Detecting and Quantifying Man-Made Changes in Earthquake Activity
In 2011, a 5.6-magnitude earthquake struck Oklahoma, damaging 14 homes and injuring two people. The cause of the earthquake has been linked to activities related to oil and gas production, specifically wastewater disposal.
The occurrence of such quakes by man-made causes is called “induced seismicity.”
Among other efforts to reduce the risk associated with induced seismicity, PhD student Pengyun Wang, along with CEE professors Mitchell Small and Matteo Pozzi and University of Pittsburgh professor William Harbert, have been developing approaches for detecting and quantifying increases in the frequency of man-made earthquakes.
Wang and the professors are using data such as magnitudes and epicenters from Oklahoma quakes for their work. Through one line of research, Wang and his advisers have established a statistical model that tells them when a critical shift in the frequency of seismic activity occurs.
By monitoring for these shifts, the researchers can know early when an area is exposed to an increased risk of earthquakes. For Oklahoma, the model could have detected the shift leading up to the 2011 quake as early as 2009.
“We want to prevent a severe consequence from happening by detecting induced seismicity when it just begins,” Wang said. Through a related line of research, Wang and the professors can determine the level of risk a detected shift in activity poses. Using a different statistical method and the Oklahoma data, they can determine not only when seismic activity increases, but how much that activity increases over time.
“Mere detection isn’t good enough,” Wang said. “Induced seismicity can be of small consequence—it can also be big. We need to quantify the [rate] of induced seismicity because a higher rate means a higher risk.”
By knowing the rate of seismic activity, the researchers can determine the probability of an earthquake above a certain magnitude striking an area in the upcoming months. With methods established for better detecting increases in seismic activity, the researchers’ now are working to break down that activity even further by mapping the locations and frequency of earthquakes across Oklahoma.
Although currently they can determine risk for large swathes of the state, they are developing ways to determine the varying levels of risk for different areas within it.
“We have been studying the entire region as a system, but the system has a lot of connected components,” Wang said.
By determining risk associated to different parts of the state, the researchers can produce hazard maps for the region that show decision-makers which area is subject to the largest increased hazard of earthquakes.
Through all of their work, the researchers are making it easier to determine when and where seismic activity presents danger. If they can detect increases in activity early, they can save time for implementing strategies to mitigate the effects of earthquakes.
For example, people using the models could alert oil and gas companies to operate wells differently in order to potentially prevent a destructive quake. They can also alert decision-makers like government agencies and building engineers to reassess earthquake codes and building requirements in affected areas. The models, although tested in Oklahoma, can be applied to any area.
If they can successfully map risk, Wang and the professors at CMU and Pitt can potentially reduce the occurrence and damage caused by earthquakes like the ones they are studying.