We demonstrate the value of using the self-potential method to study volcanic environments, and particularly fluid flow in those environments. We showcase the fact that self-potential measurements are a highly efficient way to map large areas of volcanic systems under challenging terrain conditions, where other geophysical techniques may be challenging or expensive to deploy. Using case studies of a variety of volcano types, including tuff cones, shield volcanoes, stratovolcanoes, and monogenetic fields, we emphasize the fact that self-potential signals enable us to study fluid flow in volcanic settings on multiple spatial and temporal scales. We categorize the examples into the following three multiscale fluid-flow processes: (1) deep hydrothermal systems, (2) shallow hydrothermal systems, and (3) groundwater. These examples highlight the different hydrological, hydrothermal, and structural inferences that can be made from self-potential signals, such as insight into shallow and deep hydrothermal systems, cooling behavior of lava flows, different hydrogeological domains, upwelling, infiltration, and lateral groundwater and hydrothermal fluid flow paths and velocities, elevation of the groundwater level, crater limits, regional faults, rift zones, incipient collapse limits, structural domains, and buried calderas. The case studies presented in this paper clearly demonstrate that the measured SP signals are a result of the coplay between microscale processes (e.g., electrokinetic, thermoelectric) and macroscale structural and environmental features. We discuss potential challenges and their causes when trying to uniquely interpret self-potential signals. Through integration with different geophysical and geochemical data types such as subsurface electrical resistivity distributions obtained from, e.g., electrical resistivity tomography or magnetotellurics, soil CO2 flux, and soil temperature, it is demonstrated that the hydrogeological interpretations obtained from SP measurements can be better constrained and/or validated.
Volcanic environments are fascinating and complex geological settings. Besides activity that is directly visible at or above the surface, a lot of complicated, dynamic processes happen below the surface that determine the subsurface characteristics of the volcano, which in turn control, e.g., groundwater flow and local hydrology [
Volcanic settings are particularly unique in terms of heat anomalies affecting various subsurface fluid-flow processes. For example, both in active and extinct volcanoes, subsurface heat and remnant heat may generate hydrothermal fluid motion [
Permeability in volcanoes is controlled by several factors. These include the inherent formation properties of the volcanic deposits and possible chemical and/or mechanical alteration (weathering) of formations. The permeability is also heavily influenced by the presence of cracks, faults, and fracture networks. These features can strongly control groundwater flow paths. In volcanic systems, a variety of mechanisms can develop cracks, faults, and fracture networks, which can enhance the permeability and establish preferential fluid flow paths (although faults may also have a sealing-nature). At least four general scenarios seem plausible for developing crack/fault/fracture networks in volcanic environments:
The main processes in volcanic environments that reduce rock formation permeability (or hydraulic conductivity) are chemical alteration processes [
The clay minerals, besides reducing the permeability of the rock formation, have a pronounced effect on the electrical conductivity of the medium as well. Due to the relatively high specific surface area of clay minerals, combined with the CEC (cation exchange capacity) of clays, chemically altered formations can possess high surface conductivities [
In the above, we have briefly introduced various levels of complexity that exist in volcanic environments, affecting hydrology: complex geology, a strongly spatially varying and dynamically changing formation permeability/hydraulic conductivity, together with the presence of heat and heat-driven fluid motion. The systems are further complicated by the variability between different types of volcanoes and volcanic features, such as stratovolcanoes, shield volcanoes, tuff cones, and monogenetic volcanoes.
These features make volcanic environments very interesting and challenging study sites, especially in terms of groundwater hydrology and geothermal energy.
