Energy Technologies: How They Work and Their Induced Seismicity Potential
Much of the energy used in the United States comes from fluids pumped out of the ground. Oil and gas have been major energy sources in the country for over 100 years, and new developments in the production of natural gas indicate that it may provide a significant source of energy for the nation during the twenty-first century. Geothermal power has been used to supply energy in the United States for almost as long as oil, although major electricity generation from geothermal energy sources began only in the 1960s at The Geysers in Northern California. A 2006 report on the potential of geothermal energy (MIT, 2006) suggested it could be a major contributor to the nation’s energy supply in the coming decades. Efforts to reduce concentrations of carbon dioxide (CO2) in the atmosphere have spurred development of technologies to capture and store (sequester) CO2. Projects to accomplish carbon capture and storage (CCS) from industrial facilities are currently being piloted in the United States and elsewhere in the world. Underground injection of CO2 has also been commonly used to enhance oil and gas recovery.
This chapter reviews the potential for induced seismicity related to geothermal energy production, conventional oil and gas development (including enhanced oil recovery [EOR]), shale gas development, injection wells related to disposal of wastewater associated with energy extraction, and CCS.
Geothermal energy exists because of the substantial heat in the Earth and the temperature increase with depths below the Earth’s surface. Depending upon the regional geology—including the composition of the rocks in the subsurface and any of the fluids contained in the rocks—the temperature increase with depth (the thermal gradient) may be fairly steep and represent the source of sufficient geothermal energy to allow commercial development for electricity generation. The largest actively producing geothermal field in the United States at The Geysers in Northern California generates approximately 725 megawatts of electricity per year (“megawatts electrical” or MWe). This is enough to power 725,000 homes or a city the size of San Francisco. Currently this geothermal field supplies nearly 60 percent of the average electricity demand of the northern coastal region of California.
The most likely regions for commercial development of geothermal power are generally the same regions that have experienced recent volcanism (Figure 3.1). Such areas are concentrated in the western portion of the country. The U.S. Geological Survey (USGS) estimates that the total power output from the hydrothermal (vapor- and liquid-dominated) geothermal resources in the United States can probably be increased to 3,700 MWe per year, and a 50 percent probability exists that it can be increased to about 9,000 MWe per year (Williams et al., 2008). Two recent studies have produced nationwide estimates of the electric power potential that might be achieved by a successful implementation of enhanced geothermal systems (EGS) technology, perhaps contributing 100,000 MWe of electrical power per year (MIT, 2006). More recently the USGS (Williams et al., 2008) has published a mean estimate for potential EGS development on private and accessible public land at
FIGURE 3.1 The location of the geothermal provinces in the United States. Within the United States the regions of relatively high thermal gradients, shown in red, exist only in the West. The typical local geologic setting for these high-geothermal-gradient areas is within sedimentary basins located near or intruded by recent volcanics, or within (as part of) the buried volcanic rocks themselves. Only one vapor-dominated reservoir has been developed in the United States (The Geysers); the remainder of the areas in red and orange may host viable liquid-dominated or enhanced geothermal system reservoirs. SOURCE: SMU Geothermal Lab; Blackwell and Richards (2004).
517,800 MWe. This is approximately half of the current installed electric power generating capacity in the United States.1
The three different forms of geothermal resources are recognized: (1) “vapor-dominated,” where primarily steam is contained in the pores or fractures of hot rock; (2) “liquid-dominated,” where primarily hot water is contained in the rock; and (3) “hot dry rock,” where the resource is simply hot and currently dry rock that requires an EGS to facilitate development (see Figure 2.1). Vapor- and liquid-dominated systems are collectively termed hydrothermal resources. The vast majority of known hydrothermal resources are liquid dominated.
The different forms of geothermal resources result in significant differences in the manner in which they are developed and particularly in the manner that liquids are injected to help stimulate energy development. Different injection practices can cause induced seismicity through different processes. The nature of and differences among the induced seismicity that may result from each of the three geothermal resources are summarized here.
Vapor-Dominated Geothermal Resources
A limited number of localities in the world exist where the geothermal resources naturally occur as steam. Despite their rarity, the two largest geothermal developments of any kind in the world are both vapor-dominated geothermal reservoirs. The Larderello geothermal field in the Apennine Mountains of northern Italy became the first of these and has generated electricity continuously since 1904, except during World War II. However, the most productive geothermal field development in the world is The Geysers (Figure 3.2), located about 75 miles north of San Francisco. The Geysers also has the most historically continuous and well-documented record of seismic activity associated with any energy technology development in the world.
The first commercial power plant at The Geysers came online in 1960 with a capacity of 12 MW (Koenig, 1992). Over the next 29 years the installed generation capacity was increased to a total of 2,043 MW through building 28 additional power plant turbine-generating units (CDOGGR, 2011). The basic elements of the process to generate electricity in this type of power plant are illustrated in Figure 3.3.
These plants were supplied with steam from 420 production wells, with the steam capable of flowing up the production wells under its own pressure. The condensed steam not evaporated at the power plant cooling towers was being reinjected into the steam reservoir by using 20 injection wells drilled to similar depths. The area of development had been expanded from the original 3 square miles to about 30 square miles. Because the generation of energy from the field consumes natural steam originally in the reservoir, by 1988
FIGURE 3.2 Ridgeline Unit 7 and 8 Power Plant (rated at 69 MW) in the left foreground at The Geysers in California. The turbine building, housing the two turbine-generator sets, the operator’s control room, and various plant auxiliaries are on the left. The evaporative cooling tower with steam emanating from the top is on the right of the main complex. The beige pipelines along the roads (with square expansion loops) are the steam pipelines that gather the steam from the production pads and bring it to the plant. A high-voltage transmission line (denoted by lattice towers) is in the middle foreground of the picture. SOURCE: Calpine.
the production of steam had started to decline; this decline was marked by a significant decrease in reservoir pressure from an original pressure of about 500 pounds per square inch (psi)2 to levels as low as 175 psi (Barker et al., 1992). For years the annual injection volumes returned to the geothermal reservoir were less than a third of the amount of steam being produced, so the reservoir was drying up. New sources of water were established by constructing two pipelines that currently deliver about 25 million gallons of treated wastewater a day for injection, increasing the current annual mass replacement to 86 percent compared to 26 percent back in 1988 (CDOGGR, 2011).
Early reports of induced seismicity at The Geysers, begun by USGS researchers (Hamilton and Muffler, 1972), described microseismicity that was observed close to where
2 A car tire for a standard, midsized automobile is usually inflated to a pressure of about 30-35 psi for comparison.
FIGURE 3.3 Elements of the power plant cycle for vapor-dominated geothermal resources. The steam is directed by the main steam line into a turbine that spins the connected generator unit, typically generating electricity at 13.8 kilovolts (kV), which a transformer increases to 230 kV for distribution by a transmission line. The steam leaving the turbine enters the condenser that contains a network of tubing through which cool water is circulated, facilitating the condensation process. The condensate is then pumped to the cooling tower where it is cooled by evaporation, with the cooled water being in part recirculated by the circulating water pumps back to and through the condenser. Because some noncondensable gases usually occur naturally in the steam, those gases are removed from the condenser by the gas ejector system that creates a partial vacuum by the flow of a small amount of steam delivered by the auxiliary steam line. Those gases, in particular H2S, are chemically processed commonly by a Stretford System before delivery to the cooling tower where they are vented. SOURCE: Adapted from the Northern California Power Agency.
the geothermal development operations were taking place. As the area of steam field development expanded, the areal distribution of seismic events similarly expanded, and the number of the events progressively increased (Figure 3.4).
With the addition of more seismometers of increased sensitivity distributed throughout the expanded development area, a clear association became evident between these induced events and the active injection wells and volume of water being injected. Figure 3.5 shows where injection took place in the southeastern part of The Geysers in 1998, the year following the startup of the first wastewater pipeline that more than doubled the injection volume. During 1997-1998, 1,599 events of M ≥ 0.6 were recorded, an increase of just over 50 percent compared to the prior 12 months.
The history of steam production, water injection, and seismic history at The Geysers since 1965 is shown in Box 3.1. Steam production and therefore electricity generation reached a maximum in 1987, followed by a fairly rapid decline until the wastewater pipelines
FIGURE 3.4 Geysers seismicity maps in 10-year intervals show the expanding distribution of development as illustrated by the increased numbers of green squares that indicate the locations of the operating power plants. SOURCE: Preiss et al. (1996).
FIGURE 3.5 The locations of injection wells and the location and depth distribution of seismic events in the southeastern part of The Geysers area during 1997-1998. Map on the left shows injection wells in 1998. The middle map shows the total number of recorded seismic events from the period 1997-1998 with the line of cross section (figure on the right). The cross section shows the positions of three geothermal wells with the location at depth of the seismic events (red dots). SOURCE: Beall et al. (1999).
began deliveries in 1997 and 2003. The annual amount of water injected followed the same trends until new sources of water other than condensate were developed, allowing recent injection to become nearly equal to the annual production levels.
The method of injection at The Geysers is unusual because of the extremely low fluid pressures in the deep underlying reservoir. No surface pressure is needed to inject; the water simply falls down the injection well as though through a partial vacuum because the fluid pressures in the reservoir are incapable of supporting a liquid level to the surface. Consequently, without elevated bottom-hole pressures, the primary cause of the induced
Geysers Annual Steam Production, Water Injection, and Observed Seismicity, 1965-2010
Figure The history of induced seismicity at The Geysers is shown in three forms. First, the number of recorded events of M 1.5* and greater is shown to have increased from almost none in the 1960s to 112 in 1975 and then to as many as 1,384 in 2006 (thick green line). Second, the annual number of earthquakes of M 3.0 and greater is shown along the bottom of the graph (pale green line). By 1985, 25 such events occurred annually, and that rate of about two events of M 3.0 and greater per month has continued to the present. Third, events of M 4.0 and greater are shown near the top (green dots). The first such event occurred in 1972, and more recently about one to three of these have occurred per year. The maximum magnitude was a M 4.67 event in May 2006. SOURCES: Adapted from Smith et al. (2000) and Majer et al. (2007).
