Hydrochemistry of produced water from the Pohang EGS project site, Korea: implications for water-rock reactions and associated changes to the state of stress accompanying hydraulic fracturing of granite

Rob Westaway, Neil Burnside, David Banks

Research output: Contribution to conferencePaperpeer-review

16 Downloads (Pure)

Abstract

During August 2017, ~1700 m3 of water was injected into, and subsequent allowed to flowback from, the ~4 km deep PX-1 well, drilled into Permian granite at Pohang, Korea. Sampling of this produced water and its hydrochemical analysis, for major elements, trace elements, and stable isotopes, has provided an important dataset for the investigation of water-rock reactions in granite. As a first approximation, the variations in composition of this produced water are consequences of mixing of end-member compositions representing the injected surface water and in situ groundwater. The majority of the produced water was in situ groundwater, indicating that only a small proportion, estimated as only ~400 m3 , of the injected surface water was recovered; the rest remained in the subsurface. Superimposed on these variations are increases in concentrations of ionic species, which reflect chemical reactions between the injected surface water and the granite. These include silica from hydrolysis of quartz, sodium from hydrolysis of feldspar, and sulphate probably from hydrolysis of pyrite, this latter mineral forming deposits that line fractures within the Pohang granite. The MW 5.5 Pohang earthquake occurred on 15 November 2017, three months after this injection experiment, in close proximity to the EGS site, initiating discussion regarding the possibility of a cause-and-effect connection. This oblique reverse-faulting earthquake occurred on the Namsong Fault, which passes above the injection point and transects the wellbore, its hypocentre being within a kilometre of the injection. We explore the possibility that injection at the EGS site might have caused this earthquake. Assuming that the injected surface water that was not recovered entered this fault, and given the ~160 °C ambient temperature, we calculate timescales for chemical re-equilibration of different ionic species. Re-equilibration of silica, for example, is thus estimated to require roughly three months, the time difference between the injection and the earthquake. We thus derive a quantitative model for this process, in which hydrochemical re-equilibration is envisaged as occurring by dissolution of asperities, where patches of fault are in mechanical contact, thus ‘unclamping’ the fault. This model leads to a new scaling law between injected volume and the maximum seismic moment, or magnitude, of resulting induced seismicity. This new scaling law is less conservative than existing theory, which assumes that pore space throughout a rock volume is flooded with injected fluid, because only the volume of the fault is flooded, the surrounding granite having extremely low porosity. It is thus shown that net injection of ~1000 m3 of surface water can account for an induced earthquake comparable to that observed, provided that the seismogenic fault was already critically stressed. This analysis does not prove, of course, that the August 2017 injection experiment caused the MW 5.5 Pohang earthquake; this earthquake might, for example, have been a consequence of the combined effect of all five injection experiments that took place at this site during 2016-2017 (three in well PX-2, including one in September 2017, and two in well PX-1), before the Korean authorities forced suspension of the EGS project. It nonetheless provides a salutary warning of a potential adverse consequence of injection of surface water into granite. A conceivable remedy, albeit at considerable cost, would be to inject water containing dissolved ionic species that match as closely as possible, or indeed exceed, the concentrations in the local groundwater.
Original languageEnglish
Number of pages12
Publication statusPublished - 21 Oct 2021
EventWorld Geothermal Congress 2020+1 - Harpa, Reykjavík, Iceland
Duration: 24 Oct 202127 Oct 2021
https://www.wgc2020.com/

Conference

ConferenceWorld Geothermal Congress 2020+1
Abbreviated titleWGC
Country/TerritoryIceland
CityReykjavík
Period24/10/2127/10/21
Internet address

Keywords

  • Pohang
  • Korea
  • EGS
  • granite
  • anthropogenic seismicity
  • hydrochemistry
  • geomechanics

Fingerprint

Dive into the research topics of 'Hydrochemistry of produced water from the Pohang EGS project site, Korea: implications for water-rock reactions and associated changes to the state of stress accompanying hydraulic fracturing of granite'. Together they form a unique fingerprint.

Cite this