Numerous factors influence how hydraulic fracturing and wastewater disposal activities during natural gas development may cause earthquakes, a phenomenon known as induced seismicity. These factors are either geological (related to the rock formations that host the natural gas resources) or related to completion practices (the process of hydraulically fracturing a drilled well to produce natural gas from zones that would otherwise be uneconomic to produce), or a combination of these factors.
Geoscience BC has been funding induced seismicity research projects in BC’s Northeast Region since 2012. Additional research was needed to better understand when, where and why seismicity occurs. The results will inform industry, government and community decisions so that the likelihood and magnitude of future induced seismicity can be reduced.
This Energy (Oil and Gas) project fits under Geoscience BC’s Strategic Objective of Facilitating Responsible Natural Resource Development and our goal to:
- Maintain joint research with partners examining seismicity induced by hydraulic fracturing in northeastern BC to provide new science to better understand induced seismicity, mitigate risks and further improve regulation and industry practices.
Specifically, this project aimed to:
- Compile a database of well completion parameters and a limited database of geologic risk factors for wells in the study area (Montney Play);
- Create a Montney Play-specific predictive model for probability and magnitude of induced seismicity;
- Develop a publicly available methodology for the statistical interpretation of seismicity risk; and
- Make recommendations for Montney Play-specific mitigation measures to reduce the influence of high-importance features on induced seismicity.
Benefits of this research include:
- Development of new technologies and methodologies to aid hydraulic fracturing management;
- New public peer-reviewed research to inform future natural gas development and regulation;
- Increased number of induced seismicity subject matter experts; and
- Legacy methodology, models, and databases for other researchers.
The project focussed on the Montney Play, a gas-prone area in the Western Canadian Sedimentary Basin in BC’s Northeast Region that contains some of North America’s most significant natural gas deposits.
What Was Found?
Data for the study came from several earthquake catalogues, published and proprietary maps, and two collections of hydraulic fracturing data.
The study used a multivariate statistical approach (observing or analyzing more than one outcome variable) to assess publicly available data on well completion and geologic factors to help identify which hydraulic fracturing and geologic parameters show the strongest correlation to induced seismic events. Portions of the data analysis were completed using machine learning algorithms.
The researchers concluded that:
- The analysis does not single out one or more clearly causal features that are responsible for induced seismicity from hydraulic fracturing in the Montney Play.
- Classifying wells as seismogenic and predicting induced event magnitude are highly dependent on such factors as the specific dataset used; the model used; and the well, completion and geologic features selected for inclusion in models.
- Top of Montney structure and distance to faults (normal and thrust) are important factors in magnitude regression models. The number of stages and a variety of structural geology proxies appear to relate to increased magnitude seismic events. As expected, geological features showed a dominate role in the modeling.
- Development density, distance to listric faults, total fluid injected, geothermal gradient and Paleozoic structure appear to be keys factors in positive seismogenic classification.
- Machine learning alone cannot decide which features to include in final modeling – a considerable amount of thought is required to apply appropriate filters and guide the machine learning process.