*Note that this report uses M 2.0 as the general limit below which earthquakes cannot be felt by humans; however, at The Geysers M 1.5 is the lowest magnitude that the USGS can report faithfully year after year. Furthermore, residents in Anderson Springs may feel events as low as M 1.5 because the events are spatially quite close to the community.
seismicity is the fact that the hot subsurface rocks are significantly cooled by the injected water, and the resulting thermal contraction reduces the confining pressures and allows the local stresses to be released by limited movement on fracture surfaces.
The two strong motion recording instruments installed in 2003 near the neighboring communities of Anderson Springs and Cobb commonly record moderate shaking, plus about a dozen Mercalli VI (strong shaking) events each year (see also Chapter 1 for a definition of the Mercalli scale). The one event of Mercalli VII intensity caused an average acceleration of 21.0%g3 at Anderson Springs and was related to a M 3.03 seismic event located at a depth of 4,750 feet only 1.2 miles west of the recording instrument.
The operators at The Geysers meet regularly with representatives of these two communities, county government, federal and state regulatory agencies, the USGS, and the Lawrence Berkeley National Laboratory to discuss the field operations and the recently observed seismicity. Minor damage is occasionally caused by the induced seismicity at The Geysers, generally as cracks to windows, drywalls, or tile walls or flooring in these communities. A system for receiving, reviewing, and approving such damage claims attributed to the local seismicity was established 6 years ago, and the homeowners are reimbursed for their costs to have the home damage repaired. To date these reimbursements for home repairs total $81,000, and this system appears to be resulting in mutually satisfactory relationships.
Liquid-Dominated Geothermal Resources
In contrast to the development of the vapor-dominated geothermal resources, liquid-dominated resources commonly use downhole pumps in the production wells to deliver the thermal waters to surface facilities. Surface pumping facilities are needed to force the injected waters back down into the reservoir. The liquid-dominated geothermal reservoirs that have been commercially developed to produce electricity in the western United States are listed in Table 3.1 (sources include the California Division of Oil, Gas and Geothermal Resources [CDOGGR], the Nevada Commission on Mineral Resources, the Imperial Irrigation District, and various operators).
Several different methods are used to generate electricity in liquid-dominated geothermal systems depending primarily on the temperature of the produced fluids; the flash steam power cycle process and the binary cycle process are the most common (Figure 3.6).
The cause and extent of the induced seismicity related to the development of liquid-dominated geothermal resources are different from those in the vapor-dominated resources (Box 3.2). From the start of operations the amount of fluid produced from a liquid-dominated reservoir is almost fully replaced by injection, which prevents a signifi-
3 “%g” is motion measured as acceleration by an instrument, expressed as a percent of the acceleration of a falling object due to gravity.
cant decline in reservoir pressure. The temperature difference between the produced and reinjected waters is also relatively limited, so less cooling of the reservoir results. Consequently, if the surface and resulting bottom-hole pressures in the injection wells are limited to be less than that necessary to induce fracturing, little cause exists for the operations to produce significant induced seismicity. Monitoring at many of the liquid-dominated geothermal fields has demonstrated a relative lack of induced seismicity. However, as described below, the Coso geothermal field began as a strictly liquid-dominated field and has evolved during extended production to become partly vapor dominated. This evolution has resulted in reduction in fluid replacement and has caused the introduction of induced seismic events.
The Coso geothermal field provides a well-documented example of a complex resource area that was liquid dominated before the start of development 25 years ago and that may have evolved, following extensive production, into a resource that is now in part vapor dominated (see Box 3.2). Coso near Ridgecrest, in southeast-central California, is in a region of recent volcanism that is also seismically active. The first commercial geothermal power plant began operating in 1987; since 1989 three plants have been in operation with a total generating capacity of 260 MW, with about 85 production and 20 injection wells currently in use (CDOGGR, 2011). The geothermal fluids (dominantly water) are at temperatures in excess of 300°C (572°F) at depths of 1.5-2 km (~0.9-1.2 miles) (Feng and Lees, 1998).
The areal coincidence of the local seismicity at Coso with local surface subsidence, identified by using synthetic aperture radar data, suggest that the Coso field operations have caused reservoir cooling and thermal contraction, resulting in induced seismicity (Fialko and Simons, 2000). More recently, Kaven et al. (2011), based in part on their investigation of local changes in seismic velocities (Vp:Vs ratios), attribute the induced seismicity at Coso to decreases in fluid saturation and/or fluid pressure within the active geothermal reservoir.
An important issue to emphasize with regard to potential changes in pore pressure at vapor- and liquid-dominated geothermal power plants is the selection of conversion cycle—whether flash cycle or binary cycle (see Figures 3.2 and 3.6). The cycle selection is determined by the temperature and nature (physical state) of the geothermal fluids produced to the surface. Those power-cycle differences are important to explain why evaporative losses are significant at vapor-dominated resource power plants and moderate at flash cycle power plants. Evaporative losses can result in pore pressure and thermal losses that in turn can result in significant or moderate levels of induced seismicity. Equally important is to explain why in the case of binary cycle power plants there are no evaporative losses and generally little if any loss of pore pressure or fluid temperature, and therefore little if any associated induced seismicity.
TABLE 3.1 Liquid-Dominated Geothermal Fields in the United States with Operating Power Plants
|Plant Start Year||Power Cycle Used||Power Plant Capacity (MWe)||Average Generation (MWe)||Average Resource Temperature (°F)||Owner/Operator|
|East Mesa||1987||Binary & Flash||105||59||306||Ormat|
|Heber||1985||Binary & Dual Flash||92||75.9||324 to 350||Ormat|
|Salton Sea||1982||Single, Dual, & Triple Flash||352||314.6||480 to 690||Cal Energy|
|Coso||1987||Dual Flash||260||48||480 to 580||TerraGen|
|San Emidio||1987||Binary||3.6||2.6||275 to 290||U.S. Geothermal|
|Soda Lake||1987||Binary||26.1||10.4||360 to 390||Magma|
|Steamboat||1988||Binary & Flash||139.5||105.5||300||Ormat|
|Blue Mountain||2009||Binary||49.5||40||375||Nevada Geo|
|Dixie Valley||1988||Dual flash||67.2||41.2||400 to 480||TerraGen|
|Roosevelt||1984||Binary & Flash||37||34||510||Pacific Corp|
|Thermo||2008||Binary||10.0||6.6||250 to 390||Raser|
|Power subtotal 47 40.6
|Raft River||2008||Binary||13||8.4||275 to 300||U.S. Geothermal
|Chena Hot Springs||2006||Binary||0.73||0.5||165||Chena Energy|
FIGURE 3.6 (a) The fluids delivered to the surface by the production wells in a flash steam power cycle are passed through a flash vessel or separator; the separated steam that flows out of the top is directed into a power plant where it is used to spin a steam turbine connected to a generator that produces an electrical output. The spent steam travels through a condenser, and the condensate is then pumped to the cooling tower, where the liquids are cooled before some of the fluids are pumped back inside the condenser and some are combined with the water drained from the bottom of the separator and sent to the injection wells. (b) The produced fluids for binary cycle power plants are first passed through a heat exchanger to heat a secondary liquid, usually an organic fluid such as isopentane, which vaporizes (boils) at a lower temperature than does water. That vaporized secondary fluid is then used to spin a turbine generator to make electricity. Similarly, that vapor is then condensed and returned directly to the heat exchanger to be reheated, revaporized, and recycled without any fluid loss. The produced geothermal water that has passed through the heat exchanger is then delivered to the injection wells. SOURCE: Idaho National Laboratory.
Induced Seismicity at the Coso Liquid-Dominated Geothermal Field
Locally induced seismicity recorded in the area of the Coso geothermal field development between 1996 and 2008 in map view (Figure 1, top) and cross section (Figure 1, bottom) shows clustering relative to the location and depth of the geothermal wells shown in blue. The number of seismic events of magnitude 0.5 and greater is plotted; these events total 10,200.
The history of geothermal fluid (dominantly water) production, water injection, and recent seismic history at the Coso field from 1977 through 2011 is shown in Figure 2. Starting in 1987, annual production reached a maximum of 121 billion lb* in 1990 and had decreased to 68 billion lb by 2009, while annual injection has declined from a maximum of 80 billion lb to 27 billion lb (CDOGGR, 2011). The relatively low reinjection rate for a liquid-dominated resource is because of cooling tower evaporative losses that result from the produced fluids containing an increased steam fraction as reservoir pressures have declined over the almost 25 years of operation.
Using the catalog of data available from the Southern California Earthquake Data Center, the history of local seismicity at the Coso field from 1977 to 2009 is shown in Figure 2.
With reference to Figure 2, the number of events of M 1.5 and greater averaged 5 per year during the 10 years prior to development, then doubled in the first 5 years after 1987, reaching maxima of 51 in 1995, 55 in 1998-1999, and 64 in 2001 before declining to a current level of about 20 per year. The peaks in 1995 and in 1998-1999 were attributed by Bhattacharyya and Lees (2002) to triggering in response to significant (M > 5.0) nearby earthquakes at Ridgecrest and in the Coso range. Additionally, the number of earthquakes of M 3.0 and greater is shown near the bottom of the chart. Single events occurred in 1978, 1995, 1998, 1999, and 2007, with three in 2009. The single earthquake in 2007 was a M 4.11 event, as shown near the top of the chart.
*Note that where at least part of the production is in the form of steam as well as liquid water, “pounds” is needed as the single unit to describe both the quantity of production and injection because gallons or cubic meters cannot be used in reference to steam.
