Pinpointing Deep Geothermal Upflow in Zones of Complex Tectono‐Volcanic Degassing: New Insights from Aluto Volcano, Main Ethiopian Rift

Active rifts release large amounts of gases from deep sources to the atmosphere by advection and diffusion processes along permeable fracture zones. The objective of this study is to develop geothermal exploration concepts for areas with little or no hydrothermal surface expressions suitable for fluid sampling and analyses (e.g., hot springs, geysers, and fumaroles). In such areas, soil gas surveys can complement established geophysical and geochemical exploration. We report CO2, 222Rn (radon), and 220Rn (thoron) emission data and ground temperatures from the Aluto volcanic complex in the Main Ethiopian Rift to improve understanding of tectonic and volcanic controls on the existing geothermal system. This suite of gas emission measurements allows us to identify major, deep‐rooted permeable structures with active fluid circulation and identify suitable drilling targets for geothermal production wells on Aluto. We show that significant differences in gas signatures (i.e., efflux and spatial pattern) can be used to identify predominantly volcanically and/or tectonically influenced compartments. Major gas emissions indicate significant fluid circulation at depth, which is typical for magmatic systems. Such high gas emissions have been observed in areas affected by major tectonic structures interacting with magmatic bodies at depth (tectono‐volcanic). Predominantly fault‐controlled sectors also show hydrothermal fluid circulation, but to a lower extent compared to tectono‐volcanic sectors. Within the Aluto volcanic complex, geothermal production wells mainly target such fault‐controlled domains, whereas results of the study indicate strongest fluid circulation in tectono‐volcanic sectors. This result should be considered for the future exploration and development strategy of the site.


Introduction
The East African Rift System has numerous high-temperature geothermal fields, which offer huge potential for power generation and opportunities for economic development; however, the heterogeneity and complexity of the highly fractured geothermal reservoirs pose major difficulties for their exploration (Younger, 2014). Conventional geophysical and geochemical exploration technologies are useful to prove a resource and understand the large-scale architecture of geothermal systems but do not pinpoint preferential locations for targeting expensive production wells into most permeable pathways. Additionally, geophysical data (e.g., resistivity data) often cannot distinguish active from paleo-hydrothermal activity as the resistivity structure will be preserved in the subsurface, even if systems are cooling (Bibby et al., 1992).
For a successful exploration strategy in these environments, information on the system's current activity and fault permeability (e.g., upflows and outflows) is required at a high spatial resolution. The best geochemical ©2019. The Authors. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.
Main Ethiopian Rift is reported • Volatile emissions focus along volcanic and tectonic structures and allow mapping of deep gas sources across the Aluto volcano, Ethiopia • Emissions increase toward the volcanic centre, implying a degassing magmatic body Supporting Information: • Supporting Information S1 • Figure S1 • Figure S2 • Figure S3 • Figure S4 approach for obtaining such high spatial resolution observations are soil gas surveys (e.g., CO 2 , 222 Rn, and 220 Rn), which help to identify the most permeable structures with highest hydrothermal fluid flow (Jolie, Klinkmueller, et al., 2015;Neri et al., 2016). In this paper we prove the suitability of these techniques for geothermal exploration projects.
Major faults in the East African Rift System are controlled by tectonic and volcanic mechanisms. Surface gas release and hydrothermal fluid circulation in active rift systems not only are related to its volcanic heat source but also reflect local and regional tectonic controls in the subsurface (Brune et al., 2017;Goerner et al., 2009;Jolie et al., 2016). Increased degassing along fracture zones outlines deep-rooted permeable segments of active faults and points to major upflow and outflow zones of hydrothermal fluids. Depth of faults is at least 5 km, but no distinct structure could be identified below 8 km based on Magnetotelluric (MT) studies by Samrock et al. (2015). In Tule Moye, a comparable site to Aluto, faults channel magmatic fluids from a depth of~14 km to the surface (Samrock et al., 2018). Fault segments with strong gas emissions are characterized by increased slip and dilation tendencies Jolie et al., 2016). Analyses of a variety of soil gases have proven their suitability for assessing the tectonic and volcanic controls on geothermal systems (Hunt et al., 2017;Jolie, Klinkmueller, et al., 2015;Jolie et al., 2016;Lee et al., 2016). These gases provide vital information for the exploration of geothermal fields and help in targeting wells in the most permeable structures with the greatest flow rates of high-temperature hydrothermal fluids. Soil gas surveys do not require heavy equipment and can be performed in a relatively short time even in some of the most difficult terrain, where other technologies are limited. Many techniques even allow in situ analysis in the field (e.g., accumulation chamber method, alpha spectroscopy, and gas chromatography), which gives much flexibility in the planning of field campaigns (Jolie, Klinkmueller, et al., 2015).
Aluto offers an ideal location to address some of these issues and understand the architecture of a rift-related alkaline, fault-controlled geothermal system. The volcanic complex has been a target for geothermal exploration and exploitation for several decades (Gebregzabher, 1986;Gianelli & Teklemariam, 1993;Gizaw, 1993;Hochstein et al., 2017;Teklemariam et al., 1996;Valori et al., 1992) as well as recent geological, geochemical, and geophysical investigations (Biggs et al., 2011;Braddock et al., 2017;Hutchison et al., 2015;, Hutchison, Fusillo, et al., 2016Iddon et al., 2018;Nowacki et al., 2018;Saibi et al., 2012;Samrock et al., 2015;Wilks et al., 2017). Although Aluto is wellstudied compared to many of Ethiopia's geothermal prospects, major uncertainties remain about how geothermal fluids are distributed in the subsurface and how they ascend along the mapped fault zones. Answering these questions has implications for understanding Aluto's geothermal system and for characterizing the fault network that might be exploited by future dike intrusions and volcanic events (Hutchison et al., 2015); more generally, it can provide important lessons on how to exploit the complex, highly-fractured alkaline geothermal systems that typify the East African Rift System.
In this study, we show the results of a multi-method approach that focuses on the identification and characterization of degassing fluids that ascend from a magmatic system along tectonic faults (tectono-volcanic interaction). Tectono-volcanic interactions have an effect on soil degassing within the Aluto volcanic complex and result in more intense gas emissions. Our approach consists of a combination of systematic alpha-spectroscopic measurements as well as CO 2 efflux measurements based on an area-wide, regular grid (Jolie, Klinkmueller, et al., 2015;Jolie et al., 2016). Results illustrate that substantial information on permeable subsurface structures and lateral outflows can be obtained by the analysis of surface degassing. This adds more details to the results of conventional structural mapping at the surface, which is often limited in particular in areas with fast geomorphological changes due to strong erosional activity. Soil gas surveys provide useful solutions for a comprehensive analysis of geothermal systems and the identification of suitable areas for geothermal development.

