et al. [85], serves as the foundation for the 2D Euler deconvolution solution. The following equation is used to compute the 2-D Euler deconvolution:

$$x_o\frac{\partial \Delta T}{\partial x}+z_o\frac{\partial \Delta T}{\partial Z}=x\frac{\partial \Delta T}{\partial x}+N\Delta T\left(x\right)$$ (1)

where (x_o, z_o) describe the depth and location of the point-analogous source along the profile, ΔT is the observed total field and N is the type of source that best characterizes the anomaly. The z-coordinate represents the source's vertical position.

In equation (1), N denotes the Euler's structural index (SI), which may take on values between 0 and 3. According to Yadav and Sircar [86]; Mandal et al. [69]; Fregoso et al. [78]; Durrheim and Cooper [82]; Mushayanebyu et al. [87] and Reid et al. [79], the horizontal derivative and vertical derivative gradient of a magnetic field can be calculated from $\frac{\partial \Delta T}{\partial x}$ and $\frac{\partial \Delta T}{\partial Z}$, respectively.

The software Euler 1.15, created by Cooper [88] was used to produce the 2D Euler deconvolution solution displayed in Fig. 9, Fig. 10, Fig. 11B. The total magnetic anomaly grid was used to extract 205, 180 and 217 data points with a spacing of 100 m along profile lines A-A′, B–B′ and C–C′, respectively. The geomagnetic field, inclinations and declinations of the survey locations are required by the software as magnetic parameters. The '.dta' files with two data columns are fed into the program-the first column contained the locations of the magnetic stations and the second column contained the corrected magnetic field values. As an input, a geomagnetic field strength of 355517 nT, with an inclination of 1.24° and a declination of 2.16° was employed. The procedure utilized a window size of 21, 100 m x-separation, 1000 m y-separation and a structural index of 1.0.

Fig. 9.

(A) Topography elevations from Aster DEM along profile A-A′ and showing surface faults. (B) 2D Euler deconvolution solutions along profile A-A’; the yellow lines representing the estimated fault locations.

Fig. 10.

(A)Topography elevations from Aster DEM along profile B–B′ and expected fault area, and (B) 2D Euler deconvolution solutions along profile B–B’; the yellow lines representing the estimated fault locations.

Fig. 11.

(A) Topographic elevations from Aster DEM along profile C–C′ and expected fault area and (B) 2D Euler deconvolution solutions along profile C–C; the yellow lines representing the estimated fault locations.

3.3. Radially averaged power spectrum

The 2D Radially averaged power spectrum method is based on spectral Fourier analysis sets of gravity-magnetic data. Fast Fourier Transform (FFT) is used to calculate the energy spectrum in MAGMAP of Oasis Montaj (v 8.4) using pre-processed gravity and RTE magnetic grid files that have been extended by 10% as a square grid. The resulting graph (Fig. 12) shows the natural logarithm power of the corresponding data against wave number. The depth and the power spectrum are related by the following mathematical relation:

$$S\left(w\right)=\sim e^{-2wh}$$ (2)

where S(w) is the power spectrum, w is the wavenumber and h is the depth of the source [89]. The energy spectrum curve is then split up into several straight-line segments to provide a satisfactory fit for the curve. The following mathematical formula was used to calculate a source depth from the slope of the energy spectrum curve [90,91]:

$$h=\frac{S\left(x_1\right)-S\left(x_2\right)}{4\pi \left(x_1-x_2\right)}$$ (3)

where x₁ and x₂ are the start and end points of the radial frequencies, and s(x₁) and s(x₂) are the corresponding values of the radially averaged power spectrum components [91]. Fig. 11 A and B show the 2D radially averaged power spectra analysis of gravity and magnetic anomaly data sets using MAGMAP of Geosoft software. According to Fig. 12 A, the depth of the gravity source is separated into three unique layers, whereas Fig. 12 B shows that the magnetic anomaly source is divided into four distinct layers. Depth 1 denotes a higher slope (lower frequency, greater depth), whereas shallower depth denotes a lesser slope (higher frequency).

Fig. 12.

2D radially averaged power spectrum used for estimating source depth of A) Residual gravity anomaly and B) Magnetic anomaly. Y₁ to Y₄ are the spectrum zone used to calculate the depth to the deeper (Y₁) and shallower (Y₄) depths.