Enhanced Geothermal Systems
In addition to the vapor- and liquid-dominated resources already described, some regions have sufficiently high temperature at reasonably shallow depths for potential commercial development of EGS. To develop EGS some form of engineering is required to generate the permeability necessary in generally impermeable rocks to promote the circulation of hot water or steam for delivery to the surface at adequate rates to sustain operations. Previously referred to as “hot dry rock” projects, these systems are now referred to as “enhanced geothermal systems” or EGS (Figure 3.7).
The primary method employed to enhance rock permeability is hydraulic fracturing. This process, often termed “stimulation,” requires the injection of a liquid at sufficient pressure in one well to overcome the confining pressures at depth and to thereby force open incipient fractures and planes of weakness or to create new fractures to allow fluids
FIGURE 3.7 Schematic of an EGS development with an injection-production well pair and a power plant. The injection well (blue) is accompanied by a second (production) well (red) that is drilled to intersect the fractures generated by the injection well at a depth and appropriate lateral distance from the injection well. The distance allows the injected water to be sufficiently heated by the hot surrounding rock as it is circulated to the production well and pumped to the surface. Once at the surface the hot water can be flashed to steam or used to heat a secondary fluid that can be used in a binary cycle process. SOURCE: U.S. Department of Energy.
to flow more freely through the subsurface rock. The location of the new fractures can be determined by monitoring the microseismic response at the surface or downhole.
The history of the development of EGS projects in the United States began near the Los Alamos National Laboratory in New Mexico during the 1970s. That project provided a base for gaining experience in conducting hydraulic fracturing operations at high temperatures in low-permeability crystalline rocks. Data from this project have led to a series of similar EGS experiments in England, France, Germany, and Japan, followed more recently in Australia, Sweden, and Switzerland. In each case of active EGS development some induced seismicity has been registered. One recent example in Basel, Switzerland, generated an increased level of public awareness of the existence of induced seismicity (Box 3.3).
This Basel incident has become one of the best-known international induced seismic case studies, not because of local damage (which was minimal) but because of the immediate negative impact to the project due to the risk liability of induced seismicity. The urban setting for the project combined with the fact that this region is tectonically unstable and with a history of natural seismicity proved decisive in the project being terminated.
The occurrence of some post-shut-in seismicity at Basel and at another EGS project in Soultz-sous-Forêts, France, is a phenomenon that is not yet completely understood and can create added concern from the public standpoint in that some events are beyond the control of the operator. Understanding these post-shut-in events involves development of subsurface models with numerical simulations that can track the progress of the injected fluids through the rock and can calculate potential for further seismic activity. Development of coupled reservoir fluid flow and geomechanical simulation codes has been suggested as a way to advance this understanding (Majer et al., 2007) and may also have an impact on understanding post-shut-in phenomena related to other energy technologies (see also below).
CONVENTIONAL OIL AND GAS PRODUCTION INCLUDING ENHANCED OIL RECOVERY
In a conventional oil or gas reservoir, the reservoir rocks are generally pressurized above hydrostatic pressure due to compaction of sedimentary rocks over geologic time. The use of the term “reservoir” is common but may be misleading: the gas or oil does not exist in a single, large pool in the rocks, but in the pores of a rock formation. Compaction reduces the naturally occurring pore space in the rock (reduces the porosity) and either displaces reservoir fluids (hydrocarbons and water) or increases the pressure in the reservoir, or both. When penetrated by a well bore with the aid of pumping, fluids in the pressurized layer flow to the surface until the pressure in the reservoir is reduced to hydrostatic pressure. The reduction in pressure also causes gas to come out of the fluid, much like a bottle of soda when the cap is removed. The released gas can also help to drive the oil to the surface until the pressure is reduced to hydrostatic conditions.
Induced Seismic Activity in Basel, Switzerland
Basel, Switzerland, is in the southeastern region of the Upper Rhine Graben, a fault-bounded trough, and was selected as the site of a planned geothermal cogeneration plant. Basel is known to be an area of potential seismic risk but had not suffered a damaging earthquake since a M 6.2 earthquake in 1356 that destroyed much of the city. Due to awareness of historical seismicity, the geothermal project operators and planners had installed both borehole and surface seismic sensors that formed a network for monitoring any seismicity, whether natural or induced. The monitoring efforts included the drilling of six monitoring wells, ranging in depth from 300 m (~980 feet) to 2,750 m (~9,000 feet) in addition to a surface array of both weak and strong motion detectors. Recording of seismic activity began in early 2006 to record background seismicity.
The seismic monitoring arrays served several purposes. They recorded the background seismicity before well stimulation began and they were used to monitor the fracturing of the geothermal reservoir (the objective of the stimulation). Finally they could provide information (magnitude and location if possible) of any induced seismicity that might occur as a result of the stimulation. All monitoring stations were connected so that real-time data could be recorded and quickly analyzed.
The drilling of a deep geothermal well near the center of Basel (Figure 1) began in May 2006 and was completed some months later. Stimulation of the well to induce fractures for heat exchange with the geothermal source at 5,000 m (~16,400 feet) began on December 2 and was accompanied by a significant increase in the number of small seismic events (Figure 2). In accordance with the traffic light procedure—a procedure where increases in seismic activity beyond a certain, predetermined level trigger reactions by the operator to
Figure 1 Drilling activity in the middle of the city of Basel. SOURCE: KEYSTONE/Georgios Kefalas.
mitigate the occurrence of further events—injection was stopped in the early morning hours of December 8 after approximately 11,500 m3 (~3 million gallons) of water were injected (Deichmann and Giardini, 2009) and after the recording of M 2.6 and M 2.7 seismic events. During this injection period, more than 10,500 seismic events were recorded (Häring et al., 2008). While the well was shut in (operations terminated), seismic activity continued, so it was decided to “bleed off” the pressure (reduce pressure through controlled release). On December 8, an earthquake of M 3.4 occurred in Basel and was clearly felt by the local population. This was followed by three more events greater than M 3.0. The project, operated by Geopower Basel AG as a partnership of both public and private companies, was immediately suspended and then ultimately abandoned almost 3 years later following further study and risk evaluation after these seismic events. However, increased seismicity activity over historical levels is likely to continue for 7 to 20 years based on Bachmann et al.’s (2009) model for induced seismicity.
Figure 2 Seismic events and wellhead pressure at Basel. SOURCE: Kraft et al. (2009).
Flowing and pumped wells are considered “primary recovery” from the well and about 12 to 20 percent of the original oil in place in the reservoir is recovered in this manner. This relatively low rate of recovery results from several factors: (1) the decrease in natural reservoir pore pressure over time; (2) the natural porosity and permeability of the rock formation (which is an indication of how easily the oil can move through the formation to the well bore); and (3) the viscosity of the oil, which, when combined with porosity and permeability, is also an indicator of the ease with which oil can migrate through the rock. Recovery rates for natural gas are generally higher than for oil (up to 50 to 80 percent may be recovered through primary production methods) because gas expands naturally upon release of pressure and has a lower viscosity than liquid petroleum, contributing to the natural movement of gas up the well bore (Shepherd, 2009).
When primary recovery is no longer viable, petroleum companies may use a variety of technologies to extract the remaining oil and gas. These technologies include what are termed secondary and tertiary recovery methods; tertiary recovery is generally also referred to as enhanced oil recovery (EOR) (Shepherd, 2009). Figure 3.8 shows the differences between primary, secondary, and tertiary recovery methods.
FIGURE 3.8 Schematic showing the progression of oil production from primary to tertiary recovery. IOR, improved oil recovery; EOR, enhanced oil recovery. SOURCE: Al-Mutairi and Kokal (2011).
Primary Oil and Gas Production
Although felt seismic activity known to be related to primary petroleum production is uncommon relative to the large number of operating oil and gas fields worldwide, withdrawal (extraction) of oil and gas has been linked to felt seismic events at 38 sites globally, 20 of which were in the United States (Appendix C; Box 1.1). These have included events in Texas, Oklahoma, California, Louisiana, Illinois, and Nebraska, the majority of which have been of M < 4.0 (Appendix C; see also Chapter 1); the well-documented events at the Lacq gas field in southwestern France (see Box 2.5); and the large events in the Gazli gas field in Uzbekistan (Box 3.4). Withdrawal of oil or gas from the subsurface can result
Induced Seismicity Related to Natural Gas Extraction:
A Case from Gazli, Uzbekistan
The Gazli gas field is located about 500 miles (800 km) east of the Caspian Sea in a generally aseismic region of Uzbekistan. The gas deposits were discovered in 1956 and gas production began in 1962. The gas field lies within a large (38 km by 12 km [22.8 mile by 7.2 mile]) asymmetrical anticline over crystalline rocks. Large volumes of water were injected between 1962 and 1976 to enhance production, but subsidence and reduced gas pressures were reported despite this injection; the initial pressure in the gas field of about 70 atm (~71 bars or 1030 psi) in the 1960s decreased to about 30-35 atm (~30.4-35.5 bars or 435-515 psi) by 1976 and to about 15 atm (15.2 bars or 218 psi) by 1985. This pressure decrease indicates a net removal of mass, even with injection of large volumes of water. Thus, although the field operators had begun to use secondary recovery techniques (waterflooding), the cause of the earthquakes is attributed to pressure decrease due to fluid withdrawal.
On April 8, 1976, a M ~ 7 earthquake occurred about 20 km (12 miles) north of the gas field boundary. This was followed by another M ~ 7 earthquake on May 17, 1976. A third large earthquake (also M ~ 7) occurred on March 20, 1984. All three earthquakes had epicenters 10-20 km (6-12 miles) north of the gas field boundary, over an east-west distance of about 50 km (30 miles). Reported hypocentral depths of these large earthquakes were 10-15 km (6-9 miles). Geodesic measurements indicated surface uplift of some 70-80 cm (~28 to 31.5 inches) north of the gas field at the epicentral locations of the three large earthquakes; this uplift is consistent with thrust movement on faults dipping to the north. However, source modeling indicates that the ruptures progressed downward, which is uncommon for thrust mechanism earthquakes. The locations and magnitudes of these large earthquakes were determined from worldwide seismographic data and are therefore somewhat uncertain, leading to some uncertainty on the causal relationship between gas extraction and earthquake activity. Nonetheless, observations of crustal uplift and the proximity of these large earthquakes to the Gazli gas field in a previously seismically quiet region strongly suggest that they were induced by hydrocarbon extraction.