Regional Geology-The Main Ethiopian Rift
The~600-km-long Main Ethiopian Rift (MER) is an actively extending magmatic rift that forms the northernmost segment of the East African Rift System (Corti, 2009). The MER shows clear along-axis variations in rift morphology (Molin & Corti, 2015), tectonics (Agostini et al., 2011), and crustal structure (Keir et al., 2015;Maguire et al., 2006) and is usually divided into northern, central, and southern sectors (Figure 1a).
Rift magmatism has led to surface volcanism across the MER and Quaternary-Recent volcanoes are mostly concentrated along the rift axis (Di Paola, 1972). Although the styles and geochemistry of MER volcanism is diverse , the most attractive geothermal prospects are hosted by chemically evolved, silicic shield volcanoes and calderas, for example, Corbetti, Aluto, or Gedemsa (Kebede, 2012;Samrock et al., 2018). Satellite remote sensing surveys (e.g., Biggs et al., 2011;Lloyd et al., 2018) have highlighted episodes of ground deformation at a number of these silicic volcanoes (e.g., Aluto and Corbetti). These observations provide compelling evidence for subsurface fluid movement, although contrasting models have been proposed to explain the causes of deformation, including magmatic intrusions, gas and magmatic fluid pulses, and clay swelling Samrock et al., 2015). Despite this uncertainty, there is a consistent view that pre-existing structures (both tectonic and volcanic in origin) play an important role in controlling the movement of magma, hydrothermal fluids, and gases at these volcanoes (Goerner et al., 2009;Hutchison et al., 2015;Lloyd et al., 2018).