In terms of (hydro)geophysics, volcanic environments are highly challenging environments to image and characterize. For example, basaltic rocks typically result in very high seismic velocities, and seismic imaging suffers from high amounts of scattering due to the abundance of fractures, cracks, and faults [
Ambient noise surface wave tomography has been proven a valuable tool to identify structural subsurface characteristics of volcanic systems, both laterally and with depth [
Electrical Resistivity Tomography (DC resistivity) can provide good quality results for studying volcanic environments [
The image resolution that can be obtained through static ERT imaging is often insufficient to distinguish individual fractures [
An alternative to ERT is the Time-Domain Electromagnetic (TDEM) method (or transient electromagnetics TEM). Electrical currents are induced in the Earth via electromagnetic induction. TDEM often offers better depth resolution than ERT and often displays a higher sensitivity to conductive layers [
In volcanic systems, groundwater and geothermal resources can reside at depths > 2 km (e.g., 2800 meters below sea level [
The magnetotelluric (MT) method is one of the methods, besides seismics, that can be used to study these deep fluid systems. MT allows for great depths of investigation, but typically at the cost of resolution [
In recent years, we have seen the emergence of airborne geophysical methods, such as airborne electromagnetics (e.g., [
Throughout the years, the self-potential (SP) technique has been proven to be a cheap, efficient, and quite successful geophysical method for studying subsurface fluid behavior in volcanic environments (e.g., [
In this paper, we present an overview of a variety of self-potential field studies carried out in different volcanic environments, including shield volcanoes, monogenetic volcanoes, tuff cones, and stratovolcanoes. Based on these different case studies and the interpretation of the measured SP signals, we illustrate and discuss the variety of volcanic subsurface knowledge, as well as the breadth of information on porous medium fluid processes that we can extract and infer from these datasets, as well as highlight the demonstrated successes of using the SP method for studying volcanic environments. Like other geophysical methods applied to volcanic systems, the SP method has its own benefits and drawbacks. We will discuss several of these in greater detail below, thereby also discussing some of the challenges and pitfalls for SP in volcanic environments.
The self-potential method measures the naturally occurring difference in electrical potential between two electrodes coupled to the Earth. It is a passive geophysical method; i.e., it does not use human-made sources for the geophysical data acquisition. Different types of nonpolarizable electrodes have been and are currently used, consisting of a metal rod immersed in a solution of this metal. Currently, the most frequently used types are the Pb-PbCl2 [
The self-potential method offers very fast and low-cost data acquisition, enabling surveying of large areas of land [
Several transport processes can occur in porous media, including coupled processes, i.e., flow as the result of an actuating gradient of a different type than the flow phenomenon itself [
Onsager [
Onsager matrix displaying the types of coupled flow that can occur in a porous medium. The main diagonal represents direct flow phenomena, i.e., the actuating force and the corresponding flow are of the same type. The off-diagonal entries represent the coupled flow phenomena. Onsager reciprocity is reflected by the symmetrical entries across the main diagonal (e.g., streaming potential vs. electroosmosis). Bear in mind that this table is a highly simplified representation. Table
| | |||
---|---|---|---|---|
| | | | |
| Darcy’s law | Electroosmosis | Chemical osmosis | Thermoosmosis |
| ||||
| Streaming potential | Ohm’s law | Dorn potential | Seebeck effect |
| ||||
| Ultrafiltration | Electrophoresis | Fick’s law | Soret effect |
| ||||
| Isothermal heat | Peltier effect | Dufour effect | Fourier’s law |
Self-potential signals can arise from a variety of these coupled flow phenomena.
Often, the electrical double layer (EDL) that exists at the surface between the solid particles (e.g., grain surfaces in a rock formation) and the pore water [
When the grains and the pore water are moving, the place of zero relative velocity is called the shear-plane. This shear plane occurs somewhere in the diffuse layer. For the sake of simplicity, it is often considered to be at the interface between the Stern and diffuse layer. The electrostatic potential on this plane is referred to as the zeta potential (see, e.g., [
As a consequence of the presence of an EDL, electrokinetic phenomena can occur, such as the electroosmotic phenomenon, where an electric field drives fluid flow, and the streaming potential phenomenon, where a hydraulic gradient creates an electric field (see Table
Various electrochemical effects can induce SP signals. A common mechanism involves the coupling between chemical gradients and electric current density (diffusion of ions, Fick’s law) [
Besides streaming potentials and electrochemical effects, thermoelectric effects (such as the Seebeck effect) can occur, where direct conversion from temperature gradients to electric voltages takes place. The electrical double layer has its own contribution to the total thermoelectric coupling [
When we interpret SP signals, we often assume that they are predominantly the result of the electrokinetic streaming potential phenomenon [
When only considering SP signals of electrokinetic origin, different
In active volcanoes, shallow thermal energy can trigger hydrothermal circulation that will influence the topographic effect [
When measured in the field, especially in volcanic environments, SP signals are likely the result of a mixture of various microscopic and macroscopic mechanisms, where the relative importance of individual mechanisms varies for different scenarios. Think for example of SP anomalies related to geothermal systems: strong thermal gradients exist in the subsurface, and much of the geothermal energy is released through groundwater circulation, hot gas emission, and thermal conduction [
The above is to emphasize that besides the variety and/or combination of microscale physical processes that can generate SP signals (e.g., electrokinetic, electrochemical, and thermoelectric), the actual macroscale structural features play a major role in the characteristics of the measured SP signals. Measured SP signals are the result of the coplay between microscale processes and macroscale structural and environmental features.