SOURCES: Adushki et al. (2000); Grasso (1992); Simpson and Leith (1985).
in a net decrease in pore pressure in the reservoir over time, particularly if fluids are not reinjected to maintain or regain original pore pressure conditions (see also other technology descriptions, below, and Chapter 2). This change in pore pressure can cause changes in the state of stress of the surrounding rock mass and of nearby faults, with the potential to result in induced seismic events.
Secondary Oil and Gas Recovery
Secondary recovery is the process of injecting water (often described as a “waterflood”) or gas (also known as pressure maintenance) into a petroleum reservoir. The water or gas replaces the produced hydrocarbons and water in order to maintain the reservoir pressures and is used to “sweep” an oil reservoir; injected gas may become dissolved in the oil, reducing the oil’s viscosity. Secondary recovery processes drive hydrocarbons trapped in the rocks from the injection well toward production wells (Shepherd, 2009; Figure 3.9). Waterflood or pressure maintenance projects can result in recovery of up to 40 percent of the initial petroleum in the reservoir (DOE, 2011). The number of permitted wells that use
FIGURE 3.9 Diagram illustrating waterflooding method of secondary recovery. SOURCE: NETL (2010).
waterflooding in the United States is about 108,000; in Texas alone, current data from the Railroad Commission of Texas indicate that more than 36,000 wells are currently permitted to use saltwater injection for the purposes of secondary recovery.4
Injection pressures and volumes in waterflooding projects are generally controlled to avoid increasing the pore pressure in the reservoir above the initial reservoir pore pressure. Nonetheless, reservoir pore pressure can increase as a result of waterflooding, and felt induced seismic events at 27 sites globally (18 of which have been in the United States) have been caused by or likely related to waterflooding (Chapter 1, Box 1.1; Appendix C). Waterflooding at the Rangely Field in Colorado induced seismic events with magnitudes up to M 3.4 (Chapter 2, Box 2.4). Near Snyder, Texas, seismic events with magnitudes as large as M 4.6 occurred in 1978 after the initiation of a large (25 million barrel per year [10.2 trillion gallons per year]) waterflooding project in Cogdell Field (Davis and Pennington, 1989; Nicholson and Wesson, 1990; see also Appendix C).
Tertiary Oil and Gas Recovery (EOR)
Tertiary recovery is the process of recovering greater amounts (often greater than 50 percent) of the original oil and gas contained in a reservoir (DOE, 2011) and is generally, though not exclusively, initiated after the use of secondary recovery operations.5 In addition to maintaining reservoir pore pressure, EOR methods help displace the hydrocarbons toward the production well. These methods can be broadly grouped into three main categories: thermal, miscible displacement, and chemical injection (polymer flooding) (Shepherd, 2009). Chemical injection methods are primarily used in California but are not commonly used elsewhere in the United States and are not discussed further. Note also that “other” methods in Figure 3.8 include microbial, acoustic, and electromagnetic methods; these are not frequently used and are not discussed further.
Thermal techniques change the viscosity of oil in the reservoir by heating it through the injection of steam or air (Shepherd, 2009). Heating lowers the viscosity of the fluid and allows hydrocarbons to flow more easily through a reservoir toward a production well. Over 40 percent of EOR operations in the United States use this method; it is most commonly employed in fields with high-viscosity oils (DOE, 2011). Miscible displacement is generally used for lower-viscosity oils and involves injecting gases such as nitrogen or CO2 that can reduce the viscosity of the oil and physically displace it toward production wells (Figure 3.10). Nearly 60 percent of EOR projects in the United States use this gas injection technique (DOE, 2011). In the United States, over 600 million tons of CO2 (11 trillion standard cubic feet; ~540 million metric tonnes) have been injected in ~13,000 wells for
4 See www.rrc.state.tx.us/data/wells/fluids.php.
5 See www.glossary.oilfield.slb.com/Display.cfm?Term=enhanced%20oil%20recovery.
FIGURE 3.10 Enhanced oil recovery through CO2 injection. SOURCE: NETL (2010).
EOR as of 2007 (Meyer, 2007). Current records from the Railroad Commission of Texas indicate that more than 9,400 wells are permitted in Texas alone for CO2 injection for EOR.6 Among the many thousands of wells used for EOR in the United States, the committee did not find any documented instances of felt induced seismicity in the published literature or from experts in the field with whom the committee communicated during the study.
One reason for the apparent lack of induced seismicity with EOR may be that EOR operations routinely attempt to maintain the pore pressure within a field at levels near preproduction pore pressures. This “balance” of the pore pressure means only a minimum pressure change occurs in the reservoir, reducing the possibility of induced seismic events; this maintenance of pore pressure is achieved broadly by maintaining balance between the amount of fluid being injected and the amount being withdrawn. EOR using CO2 injection is also considered one form of CCS, a technology under broader development in several other geological settings as part of the effort to reduce greenhouse gas emissions. CCS is discussed in detail later in this chapter.
UNCONVENTIONAL OIL AND GAS PRODUCTION INCLUDING SHALE RESERVOIRS
The permeability of rock in the subsurface varies tremendously (see Figure 2.1). Mudstone, siltstone, or shale formations that are high in organic content may contain significant amounts of natural gas and oil but have very low permeability; a shale formation that contains predominantly gas and/or oil is called a shale reservoir. Shales that are actively drilled for both oil and gas development in the United States are, for example, the Barnett, Marcellus, Eagle Ford, and Bakken formations (Figure 3.11).
Unlike conventional oil and gas fields, where the hydrocarbons were formed in source rocks high in organic content and then migrated over geologic time into porous rock such as sandstones and limestones that serve as the reservoirs today, the hydrocarbons in shales have developed from and remained for the most part trapped in their original source rock (organic-rich fine-grained sediments) because of the very low permeability of the shales. The shale gas resides in the microporosity in the shale layers and is held in place by a combination of cap rock, adsorption of gas onto the shale grains, and low permeability. The last of these effects is primarily responsible for the low production rates of drilled shales before being hydraulically fractured. Hydraulic fracturing creates additional pathways among the micropores for the gas to flow to the wellbore (see, e.g., NRC, 1996). This type of hydrocarbon reservoir, which requires additional engineered solutions for extraction of hydrocarbons, is often called an unconventional reservoir.
Extraction of gas and oil from these unconventional reservoirs has been made feasible through the combined application of horizontal drilling and hydraulic fracturing, technologies developed by the petroleum industry and through research supported by the Department of Energy (EIA, 1993, 2011; NETL, 2007; NRC, 2001). Hydraulic fracturing has been used for over 50 years to stimulate some conventional reservoirs (EIA, 2011) but is required to produce from low-permeability reservoirs such as shales for which commercially viable technology was developed by Mitchell Energy during the 1980s and 1990s (EIA, 2011). A large upswing in the use of horizontal drilling and hydraulic fracturing
FIGURE 3.11 Location of areas of active exploration and/or production for shale hydrocarbons (oil and gas) in the contiguous United States. Light pink areas are major sedimentary basins; dark pink areas (e.g., Eagle Ford, Barnett) are under active development and production for gas or oil from shale; orange areas are prospective regions currently being explored for potential oil or gas development from shale. Several shale units of different ages may overlie one another, and these units are outlined in thick red, blue, and purple lines representing youngest to oldest shale units, respectively. A "play" is a set of oil or gas accumulations that share similar geologic, geographic, and time characteristics. SOURCE: EIA (2011). Available at www.eia.gov/pub/oil_gas/natural_gas/analysis_publications/maps/maps.htm.
occurred in the late 1990s and continues to the present day; estimates suggest that today approximately 60 percent of the wells drilled are hydraulically fractured (Montgomery and Smith, 2010).
A typical production well in shale is drilled vertically to an appropriate depth and then turned horizontally to extend the well bore through the target shale formation. The horizontal segment (or “lateral”) of the well typically extends over 1-2 miles (~1.8-3 km) (Box 3.5). To facilitate the flow of the gas or oil into the well bore, the permeability through the shale reservoir is increased by the creation of artificial fracture networks in the shale around the horizontal portion of the well bore through the process of hydraulic fracturing (Box 3.5). Microseisms generally of M < 0 are induced during a hydraulic fracture treatment, and the locations of these microseisms are used to help understand the location of
the artificially created fractures and can be used as stress measurement tools (Appendix I describes this kind of microseismic monitoring; see also Engelder, 1993).
After the hydraulic fracturing is completed, a process known as flowback occurs. The well is opened and injected hydraulic fracture water is allowed to flow back from the formation into the well. For tight shale formations, between 10 and 50 percent of the hydraulic fracture water is returned (King, 2010). The flowback water may be reused as fracturing water for another hydraulic fracture procedure, may be disposed of in a wastewater injection well (see next section), may be stored, or may be treated to a purity that would allow for its safe release to the environment or for its use for other beneficial purposes. Two National Research Council reports (NRC, 2010, 2012) describe in some detail the potential options for management and beneficial use of wastewater from industrial activities.
The process of hydraulic fracturing a well as presently implemented for shale gas recovery does not pose a high risk for inducing felt seismic events (M > 2). Estimates suggest that over 35,000 wells for shale gas development exist in the United States today (EPA, 2011). Only one case has been documented worldwide in which hydraulic fracturing for shale gas development has been confirmed as the cause of felt seismic events. This event occurred in Blackpool, England, in 2011 (De Pater and Baisch, 2011; Box 3.6). Three other possible earthquake sequences have been discussed in the literature that may be associated with hydraulic fracturing in Oklahoma, only one of which was related to shale gas production. In the most recent case, in 2011, hydraulic fracturing for shale gas production was cited as the possible cause of felt induced seismic events, the largest of which was M 2.8 (Holland, 2011; Appendix J). The close proximity and timing of the earthquakes to the hydraulic fracturing well suggested a possible, but not fully established, link. However, the quality of the event locations was not adequate to fully establish a direct causal link to the hydraulic fracture treatment.