Aluto Geology and Geothermal Field
Aluto volcano, the focus of this contribution, is located in the central MER (Hutchison et al., 2015, Hutchison, Fusillo, et al., 2016 and is Ethiopia's only exploited geothermal field (Kebede, 2012). Its geological and eruptive history has been the focus of work by Hutchison et al. (2015), , Hutchison, Fusillo, et al., (2016) who have shown that the complex was initially built up as a trachytic shield before undergoing a series of large caldera-forming eruptions between 320 and 300 ka. Since~60 ka, Aluto has been in a post-caldera phase typified by the eruption of small volume (<250 × 10 6 m 3 ) pumice cones and obsidian domes across the complex. The key structural feature of Aluto is a NNE-SSW (Wonji-aligned) fault known as the Artu Jawe Fault Zone (AJFZ; Hutchison Figure 1. (a) Regional topographic map. The black lines represent major plate boundaries, while pink triangles mark all Holocene volcanoes (after Siebert & Simkin, 2002). The Main Ethiopian Rift (MER) is divided into northern (NMER), central (CMER), and southern (SMER) segments. Aluto volcano, the focus of this study, is located in the CMER and is shown in red. White arrows show current extension vectors relative to a fixed Nubian Plate (after Saria et al., 2014). In (b), a detailed map of the Aluto volcanic complex with a summary of the geology, hydrothermal manifestations, and degassing studies is shown (coordinate system WGS 1984 UTM Zone 37N). The latest version of the geological map of Aluto is available in . The large blue arrow shows the presumed inflow of cool groundwater to the geothermal reservoir (after Braddock et al., 2017;Darling et al., 1996), while the smaller orange arrows show the shallow outflow of hot geothermal fluids that feed fumaroles and hot springs on the West and South of the complex. In (c), a detailed map of the study area including all sampling sites for 222 Rn, 220 Rn, CO 2 , and T S is shown. et al., 2015), and deep well observations suggest that the AJFZ predates the first eruptions of the silicic complex. Young obsidian lava flows seem to be co-located with this major structure and represent a record of the most active structures across Aluto, which channeled magma during eruptive periods and now channel magmatic gases. The AJFZ together with an elliptical ring fault, assumed to have developed during Aluto's caldera collapse, strongly influences magma ascent and also direct hydrothermal fluids toward the surface. Geothermal surface manifestations (e.g., steam vents, alteration, and warm ground) are widely distributed within the volcanic complex, but strong fumaroles and boiling hot springs only occur outside the complex (Figure 1b).
Aluto's geothermal field has also been the focus of numerous surface and deep well investigations (e.g., Gianelli & Teklemariam, 1993;Gizaw, 1993;Teklemariam et al., 1996). Observations from the wells have been used to infer that the main part of the geothermal reservoir is >2,000 m beneath the surface and is sealed by a sequence of intensely altered basalts and tuffs that form a low-permeability, low-resistivity clay cap at~500-1,500 m (Samrock et al., 2015). The geothermal fluid appears to be primarily sourced from rainfall on the rift margin (Darling et al., 1996;Rango et al., 2010) rather than the surrounding lakes, and there is good evidence from resistivity soundings for major outflows to the South and West (Hochstein et al., 2017). This and the high temperatures indicated by water chemistry, mineralogy, and downhole temperatures (>300°C; Gizaw, 1993) all suggest that the AJFZ represents a major upflow zone from the deep geothermal reservoir toward the surface. To date, geothermal production wells only target the AJFZ, specifically its Northern Sector ( Figure 1).
Previous studies of Aluto's gas emissions (Hutchison et al., 2015 focused mainly on CO 2 and involved a couple of transects of major structures and a small-scale study of the main AJFZ area (Figure 1). These results supported findings from earlier geothermal investigations that identified the AJFZ as the key structure for fluid upflow and geothermal utilization. Hutchison et al. (2015) also found significant anomalous degassing in other areas of the complex (specifically around the caldera rim and ring fault on the east of the complex, Bobesa; Figure 1). Owing to the limited number of sample sites and the focus on CO 2 data, major uncertainites remain about these degassing regions as well as the upflow along the AJFZ. In particular, it is not clear whether (1) all mapped faults are connected to a deep geothermal reservoir; (2) fluid circulation extends along NNE-striking fault zones beyond the boundaries of Aluto volcanic complex; (3) fault intersections provide enhanced permeability and hence good targets for future exploration; and (4) there is any evidence for lateral fluid outflow. By extensive and systematic soil gas measurements (involving a wide range of parameters, CO 2 , 222 Rn, and 220 Rn), we aim to address these challenges by assessing the variability of gas emissions along fracture zones and identifying further permeable but less prominent or maybe even hidden structures.

Methods
A large-scale soil gas survey was performed in May 2016 covering a significant area of the two dominating structures (i.e., Artu Jawe and Ring Fault; Figure 1c) as well as the area in between the two structures (referred to as "Central Area"), building up on previous efforts by Hutchison et al. (2015), . A second small-scale survey was performed in early 2017 to extend the survey area to the North and Northwest for a comprehensive understanding of the NNE continuation of the AJFZ and the continuation of the Ring Fault to the West. The NE-SW extension of the study area is 5.7 km with a maximum NW-SE extension of 3.5 km. The following parameters were determined during the field survey-CO 2 efflux by the accumulation chamber method, 222 Rn and 220 Rn activity concentration by alpha-spectroscopic measurements, and soil temperatures (T S ) at 50 cm depth (Table 1; Fridriksson et al., 2016;Jolie, Klinkmueller, et al., 2015). The accumulation chamber method is a technology where a chamber is placed on the ground and the increasing CO 2 concentration in the chamber is measured (2-3 min) by an infrared gas analyzer (LI-COR 820, accuracy <3% of reading). The increase in CO 2 concentration correlates to the CO 2 efflux at the interface from geosphere to atmosphere. Throughout the survey, we used a WEST Systems Type A accumulation chamber (WEST Systems, 2019). Temperature measurements have been performed using a Greisinger GMH 285-BNC thermometer with Pt1000 sensor in a 620-mm-long stainless steel probe (accuracy ±0.1 K). 222 Rn and 220 Rn activity concentrations are determined by a radiometric measurement (~15 min), which determines their short-living radon daughter nuclides 218 Po and 216 Po from soil 10.1029/2019GC008309 Geochemistry, Geophysics, Geosystems gas samples collected 1 m below the surface (Jolie, Klinkmueller, et al., 2015) and pumped to the analyzer. Measurements have been performed by a Sarad RTM 1688 and RTM 2200 using a high-voltage measuring chamber with electrostatic precipitation of ionized nuclides on the surface of a 2 cm 2 semiconductor silicon detector (accuracy: 3 counts/min).