3.4. 2D forward modeling

For a final model to be reasonable and acceptable with the least tolerable error, joint gravity/magnetic 2D forward modeling needs apriori constraining information. The two NW-SE profiles (profiles A-A′ and B–B′) and one N–S profile (profile C–C′), which pass through the Artu and Salen ridge, were utilized to perform the 2D forward modeling. The modeling was carried out using a seven-layer configuration, as 2D forward models were estimated by different researchers in the CMER [30,41]. The depths of the interfaces were considered 400, 1300, 1600, 2400, 3300, 3937 and 4500 m below the surface by roughly correlating with the depth determined from radially averaged power spectra of the magnetic anomaly. From top to bottom, young basalt, silicic lavas, Pliocene Bofa basalts, silicic products and rift floor ignimbrite, flood basalts and rhyolites, Mesozoic sediments and metamorphic crystalline basement are the proposed rock units in the subsurface, respectively. The forward model clearly shows a seven-layered subsurface with rock densities of 2750, 2550, 2810, 2580, 2850, 2500 and 2750 kg/m³ and magnetic susceptibilities of 0.775, 0.413, 0.747, 0.465, 0.821, 0.134 and 0.661 nT from the top to the deeper layers, respectively (Fig. 13 A, B and C, Table 1). The start-up model depth was derived from previously known information about the local geology and different layering, such as that found in the schematic stratigraphic column [51], seismic survey [92], 3D model based on MT data [24], resistivity modeling [41], radial averaged power spectra and the depths determined from the Euler deconvolution of the gravity and magnetic data of this study is also considered. The crystalline basement horizon is predicted to be at a mean depth of 3 km using the radial averaged power spectrum approach, which is in good agreement with Kebede et al. (2021). For the 2D forward modeling procedure, Oasis Montaj's GM-SYS module was utilized. The joint 2D forward models of the 3 profiles were estimated using the calculated magnetic and gravity anomaly values of the seven horizons (Fig. 12 A, B and C). Under the parameters imposed by the geological constraints, the model was interactively run on a computer screen until a reasonable degree of agreement (minimal RMS error) was achieved between the computed and observed values.

Fig. 13.

2D forward gravity models along the profiles passing through the Salen Mt.: A) profile A-A′ passing through Salen Mt. and Dima mount (fractured area), B) profile B–B′ passing through Giano Mt. (recent volcano), and C) profile C–C′ passing through Artu and Gnaro obsidian.

Table 1.

Regional stratigraphy, estimated thickness, density, and susceptibility values for Tulu Moye geothermal prospect area.

Layer	Horizon geology	Thickness range (m)	Density (gcm−3)	Susceptibility (m3Kg−1)10−3	Source
1	Early Pleistocene-Holocene Young basalts	400	2750	0.775	[51,93,94]
2	Early Pleistocene-Holocene silicic lavas, (unwelded pumice and ignimbrites)	1300	2550	0.413	[51,94]
3	Pliocene Bofa flood basalts	1600	2810	0.747	[51,95]
4	Mio-Pliocene Silicic lavas and ignimbrites	2400	2580	0.465	[51,95]
5	Oligo-Miocene Flood basalts and rhyolites	3300	2850	0.821	[96,97]
6	Mesozoic sediments	3900	2500	0.134	[[93], [94], [95]]
7	Precambrian crystalline basement	4500	2750	0.661	[51,[93], [94], [95],98]