The two other possible cases in Oklahoma discussed by Nicholson and Wesson (1990) are listed under “Less Well Documented or Possible Cases” in their original paper (see also Appendix C). Both cases were associated in time with hydraulic fracturing related to stimulation of a conventional oil and gas field, not for shale gas production. The older of the two cases relates to a series of earthquakes that occurred on June 23, 1978, near the commercial stimulation of a 3,050-m (10,000-foot) well near Wilson, Oklahoma. Seventy earthquakes occurred in 6.2 hours (Luza and Lawson, 1980; Nicholson and Wesson, 1990). In the third case, two earthquakes were felt in a sequence in Oklahoma in May 1979, during the time that a well was vertically stimulated in three different zones, ranging from deep to shallow (ranging from 3,700 to 3,000 m depth [~12,000 to 10,000 feet]). The largest event in this third case was M 1.9. The well was located 1 km (3,280 feet) from a seismic monitoring station. The first hydraulic fracture treatment at 3,700 m depth was followed 20 hours later by about 50 earthquakes that occurred over a 4-hour time period. Forty earthquakes immediately followed the second hydraulic fracture treatment at 3,400 m, over a time period of 2 hours. No earthquakes were recorded during the third hydraulic fracture
A hydraulic fracture is a controlled, high-pressure injection of fluid and proppant into a well to fracture the target formation (see Figure). “Proppant” refers to sand or manmade ceramics used to keep the fractures open after fluid injection stops. The injected fluid is usually a combination of water and small amounts of chemical additives that reduce pipe flow friction, minimize rock formation damage, and help carry proppant into the fractures (see also Box 2.3; DOE, 2009; King, 2012). Horizontal wells are hydraulically fractured in multiple pumping “stages,” starting at the far end of the horizontal well and progressing toward the wellhead. Each fracture stage is isolated within the horizontal well with packers or mechanical sleeves that open and close each zone. After the entire hydraulic fracture procedure is completed, the injected fluid is allowed to flow back into the well, leaving the proppant in the newly created fractures. The amount of fracturing fluid used in one horizontal well fracturing stage varies, depending in large part on the geologic formation, and is on the order of millions of gallons per well. Generally, water volumes are estimated from 2 to 5.6 million gallons per well (DOE, 2009; King, 2012; Nicot and Scanlon, 2012; Soeder and Kappel, 2009). Horizontal wells can be hydraulically fractured in one to more than 30 stages depending on the length of the horizontal well.
The distance and direction of the manmade fractures propagating from the well vary depending on the type of hydraulic fracture treatment and the geologic properties near the well, including the rock toughness and stress state in the formation. In general, the fractures are observed from geophysical surveys such as microseismic (Appendix I) and tiltmeters (Cipolla and Wright, 2002) to propagate perpendicular to the direction of the minimum in situ stress. The induced fractures can form a complex fracture network in areas of low horizontal stress differences or simple fracture geometry in higher differential stress areas. Although the extent and direction of the fractures are not known precisely, hydraulic fractures may extend on the order of one hundred to over a thousand feet from the well. The upward growth of the hydraulic fracture tends to be limited by the horizontal layering (bedding) of the shale formations and by the vertical stress exerted by overlying rock and rarely extends up more than a few hundred feet (less than 100 m) from the wellbore (Fisher, 2010; Fisher and Warpinski, 2011). The geometry of hydraulic fractures can be estimated using a special seismic monitoring technique termed microseismic mapping (see Appendix I), although this geophysical procedure is completed on only a small percentage of hydraulically fractured wells, largely due to the cost.
Felt Earthquakes Near Blackpool, England, Related to Hydraulic Fracturing
Hydraulic fracturing of the Preese Hall-1 well in the Blackpool area of England caused seismicity in April (M 2.3) and May (M 1.5) 2011. The April earthquake was felt in northern England and was widely reported in the press. The well was drilled and hydraulically fractured by Cuadrilla Resources to explore the gas potential of the Bowland Shale Formation.
The Preese Hall-1 exploration well was stimulated vertically to 9,004 feet measured depth with five hydraulic fracture stages. The April M 2.3 event occurred during stage 2, and the May M 1.5 occurred during stage 4; in
Figure Schematic diagram of a horizontal well following a 10-stage hydraulic fracture treatment. Upper right inset shows a magnified view of the induced fractures (yellow) created during the hydraulic fracture treatment. The relative depths of local water wells is shown near the surface for scale, labeled “domestic well.” The formation depth and horizontal well length vary from area to area; the depth and well length numbers shown are approximate averages for North America. The well is fractured in stages from the end of the well (stage 1) to the start of the well (stage 10). Each hydraulic fracture stage is isolated within the wellbore as discussed in the text. Depths and distances of 2,000-10,000 feet correspond to about 600-3,000 m. SOURCE: Adapted after Southwestern Energy, used with permission.
addition approximately 50 weaker events were detected after additional seismic stations were deployed (De Pater and Baisch, 2011). Cuadrilla Resources initiated an extensive study of the incident, including installing portable seismic stations and a detailed seismic analysis as well as geomechanical studies and core studies, which were released to the public on their website. The research demonstrates that the hydraulic fracturing induced the seismic events. A report by Geosphere Ltd. (Harper, 2011) suggests the propagation of the fracturing fluid and pressure went farther than expected along the bedding planes. A nearby, apparently unstable fault was reactivated by the increase in fluid pressure, which caused the seismic events (Harper, 2011).
SOURCES: De Pater and Baisch (2011).
treatment at 3,000 m. All three Oklahoma cases demonstrate a reoccurring problem in induced seismicity studies: the seismic events are small, the regional networks are sparse, and the data quality is often too poor to fully confirm a causal link to fluid injection for energy development (see also Chapter 1).
INJECTION WELLS USED FOR THE DISPOSAL OF WATER ASSOCIATED WITH ENERGY EXTRACTION
In addition to fluid injection for specific kinds of energy development (e.g., water injection to produce steam for geothermal energy recovery, or fluid injection for waterflooding [secondary recovery]), water injection to dispose of water generated as a result of geothermal and oil and gas production operations is very common in the United States. Water that must be disposed of originates from production (see, e.g., NRC, 2010) or from flowback. Hereafter we refer to this kind of water broadly as wastewater; Chapter 4 clarifies the different kinds of water from energy production that are disposed of and the different classes of wells that are designated in the United States for this purpose. A recent study by Argonne National Laboratory estimated the total oil and gas fluid recovered from flowback after hydraulic fracturing operations and waste fluid produced during daily oil and gas production in the United States to be 20.9 billion barrels (about 878 billion gallons) of water per year (Clark and Veil, 2009). The majority (95 percent) of this water was managed through underground injection and more than half (55 percent) was injected for the purpose of enhanced recovery (Clark and Veil, 2009) (see the section Tertiary Oil and Gas Recovery [EOR] in this chapter). Just over one-third of the total wastewater volume (39 percent) or 6 billion barrels (252 billion gallons) was injected in disposal wells. Table 3.2 shows the water volumes produced in conjunction with oil and gas operations for various states. Importantly, other types of fluid may also be disposed of through underground injection (industrial wastes, for example, from manufacturing unrelated to energy production); these different kinds of underground injection are also discussed in Chapter 4.
The annual volume of wastewater in the United States is disposed of in many tens of thousands of injection wells. For example, in Texas, over 50,000 Class II7 injection wells were permitted as of 2010 (of which approximately 40 percent would be associated with disposal of wastewater and the remainder associated with waterflooding for secondary recovery; Texas RRC, 2010) (Figure 3.12).
Felt induced seismicity potentially related to Class II water injection wells has been identified at individual sites in Arkansas (see Chapter 4), Ohio, and Texas (Box 3.7). USGS
7 Wells in the Environmental Protection Agency’s (EPA’s) Underground Injection Control (UIC) program are described and regulated under one of six “classes.” Class II wells are specifically those that address injection of brines and other fluids associated with oil and gas production and hydrocarbons for storage. EPA’s well class system and the UIC program are described in more detail in Chapter 4.
TABLE 3.2 U.S. Onshore and Offshore Oil, Gas, and Produced Water Generation for 2007
|State||Crude Oil (bbl/year)||Total Gas (Mmcf)||Produced Water (bbl/year)||Data Source|
|Illinois||3,202,000||No data||136,872,000||1, 5|
|Kentucky||3,572,000||95,000||24,607,000||1, 3, 6|
|North Dakota||44,543,000||71,000||134,991,000||2, 4|
|West Virginia||350,000||1,000||2,263,000||4, 6|
|Tribal Lands||9,513,000||297,000||149,261,000||2, 6|
NOTE: 1, provided directly to Argonne by state agency; 2, obtained via published report or electronically; 3, obtained via electronic database; 4, obtained from website in form other than a published report or electronic database; 5, obtained from EIA; 6, produced water volumes are estimated from production volumes. SOURCE: Clark and Veil (2009).
FIGURE 3.12 Map of oil and gas wells (red dots) and saltwater disposal wells (green boxes) in Tarrant and surrounding counties in Texas. The approximate location of the Dallas-Fort Worth (DFW) airport is marked with box (as labeled), along with the injection wells near the airport. SOURCE: Modified from Frohlich et al. (2010).
researchers are investigating whether a recent increase in the rate of M > 3.0 earthquakes in the state of Oklahoma (see Figure 3.13) might be attributed to wastewater injection (Ellsworth et al., 2012). One of the best-documented cases of induced seismicity from fluid injection is in the Paradox Basin, Colorado, where brine from a natural seep has been reinjected in one disposal well at 14,000 to 15,000 feet (4,300 to 4,600 m) depth since 1996 to prevent brine flow into the Colorado River (Appendix K). To date over 4,600 induced seismic events (M 0.5 to M 4.3) as far away as 16 km (9.9 miles) from the injection well have been documented in the Paradox Basin (Block, 2011). Although the number of felt induced seismic events relative
Dallas–Fort Worth Earthquake Swarm October 2008 to May 2009
A series of M 2.5 to M 3.3 earthquakes occurred in the Dallas–Fort Worth (DFW) area of Texas, where earthquakes were felt and reported by local residents in October 2008 and May 2009. The National Earthquake Information Center (NEIC) located the earthquakes in the vicinity of the DFW airport.