Radon Versus Thoron
Radon ( 222 Rn) and thoron ( 220 Rn) activity concentrations have been used for the assessment of hydrothermal upflow zones in geothermal resources (Jolie, Klinkmueller, et al., 2015). 222 Rn activity concentrations above and in the vicinity of hydrothermal upflow zones usually result in increased gas concentrations. In general, measurable activity concentrations of both parameters depend primarily on (1) the depth of the uranium/thorium source in the subsurface, (2) the fluid migration velocity from the subsurface to the surface, and (3) the different decay velocities for 222 Rn (T 1/2 = 3.8 days) and 220 Rn (T 1/2 = 55.6 s). This means that in areas with fast fluid migration velocities from the geothermal reservoir to the surface 220 Rn values can also reach peak values. Assuming that the approximate source for uranium and thorium is at a constant depth level throughout the study area, different fluid migration velocities toward the surface have a strong impact on the 222 Rn-220 Rn ratios (RTR). In practical terms, it means that multiple RTRs can indicate geothermal upflows from depth, instead of one specific ratio. This requires a careful interpretation. A 222 Rn-220 Rn plot from the five sectors with the main degassing features allows one to identify major upflow zones. During the migration of fluids to the surface, a lot more 220 Rn than 222 Rn is lost due to decay as a result of its short half-life of 55.6 s, compared to 3.8 days for 222 Rn. This effect causes significantly increased RTRs at medium fluid migration velocity. However, if the transport velocity is high enough, both 222 Rn and 220 Rn will result in maximum values and consequently intermediate RTRs. Conversely, shallow uranium/thorium sources could also result in increased 222 Rn and particularly high 220 Rn activity concentrations (low RTRs); however, this would require high-temperature geothermal activity for fluid-rock interaction, which mainly occurs in deep geothermal reservoirs. In that sense, it is possible to differentiate between (1) deep sources/medium upflow velocity (high RTR: 6.7-33), (2) deep sources/fast upflow velocity (intermediate RTR: 1.5-3.9), and (3) shallow sources (low RTR: 0.3-0.7).

Soil Gas Fingerprint
Fingerprint analyses are widely used in geochemistry, but fingerprints for multi-parameter soil gas datasets are still a novel processing and visualization technique and presented herein for the first time. Soil gas fingerprints give a more comprehensive understanding and classification of the spatial variability of gas emissions by plotting maximum (mean) values of each analyzed parameter from selected compartments of a study area, normalized to the maximum (mean) values observed in the entire study area. These compartments define themselves by particularly high or low soil gas emissions and other geological or hydrological conditions (e.g., terrain, known faults, rivers, and geothermal surface manifestations). In some cases, multiple scenarios for compartments need to be defined and tested for one area. The method allows fast comparison of spatial trends and differences for an unlimited number of parameters. Soil gas fingerprints can be determined for any specific area and aerial extent. Ongoing research activities will further improve this novel approach for a fast assessment and comparison of gas emissions from geothermal fields by including more Note. The maximum encountered CO 2 efflux value was above the detection limit (DL) of the device and therefore set to the maximum measurable CO 2 efflux value.

10.1029/2019GC008309
Geochemistry, Geophysics, Geosystems parameters. Based on this technique, we aim to differentiate deep from shallow sources, tectonic from volcanic degassing, and other characteristics. In this study, we present a soil gas fingerprint analysis for the key structures.

Results
The two major structures in the study area (i.e., Artu Jawe and Ring Fault) have been subdivided (Figures 1c  and 2). Artu Jawe Fault was divided into Northern (A1) and Southern AJFZ (A2), where higher gas emissions were determined. Along the Ring Fault, two major degassing clusters were identified; therefore, the area was subdivided into R1 and R2.

Carbon Dioxide
The CO 2 efflux dataset was statistically separated into different classes based on two major inflection points (CO 2 efflux: 370 and 8,500 g·m −2 ·day −1 ) in the probability plot (Figure 3). Peak values above 8,500 up to 25,084 g·m −2 ·day −1 (detection limit of the device) are illustrated as points (Figure 2). All values below 350 g·m −2 day −1 are considered as background efflux. A more detailed assessment of the background population identified two further inflection points at 35 and 160 g·m −2 ·day −1 , splitting the background into two distinct populations. Values below 35 g·m −2 ·day −1 are likely to be related to a biogenic source with a mean of 24.9 g·m −2 ·day −1 , whereas values up to 160 g·m −2 ·day −1 and above are considered as volcanic background values (Hutchison et al., 2015).
Increased CO 2 efflux values have been measured along Artu Jawe and the Ring Fault, as well as in one well-defined zone in the Central Area (Figure 2). Peak CO 2 efflux values (>8,500 g·m −2 ·day −1 ) occurred only along the Ring Fault (R1 + R2) and the Southern AJFZ (A2), but not in Northern AJFZ (A1). A strong segmentation of the degassing pattern was observed along the AJFZ, which resembles partly the location and dimension of volcanic eruption craters (A1).
The Central Area (C) hosts two areas with increased degassing rates. A larger anomaly (C1) is located in the Northern part between Artu Jawe and the Ring Fault (referred to as Bobesa) and correlates with increased 222 Rn values (without geothermal surface manifestations). Another small anomaly (C2) sits further Southeast where increased T S and 222 Rn were measured (with geothermal surface manifestations).
Existing CO 2 efflux data from 2012 to 2014 and 2016 indicate stable degassing features in space and time for a number of years ( Figure 4). This area also represents the transition zone from a structurally dominated segment in the North to a volcanically dominated segment in the South.

