Induced Polarization Images the Plumbing System of Hydrothermal Vents in an Intracontinental Rift, Lake Abhé, Republic of Djibouti

Recent developments in induced polarization allow for the characterization of alteration halos within hydrothermal systems. We explore the possibility of using electrical conductivity and normalized chargeability tomograms in concert to image hydrothermal conduits thanks to their high cation exchange capacities associated with alteration. The hydrothermal plumbing system of the late‐stage rift area of Lake Abhé (Republic of Djibouti) is used to test the ability of induced polarization to reveal such plumbing system associated with the prominent hydrothermal chimneys serving as vents for the hydrothermal fluids. We show that induced polarization can be used to provide an alteration tomogram, which highlights the flow path toward 2 chimneys in the top 70 m below the surface of the sediments at Lake Abhé.


Introduction
The development of renewable energies is currently of paramount importance to develop a sustainable civilization.Accessing the renewable energy of geothermal systems requires the development of sophisticated imaging techniques to obtain subsurface information of their plumbing system (Domra Kana et al., 2015).In this context, the objective of the LEAP-RE (Long-term joint European union-African union research and innovation Partnership on Renewable Energy) consortium, is the development of stand-alone energy production plants in African countries (Varet et al., 2014).The Geothermal Village program, part of this consortium, aims to develop geothermal energy with the LEAP-RE approach, in the East African Rift System, where the geothermal resource is of high value, while being under-explored and under-exploited.Lake Abhé is an alkaline and hypersaline closed lake located at the border between the Republic of Djibouti and Ethiopia (Figure 1).The lake is located in the Afar Depression at the triple junction between the Somalian, Nubian, and Arabian plates, in an area of intense heat flow and hydrothermal activity (Abbate et al., 1995;Barberi & Santacroce, 1980;Hochstein, 2005;Michon et al., 2022;Saemundsson, 2010).It appears as a promising site for geothermal energy exploration (Hochstein, 2005).Indeed, the lake's eastern margin is known for its massive carbonate-rich chimneys over flat-lying sediments (DeMott et al., 2021).This hydrothermal activity, associated with volcanism within the East African Rift, manifests itself on the site of Lake Abhé as hydrothermal fluid discharges through hotsprings leading to the formation of travertine chimneys at the Lake bottom.

10.1029/2023GL105145
2 of 7 The geophysical panel of methodologies lacks innovative imaging methods to image the plumbing system of hydrothermal systems.The plumbing systems of geothermal fields are characterized by alteration halos.Induced polarization is a geophysical technique imaging the conductivity and chargeability of rocks (Vinegar & Waxman, 1984) including at the kilometer scale (Gross et al., 2021;Revil et al., 2023).Induced polarization tomography has been shown recently to provide an efficient approach to image alteration in active volcanoes (Revil & Gresse, 2021).Revil and Gresse (2021) demonstrated that smectite plays a key role in the induced polarization response to alteration in volcanic rocks and that this effect (through the high cation exchange capacity of this mineral) is temperature-controlled.
From this knowledge, the question of this research work will be addressed as the following.What is the efficiency and suitability of the induced polarization method to image the plumbing system associated with hydrothermal discharges in a sedimentary system?A 2D induced polarization profile crossing two surface alignments of chimneys has been performed in the context of Lake Abhé to image their plumbing system in order to see the usefulness of the analysis of alteration by induced polarization.This geophysical survey is located at the contact between the Stratoïd Series basalts and the rift extension infilling sediments (Profile AB in Figure 1).The Stratoïd Series are composed of a series of overburdened lava flows and  (Chorowicz, 2005), black lines: main faults; white surfaces: lakes; gray levels from dark (low elevations) to light (high elevations), (b) Extract of the geological map of the survey area modified after Le Gall et al. (2014), (c) satellite view of the location of the geophysical profile, its displayed section AB, photographic view location, and fluid conductivity sample locations from Awaleh et al. (2015), (d) Photographic view of a representative chimney structure.The main contact between the clays and Stratoïd Series basalts is the main feature guiding the resistivity results of the area.The profile intersects several chimney structures visible on the satellite map, which are not pointed out on the larger-scale geological map.The survey site is represented as a circle as its symbol does not imply any scaling.
10.1029/2023GL105145 3 of 7 pyroclastic materials.The sediments are mainly composed of clayey materials with low fraction of carbonates plus terrigenous material, mainly coming from the erosion of nearby basaltic grabens.The sediment layer in the basin is thickening westwards, with N110 and N10 faulting accommodating for the extension deformation.In this context, low-frequency polarization processes might occur associated with the presence of magnetite, pyrite, clays coming from the hydrothermal alteration (Fontboté et al., 2017), and clays deposited as lacustrine sediments.
Even though this geothermal system is located in a sedimentary terrane, the conclusions obtained in Revil and Gresse (2021) can be applied to image hydrothermal conduits as the geothermal fluid is in both cases the main driver of mineral alteration.Within the studied hydrothermal system of Lake Abhé, the fluid that drives alteration shows surface temperatures of nearly 100°C and is of the Na-Cl type (Awaleh et al., 2015) with Local Meteoric Water Line (LMWL) isotope alignment.Major ions show at the nearest geographical water sampling site named SCH2 concentrations of 225 ppm (Ca), 0.93 ppm (Mg), 1,074 ppm (Na), 27 ppm (K), 0.15 ppm (CO 3 ), 15.26 ppm (HCO 3 ), 1,737 ppm (Cl), and 349 ppm (SO 4 ).