4. Result and discussion

4.1. Interpretation of the gravity anomaly

Bouguer anomaly map of the Tulu Moye-Bora-Berecha volcanic complex (Fig. 5A) reveals an anomaly range from a minimum of −235.1 to a high of −185.31 mGal and the elevation grid of the area shows as the elevation ranges from 1598 m to 2370 m (Fig. 5B), however, the plot covers a broader area than the research region. It is noted that an E-W trending low-gravity zone (between −216.5 and −235.1 mGal) in the south-eastern half has widened and shifted its direction to almost southwest. The broad gravity low in the southeast is assumed to represent a component of the Asela/Chilalo eastern side of the rift, whereas the west-southwest flank of this low gravity zone is assumed to be the Bora-Berecha volcano. Lower gravity anomalies are seen at higher altitudes in the southern and southeastern areas while the WFB segment shows high gravity values. In the northeast, center and southwest of the research region, there is a more pronounced gravity high (−191.8 to – 185.1 mGal), which is associated with the occurrence of Quaternary basaltic rocks (Wonji group basalts) in the near subsurface. According to Ebinger and Casey [99], the study region is considered to be a component of the Aluto-Gedemsa magmatic segment and also contains recent lava flows in the vicinity (such as the Gnaro obsidian area; see Fig. 2, Fig. 4). As a result, zones with medium to high amplitude gravity anomalies may be linked to pyroclastic and Chilalo volcanic deposits; whereas zones with high amplitude gravity anomalies may be caused by the presence of high-density rocks such as basalts (see Fig. 2, Fig. 4). Due to the intrusion of basalts (diking) at the center of the Gedemsa and Tulu Moye-Salen calderas, the study area's gravity highs and lows have undergone significant alterations to the east, west and the center of the region (Fig. 2, Fig. 4). The Wonji Fault Belt and the transversal faulted network, as well as the regional variation in gravity anomaly, suggest that the region is impacted by the rift system's lateral expansion in the faulted zone and the buildup of sediments on the graben at the rift floor.

Fig. 5.

Bouguer anomaly over the study area A) Complete Bouguer anomaly grid and B) Elevation grid. The blue triangles indicate fumaroles located in the region while the black solid lines represent the profile lines.

The scope of the Bouguer anomaly map in Fig. 6 represents the research region under examination and the adjacent area. There is an anomalous gravity field with a maximum density value of −185.1 to −189.4 mGal that is present to the west of Salen fissure-fed basalt, south of Gedemsa caldera, and the west of the research region. Low-gravity anomalies are found southeast of Iteya town or southeast of the rift border, as illustrated in Fig. 6A, and they range in size from −233.8 to −221.3 mGal. The high gravity anomalies of the WFB contrast with the low gravity anomalies of the Tulu Moy-Salen-Giano silicic centers that lie within it. It is assumed that the high gravity anomaly in the northeast is a southward continuation of the Gedemsa caldera and is driven by magmatic intrusion along the fault lines. It is possible to interpret the fumaroles near Artu as resulting from emanations through faults and fissures of the WFB (Fig. 2, Fig. 6A). The regional gravity anomalies show that the central and northwest high gravity anomaly zone is more closely related to the north-northeast high gravity anomaly associated with Gedemsa volcanic center (Fig. 6C). This may be explained by the fact that the two high gravity anomaly sites (Tulu Moye and Gedemsa) are the consequence of comparable magmatic segments [99].

Fig. 6.

Surface gravity anomaly maps of the study area: A) 100 m Upward continuation of CBA; B) Regional anomaly map and C) Residual anomaly. A-A′, B–B′, and C–C′ are three forward modeling profiles.

The residual gravity map (Fig. 6B) shows that the estimated anomalies range from −8.797mGal to +6.992mGal, derived after filtering out the regional impacts. After removing the regional gravity anomaly (Fig. 6C), it was observed that the shallow subsurface features were more noticeable (Fig. 6B). The residual gravity anomaly map reveals a NE-SW trend along the WFB system of faults and is also clearly aligned with the system of transversal faults found southwest of Tulu Moye-Salen ridge (Fig. 6B). The localized low amplitude gravity anomalies in (Fig. 6b) that are less than +0.61 mGal over the rift basin regions and central to west areas following the transversal system of faults can be explained by the low-density source rocks, or by recent rift sediment depositions (like lacustrine sediments and alluvial deposits). It is possible to assume that the locations with the sharp fall in gravity values from high (+6.992 mGal) to low (−8.797 mGal) are the consequence of faults and newly intruded material (see Fig. 2, Fig. 4, Fig. 6B). The largest amplitude residual anomaly zones (from 2.951 to 6.992 mGal) on the residual anomaly map can be explained by the existence of volcanic rocks that have recently erupted basaltic rocks. Low-to-moderate anomalies (−8.797 to + 0.61 mGal), found in the NE, SE and western zones, outside of the transversal and WFB, may be caused by the presence of low density pyroclastic and lake sediments (Fig. 6C).