Soil Temperature
Thermal anomalies occur mainly along Artu Jawe (A1 + 2) and the Ring Fault (R1 + 2), plus two areas (C2 + C4) in the eastern Central Area (i.e., steaming ground; Figure 2). C2 seems to be connected to R2. Highest T S was measured in Southern Artu Jawe and the Ring Fault (R2). Soil temperatures along the Northern AJFZ do not exceed 75°C. A wide area with slightly increased soil temperatures (up to 32°C) was identified in the northern part of the study area.

Radon
The 222 Rn activity concentration dataset was statistically separated into different classes based on major and minor inflection points in the probability plot (Figure 3). Peak activity concentrations (>120.4 kBq·m −3mean of PIII) occur along the Southern AJFZ with decreasing values toward the North (Figure 2). Peaks in the northern sector are located within volcanic eruption craters and appear as elongated anomalies. All values below 15.5 kBq·m −3 are considered as pure background. Values between 15.5 and 50.1 kBq·m −3 are expected to be a mixture of background and peak values.
The map (Figure 2) illustrates a segmented, fault-controlled pattern of increased 222 Rn activity concentration along Artu Jawe (mainly characterized by dextral offsets). Compared to CO 2 efflux and T S , 222 Rn gives well-defined anomalies. Two major anomalous, but rather circular areas are located along the Ring Fault (R1 + R2). Three distinct anomalous areas, but with lower magnitudes, were identified in the center (C1 + C2 + C3). The anomaly to the East (C2) coincides with increased soil temperatures. In the northern part of Artu Jawe at the intersection of the Ring Fault and the AJFZ 222 Rn values and T S have only 10.1029/2019GC008309 Geochemistry, Geophysics, Geosystems

10.1029/2019GC008309
Geochemistry, Geophysics, Geosystems increased minimally. East of the intersection, slightly increased soil temperatures have been measured along a NNE-oriented zone. Anomalous 222 Rn values also appear beyond the intersection of the Ring Fault and AJFZ to the NNE in continuation of the AJFZ.

Thoron
The 220 Rn activity concentration dataset was separated into different classes based on major and minor inflection points in the probability plot ( Figure 3). Additional class breaks have been defined for a finer discretization of the map (Figure 2). The majority of peak 220 Rn values (>71.8 kBq·m −3 ) was measured at Bobesa. This anomaly extends to the West with decreasing values. Another peak 220 Rn cluster occurs in the Southern AJFZ and some medium 220 Rn values (17.6-71.8 kBq·m −3 ) in Northern AJFZ. Activity concentration along Artu Jawe is increasing from the North toward the South. All values below 17.6 kBq·m −3 are considered as background.

Along Fault Variations
A profile of gas emissions along the AJFZ (X-X′; Figure 1) is shown in Figure 5. Fault-controlled emissions occur over a broad zone orthogonal to the fault (Figure 2), and so we filtered the emission data so that data within a fixed distance of the line were included in the plot (this is referred to as the buffer zone). For 222 Rn, 220 Rn, and T S , a buffer zone of 500 m either side of the profile line was used, while for CO 2 a narrower buffer zone of 300 m was used. The narrower buffer zone for the CO 2 data was selected because data density is greater (as we included the high-spatial resolution survey of Hutchison et al., 2015; Figure 4). It is important to note that changing this buffer zone does not significantly affect the results and interpretations. We also calculated a running average ( Figure 5, grey lines), by binning the data in 1 km segments along the fault so that we could identify whether there was any broad (kilometer-scale) change in emissions.
The along-fault plot ( Figure 5) illustrates the trend of increasing gas emissions from the Northern AJFZ toward the Southern AJFZ. This correlating trend is reflected in all parameters (CO 2 , 222 Rn, 220 Rn, and T S ). A possible explanation for this is the combination of volcanic (vicinity to a magmatic volatile source) and tectonic (structural permeability) impacts in Southern AJFZ, whereas the northern part of Artu Jawe is mainly tectonically influenced (permeable, but distant from magmatic volatile source). For both areas, gas emissions are highly variable over short (1-100 m) length scales, which can be explained by a fault segmentation, non-optimal orientation of the fracture zones in the acting stress field, sealing effects, and multiple upflow zones of geothermal fluids.

Soil Gas Fingerprint
A first attempt of a soil gas fingerprint analysis is presented in Figure 6. For each of the five analyzed key compartments (A1, A2, R1, R2, and C), a stacked column is plotted illustrating the maximum and mean values for 222 Rn, 220 Rn, CO 2 efflux, and T S normalized to the maximum value in the entire study area. Such plots allow a fast spatial evaluation of the collected data. Here the significant differences in gas emissions originating from A2 and R2 versus A1, R1, and C can be recognized.

