Using satellite data and computer modelling, geoscientists may now be able to analyse and predict seismic events caused by wastewater injection, writes Ros Davidson
Increasingly, the US is feeling the effects of earthquakes linked to wastewater injection by oil and gas operators. In the central US, in states such as Texas and Oklahoma, between 1973 and 2008 there was an average of 24 earthquakes a year in the US with a magnitude of 3.0 or larger on the Richter scale. But from 2009 to 2015, there were 318 yearly, an increase of more than thirteen times.
While some are unconvinced, the United States Geological Society (USGS) has unequivocally stated on its website: “Wastewater disposal is the primary cause of the recent increase in earthquakes in the central United States.” Fracking can also cause induced quakes, but because the volume of liquid is much smaller, the number of quakes is much smaller.
Wastewater comes from a variety of sources, whether as a by-product of extraction, or left over from high-volume fracking. Whatever its origins, about 2 billion gallons (7.5 billion litres) of it are injected daily into an estimated 180,000 wells in the US, mostly in Texas, Oklahoma, Kansas and California.
The accelerating seismicity means that an estimated 7 million of people and their communities are at risk, the USGS warned in its annual seismic outlook published in March. Insurance rates are higher for the oil and gas sector in the affected areas, and in some cases operators are being sued for quake damage. The publicity is also tainting the industry: the current popular buzz is that only fracking causes quakes, even although wastewater injection has been identified as the most frequent culprit.
But a new way to look at the issue is now emerging. A team of scientists led by Arizona State University has devised a method which could be used to predict or prevent induced quakes. Combining satellite radar measurements and a new computer model which predicts underground “pore elasticity,” the tools could prove useful in aiding operators and state or federal agencies to minimise the danger.
The research, conducted by geophysicist Manoochehr Shirzaei of Arizona State University and colleagues, is more than ground-breaking, so to speak. Co-authored between William Ellsworth of Stanford University, Kristy Tiampo of the University of Colorado Boulder, Pablo González of the University of Liverpool and Michael Manga of UC Berkeley, and published in the respected journal Science, the work is among the first to measure – definitively – uplift on the ground above wastewater wells.
Their model also shows how water pressure in pores in underground rock can radiate out from wastewater wells to reach fault zones, where it could trigger small or moderate quakes. Even small quakes can lead to larger seismic events as they destabilise a region’s once finely balanced geology. The implication, as Shirazei told InnovOil, is that better monitoring would enable safer injections: “We could minimise the probability of large quakes – that’s a huge thing,” he said.
The research could guide where wastewater wells should be located – such as in what sort of geology and at what depth – and when injection should be slowed or halted. It is crucial to determining exactly how slowly injection should be tapered off.
In September, Oklahoma regulators ordered the shut-down of 37 wastewater wells connected to the Arbuckle formation after a tremor that tied as the state’s worst ever. The closures were ordered over a period of 7 to 10 days, for fear of causing more earthquakes if the volume of wastewater injected along a faultline were suddenly reduced. The regulators‘ move was based on current knowledge, but Shirazei and his co-authors have found that there could be greater pore pressure underground even after injection was halted for some time.
So SAR, so good
To achieve these results, the researchers analysed more than three years of radar data from the Japanese Advanced Land Observing Satellite (ALOS), which contains interferometric synthetic aperture radar (InSAR), a remote satellite-based sensing technique.
Synthetic aperture radar uses pulsed radio waves which are transmitted over a target area. The echoes of these pulses are then received and recorded, and the process is repeated. Multiple signals are sent from a single moving antenna and the successive signals, when processed, offer more detailed, higher-resolution images than with a single pulse. InSAR uses two or more of these images, using the differences in wave phases to measure surface deformation in millimetre-detail.
Using this technology, researchers looked at the highly accurate measurements of the ground from May 2007 to November 2010 in and around Timpson, Texas, an area which also contained four high-volume disposal wells. They chose that location in large part because they were able to access continuous satellite readings, as well data from the state on injection volumes. There were wells at different depths, and there was variable geology. They found that the earth’s surface around two of the wells bulged by as much as 3mm a year. Uplift was detected up to 8 km from the wells. “For this Timpson study area we got lucky and could see the deformation,” Shirazei said. Many previous studies have found that ground can subside when oil, gas or water is extracted, but this research is among the first pieces of research to find uplift after fluid is pumped into the ground.
The wells in Timpson where there was more deformation were shallower, he told InnovOil, and the surrounding rocks were more compliant and softer, suggesting they may have deformed more. The other wells nearby were in denser rock, and pore pressure could thus have been prevented from reaching crystalline basement rock and triggering quakes, he added.
The researchers also plugged the radar data into their new poro-elastic model, along with histories of injected water using data from operators that must be submitted to the Texas Railroad Commission. They found that a front of pore pressure – starting at the injection site – had moved underground, eventually affecting faults and possibly triggering quakes between depths of 3.5 and 4.5 km, depending upon the specific geology of a location. The model uses geological and hydrogeological information.
Indeed, two years or more after the data were gathered, in 2012 and 2013, a swarm of small quakes hit the Timpson area, including a moderate one of 4.8 magnitude, the largest ever recorded in east Texas. While it would support the team’s theory, the results are not definitive: the quakes were adjacent to the wells where there was less bulging in the earth’s surface, and about 25 km from the wells where there was more.
One of their findings – that seismic activity can still increase even after a cessation of injection operations – could be because it takes time for the wastewater to travel underground, or because there could be new stresses arising if injection is suddenly stopped. “If you stop injection today, it’s possible that earthquake activity goes on for the next decade or so,” Shirzaei commented, drawing on the Timpson study in particular.
Of wastewater wells in general, he told InnovOil that ideally injection should not be shut down suddenly, as it may cause a sudden reduction of normal stress on optimally oriented faults and could even trigger earthquakes. He also said that, in general, pore pressure in wastewater wells can remain high for a “long time after shut-down, thus the probability of the large earthquakes remains high in the region for a long period of time.” Equally, quite often the pressure can return to normal relatively quickly and injection can resume.
Model for monitoring
The research has been well received, especially for successfully extracting the earth’s “crustal strain” from the complex InSAR data, which also picks up factors that can change year-round, such as vegetation. Neither are the results fully conclusive. Some scientists have noted, for example, that rock stiffness can mean there is more pressure, not less. Others have also commented that the example studied might have been more satisfying had more deformation been detected above the deeper wells specifically.
Nevertheless, if the research can be replicated, its applications could be wide-ranging. Not only could the location of wells be fine-tuned, but the injection itself could also be customised, based on the findings of real-time monitoring, so that a critical stage of deformity is never reached.
Perhaps not surprisingly, Shirazei has meetings set up with oil and gas operators and government agencies so far in Texas, Oklahoma and Colorado and has begun speaking to the US Department of Energy (DoE) regarding additional funding to expand the team’s research.
He says the monitoring itself not expensive. For example, some space agencies may even offer similar satellite data for free. States also collect injection data – daily in the case of Oklahoma, monthly in Texas – meaning most of the raw information should be readily available. The researchers, in addition, have made their poro-elastic model public (although it does require a supercomputer to handle it).
With further work, and hopefully increased co-operation between regulators, data providers and the industry, techniques like InSAR could become a powerful new tool in the monitoring and safety arsenal.