Hydrothermal fluid flow triggered by an earthquake in Iceland

Fluids in the Earth's crust play a crucial role in the release of seismic and thermal energy, particularly at divergent plate boundaries. However, the interaction between large-scale fluid circulation and active fault tectonics remains poorly understood. Conceptual models, such as the fault-valve model, highlight the influence of fluid pore pressure variations during the seismic cycle. Aseismic displacements along normal faults have also been linked to fluid-related processes like pressure solution creeps. The unexpectedly low heat flow at spreading ridges underscores the significance of hydrothermal convection. At slow-spreading ridges, this convection is primarily concentrated above heat sources (magma chambers, intrusions) in the middle of oceanic spreading segments. This hydrothermal convection is often accompanied by seismic activity distinct from aftershock sequences, typically forming clustered swarms of low-magnitude double-couple events. Similar swarms are observed during fluid injections in deep boreholes, as in hot dry rock experiments. While high fluid pressures can facilitate shear activation in a fractured medium, the reason for the clustering of seismic activity within a large crustal volume during short-duration sequences remains unclear. Processes of pore-pressure diffusion within a fractured rock mass, potentially triggered by distant or local seismic events, have been proposed. Pore pressure diffusion is primarily governed by fluid diffusivity, which depends on the medium's permeability and fluid properties. At high temperatures, highly dynamic fluids can also propagate new cracks through stress corrosion. In Iceland's hydrothermal systems, fluids largely originate from seawater or meteoric sources, with the hydrosphere serving as the primary reservoir. While analytical and numerical modeling has illustrated hydrothermal fluid flow in fractured media, these models struggle to capture the complexity of thermally driven fluid circulation with unknown properties in an anisotropic fractured medium. A clear geophysical time-lapse imaging of hydrothermal fluid circulation has been lacking, despite three-dimensional electrical resistivity tomography being able to image the structure of such systems. The temporal dimension, crucial for understanding the space-time pattern of hydrothermal fluid circulation, is often missing. Although the general bottom-to-top upwelling of hydrothermal fluids from heat sources is well understood, the top-to-bottom fluid migration from the shallow crust to recharge deeper regions is critical to understanding fluid flow in active volcano-tectonic settings. Crustal faults are suspected to facilitate this downward migration. This study compares high-resolution seismic experiments in Krýsuvík, SW-Iceland, a region with significant hydrothermal activity, to investigate these processes.

Previous research has emphasized the role of fluids in seismic and thermal energy release at plate boundaries, particularly in the context of the fault-valve model, which explains the influence of fluid pore pressure changes on the seismic cycle. Studies have also shown the connection between fluid-related processes such as pressure solution creep and aseismic fault displacements. The importance of hydrothermal convection in influencing heat flow anomalies at spreading ridges has been established, especially at slow-spreading centers where convection is often located above magma chambers or intrusions. These studies have noted the association between hydrothermal convection and distinct seismic swarms, different from typical aftershock sequences. While the influence of high fluid pressure on shear activation in fractured media is accepted, the mechanism behind the clustering of seismic activity in short, intense bursts remains debated. Previous research has explored pore-pressure diffusion in fractured rock masses as a potential trigger for seismicity, influenced by factors like fluid diffusivity, permeability, and fluid properties. At high temperatures, the role of stress corrosion in crack propagation has also been recognized. Studies of Icelandic hydrothermal systems have described the origin of fluids (seawater and meteoric), while modeling efforts have attempted to simulate fluid flow in fractured media. However, these models often lack the complexity to capture the real-world behavior of fluids with variable properties in anisotropic fractured systems. Geophysical imaging, especially using time-lapse techniques, has been limited in elucidating the space-time dynamics of hydrothermal fluid circulation. Prior research has highlighted the downward migration of fluids through crustal faults, a process critical to understanding fluid flow in active volcano-tectonic settings. Earlier work in Krýsuvík showed evidence of deep downward fluid percolation driven by localized crustal dilatation, but lacked the comprehensive time-series data to fully understand the subsequent upward migration.

Two passive seismic experiments were conducted in the Krýsuvík geothermal area, SW-Iceland, one from April to September 2005 and another from May to October 2009. The 2005 experiment utilized a smaller network which allowed for high-precision earthquake relocation (estimated 200 m accuracy). The 2009 experiment employed a much denser network (32 stations vs. the 2005 network) covering a 30 km² area, providing significantly improved resolution. More than 10,000 events were detected in the 2009 survey. Seismic data were processed using established techniques for earthquake location and velocity model determination. The 2009 data set enabled a more refined 3D velocity model using TomoDD software which performs simultaneous inversion of absolute hypocenter locations and 3D P- and S-wave velocity structure, improving upon the 2005 results. The post-processing weighted average model (WAM) method was used to increase the robustness of the velocity model and reduce dependence on the initial model and grid parametrization. The final uncertainty of hypocenter locations was estimated to be less than 140 m in all three directions. Focal mechanisms were determined using FPFIT software, considering events with sufficient picks and azimuthal gaps. Two approaches were used to estimate crustal permeability: 1) using the mean flow velocity inferred from the vertical migration of seismicity between the 2005 and 2009 surveys; and 2) analyzing pore-pressure diffusion within a single seismic swarm from the 2005 data using a diffusion model. Heat flux was estimated using a model that accounts for fluid properties (density, viscosity, specific enthalpy) at various pressures and temperatures.