From Data Acquisition to Tomography
The induced polarization survey was done with an ABEM-SAS4000 and two ABEM Advanced 4/81 cables with 42 electrodes with a non-regular spacing of 15 and 10 m in the center of the profile.The data acquisition was done with a Wenner protocol, 1 s of current time injection, giving a total number of 224 quadrupoles.The secondary voltage was recorded after a dead time of 0.2 s using 3 windows of 100 ms each.Topography was obtained with a Garmin GpsMap 62s.The position of the profile AB was chosen to cross two alignments of hydrothermal chimneys with the goal of imaging their conduits at depth.The raw data set was filtered by removing negative apparent resistivity and chargeability data.As it is impossible to attain a medium with negative resistivity or chargeability, it has been chosen to reject those values, even if some particular geological objects can lead to negative apparent readings (Hyun-Key et al., 2018).The apparent resistivity and chargeability were inverted using a smoothness-constrained least-squares inversion method (de Groot-Hedlin & Constable, 1990) with the Res2DInv software.Both resistivity and chargeability data sets have been inverted with a smoothness constraint on the model perturbation vector only.The inverted resistivity model shows a RMS (Root Mean Square deviation of the model) of 3.6% and an average absolute error of 4.0%.The chargeability model shows a RMS of 5.8% and an average absolute error of 4.3%.This data set leads to a model made of 616 cells, 610 m long and 92.9 m deep.It is then used for the calculation of the normalized chargeability and CEC as explained in the next section.
In addition, some chimney samples have been collected and put under an independent measurement of their Cationic Exchange Capacity.It is measured through the cobalthexamine method, for which samples are placed in a 0.05N colbalthexamine chloride solution for a test of absorption after centrifugation and appropriate equilibrium.The test of absorbance is performed at 472 nm and allows for the CEC determination of surface samples (Aran et al., 2008).