The state of Texas historically experienced a low rate of natural seismicity at the time of these earthquakes and the entire state has only two permanent seismographic stations operated by the NEIC. Because of the sparse seismographic station coverage, the NEIC can only locate events in Texas that are greater than about M 2.5 with location accuracy of plus or minus 6 miles or 10 km. Researchers from the University of Texas (UT) and Southern Methodist University (SMU) deployed a temporary network of six seismographic stations in the DFW area to locate seismic events more precisely. The UT-SMU seismic array ran from November 9, 2008, to January 2, 2009, and located 11 earthquakes that spanned a 1-km-long, north-south trending zone in close proximity to a saltwater disposal (SWD) well used for wastewater injection by Chesapeake Oil and Gas Company. The wastewater originated from wells in the vicinity of the DFW airport producing from the Barnett Shale (Figure 3.11). The first felt DFW earthquakes started about 6 weeks after injection into the disposal well was initiated. The close correspondence of the earthquakes with the location and depth of the well, together with the close timing of the start of injection and the start of seismic activity, strongly suggest that injection was the cause of the seismic activity.
A state tectonic map compiled by the Texas Bureau of Economic Geology shows a northeast trending normal fault in the subsurface in close proximity to the SWD injection well. The earthquake swarm continues in the DFW area to this date, with M 2.6 or less events occurring prior to August 2011, over 2 years after shutdown of the injection well (Eisner, 2011). The persistent seismicity after the nearby injection wells were shut in demonstrates the difficulty in assessing whether the seismic activity is induced or natural. Similar to the post-shut-in events that have occurred in relation to EGS projects in France and Switzerland, understanding the cause and magnitude of these events through time requires further research that combines field observations and data with fluid flow and geomechanical simulation codes.
FIGURE 3.13 Graph showing the cumulative number of earthquakes M > 3.0 in the central Oklahoma region (34-37°N, 94-100°W) from 1900 to present day, showing a dramatic but as yet unexplained increase in seismicity since 2009. SOURCE: Ellsworth et al. (2012).
to the tens of thousands of produced water injection wells is small, the events themselves can cause considerable public concern. Addressing the causes and conditions for these events is useful for understanding induced seismicity potential for future wastewater injection projects.
Water injection wells only inject (dispose of) fluid, in contrast to injection wells for EOR or liquid-dominated geothermal systems where the fluid injected is approximately equivalent to the fluid extracted. Fluid injection in proximity to a favorably oriented fault system with near-critical stresses has an increased potential to generate felt induced seismic events in the absence of nearby extraction that could help maintain reservoir pressure. Class II injection wells used only for the purpose of water disposal normally do not have a detailed geologic review performed, and often data are not available to make such a review. Thus, although fluid pressure in the injection zone and the fracturing pressure of the injection zone can be measured after the disposal well is drilled, the location of possible faults is often not known as part of standard well siting and drilling procedures. Importantly, the mere presence of a fault does not always correlate to increased potential for induced seismicity. Chapter 6 discusses potential steps toward best practices with these challenges in mind.
Introduction of large amounts of CO2, a greenhouse gas, into the atmosphere is considered a likely driver in climate change (NRC, 2011). In 2010 approximately 33.5 billion metric tonnes of CO2 (~37 million tons) were introduced to the atmosphere by industry, transportation, and agricultural production globally (Boden and Blasing, 2011; Friedlingstein et al., 2010). For a number of years research has explored various methods for reducing carbon emissions to the atmosphere, including methods that can capture CO2 from point sources (e.g., fossil fuel burning power plants, industrial plants, and refineries), transport it to a geological storage site, and inject it into the ground for permanent storage (sometimes called sequestration) and monitoring (shown schematically in Figure 3.14). If successful and economical, CCS could become an important technology for reducing CO2 emissions to the atmosphere.
FIGURE 3.14 Illustration of the concept of carbon sequestration. SOURCE: USGS; Duncan and Morrissey (2011).
Geologic formations considered suitable for underground storage of CO2 include oil and gas reservoirs, unmineable coal seams, and deep saline rock formations (Kaldi et al., 2009). Naturally occurring CO2 has been trapped in geologic formations for millions of years, which indicates that retaining injected CO2 in the Earth under the right geological conditions is possible. Injection of CO2 for EOR has been used in the oil and gas industry for many decades with no obvious adverse effects (see the section Conventional Oil and Gas Production Including Enhanced Oil Recovery, this chapter); CO2 has also been injected in small volumes into saline rock formations in the western United States and Canada since 1989 without negative consequences (NETL, 2012; Price and Smith, 2008). Saline rock formations used for this purpose are sedimentary rocks that are naturally saturated with highly saline water that is otherwise unsuitable for humans, livestock, or agriculture.
Individual large, coal-fired power plants in the United States produce CO2 emissions that amount to up to 25 million metric tonnes (~27 million tons) per year.8 Capturing and transporting CO2 from industrial plants is technologically possible but is currently expensive, though a significant amount of research is exploring ways to bring costs down (Melzer, 2011). The United States as a whole accounted for approximately 1.5 billion metric tonnes (~1.7 billion tons) of CO2 emissions in 2010 (EIA, 2012). Storing even a portion of this amount of CO2 would require capturing the gas at many locations around the country and transporting it to facilities that could inject the CO2 into appropriate subsurface rock formations.9
Efficient underground storage of CO2 requires that it be in the supercritical (liquid) phase to minimize required storage volume.10 For CO2 to remain in a supercritical phase, the confining pressure in the reservoir must be greater than 7.3 MPa (about 73 atm11) and temperatures greater than 31.1°C, which can be achieved at depths greater than about 2,600 feet (790 m) (Buruss et al., 2009). These conditions require that the CO2 be injected at high pressures (62-64 bars [6.2-6.4 MPa or 900-930 psig] at the well head) so that the CO2 stays as a liquid. The density of supercritical CO2 is in the range of 0.60-0.75 g/cm3
8 See Carbon Monitoring for Action, available at carma.org/.
9 EOR operations do pump CO2 underground. However, EOR operations are designed to roughly balance the natural pressure in a reservoir from pumping out of hydrocarbons with pumping in of CO2. EOR using CO2 injection currently accounts for approximately 6 percent of U.S. crude oil production (Koottungal, 2010). Natural CO2 fields are currently the dominant source of CO2 for U.S. EOR and provide approximately 45 million metric tonnes (~50 million tons) per year, whereas anthropogenic sources, such as CO2 captured from industrial facilities, account for approximately 10 million metric tonnes (~11 million tons) per year (Kuuskraa, 2010). One of the biggest challenges for EOR projects that wish to use CO2 injection is being able to secure enough CO2 consistently at an acceptable cost (Melzer, 2011).
10 One pound of liquid CO2, which is about the volume of a typical fire extinguisher, will expand to approximately 8.8 cubic feet (0.25 m3) at normal room temperature and pressure.
11 One unit of atmospheric pressure or 1 atm is equivalent to the pressure exerted by the Earth’s atmosphere on a point at sea level.
(Sminchak and Gupta, 2003), whereas the density of most formation fluids within potential reservoirs is higher, typically 1.05-1.30 g/cm3. Supercritical CO2 is also less viscous than saline formation fluids. These differences in density and viscosity mean that the liquid CO2 will behave buoyantly within the reservoir. This buoyancy is what makes CO2 an effective fluid for EOR (Szulczewski et al., 2012).
For CCS, however, the buoyancy of CO2 means that the geologic reservoir must have a covering of impermeable rock (a “seal”) to ensure that the CO2 will not escape upward (Szulczewski et al., 2012). Depending on the composition of the geologic reservoir for the injected CO2, some potential exists for supercritical CO2 either to dissolve, weaken, or transform existing minerals or to precipitate new minerals in the geologic reservoir. For these reasons, selection of a suitable reservoir in which to inject and store CO2 is critical.
The effects of supercritical CO2 on geologic materials and the potential impacts of geochemical reactions with brines, cements, casing materials in injection wells, and materials that may seal faults and fractures in the reservoir have been topics of research supported by the Department of Energy (DOE) at academic institutions and national laboratories, and also by the petroleum industry. For example, in 2009 DOE supported 11 projects to conduct site characterization of promising geological formations for CO2 storage.12 Research at DOE’s National Energy Technology Laboratory (NETL) is based on developing efficient injection techniques, protocols that assess and minimize the impacts of CO2 on geophysical processes, and remediation technologies to prevent or reduce CO2 leakage. Currently NETL lists 37 active projects that address the critical geologic barrier for CO2 storage.13
The volumes of supercritical CO2 discussed for CCS are extremely large. An Intergovernmental Panel on Climate Change special report on CO2 capture and storage suggests that between approximately 97 and 306 million m3 per year (converted from 73 and 183 million metric tonnes)14 of CO2 could be captured and stored worldwide from coal and a similar amount from natural gas energy plants (Metz et al., 2005). This amount is equivalent to approximately 40,000 to 120,000 Olympic size swimming pools. For comparison, over 300 million m3 of crude oil were produced in the United States in 2010 (over 4 billion m3 were produced worldwide) (see Table 3.3). It is anticipated that CCS would take place at a number of locations, ideally places near power plants that produce CO2 so as to avoid long transportation distances. Many of the facilities would be expected to inject CO2 volumes on the order of several million tonnes (equivalent to several million cubic meters) or more into the ground each year (e.g., Szulczewski et al., 2012). Globally, only a few small-scale commercial CCS projects (the committee defines small-scale as about
12 See www.fossil.energy.gov/recovery/projects/site_characterization.html.
13 See www.netl.doe.gov/technologies/carbon_seq/corerd/storage.html.
14 As the density of supercritical CO2 ranges from 600 to 750 kg/m3, the volume of 1 million metric tonnes (~1.1 million tons) of supercritical CO2 ranges from 1.33 to 1.67 million m3. In-ground storage volume will depend on the effective porosity (i.e., the porosity times the storage efficiency).
