Harnessing Geothermal Energy in Hengill:

Situated at the convergence of the Reykjanes Ridge, Western Volcanic Zone, and South Iceland Seismic Zone, Hengill boasts a distinctive landscape shaped by these tectonic forces and the drifting of the Eurasian and North American tectonic plates. Abundant in geothermal resources, the Hengill area hosts two major power plants operated by Orkuveita Reykjavíkur  (OR, https://www.or.is/en/). With boreholes reaching depths of 2 km, hot fluids are extracted with a rate of 300 Kg/s/m² generating 303 MW of electricity  and 400 MW of thermal energy. Approximately 60% of the extracted fluids are reinjected back into the subsurface in order to mitigate the increasing fluid deficit within the reservoir.

Despite the partial re-injection, the reservoir suffers a pressure decline that, by decompression boiling, leads to steam formation. While subsurface steam accumulations are clean and valuable energy resources, they alter the thermodynamic evolution of the reservoir and entail diverse risks, such as reservoir roof collapse and flooding of cold adjacent fluids. The pressure decline also dpromotes rock compaction and, ultimately, land subsidence.

Challenges in Monitoring Geothermal Reservoirs:

Steam coexist with liquid water at reservoir depths in Hengill, forming what is called a two-phase fluid. The relative amount of each phase is quantified by the steam fraction, whose estimation is an inherently challenging task. However, monitoring the steam fraction is of special interest for both operational and economic perspectives.

To achieve accurate measurements, production boreholes must be temporarily halted for approximately two weeks, allowing the reservoir to reach a steady state. Additionally, when opening the borehole to take measurements, the ascent of liquid and gas phases at different velocities, coupled with flashing, complicates the process. The steam fraction is then roughly estimated based on temperature and pressure measurements. For these reasons, steam fraction estimations are very scarce both in time and space. Quantifying the steam fraction present in the subsurface from indirect and surface-based measurements is still an unsolved matter.

Innovative Monitoring Technique for Geothermal Reservoirs:

The interferometry of seismic noise offers a passive and non-intrusive method for subsurface imaging and monitoring. This method allows studying the mechanical properties of the subsurface by estimating the time evolution of seismic velocities. We apply this approach to the seismic recordings in Hengill, using data from the COSEISMIQ project (http://www.coseismiq.ethz.ch). We also estimate the surface deformation in the area via Interferometric Synthetic Aperture Radar (InSAR) and work with in situ borehole data. We observe a clear correlation between the long-term seismic velocity evolution, the steam fraction present in the reservoir and the land subsidence overall. The evolution of these observables, stemming from geothermal energy harnessing, exhibit significant variations near the Hellisheidi power plant and diminish with distance.

Using in situ borehole data, we perform a rock physics model for both P- and S-waves to analyse the expected seismic velocity evolution and understand the role of the different thermodynamic observables on seismic velocities. The model reveals that the transition from liquid water to steam results in a 20% decrease in bulk modulus accompanied by 1% reduction in density over a span of 10 years. Consequently, the modeled seismic velocity for P-waves decreases as a function of the steam buildup while for S-waves, the velocity increases as a function of the density, which, in turn, depends on the pressure and temperature conditions. Notably, we find a striking resemblance between the P-wave velocity model and the observations from a surface-deployed seismic station.

Future Implications and conclusions:

This study introduces seismic noise interferometry as a powerful tool for monitoring subsurface two-phase fluids with minimal infrastructure, only one seismic station.
 Furthermore, distinguishing P- and S-waves in seismic noise correlations would enable monitoring phase changes as well as pressure and temperature variations within the reservoir. Beyond geothermal sites, the methodology could extend to diverse geological settings, such as volcanoes, CO₂ storage sites, and hydrocarbon reservoirs, among others.