4.2. Interpretation of the magnetic anomaly

According to the total magnetic field (TMF) map (Fig. 7A), there are variations in the TMF from a low of 34854.8 nT to a high of 36143.40 nT. The maximum value of the TMF is observed at the center of the study area which has an E-W trend and a lower anomaly TMF is observed along with a prominent south-southeast trend. The magnetic anomaly map and the TMF map are quite similar, as can be seen in Fig. 7A and B. Due to the study area's lower latitude (<15°), the Reduced to Equator (RTE) transformation technique was used to further enhance the magnetic anomaly data (Fig. 7C). Note that in the maps of RTE anomaly, such as Fig. 7C, the anomaly values are reversed due to this operation. The RTE magnetic anomaly map shows a large variation in magnetic anomaly value (−481.5 to 648.1 nT), where high magnetic anomaly (associated with high rock demagnetization but shown in reverse) is found at center to NE, and a broad low southern part of the research region next to Tulu Moye-Salen-Giano centers (Fig. 7C). In the research area's central region, close to WFB and fissure-based fumaroles, there is a linear E-W trending high magnetic response region. The high-to-low magnetic anomaly transition zone is where the fumaroles are located. In the figure, the high gravity anomaly (Fig. 5A) and the low magnetic anomaly (Fig. 7C) are in perfect alignment.

The high RTE, analytic signal and the regional magnetic anomaly (Fig. 7C–F) are mostly centered in the middle section of the research region around Salen ridge, in agreement with the high gravity anomaly (Fig. 5A), which is assumed to represent an extension of the Gedemsa volcanic center due to the high-density present at depth. The combination of a high temperature (which causes loss of magnetization) and high-density body (basaltic material) located close to the surface might be understood as the possible cause of this phenomenon. Its natural setting is a reliable predictor of the target area's geothermal energy potential.

The Complete Bouguer Anomaly (CBA) plot Fig. 5A, gravity anomaly maps of Fig. 6 and the various magnetic anomaly and enhanced plots of Fig. 7 (A-F) all indicate towards the possibility of the existence a major geologic feature of interest-a possible caldera-encompassing the Bora-Berecha-Tulu Moye volcanic centers.

4.3. Interpretation of 3D Euler depth

In order to compute the Euler depth solutions to figure out the source depth and extent of faults, dykes and fractures, the structural index (SI) 1 for the gravity and 2 for the magnetic anomalies was taken [79,80], coupled with a window size of 10 and a depth tolerance of 15%. The RTE magnetic anomaly and gravity anomaly maps (Fig. 8A and B) were overlain with these findings for close examination. The depth solutions for the magnetic anomaly were computed at depths of 690–1682 m, and these were seen to be accumulating near the magnetic low-high transition. The source depth for the density anomaly was computed at depths between 263 and 1747 m, and it was seen that they were accumulating along the transversal fault and WFB. This may have occurred as a result of the intrusion of basaltic material. The Euler depth difference between the magnetic and gravity anomaly are the result of demagnetization of the rocks arising from high temperatures at a relatively shallow depth in the area. The maps show that the causative bodies are considerably shallower in the central part of the research area and are deeper outside of the WFB at the eastern and western edges. Most of the gravity and magnetic depth sources are oriented along an NE-SW direction, following the WFB, and have depths greater than 1500 m. These occurrences can be attributed to high density basaltic material entering through the WFB as the cause of these phenomena.

4.4. Interpretation of the 2D Euler deconvolution

Fig. 9(A and B), 10(A, B), and 11(A, B) validate the elevation profile and magnetic anomaly data that have been analyzed for profile lines A-A′, B–B′, and C–C′ respectively. The magnetic sources are denoted by the plus sign (+) and are dispersed across the profile in accordance with the dispersion of subsurface magnetic sources. The reflection of the magnetic source can take the form of an ellipse or other distinct structures that can be used to detect the existence of feeder dykes, fissures, or sills. At certain locations of the 2D Euler solution, the magnetic source was observed to be randomly spread (Fig. 10, Fig. 11B), and this may be interpreted as the presence of active seismic activity which extends to 9 km along profile B–B′ and to 7 km along profile C–C’ [100]. The region where magnetic signals are concentrated in the Euler depth solution corresponds most closely to the field faults. The absence of magnetic signals between the two major faults may be an indication that hydrothermal fluids are present in the region. Moreover, the presence of a wide space between the two faults along the profile is a probable indicator of a heat source at shallow depths (profile B–B′ between faults 2 and 3; Fig. 10B). The lack of a magnetic signal at the middle section of profile A-A' (Fig. 9B), the middle portion of profile B–B' (Fig. 10B), and the left portion of profile C–C' (Fig. 11B) below 5 km suggests that the heat source may be present at shallow depths (<6 Km) particularly below Tulu Moye-Salen-Giano volcanic field.