For the volcanic case studies presented in this paper, the measured SP signals are primarily interpreted as being a consequence of macroscale features and flow behavior and assume the electrokinetic (streaming potential) mechanism to be the major microscopic origin for these signals.
We present several self-potential case studies and their accompanying hydrogeological interpretations for a variety of volcanic environments, including tuff cones, shield volcanoes, stratovolcanoes, and monogenetic fields.
These studies showcase the success and value of SP data acquisition for studying volcano hydrogeology and dynamics and highlight the applicability of the method for these various volcanic environments.
We categorize the examples into the following three multiscale fluid-flow processes: Deep hydrothermal systems: La Fossa, Piton de la Fournaise, and Teide Shallow hydrothermal systems: Piton de la Fournaise and Stromboli Groundwater: Teide and Garrotxa.
These examples display SP signals associated with several of the mechanisms discussed above, including Volcanic hydrothermal systems displaying the “W”-shape SP signal pattern Shallow hydrothermal anomalies, from the discharge of the central hydrothermal system and from recent volcanic deposits Thermoelectric effects related to the cooling of lava flows Structural and hydrogeological features: crater limits, regional faults, rift zones, incipient collapse limits, different hydrogeological domains, elevation of the water level/aquifer, and buried caldera.
Our first example is the stratovolcano La Fossa (Vulcano, Aeolian Islands, Italy). The volcanic activity of Vulcano is caused by magmatic activity that is controlled by regional tectonics, such as strike-slip faults [
Active for ~6000 years [
Figure
2D measured and modeled self-potential profile (top) and 2D electrical resistivity tomography transect with modeled groundwater flow velocities (bottom), La Fossa di Vulcano, Vulcano Island, Italy. We can clearly observe the positive SP anomaly coinciding with the electrically conductive zone in the center of the crater. Finite-element modeling of the groundwater flow using Comsol Multiphysics 3.3 confirmed that the observed SP anomalies are very likely to be caused by the combination of a hot, hydrothermal system in the center, causing hydrothermal upwelling and a positive SP anomaly, and downward flow through ashes and tuff materials along the flanks of the volcano. The figure is modified from [
As a consequence of hydrothermal upwelling, a positive SP anomaly can be observed (Figure
On the flanks of the cone, the negative SP anomalies are interpreted to be the result of downward flow and infiltration/percolation of cooling groundwater at relatively shallow levels (groundwater flow paths are controlled by structural and geological features of the volcano). Both interpretations have been validated by finite element modeling of the groundwater flow paths, using Comsol Multiphysics 3.3 [
Figure
Self-potential (top) and soil temperature (bottom) maps of La Fossa cone overlain on the Digital Elevation Map (DEM) of the area. White dots represent the measurement points. Areas with steady hydrothermal activity are well delineated by self-potential maxima and correlated to high temperature anomalies inside the craters and along their limits. PN (Punte Nere), Pa (Palizzi), FV (Forgia Vecchia), PC (Pietre Cotte), and GC (Gran Cratere) craters are localized with black-dashed lines. Profile
Soil temperature measurements (top), self-potential values (middle), and electric resistivity tomography results (bottom) along Profile
The Fossa cone example demonstrates that SP measurements are capable to (1) detect the presence of hydrothermal systems on active volcanoes, (2) identify the percolation and infiltration of groundwater along the flanks and in shallow regions of volcanic edifices. Similar observations have been made in Stromboli [