The 2005 seismic survey revealed a swarm-like seismicity distribution near Lake Kleifarvatn, with most hypocenters located between 4 and 5 km depth, above a low Vp/Vs ratio anomaly. This anomaly, extending to at least 6 km depth, was attributed to compressible fluids in a fractured heat and fluid reservoir, likely in supercritical conditions. The 2009 survey, with its denser network, provided higher resolution, showing a deeper low-Vp anomaly extending to 7 km depth. The Vp=5 km/s contour surface revealed an apparent continuity from the deep to uppermost crust, with a curved area extending eastward and upward beneath Lake Kleifarvatn. The Vp/Vs ratio anomaly showed a similar shape to the 2005 data. Seismicity in the Krýsuvík segment was concentrated southwest of Lake Kleifarvatn, mimicking the geothermal area. Several short-duration clusters were identified, with some shallow swarms displaying a NS trend. A vertical cross-section showed a dome-like shape for the deepest hypocenters, mimicking the top of the Vp/Vs anomaly. Above this surface, seismicity displayed a 3D organization with several vertical peaks corresponding to NS-trending clusters. The short duration of seismic bursts (≈2 days), low magnitude, and clustered space-time patterns strongly suggest a fluid-pressure origin rather than tectonic activity. The decay rate of seismicity didn't follow the modified Omori-Utsu law, further supporting a fluid-driven mechanism. Comparing the 2005 and 2009 data sets revealed a clear upward migration of seismicity, interpreted as upward fluid flow. The NS orientation of some swarms suggests interaction with NS-trending sub-vertical dextral fault zones. The pipe-like pattern of seismicity suggests multiple hydrothermal convection cells, with estimated spacing and height. The upward fluid migration velocity was estimated to be ~2 × 10⁻⁵ m s⁻¹. Permeability estimates, using both the overall seismicity migration and diffusion within a single swarm, yielded a consistent value of ~10⁻¹³ m², much higher than typical oceanic crust but consistent with models of hydrothermal convection. The estimated advective heat flux, using the permeability estimate, was ~500 kW m⁻¹, comparable to large deep-sea hydrothermal systems. Geodetic data (GPS and InSAR) showed local uplift (~20 mm/yr) between 2006 and 2009, correlating with the intense seismicity during the 2009 experiment, further supporting the fluid migration interpretation. A low-Vp anomaly extending from the surface downward suggests a large fractured area with open cracks, likely facilitating fluid migration.

The observed upward migration of seismicity between 2005 and 2009 strongly indicates a bottom-to-top fluid flow in the Krýsuvík area, consistent with the earlier proposed top-to-bottom model. This model suggests a two-reservoir system: a shallow hydrostatic reservoir and a deeper over-pressurized reservoir. The 2000 earthquake likely increased crustal permeability, allowing fluids from the shallow reservoir to migrate downward and recharge the deeper reservoir. The subsequent upward fluid flow, evidenced by the seismicity migration, suggests that the pressure in the deeper reservoir eventually exceeded the confining pressure, forcing the fluids upward through the high-permeability fractured zone. The variation in seismic activity patterns appears to reflect changes in fluid pressures within the upper crust, influenced by changes in fracture geometry and/or fluid volume. The estimated high permeability (~10⁻¹³ m²) supports the rapid upward migration of fluids. The high heat flux (~500 kW m⁻¹) is also consistent with observations in large deep-sea hydrothermal systems. The observed uplift corroborates the inferred increase in upper crustal pore pressure associated with the fluid diffusion. The observed diversity of focal mechanisms suggests that the fluid flow reactivate pre-existing fault zones. The study's findings contribute to our understanding of the interaction between active tectonics and hydrothermal fluid flow in extensional settings. The interplay between regional tectonics (NS-trending dextral faults) and the dynamics of fluid convection in the upper crust is highlighted.

This study provides valuable insights into hydrothermal fluid dynamics within active volcano-tectonic systems. The temporal changes in seismicity patterns in the Krýsuvík-Kleifarvatn area reflect variations in fluid pressures within the upper crust. The upward fluid migration from a deep reservoir, confirmed by seismic and geodetic data, suggests a complex interplay between tectonic events and hydrothermal systems. The high permeability and heat flux estimations highlight the significant role of fluids in shaping the area's geological and geophysical characteristics. Future research could focus on refining the fluid properties and further investigating the interaction between different fault systems and the fluid flow.

The study is based on two snapshots in time (2005 and 2009), limiting the ability to continuously track the evolution of the fluid pressure front. The precise composition and properties of the fluids involved remain uncertain. The permeability estimations, while consistent across different methods, rely on assumptions about fluid viscosity and rock porosity, which could influence the results. The model assumes a homogenous isotropic permeability which may not be entirely accurate, reflecting the complexity of the fractured rock mass. The generalizability of these findings to other geothermal areas needs further investigation.