Application of the Petrophysical Model
A petrophysical model called the dynamic Stern layer model (Revil et al., 2017(Revil et al., , 2022) ) is used to interpret the conductivity and normalized chargeability tomograms.In saturated conditions, the electrical conductivity σ (in S m −1 ) and the normalized chargeability M n (in S m −1 ) of a rock or sediment are given by Revil et al. (2017Revil et al. ( , 2022) (1) where σ w denotes the pore water conductivity (in S m −1 ), ϕ is the porosity (dimensionless) and m is called the cementation or porosity exponent of Archie' law (typically m = 2.0 ± 0.5, Vinegar & Waxman, 1984).ρ g denotes the grain density (∼2,710 kg m −3 for calcite as we study a carbonated chimney system), CEC denotes the Cation Exchange Capacity of the porous material (expressed in C kg −1 or in meq/100 g with 1 meq/100 g = 963.20 C kg −1 ), B (in m 2 s −1 V −1 ) denotes the apparent mobility of the counterions for surface conduction and λ (in m 2 s −1 V −1 ) 10.1029/2023GL105145 4 of 7 denotes the apparent mobility of the counterions for the polarization (see Vinegar & Waxman, 1984).We have Β(Na + , 25°C) = 3.1 ± 0.3 × 10 −9 m 2 s −1 V −1 and λ(Na + , 25°C) = 3.0 ± 0.7 × 10 −10 m 2 s −1 V −1 .A dimensionless number R was introduced by Revil et al. ( 2017) as R = λ/B ≈ 0.10 ± 0.02 independent of temperature and water content.The CEC is temperature independent but the production of smectite affecting the value of the CEC in a geothermal system is temperature dependent (Revil & Gresse, 2021).The relationship between alteration and CEC is not unique because clay minerals have different CEC values (see discussion in Revil et al. ( 2021)).
In our case, the pore water conductivity σ w (25°C) = 0.556 ± 0.008 S m −1 is obtained from in situ data by averaging ground water conductivity obtained with three water samples from three hot-springs nearby to the profile (samples SCH1 SCH2 and SCH3, see Figure 1c; Awaleh et al., 2015).The temperature dependence of σ w and B is 2% per degree Celsius (Revil & Gresse, 2021), which can be neglected here.The conductivity of the pore water being known, it is possible to compute the porosity and the CEC from Equations 1 and 2 using the tomograms of the conductivity and normalized chargeability using the following transforms: The calculation is done cell-by-cell to obtain tomograms of the porosity and CEC.The benefits of imaging near-surface hydrothermal systems with induced polarization can be seen in the dependence of the induced polarization parameters with CEC and therefore alteration.Imaging the CEC from the conductivity and chargeability is therefore more useful than just looking at these measured properties.The resulting alteration/CEC tomogram should be able to identify areas where the degree of alteration is higher than the surrounding and especially the hydrothermal conduits.