Radon-Thoron and Radon-Thoron-Carbon Dioxide Plots
Figure 7a reveals a tendency in all sectors toward increased RTRs (6.7-33) at different magnitudes for 222 Rn and 220 Rn (see also supporting information S1-S4). This indicates the presence of a significant hydrothermal upflow along permeable fracture zones from a deep source at medium fluid migration velocities. The main uranium and thorium source at Aluto is expected to be Neogene ignimbrites (below 1,400m depth), which are also hosting the geothermal reservoir where intense fluid-rock interaction occurs (Gizaw, 1993;. Stratigraphic data from well logs show a uniform distribution of the ignimbrites across the study area (e.g., Teklemariam et al., 1996). Due to the lack of chemical information from well cuttings, data from analogue outcrops have been used to assess the chemical composition of the reservoir rocks. Analyses of rock samples from Aluto ignimbrites and basalts have shown that the ignimbrites are much richer in uranium and thorium than basalts (Ignimbrites: 10-20 ppm thorium and 2-4 ppm uranium; basalt: 2-5 ppm thorium; and 0-1 ppm uranium; Teklemariam, 1996). RTRs from 1.5 to 3.9 indicate a major hydrothermal upflow from a deep source and maximum fluid migration velocities. These ratios have been identified only in A2 and R2. Lower 222 Rn and 220 Rn magnitudes in A1 and R1, but peak RTRs, are interpreted as dilution effects. Whereas A1, R1, and R2 indicate the presence of deep and shallow sources, no shallow source is indicated for A2, which highlights the presence of a major hydrothermal upflow from a deep source. Selected maximum 222 Rn and 220 Rn values ( 222 Rn > 149.6 kBq·m −3 ; 220 Rn > 80.7 kBq·m −3 ) with a trend toward a deep source and maximum fluid migration velocities (RTR 1.5-3.9) are highlighted on the map (Figure 8) and illustrate major upflows from a deep source along the Southern AJFZ (A2) and the Ring Fault (R2).
The interpretation of the interpolated maps suggests a correlation of maximum 222 Rn, 220 Rn, and CO 2 emissions in most of the geothermally active areas. Based on a ternary plot (Figure 7b), this obvious correlation was further analyzed. The principle of this concept is to identify sites with simultaneous peaks from all three parameters, even if the actual magnitudes are relatively low. As a novelty, even for areas characterized by low-intermediate gas emissions, signals from deeper sources could be identified with this approach as illustrated in Figure 7b. CO 2 efflux values were extracted from the interpolated map based on the 222 Rn-220 Rn sampling grid. This extraction procedure was necessary due to different sampling grids for the two methods. For 222 Rn and 220 Rn, inverse values are plotted, so that the peak values from all parameter plot in the same corner of the ternary plot. A cluster of data points from all compartments becomes obvious in the CO 2 -corner of the ternary plot. These values (CO 2 > 84%) have been extracted (manual selection) and plotted on a map (Figure 8). Many of the selected data points are constrained to areas with major  (Figure 2). For that reason, data within a fixed distance orthogonal to the fault line was included in the plot (this is referred to as the buffer zone). We also calculated a running average by binning the data in 1 km segments along the fault to identify any broad (kilometer-scale) changes in emissions.
geothermal anomalies. However, in addition to the well-known geothermal areas, this approach also highlights data points in the central part, which usually appeared less prominent and were therefore often considered as less important for structural permeability. Based on the ternary diagram, we found evidence for the existence of permeable structural pathways in the Central Area (Figures 7b and 8).

Main Degassing Structures
In all investigated sectors, upflowing geothermal fluids are indicated (Figure 2). The pattern analysis of soil gas emissions implies (1) elongated upflows along permeable fracture segments and (2) locally concentrated upflows. Although Hutchison et al. (2015),  identified the major structures, this new work clearly identifies and characterizes the major permeable structures, as well as the strong spatial variation of gas emissions.

Artu Jawe
The spatial pattern and variation of gas efflux and soil temperature suggest a strong fault segmentation with dextral displacements of Artu Jawe. Increased values of CO 2 efflux, 222 Rn, 220 Rn, and T S have been observed along multiple well-confined NNE-oriented segments and within areas of maximum displacement, indicating major migration pathways of geothermal fluids. In the northern part of Artu Jawe, a significant dextral displacement of about 300 m creates permeability evidenced by increased CO 2 emissions ( Figure 2; CO 2 map; marked as "X"). Maximum degassing rates occurred mainly within cluster of aligned  Rn versus 220 Rn plot for data from A1, A2, R1, R2, and Central Area (Rest) as depth indicator (see also supporting information S1-4); (b) 222 Rn-220 Rn-CO 2 ternary plot. The plot illustrates a cluster of data points with increased CO 2 , 222 Rn, and 220 Rn values. and elongated volcanic eruption craters, which have not been observed in the southern part of Artu Jawe (A2). Geothermal activity further North of the Aluto volcanic complex is decreasing. Peak CO 2 efflux and 222 Rn occurred in the Southern AJFZ, indicating a connection to the deep geothermal reservoir, but their spatial pattern appears to be less defined compared to the North. In contrast to the northern part, geothermal surface activity is indicated to continue Southward as can be seen by the presence of hot springs on the northern shore of Lake Langano in continuation of the structural trend of the AJFZ. Only minor surface alteration was observed in the North compared to the South. Less permeable segments could be affected by (1) silicic mineralization (sealing) of previously permeable fault segments (Teklemariam & Beyene, 2001) or (2) a nonpreferential fracture orientation in the acting stress field causing low structural permeability and hence less fluid circulation.
Along-fault variations on short length scales (100 m) and longer scales (kilometer; running average) are illustrated in Figure 5. Short length scale variations indicate fault segmentation and variable structural permeability, whereas long scale variations indicate regional trends, such as a general increase of gas emission to the South of AJFZ indicating a major source of magmatic volatiles.
The pattern of increased CO 2 efflux and 222 Rn values is not always congruent in the northern part of Artu Jawe. A possible explanation for this is the maturity of the geothermal manifestations. It is assumed that very young manifestations release mainly CO 2 in its initial phase. CO 2 from exsolution processes in magmatic sources travels to the surface once the magma is emplaced, for example, in dike intrusions. After some reaction time of hot and CO 2 bearing fluids with the reservoir rocks, more and more 222 Rn is released, which is ultimately resulting in a combination of increased 222 Rn and CO 2 . Surface manifestations in its extinction phase seem to emit mainly 222 Rn and less CO 2 . This could be explained by ongoing reaction processes of reservoir fluids and rocks, but an insufficient supply of new magmatic CO 2 . 5.1.2. Ring Fault CO 2 efflux, T S , and 222 Rn anomalies have been detected in two key areas (Figure 2) on the western flank of the Ring Fault (Bobesa), but not on the rim itself. It is discussed if these surface manifestations should be