The Piton de la Fournaise volcano is located on La Reunion Island, in the Indian Ocean. With 2 or more eruptions per year since 1998, it is one of the most active volcanoes in the world. Most of the eruptive activity takes place in the depression formed by the Enclos Fouqué caldera and the Grand Brûlé. This U-shaped depression is open to the East (Indian Ocean) (Figure
On Piton de la Fournaise, the whole terminal cone was mapped using the SP method, with a measurement spacing of 20 m [
Self-potential map of the whole terminal cone of Piton de la Fournaise (top) and a detailed SP map of the crater (bottom). PB (Pre-Bory), S (Soufrière), and PP (Petit Plateau) are former pit craters surrounding the Dolomieu and Bory craters, affected by major hydrothermal release as shown by the high self-potential values. White-dotted lines represent the collected SP profiles. Self-potential maxima highlight the rift zones of this volcano in consistency with the repartitioning of the fractures and fissures. Before the 2007 eruption, the self-potential data identified underground fluid flow heterogeneities highlighting structures that later controlled the collapse of the Dolomieu crater and the temporary preservation of an eastern plateau. Figure modified from [
In the case of Piton de la Fournaise, SP has shown to be a powerful tool to highlight hydrothermal circulation heterogeneities controlled by rift zone fracturing processes, crater limits, and incipient collapse faulting.
Our next example is the Central Tenerife Volcanic Complex (Canary Islands, Spain). The central area of Tenerife is dominated by Las Cañadas caldera where two stratovolcanoes, Pico Viejo and Teide, are located. The initial stages of activity of this island resulted in the formation of a basaltic shield which emerged from the ocean more than 11.9 Ma ago [
The Las Cañadas edifice eventually collapsed forming Las Cañadas caldera at 0.18-0.13 Ma [
Beyond the fumarolic activity in some parts of the caldera floor and on the two stratovolcanoes, the most recent activity of the Tenerife volcanic complex was a seismic crisis in 2004, which consisted of a swarm of more than 1000 low-magnitude earthquakes of magmatic origin [
For studying the volcano and its fluid dynamics over such a complex and huge area, the SP method is probably the most suitable land-based geophysical tool, as it offers the possibility to cover long distances with high resolution and in a small amount of time compared to other methods. Figure
SP survey in Las Cañadas caldera (Tenerife, Spain). The SP maxima highlight the main hydrothermal systems of Teide and Pico Viejo volcanoes, as well as smaller hydrothermal systems at two former volcanic vents (Montaña Blanca and Montaña Rajada). Ground water infiltration is highlighted by self-potential minima at the Diego Hernandez (DH) buried caldera and at the NW rift fractures (NWR). Figure from [
Plot of the time-lapse self-potential measurements performed at the coastline of Piton de la Fournaise. The upper graph shows self-potential data acquired in 2004 (green curve) and in 2006 (black curve) as well as the difference between both surveys (bottom black line). Recent lava flows are represented by colored bands located in the DEM figure (bottom right) adopting the same color coding. The figure is modified from [
Despite the potential impact on deep-source interpretations of SP values, shallow sources of SP signals are definitely of interest when studying hydrothermal circulations in active volcanoes. Both deep and shallow hydrothermal systems can have an impact on the measured SP signals. This can occur simultaneously, or, separately, depending on the physical target of interest and the location of the survey on the volcano.
Let us revisit the Piton de la Fournaise example that we have discussed previously in terms of deep hydrothermal systems. We now look at time-lapse SP data acquired on Piton de la Fournaise.
Figure
The second example of shallow hydrothermal activity highlighted by SP data is the Stromboli Volcano, in Italy.