4.5. Interpretation of the radially averaged power spectrum

The analysis of the residual gravity's radially averaged power spectrum yielded three depths: deep source, intermediate source and shallow source, with corresponding depths of 3384 m, 1376 m and 380 m, respectively. Similarly, for the magnetic anomaly depths of 2625 m,1013 m, 899 and 300 m values were calculated (Fig. 12 A and B). The impact of the shallower depth heat source on the magnetic body may be the cause for the difference in depth between the two datasets. It is possible to interpret the significant difference between the deep and intermediate depths (3384-1376 m) of the gravity anomaly and (2625m–1013 m) of the magnetic anomaly, as the response of the larger thickness of the Pliocene Bofa basaltic horizon in comparison with the thickness of the other layers in the area.

4.6. The joint 2D gravity/magnetic forward modeling

To choose the optimum prospective region (drilling locations) and reduce the likelihood of drilling an unsuccessful geothermal well, it is important to jointly evaluate two or more geophysical data sets [27]. For a better geological understanding of the area of study, the combined gravity/magnetic models were taken into consideration. The 2D forward modeling was produced for traverses aligned E-W (profile A-A′ and B–B′) passing through the Salen ridge and NW-SE (profile C–C’) crossing the Gnaro obsidian lava flow. The data from Table 1, the spectral analysis results and the depth solutions were used to create the starting model.

Lithologic contacts were considered at depths of 380, 1375 and 3384 m below the surface by taking the depth determined from radially averaged power spectrum and Euler depth solutions. Taking these as inputs, the forward modeling results clearly show seven layered subsurface with densities of 2750, 2550, 2810, 2580, 2850, 2500 and 2780 kg/m³ from the top to the deeper layers, respectively. These horizons are identified as: young basalts, silicic lavas, Pliocene Bofa basalts, silicic products and rift floor ignimbrites, flood basalts and rhyolites, Mesozoic sediments and metamorphic crystalline basement rocks (Fig. 13 A, B, and C).

The minimum Root Mean Square (RMS) errors obtained for the forward models (Fig. 13A, B, and C) of magnetic susceptibility and density were found to be 17.084 and 7.9 for profiles A-A′, 17.069 and 5.57 for profiles B–B′, and 18.3 and 6.7 for profiles C–C′, respectively. The inverse models were used to treat a variety of anomalies by considering two profiles across WFB that were parallel to each other, as well as one along the NW-SE direction.

As shown in profile A-A' (Fig. 13A), the flood basalts are relatively thicker along the entire profile. Moreover, the crystalline basement is observed to be raised basement morphology, up to a depth of 3 km, near Salen ridge. This may indicate the presence of a geothermal heat source at shallow depth. Near the areas encompassing the WFB system of faults and Salen ridge, the silicic ignimbrites and Bofa basalts are very thin; interspersed with and a number of faults. The faults that are found near WFB and Salen ridge may represent the geothermal source's outflow region (thin upper cap). Maximum accumulation of the Bofa basalts and flood basalts are observed along the profile between 0 and 10 Km. At a depth of approximately 3.5–4.5 km below the surface, the Mesozoic sediments are more densely concentrated close to the eastern escarpment. A formation of obsidian (rhyolite) rocks at the shallow surface was seen on the eastern escarpment of profile AA' and near Salen Ridge.

Profile B–B' (Fig. 13B) also shows reduced values of magnetic anomaly and substantially higher density values surrounding WFB and Salen ridge, in contrast to that obtained for profile A-A'. When one moves along the profile in an east-west direction, the depth of the crystalline basement is randomly decreasing to 3 Km. Beneath the WFB, a thick Bofa basalt layer and thin horizons of flood basalt and Mesozoic sediment are observed. This may suggest that Bofa basalt acts as an upper cap for the geothermal potential of the subregion. Silicic domes and flows (including obsidians and complex volcanoes) accumulated close to Giano Mount and the Eastern Escarpment. A somewhat narrow dike may also be seen in the Giano volcanic region, approximately 7 km west-east along the profile. The formation of a thin layer of Bofa basalt and young basalt beneath Giano Mountain may form a thin upper cap to allow the movement of magmatic material into the subsurface through faults. The crystalline basement is seen to be elevated up to 2.5–3 km to the surface beneath WFB. This could be a sign that magmatic segments are rising to the surface and becoming heat sources for geothermal energy.