Discussion and Concluding Statements
Our profile is located approximately 1 km westwards from the contact between the basaltic outcrop and the sediments (Figure 1b).The conductivity values are consistent with the geological sedimentary environment shown on the geological map, ranging from 10 −1 to 10 1 S/m.It implies the presence of saturated lacustrine clayey material (Figure 2a).The normalized chargeability tomogram (Figure 2b) displays subvertical anomalies suggesting a strong participation of the chimney geometry to the observed anomalies.
The CEC tomogram (Figure 3a) shows subvertical anomalies with a good correlation with the surface presence of travertine chimneys.The upper limits of these anomalies are consistent with the presence of the chimneys C1 and C2 that intersect the profile (Figures 3a and 3c).The geometry of the CEC anomaly shows a dichotomization of the fluid network, which indicates the preferable porosity and permeability paths for the pressurized fluid to escape to the surface.The tomogram values suggest a significant alteration level (∼40 meq/100g).Pure smectite is characterized by a CEC of ∼90-100 meq/100g (Christidis et al., 2004).Two samples were used for CEC laboratory measurements using the cobalthexamine method.We obtain 7 ± 1 meq/100g showing a high degree of alteration of the material of the chimneys.Note that the chimneys themselves are made also with the precipitation of calcite.Therefore the CEC value obtained with geophysics for an entire chimney is expected to be lower than the CEC of the conduits showing the fraction of high CEC minerals.In addition, the presence of pyrite in the chimneys could also be responsible for their higher polarization.The Tendaho rift, an analog to the Afar Depression, exhibits the presence of pyrite and smectite within the alteration areas (Gebregziabher, 1998).Taking into consideration the mature nature of the geothermal fluids in this location, which has attained equilibrium 10.1029/2023GL105145 5 of 7 with a primary silicatic mineral paragenesis (Awaleh et al., 2015), we can associate the fluid flow path with the alteration tomogram.This leads to the model presented in Figure 4, in which hydrothermal fluids are taking the normal fault network to rise to the surface from the geothermal reservoir.During this uprising, with fluid-rock interactions, alteration builds up within the upper sedimentary layers.While reaching the upper surface of the sediments, wherever the fluid encountered lake waters at its past levels, it led to the accretion of those characteristic hydrothermal chimneys.In this regard, alteration affects the edges of the ducts for the fluid flow, and induced polarization anomalies allow for a detection and characterization of this hydrothermal network.
In summary, Lake Abhé displays giant hydrothermal chimneys associated thanks to the strong geothermal activity of the East African Rift.Tomograms of the electrical conductivity and normalized chargeability have been performed over a depth of 90 m.We use a petrophysical model called the dynamic Stern layer model to obtain tomograms of the cation exchange capacity and porosity.The conduits associated with the hydrothermal chimneys are rich in clay minerals because of alteration.The cation exchange capacity tomogram underlines the position of the hydrothermal conduits at depth.Comparison of modeled geophysical signal and laboratory can further indicate differences between surface and depth petrophysical properties, as well as the effects of representative volume from field to laboratory scale.To our knowledge, this is the first time that such a result is obtained demonstrating how useful is induced polarization with respect to the measurement of the electrical conductivity alone.The next step would be to investigate induced polarization in 3D and at much important depths (few kilometers) and to combine the results with other geophysical methods such as magneto-tellurics, seismic, and gravity, which are all sensitive to porosity and alteration.Furthermore, large scale induced polarization tomography could be therefore used to image alteration as well as the temperature field in this type of hydrothermal systems.

Acknowledgments
This work is supported by the GeoRessources Laboratory, CNRS, University of Lorraine, Région Grand-Est (France), and LEAP-RE Program, Geothermal Village project.The authors thank the ODDEG (Office Djiboutien de Développement de l'Energie Géothermique) in Djibouti for their logistical and on-site supports.We thank S. Touzet for his help in Python and J.Richard for his help with the petrophysical measurements.We thank the Editor Dr. Daoyuan Sun and two anonymous Referees for their constructive comments.

Figure 1 .
Figure 1.Localization and geological environment of the induced polarization survey.(a) Extract of hypsographic digital elevation model of the East African Rift (Chorowicz, 2005), black lines: main faults; white surfaces: lakes; gray levels from dark (low elevations) to light (high elevations), (b) Extract of the geological map of the survey area modified after Le Gall et al. (2014), (c) satellite view of the location of the geophysical profile, its displayed section AB, photographic view location, and fluid conductivity sample locations from Awaleh et al. (2015), (d) Photographic view of a representative chimney structure.The main contact between the clays and Stratoïd Series basalts is the main feature guiding the resistivity results of the area.The profile intersects several chimney structures visible on the satellite map, which are not pointed out on the larger-scale geological map.The survey site is represented as a circle as its symbol does not imply any scaling.

Figure 2 .
Figure 2. Conductivity and normalized chargeability tomograms.Sub-horizontal anomalies are here favored in the inversion of the conductivity model as a display, in which we observe the effects of compaction on panel (a) (decrease of conductivity with depth).For the normalized chargeability, we observe and the effects of alteration on panel (b) (increase of normalized chargeability).

Figure 3 .
Figure 3. Tomogram with the application of the Stern layer petrophysical model and alignment between tomogram anomalies and surface geological features (a) Alteration tomogram with inverted and modeled values of CEC, and chimney structure interpretation (b) Chimney photographic view (c) Satellite view of the chimney and profile extract.

Figure 4 .
Figure 4. Hydrogeological model of fluid flow within the investigated geophysical volume.Lake levels have been represented for the accretion time of the chimneys and for the actual time.The IP (Induced Polarization) profile is placed conceptually within this model (no scaling implied).