Stromboli stratovolcano is the northernmost island of the Aeolian archipelago, located in the Tyrrhenian Sea. The subaerial volcanic history of Stromboli is subdivided into seven main phases, including Paleostromboli I, II, III, Scari complex, Vancori, Neostromboli, and Recent Stromboli (e.g., [
Looking at the current eruption dynamics at the summit vents of Stromboli, its activity has been almost continuous since 1932. The ordinary activity consists of frequent Strombolian explosions (approximately every fifteen minutes to every hour), as well as occasional lava flows. Most of this activity occurs in the summit area and in a restricted collapsed zone named Sciara del Fuoco, but occasional stronger explosions or paroxysms [
Soil temperature, self-potential, soil CO2 flux measurements, and electric resistivity tomography results of the multidata study carried out on the Stromboli volcano (Italy). At the location of a volcanic bomb impact crater from the March 15th 2007 paroxysmal eruption, we observe a clear maximum value for the self-potential, which correlates with the other measurements. This study uncovered shallow hydrothermal circulation which seems a common feature of active volcanoes. Figure from [
The contribution of the SP method is crucial here as the conductive layer revealed by the electric resistivity tomography alone could alternatively be interpreted as a chemically altered layer rich in conductive clay minerals.
Thus far, we have looked at the use of the self-potential method to study deep and shallow hydrothermal systems and its fluid circulation in volcanic environments. We will now provide two examples where SP signals inform us on groundwater flow in volcanic systems and the associated hydrogeology.
The Teide volcano, previously discussed to refer to its deep hydrothermal system based on an extensive SP dataset published by [
Conversion of the self-potential data (using previously acquired audio-magnetotelluric and water table data) into information on the elevation of the water table. Different colors represent different elevations of the bottom of the vadose zone in meters above sea level. Self-potential demonstrates that the underground topography of the Diego Hernandez caldera floor is a major controlling factor for the hydrogeology in this area. Figure from [
Our final study site is located in Catalonia (Northern Spain). This case is different from the rest of the examples presented in this paper for two main reasons: (1) this is a monogenetic volcanic field that consists of over 50 monogenetic volcanoes and (2) the last volcanic eruption (Rocanegra eruption) is known to be less than 11.5–13 ka old [
In this case study, SP is expected to provide information on the hydrogeological system of the area, structural information, and their relation with volcanism. Figure
Self-potential map of the Garrotxa monogenetic field (left). The map highlights the strong lateral heterogeneity of the medium resulting in a complex setting. A western zone and an eastern zone appear to be separated by a NNW–SSE lineament and a NNE–SSW lineament. A plot of self-potential vs. elevation (right) seems to confirm this separation by displaying a strong difference in SP behaviour between the east and west sides, which is likely due to lithological and hydrogeological contrasts between the two areas. Figure modified from [
We have shown that SP measurements are a highly efficient way to map large areas of volcanic systems under challenging terrain conditions, where other geophysical techniques may be challenging or expensive to deploy. Given the complex dynamics of fluid flow, heat, and permeability variations in volcanic environments, the sensitivity of SP data to fluid flow provides a unique way to identify and highlight infiltration patterns, heat-induced shallow and deep hydrothermal convection, and preferential flow paths such as fractures, faults, and impermeable zones.
The various case studies presented in this paper demonstrate that SP signals, despite the fact that they can result from a variety of microscale mechanisms and macroscopic features and consequently can be challenging to uniquely interpret as a stand-alone data type, provide valuable insight into the local hydrogeology, fluid flow dynamics, and macroscale structural features of volcanic systems, regardless of their type, and on a variety of spatial and temporal scales, such as information on Deep hydrothermal systems Shallow hydrothermal anomalies, from the discharge of the central hydrothermal system and from recent volcanic deposits Cooling behavior of lava flows Different hydrogeological domains Upwelling, infiltration, and lateral groundwater and hydrothermal fluid flow Elevation of the water level/identification of aquifers Crater limits, regional faults, rift zones, incipient collapse limits, structural domains, and buried calderas.
With respect to (7), it is important to realize that SP surveys do not directly image these limits and structural features. SP signals highlight their impact on groundwater flow and hydrothermal circulation, from which structural inferences can be made.