The WFB, the Gnaro obsidian area, Salen ridge and Rime Mountain are all traversed by profile C–C' (Fig. 13C) as one moves southeast-northwest. The crystalline basement is elevated up to 3 km below Gnaro obsidian dome, and Salen ridge, similar with the previous two profiles. Such geological formation may indicate the presence of geothermal energy in the near subsurface around the indicated area. The six horizons mapped above the crystalline basement in this area are comparatively thin and have thickness between 3.5 to 4 Km. Unlike the previous two profiles, (Fig. 13 A and B), which had raised basement morphology concentrated under the WFB and Salen ridge, the metamorphic basements deepened to 4–5 Km beneath WFB and eastern edge. The easternmost region of the profile C–C′ is dominated by the silicic product and Siliceous domes (results in good agreement with that of Tadesse et al. [101].

5. Conclusions

Satellite gravity and ground based magnetic data were analyzed to characterize the Tulu Moye geothermal prospect in terms locating areas of shallow heat source, depth to these sources, the subsurface lithological stratification and the major structures that could play a role in the movement of the heat source to shallow depths. The role of active tectonics associated with the Wonji Fault Belt (WFB) system of faults, volcanic centers of Tulu Moye, Bora and Berecha and the nearby Tulu and Gedemsa calderas are also examined.

The data have been analyzed and presented in the form of anomaly plots; regional-residual separated gravity and enhanced magnetic anomaly maps. Depths to the potential heat source were determined using the radially averaged power spectrum, 3D Euler depth and 2D Euler deconvolution techniques. Joint forward modeling of the gravity and magnetic data were done to determine the subsurface stratigraphy of the rock units and map the presence and location of major geological structures. The results have been interpreted incorporating knowledge of the local geology and tectonics as well as field observation of geologic structures and thermal manifestation centers.

In conclusion, interpretation of the results shows large residual and regional Bouguer gravity anomalies and the high RTE magnetic anomaly (high rock demagnetization) that coincide at the center of the study area could point to the presence of a hot body at shallow depth. The presence of a magmatic body at shallow depths in the region, on the other hand, may suggest that it is the primary heat source for a potential geothermal reservoir. Further, the presence of the crystalline basement at shallow depth in the 2D joint gravity/magnetic forward models, the absence of magnetic/gravity signal in the 2D Euler deconvolution plots, and the high magnetic anomaly in the RTE anomaly plot considerably match; results in agreement with Samrock et al. (2018), all indicating towards the high geothermal potential of the area.

Over all, the current study demonstrates that: (1) the fact that the high RTE magnetic anomalies are primarily concentrated in the central region of the research area indicates that the area is underlain by a shallow heat source; (2) the heat source of the geothermal reservoir is highly likely to be a shallow magmatic body; (3) the extended raised basement morphology of the crystalline basement is oriented along northwest-southeast direction oblique to the WFB (to NE-SW) direction, (4) beneath Salen ridge and the WFB, the basaltic and silicic volcanic horizons are very thin and the deep-seated regional faults possibly serve as the main conduits for the passage of hot fluid to the surface, and (5) the combined geophysical results together with geological set up suggest a broad caldera comprising the TMBB volcanic centers.

Availability of data and materials

The gravity datasets generated and/or analyzed during the current study are available at the International Gravimetric Bureau (BGI) (Via an available website, http://ddfe.curtin.edu.au/gravitymodels/GGMplus/), and the magnetic data that has been used is available on request.

Human and animal rights

This research did not involve human participants and/or animals.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not for profit sectors.

CRediT authorship contribution statement

Hilemichaeil Samson: Writing – review & editing, Writing – original draft, Supervision, Software, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Tigistu Haile: Writing – review & editing, Validation, Supervision, Software, Formal analysis. Gezahegn Yirgu: Writing – review & editing, Validation, Supervision, Formal analysis.

Declaration of competing interest

The authors declare that this manuscript is original and has no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Further, we declare that it has not been published before and is not currently being considered for publication elsewhere.