TABLE 3.3 Petroleum and Natural Gas Production in 2010
|Crude Oil||Natural Gas Plant Liquids (NGPL)||Other Liquids||Total Crude Oil, NGPL, and Other Liquids||Total Dry Natural Gas|
|United States||2.00 billion barrels||757 million barrels||391 million barrels||3.15 billion barrels||501 million m3||611 billion m3|
|World||27.0 billion barrels||3.08 billion barrels||754 million barrels||30.9 billion barrels||4.91 billion m3||3.17 trillion m3|
NOTE: 1.00000 barrel = 0.15899 m3.
SOURCE: EIA 2010 International Energy Statistics (available at www.eia.gov/cfapps/ipdbproject/cfm).
1 million metric tonnes [approximately 1.55 million m3]15 or less of CO2 stored per year in geologic reservoirs) are in operation. In the United States, no commercial CCS technologies are currently deployed, although DOE-supported research is currently exploring the most suitable technologies for CCS through regional partnerships throughout the country.16 One of these regional projects in Illinois has advanced to the stage of conducting a large-scale test to inject 1 million metric tonnes of CO2; DOE defines “large-scale” as 1 million metric tonnes [approximately 1.55 million m3] or more. Both the global, commercial projects and the Illinois test project are discussed in the sections that follow.
The Norwegian state oil company Statoil and its partners currently operate CCS projects at offshore sites in the Sleipner field on the Norwegian continental shelf and in the In Salah gas field in Algeria (Box 3.8). They had also operated a CCS project at the Snøhvit field in the Barents Sea, north of Norway, until early in 2011. At Sleipner approximately 1 million metric tonnes a year have been injected since 1996. The demonstration CO2 injection project in northern Illinois has been in development for several years; injection of CO2 began in late 2011. The project plans to inject approximately 1 million metric tonnes per year for several years (Box 3.9). Seismic activity is being routinely monitored at all of these CCS sites. Although the CO2 injection rates and volumes for these projects are
15 Volume calculated using 0.70 g/cm3 as the density of supercritical CO2; however, this density may range from 0.60 to 0.75 g/cm3.
16 See fossil.energy.gov/programs/sequestration/partnerships/index.html.
The Sleipner, Snøhvit, and In Salah CO2 Capture and Storage Projects
In 1996, the Sleipner oil and gas fields in the North Sea became the site of the world’s first and largest offshore commercial CO2 capture and storage project. Carbon dioxide is captured at a plant located on one of the field’s operating offshore natural gas platforms and is stored underground in a sandstone formation at depths of approximately 800-1,100 m below the sea bed. Motivation for the project derived from a CO2 offshore tax levied on offshore oil and gas operations by the Norwegian government in 1991. CO2 is removed from the natural gas produced at Sleipner and is reinjected into the subsurface into a very porous, permeable sandstone and saline aquifer, the Utsira Formation (Figure 1). The Utsira Formation has an unusually high porosity and permeability (porosity is between 0.35 and 0.4 and permeability is near 1,000 mD) compared to the CO2 reservoirs in the other two Statoil CCS projects (Figure 2) and to many other potential CCS reservoirs. Approximately 1 million metric tonnes (1.1 million tons) of CO2 have been stored per year since operations began—with the accumulated total CO2 in the formation at the middle of 2012 approximately 13.5 million tonnes (Eiken and Ringrose, personal communication). The project is designed for approximately 25 years of CO2 injection. Current estimates for the Utsira Formation storage capacity range from 2 to 15.7 billion tonnes of CO2 (NPD, 2011).
Figure 1 Schematic rendition of the Sleipner field with CO2 injection into the Utsira sandstone formation occurring as natural gas is extracted from the Heimdal Formation more than 1,000 m below the CO2 reservoir. SOURCE: © 2012 Schlumberger Excellence in Educational Development, Inc. All rights reserved. Available at www.planetseed.com/node/15252.
Figure 2 Comparison of porosity and permeability for the CO2 reservoirs in each of the three projects. The Utsira Formation in the Sleipner field has an unusually high porosity and permeability. SOURCE: Eiken et al. (2011).
The Snøhvit field offshore northern Norway is a natural gas field with an onshore liquid natural gas (LNG) facility. Carbon dioxide separated during the LNG process was captured at the plant and piped back to the field, where it was reinjected underground into a sandstone formation ~2,600 m (8,560 feet) below the seafloor and below the main natural gas reservoir for the gas field; the entire offshore facility is subsea and operated remotely from shore (Statoil, 2009). Carbon storage began in 2008, and CO2 injection for storage was changed from the Tubåen Formation to the gas-producing Stø Formation in March 2011. Monitoring throughout the injection phase revealed increases in reservoir pressure beyond what had been initially anticipated, indicating that the reservoir had a lower capacity to inject or store CO2 than had been calculated at the start of the project (Helgesen, 2010). Total stored CO2 through March 2011 was about 1.1 megatons (Eiken and Ringrose, 2011).
At the In Salah field at Krechba onshore Algeria, the operators began injecting CO2 in 2004 into a formation located at intermediate depths between Sleipner and Snøhvit. By early 2011, nearly 4 million tons (3.6 million metric tonnes) of CO2 had been injected. The field has five gas production wells and three CO2 injection wells. The CO2 for injection derives from both the produced gas at the field and from gas produced at other fields that is piped to the injection well (Eiken and Ringrose, 2011).
The injection histories for all three fields are shown in Figure 3. The injection at Sleipner was very smooth over 15 years with good injectivity and no evidence of pressure buildup. Very consistent injection pressures were maintained at about 64-65 bars over the course of the project. The conditions at the other two fields proved to be more challenging, with measured pressure increases and limitations on the total capacity of the storage
formations. Pressure management was deemed an important issue with downhole pressure gauges of great importance (Eiken and Ringrose, 2011).
Figure 3 CO2 injection history at Statoil’s Sleipner, Snøhvit, and In Salah fields. SOURCE: Eiken et al. (2011).
Prior to the start of all three projects, extensive monitoring was conducted to establish baseline conditions, including any microseismic activity. Monitoring during CO2 injection for possible leakage and induced seismicity has occurred in all three projects. At both offshore projects, monitoring methods have included measurements of wellhead pressure and temperature, downhole pressure, gravity, and time-lapse seismic. At In Salah, monitoring data have included time-lapse seismic; pressures, rates, and gas chemistry at the wellhead; cores, logs, and fluid samples from the subsurface; one microseismic well, five shallow aquifer wells, and an appraisal well; satellite surveys to measure surface deformation; and surface measurements to monitor for potential leakage or rock strain. Monitoring from pilot wells at this location has shown detectable microseismic events related to CO2 injection. Shallow wells with three-component seismic detectors are emerging as the preferred deployment solution to give more extensive areal coverage of the field.
SOURCES: Eiken and Ringrose (2011); Eiken and Ringrose (personal communication, June 4, 2012); Ringrose and Eiken (2011); NPD (2011); Helgesen (2010); Statoil (2009); Arts et al. (2008); and “Sleipner Vest” (available at www.statoil.com/en/TechnologyInnovation/ProtectingTheEnvironment/CarboncaptureAndStorage/Pages/CarbonDioxideInjectionSleipnerVest.aspx).
Carbon Dioxide Sequestration in the Illinois Basin:
The Midwest Geological Sequestration Consortium Project at Decatur, Illinois
The Midwest Geological Sequestration Consortium (MGSC) is one of seven regional partnerships with funding from the DOE to test methods for geological storage of CO2. The MGSC in collaboration with Archer Daniels Midland Company, Schlumberger Carbon Services, Trimeric Corporation, and supporting subcontractors has initiated the Illinois Basin-Decatur Project (IBDP), which has begun the injection of 1 million metric tonnes (~1.1 million tons) of supercritical CO2 over a 3-year period into a saline reservoir that has not had previous fluid extraction at a site near Decatur, Illinois (Figures 1 and 2).
The target reservoir is the Mt. Simon Sandstone, which lies at a depth of approximately 7,000 feet. Injection of CO2 began in fall 2011 at an initial rate of 1,000 metric tonnes/day. An active seismic surface survey completed in January 2010 prior to the start of injection and a seismic monitoring well are part of the efforts to both monitor the distribution of CO2 and assess the seismicity risk during injection. The objectives of the baseline survey were to check for faulting, assess reservoir heterogeneity, map reservoir properties, develop data for the mechanical Earth model, and record a baseline for future CO2 distribution
Figure 1 Location of IBDP. SOURCE: Illinois State Geological Survey.
Figure 2 Location of MGSC monitoring well and injection and geophone wells. SOURCE: Illinois Department of Transportation, November 8, 2010.
in the subsurface. Microseismic monitoring is accomplished in both the injection well and a specially drilled microseismic monitoring well. A network for detecting and reporting microseismic events greater than an established magnitude has been installed. The installed array at the IBDP site detected a M 3.8 event near Elgin, Illinois, in February 2010, more than a year before the first CO2 injection. As part of their efforts to develop the CCS projects, the DOE and its collaborators have undertaken a very organized campaign of public outreach and education (see NETL, 2009).
smaller than those being proposed for large power plant and industrial plant operations,17 these projects provide data for assessment of the potential for induced seismic activity associated with large-scale CCS.