Through integration with different geophysical and geochemical data types such as subsurface electrical resistivity distributions obtained from, e.g., electrical resistivity tomography or magnetotellurics, soil CO2 flux, and soil temperature, the hydrogeological interpretations obtained from SP measurements can be better constrained and/or validated. Furthermore, knowing the subsurface electrical resistivity distribution, SP data can be used to estimate Darcy flow velocities that can directly inform hydrologic models [
However, groundwater and hydrothermal fluid flow is inherently a 3D problem.
Using networks of 2D ERT acquisition lines, a pseudo-3D electrical conductivity image of the subsurface can be obtained [
Alternatively, for deeper targets, 3D resistivity models can be obtained through, e.g., 3D MT surveys [
In recent years, airborne geophysical surveys, for example, airborne electromagnetic surveys, have been proven very successful for imaging the groundwater system with large spatial coverage [
Furthermore, the latest developments in muon tomography demonstrate that muon imaging is a powerful technique, especially for complementing other geophysical methods: for example combined with gravity data to image the 3D density structure of lava domes [
Integrating SP data with any of these geophysical data types will enable a 3D understanding of groundwater flow patterns, hydrothermal systems, and volcanic structural features that control fluid flow.
For the case studies presented in this paper, the measured SP signals are interpreted as being the result of macroscale structural features and fluid flow patterns. As discussed earlier, SP signals have a microscale origin, and even though we have interpreted the case studies in this paper to be primarily the consequence of an electrokinetic mechanism, it is important to keep in mind that other microscale mechanisms may be at play (such as thermoelectric and electrochemical effects). This may be important especially in volcanic environments, where, for example, in hydrothermal systems, hot, acidic fluids circulate.
Moreover, in fractured, volcanic rock formations, heterogeneities that are smaller than the characteristic scale of the problem (such as the size of typical seismic wavelengths), but larger than the pore scale, can affect fluid flow at the so-called mesoscale [
However, an interpretation of SP signals on the mesoscale is not often made. Given the strong macroscale imprint on the measured SP signals, it is likely that mesoscopic (and microscopic) effects are often hard to recover. However, for example, in scenarios of strong preferential fluid flow through dense networks of microcracks and mesoscale fractures, which can be highly probable in volcanic environments, it is very likely that a mesoscale formulation of the SP problem and accompanying interpretation will become important. Future research in this direction will be carried out.
Despite the fact that microscale and mesoscale properties have not been inferred in this paper, the case studies clearly show that, using the self-potential method, we have a cheap and efficient way of studying volcanic systems in terms of their macroscale structural features and flow paths, regardless of the type of volcanic environment (e.g., tuff cones, shield volcanoes, stratovolcanoes, and monogenetic fields).
Given the demonstrated efficiency, success, and value of self-potential measurements to study the hydrogeology, as well as the hydrologic and hydrothermal processes of volcanic environments, we expect SP-measurements to be highly valuable in our current, multigeophysical efforts as part of the multidisciplinary NSF EPSCoR ‘Ike Wai project, aiming to better understand the hydrology of the Hawaiian Islands for a sustainable water future.
The data used in this paper are available upon request to the authors.
S. Barde-Cabusson is currently 100% funded by the NSF EPSCoR ‘Ike Wai project, and N. Grobbe is funded for 50% by the NSF EPSCoR ‘Ike Wai project and for 50% by the University of Hawai‘i at Mānoa Water Resources Research Center. Support for the Hawai‘i EPSCoR Program is provided by the National Science Foundation’s Research Infrastructure Improvement (RII) Track-1: ‘Ike Wai: Securing Hawaii's Water Future Award # OIA-1557349.
The authors declare that they have no conflicts of interest regarding the publication of this paper.
The authors would like to explicitly acknowledge the various scientists and colleagues whose time, energy, and scientific contributions resulted in the case studies displayed in this paper. N. Grobbe would like to warmly thank Steve Martel for inspiring discussions on fractures. The authors thank the editor and an anonymous reviewer for the great suggestions and comments that helped to improve this manuscript. This is contributed paper #2394 of the Hawai‘i Institute of Geophysics and Planetology, #10765 of the School of Ocean and Earth Science and Technology, and WRRC-CP-2020-03 of the Water Resources Research Center; University of Hawai‘i at Manoa, Honolulu, Hawai’i, USA.