Abstract

A pilot carbon dioxide (CO 2) sequestration experiment was carried out in the Michigan Basin in which ∼10,000 tonnes of supercritical CO 2 was injected into the Bass Island Dolomite (BILD) at 1050 m depth. A passive seismic monitoring (PSM) network was operated before, during and after the ∼17-day injection period. The seismic monitoring network consisted of two arrays of eight, three-component sensors, deployed in two monitoring wells at only a few hundred meters from the injection point. 225 microseismic events were detected by the arrays. Of these, only one event was clearly an injection-induced microearthquake. It occurred during injection, approximately 100 m above the BILD formation. No events, down to the magnitude −3 detection limit, occurred within the BILD formation during the injection. The observed seismic waveforms associated with the other 224 events were quite unusual in that they appear to contain dominantly compressional (P) but no (or extremely weak) shear (S) waves, indicating that they are not associated with shear slip on faults. The microseismic events were unusual in two other ways. First, almost all of the events occurred prior to the start of injection into the BILD formation. Second, hypocenters of the 94 locatable events cluster around the wells where the sensor arrays were deployed, not the injection well. While the temporal evolution of these events shows no correlation with the BILD injection, they do correlate with CO 2 injection for enhanced oil recovery (EOR) into the 1670 m deep Coral Reef formation that had been going on for ∼2.5 years prior to the pilot injection experiment into the BILD formation. We conclude that the unusual microseismic events reflect degassing processes associated with leakage up and around the monitoring wells from the EOR-related CO 2 injection into the Coral Reef formation, ∼700 m below the depth of the monitoring arrays. This conclusion is also supported by the observation that as soon as injection into the Coral Reef formation resumed at the conclusion of the BILD demonstration experiment, seismic events (essentially identical to the events associated with the Coral Reef injection prior to the BILD experiment) again started to occur close to a monitoring arrays. Taken together, these observations point to vertical migration around the casings of the monitoring wellbores. Detection of these unusual microseismic events was somewhat fortuitous in that the arrays were deployed at the depth where the CO 2 undergoes a strong volume increase during transition from a supercritical state to a gas. Given the large number of pre-existing wellbores that exist in depleted oil and gas reservoirs that might be considered for CO 2 sequestration projects, passive seismic monitoring systems could be deployed at appropriate depths to systematically detect and monitor leakage along them.

Highlights

  • A pilot carbon dioxide (CO2) sequestration project was carried out in Otsego County in the northern part of the Michigan Basin (Figure 1a)

  • The objectives of the demonstration project at this site were to evaluate the injectivity, storage capacity and overall suitability of the Bass Island Dolomite (BILD), a saline aquifer at 1050 m depth as a potential repository for CO2 separated from natural gas produced from the Antrim shale

  • Prior to the BILD injection, a packer was installed in the western monitoring well below the sensor array to prevent potential upward flow of CO2 from the BILD formation inside the well

Read more

Summary

Introduction

A pilot carbon dioxide (CO2) sequestration project was carried out in Otsego County in the northern part of the Michigan Basin (Figure 1a). The vast majority of the 225 seismic events occurred during the first days of the monitoring period and the seismicity rate decays to almost zero before the shallow injection into the BILD formation was even started This was completely unexpected as fluid-injection induced microseismicity is expected to start only after the onset of injection. The rapid decay of P wave amplitudes across the array indicates that these events have occurred in very close vicinity to the arrays since typically 2-3 sensors were located in the seismic near field (further discussed in Bohnhoff and Zoback, subm.). The remaining number of sensors was not sufficient to reliably calculate magnitudes

Discussion
Findings
Conclusions
Figure captions:
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call