10.1029/2019GC008309
Geochemistry, Geophysics, Geosystems considered as solely Ring Fault related features (volcanic) or if they could have evolved as secondary features due to an interaction of the Ring Fault with additional (so far unknown) NNE-oriented tectonic structures (tectono-volcanic) parallel to the NNE-oriented AJFZ.
The area to the Southwest of the Ring Fault (C2) is characterized by widespread soil alteration, a large 220 Rn anomaly with decreasing values toward the West, and slightly increased T S . This area seems to be largely affected by a geothermal outflow from Bobesa to the West. The source of the major degassing vents seems to be deep-rooted (Figure 2). Soil gas emissions show similar characteristics as A2.
At the inferred intersection of the Artu Jawe Fault and the Ring Fault, a nested eruption crater (~580 × 300 m) is located, where intermediate, but no maximum 222 Rn and 220 Rn values have been measured. CO 2 efflux values and T S also do not show significantly higher values at the inferred intersection. In conclusion, observed gas emissions do not suggest an eminent permeability structure with a connection to a deep reservoir, although measured gas emissions are above background, indicating some minor fluid circulation.

Central Area
Magnitudes of gas emissions range from low to intermediate with a few peak values. Therefore, a hydrothermal system with similar characteristics as A2 or R2 is not expected in that area, though the presence of hydrothermal fluid circulation was proven with our approach. Wide-ranging CO 2 efflux (C1 and C2) and 222 Rn activity concentrations (C1-4) of low-intermediate magnitudes were identified in the Central Area ( Figure 2). Geothermal surface activity was observed in C2 and C4. Increased CO 2 efflux in C1 can be interpreted as (1) a potential eastward directed shallow, lateral outflow from Artu Jawe (similar to a perched aquifer), and/or (2) increased structural permeability along an unknown NNE-oriented fracture zone. A biogenic source for these emissions is unlikely since biological sources (with isotopically light δ 13 C;  show significantly lower fluxes and will not correlate with the high 222 Rn in the same area. Another rather small anomaly with increased CO 2 efflux and 222 Rn was identified in C2. Here increased T S has been measured, and geothermal surface manifestations were observed. Possible explanations for this are as follows: (1) Presence of separate structure-controlled geothermal upflow, and/or (2) lateral outflow from Bobesa. An additional 222 Rn anomaly was detected in C3, which could be the result of another separate upflow of geothermal fluids along a NNE-oriented fault zone.
A large 222 Rn anomaly in the Southeast (C4) without obvious structural trends coincides with multiple, young volcanic eruption centers and lava flows. This area is also characterized by increased T S , and CO 2 efflux, and seems to be an interesting target, worth more detailed investigation, as some data points show the same trends as observed in Aluto's high-temperature areas. The area has already been drilled (LA5, Figure 8), and data indicate the presence of an outflow zone at depth (Gizaw, 1993), which is probably due to its location at the boundary of the thermal anomaly. Further investigations could help to confirm or exclude the presence of a deep heat source. In general, NNE-oriented structures are indicated in the Central Area by the orientation of areas with increased gas emissions, but magnitudes are low and a very active hydrothermal system is not expected, though evidence of hydrothermal circulation is given. While we cannot rule out deep penetrating structures in this area, no significant fault scarps are observed at the surface. Hutchison et al. (2015) hypothesized that the intersection of the Ring Fault and AJFZ could be a spot of high magmatic gas efflux. However, the results from this study do not support a high magmatic gas efflux in this region. While there is a small increase in 222 Rn, 220 Rn, T S , and CO 2 indicating a permeability structure, they do not confirm a connection to a deep source. The low gas flux at this intersection could be caused by a low permeability or limited fluids present at depth.

Intersecting Faults
The two main areas on the Ring Fault with geothermal surface activity appear to coincide with fault intersections of the Ring Fault and additional NNE-oriented fault zones inferred based on increased gas emissions in the Central Area. These intersections may be important in controlling the upflow of deep fluids.

Implications for Geothermal Exploration
The observation of increased gas flux (in particular CO 2 ) along the AJFZ to the South is a good evidence for extra degassing by the presence of magma in the crust. Whether it is a mostly cooled plutonic body or large 10.1029/2019GC008309 Geochemistry, Geophysics, Geosystems body of melt is still unclear (Iddon et al., 2018), but the along fault plot seems like good evidence for a larger magmatic contribution than typically found along other faults in the rift (Hunt et al., 2017).
From a geothermal perspective, results clearly indicate increased fluid migration from a deep source along permeable fault structures in the southern part of the AJFZ (A2) and in the southeastern segment of the Ring Fault (R2). These areas are not yet targeted for geothermal exploitation but could be interesting prospects for future exploration and development as increased hydrothermal fluid circulation is expected and indicated by soil gas analysis. At this stage, only the northern part of the Artu Jawe Fault (A1) is exploited for geothermal power generation. The gas emissions in this sector point to a deep source as well but do not indicate an area of maximum hydrothermal fluid circulation due to lower gas emission rates compared to A2 or R2. The Southeastern part of the Central Area (C4) also reveals some degassing features with elevated gas efflux in an area with lots of young volcanic landforms. This implies the possibility of a contribution by fluids from a deep source and is worth further exploration efforts.
Generally, the gas emissions from Aluto are comparable to other geothermal systems in volcanic areas where diffuse degassing surveys have been performed for geothermal system analyses. A similar range of gas emissions was also observed in other fields (Giammanco et al., 2007;Jolie, Klinkmueller, et al., 2015). As a result of our area-wide measuring approach based on a predefined regular grid, we observe a limited number of peak values, and the majority of the data shows low values. This is different in comparison to other soil gas surveys, which usually focus on areas with the strongest gas efflux and obvious volcanic edifices (i.e., steam vents, hot pools, and surface alteration).

Conclusion
The Aluto volcanic complex is characterized by variable emission rates of volcanic gases and soil gas concentrations, which can be explained by its tectonic, volcanic, and tectono-volcanic setting. Apart from the two main structural features, that is, Artu Jawe Fault and Ring Fault, further sectors with increased gas emissions have been identified in between. Gas release along the Artu Jawe Fault reflects the general NNE-trend of the Wonji faults and is increasing from the tectonically dominated North toward the volcanic center in the South. This trend can be recognized for all investigated parameter. Similar results could be demonstrated in other geothermal fields in tectono-volcanic settings (e.g., Hernández et al., 2012). Significant differences in quantity of soil gas between A1 to A2 suggest a differentiation between a tectonic and tectono-volcanic compartment. The transition from the Northern to the Southern AJFZ in the vicinity of Figure 9. The developed conceptual model attempts to simplify the complex tectonic and volcanic setting of the Aluto Volcano. The Southern Artu Jawe Fault Zone is dominated by tectono-volcanic gas emissions. Along the Northern Artu Jawe Fault, the tectonic setting dominates surface gas emissions. The Ring Fault is a volcanic feature, which is likely to be affected by NNE-oriented structural features. Geothermal outflow zones are indicated by arrows.
the Aluto-Langano geothermal powerplant is a particularly interesting area, since outcropping segments of the Artu Jawe Fault are accessible without a cover of recent eruptive deposits (Hutchison et al., 2015). Here we observe a change from tectonic (North) to tectono-volcanic mode (South). The geothermal surface expressions at the Ring Fault are considered a result of volcanic and tectonic interactions (tectono-volcanic) and are similar to the southern part of the Artu Jawe Fault.
For a comprehensive and integrated interpretation of the identified gas emissions, a conceptual model was developed (Figure 9). The existence of two additional so far unknown NNE-oriented fracture zones in the Central Area is proposed. Available soil gas data support the hypothesis of their presence. This assumption is also supported by the 222 Rn-220 Rn-CO 2 plot, on which basis data points in the Central Area with maximum values were highlighted (Figure 8). Highlighted sampling sites suggest two NNEoriented corridors of increased gas emissions. However, no obvious structural-geological evidence based on surface mapping was found. The presence of such structures could explain further geothermal surface expressions to the North and South of the study area (e.g., hot springs at northern shore of Lake Langano). They could also explain the two distinct geothermal areas along the Ring Fault as a result of intersecting NNE-oriented structures with the Ring Fault (volcano-tectonic origin). Based on this concept, we expect a continuation of areas with increased gas release toward the North and South as a result of rift-controlled degassing processes.
The results of the soil gas survey have a direct implication for future geothermal exploration and exploitation at the site, since it is indicated that exploration of the southern part of the AJFZ could also be prospective due to its combination of tectono-volcanic conditions (i.e., heat and structures). Further isotopic analyses (e.g., 3 He/ 4 He and d 13 C) will help to better understand the source of the hydrothermal fluids and allow a direct comparison of Artu Jawe and Bobesa. In addition to the major geothermal areas, hydrothermal fluid circulation is also indicated in multiple areas in the central part, which is worth further investigations.