Induced Seismicity Risks
The risk of induced seismicity from CCS is currently difficult to assess accurately. The NETL reported that no harmful induced seismicity had been associated with any of the global CCS storage demonstration projects as of February 2011.18 However, the volumes of CO2 injected at these sites so far are small in comparison to the volumes being considered for future proposed large CCS projects. Unlike most water disposal wells, CCS involves continuous CO2 injection at high rates under high pressures for very long periods of time. The potential therefore exists to increase pore pressures throughout a volume with the storage reservoir that is much larger than those affected by other energy technologies. Given that the potential magnitude of an induced seismic event correlates strongly with the fault rupture area, which in turn relates to the magnitude of pore pressure increase and the volume in which it exists, it would appear that CCS may have the potential for significant seismic risk. The combination of hydro-chemical-mechanical effects such as mineral dissolution may also exacerbate the problem (Espinoza et al., 2011). Some factors could also serve to mitigate risk such as low viscosity and lower injection pressure and limits of permanent pressure change in the reservoir depending upon variables such as reservoir thickness.
Geothermal, enhanced geothermal, oil and gas, unconventional oil and gas, and CCS technologies all involve fluid withdrawal and/or injection, thereby providing the potential to induce seismic events. The rates, volumes, pressure, and duration of the injection vary with the technology as do the potential sizes of the earthquakes, the mechanisms to which the earthquakes are attributed (Table 3.4), and the possible risk and hazards of the induced events.
Induced seismicity is commonly characterized by large numbers of small earthquakes that persist during, and in some cases significantly after, fluid injection or removal. At several sites of seismicity caused by or likely related to energy technologies, calculations based on the measured injection pressure and the measured or the inferred state of stress in
17 Approximately 3,000 million metric tonnes (~3,300 million tons) of CO2 are reported to have been emitted by the United States in 2009 from the combined activities of electricity and heat production, manufacturing and construction, and other industrial processes including petroleum refining, hydrocarbon extraction, coal mining, and other energy-producing industries. Data available at www.iea.org/co2highlights/co2highlights.pdf.
18 See www.netl.doe.gov/technologies/carbon_seq/FAQs/permanence4.html.
TABLE 3.4 Summary Information about Historical Felt Induced Seismicity Caused by or Likely Related toa Energy Technology Development in the United States
|Energy Technology||Number of Projects||Number of Felt Induced Events||Maximum Magnitude of Felt Event||Number of Events M ≥ 4.0b||Net Reservoir Pressure Change||Mechanism for Induced Seismicity|
|Vapor-dominated geothermal||1||300-400 per year since 2005||4.6||1 to 3 per year||Attempt to maintain balance||Temperature change between injectate and reservoir|
|Liquid-dominated geothermal||23||10-40 per year||4.1c||Possibly one||Attempt to maintain balance||Pore pressure increase|
|Enhanced geothermal systems||~8 pilot projects||2-10 per year||2.6||0||Attempt to maintain balance||Pore pressure increase and cooling|
|Secondary oil and gas recovery (waterflooding)||~108,000 (wells)||One or more felt events at 18 sites across the country||4.9||3||Attempt to maintain balance||Pore pressure increase|
|Tertiary oil and gas recovery (EOR)||~13,000||None known||None known||0||Attempt to maintain balance||Pore pressure increase (likely mechanism)|
|Hydraulic fracturing for shale gas production||35,000 wells total||1||2.8||0||Initial positive; then withdraw||Pore pressure increase|
|Hydrocarbon withdrawal||~6,000 fields||20 sites||6.5||5||Withdrawal||Pore pressure decrease|
|Wastewater disposal wells||~30,000||8||4.8d||7||Addition||Pore pressure increase|
|Carbon capture and storage, small scale||1||None known||None known||0||Addition||Pore pressure increase|
|Carbon capture and storage, large scale||0||None||None||0||Addition||Pore pressure increase|
aNote that in several cases the causal relationship between the technology and the event was suspected but not confirmed. Determining whether a particular earthquake was caused by human activity is often very difficult. The references for the events in this table and the ways causality may be determined are discussed in the report. Also important is the fact that the well numbers are those wells in operation today, while the numbers of events listed extend over a total period of decades.
bAlthough seismic events M > 2.0 can be felt by some people in the vicinity of the event, events M ≥ 4.0 can be felt by most people and may be accompanied by more significant ground shaking, potentially causing greater public concern.
cOne event of M 4.1 was recorded at Coso, but the committee did not obtain enough information to determine whether or not the event was induced.
dM 4.8 is a moment magnitude. Earlier studies reported magnitudes up to M 5.3 on an unspecified scale; those magnitudes were derived from local instruments.
the Earth’s crust suggest that the theoretical threshold for frictional sliding along favorably oriented preexisting fractures was exceeded (see also Chapter 2).
Figure 3.15 shows histograms of the maximum magnitudes reported for induced seismicity associated with different energy technologies: geothermal energy, hydrocarbon extraction, fluid injection for secondary and tertiary oil and gas recovery, hydraulic fracturing associated with unconventional oil and gas production, and wastewater disposal from any of the energy technologies (injection wells) (see Appendix C for data sources for this figure); note that CCS is not included in this figure due to the absence of any known significant induced seismic events associated with this technology.
The largest seismic events and most numerous reports of induced seismicity are associated with extraction activities, with magnitudes up to 7 associated with extraction of gas at the Gazli field. The next largest set of seismic events (two sites in the world, one with an event of M 5.1 and another site with an event of M 6) is associated with injection activities related to waterflooding for secondary recovery in oil and gas production. Waste and wastewater disposal activities have produced some moderate earthquakes (M ~ 4.5), notably in Denver in 1967 at the Rocky Mountain Arsenal, but these are rare. Oklahoma, Colorado, and Arkansas have experienced a recent increase in seismic activity; these events are being examined for potential links to injection (Ellsworth et al., 2012). In the New Mexico–Colorado border area, the Raton Basin is an active coalbed methane field that has experienced several swarms of seismic events, including a M 5.3 in August 2010. In light of the seismicity in the Raton Basin, the Colorado Geological Survey (CGS) is now reviewing all permit applications for water disposal wells in Colorado in regard to the possibility of induced seismicity, assisting the Colorado Oil and Gas Conservation Commis-
FIGURE 3.15 Histograms of maximum magnitudes documented in technical literature caused by or likely related to subsurface energy production globally. Note: Many gas and oil fields undergo extraction of hydrocarbons along with injection of water for secondary recovery, but if the reported total volume of extracted fluids exceeds that of injection, the site is categorized as extraction. Some cases of induced seismicity in the list above do not have reported magnitudes associated with earthquakes, and those cases are not included in the counts used to develop this figure. No induced seismic events have been recognized related to existing CCS projects. SOURCE: See Appendix C.
sion (COGCC) (CGS, 2012) in the injection well permitting process. The injection and seismicity in the Raton Basin are under close scrutiny by both the CGS and COGCC. A definitive link to injection has not been established in the Raton Basin seismicity. Enhanced seismic arrays have been installed since 2011 in the area and will continue to be studied in detail by field operators, the Colorado agencies, and the USGS.
Numerous geothermal sites report induced seismicity, but the associated maximum magnitudes are generally small, with a maximum reported M 4.6 (at The Geysers site in California). Finally, felt seismic events caused by hydraulic fracturing are small and rare, with only one incident globally of hydraulic fracturing causing induced seismicity less than M 3 (in Blackpool, England; note the description in Appendix J of the seismic event in Eola, Oklahoma).
Several authors have observed that the maximum magnitudes of seismic events induced by various causes are related to the dimension or volume of human activity. Figure 3.16 (modified from Figure 3 of Nicol et al., 2011) plots the largest earthquake magnitudes
FIGURE 3.16 Graph showing maximum induced seismic event magnitude versus volume of fluid injected into or extracted from single wells or fields that are documented to have had a seismic event directly attributed to or strongly suggested to be caused by one of the energy technologies. These are global data. Events and associated volumes are identified by technology: red triangles denote geothermal energy with most of the data points representing fields (note that the net fluid volume, injected and withdrawn, at The Geysers is actually close to or below zero; see also Figure 3.17); blue triangles denote injection for secondary recovery or waste injection (such as at the Rocky Mountain Arsenal), almost all of which represent single wells; yellow triangles denote fluid extraction (oil or gas withdrawal; note that no data were available on the amount of fluid that may also have been injected in these fields to facilitate withdrawal); and green triangles denote hydraulic fracturing for shale gas production, both of which represent single wells. Not plotted are data from some projects that do not represent maximum magnitude seismic events for that project. Geothermal, extraction, and injection data modified from Figure 3 of Nicol et al. (2011). Hydraulic fracture data have been added in this study.
strongly suggested to be associated with fluid injection or extraction versus the volume of fluid reported for the injection or extraction project. The reported data suggest a correlation between the induced earthquake magnitudes and volumes of fluid injected. McGarr et al. (2002) suggested a correlation between maximum induced magnitude and the scale of human activity by plotting the maximum induced magnitude versus the dimension of the human activity (e.g., the maximum dimension of the hydrocarbon activity). Several points are important regarding these apparent correlations between induced magnitude and fluid volume:
1. Many factors are important in the relationship between human activity and induced seismicity: the depth, rate, and net volume of injected or extracted fluids, bottom-hole pressure, permeability of the relevant geologic layers, locations and properties of faults, and crustal stress conditions. These factors, some of which are interdependent, are also described in Chapter 2. For an induced seismic event to occur, at least two criteria have to be satisfied: (1) the pore pressure change in the reservoir has to exceed a certain critical threshold and (2) a certain net volume of fluid has to be injected (or extracted) to achieve a particular magnitude. The available data suggest, but do not prove, that the net volume of fluid may serve as a proxy for these factors, which indicates what set of conditions will generate small and large earthquakes. Particularly because the other data—bottom-hole pressure, permeability of the relevant geological layers, crustal stress factors, high-resolution well data (full waveform dipole and resistivity and waveform borehole imaging logs), seismic reflection images (two- and three-dimensional surface seismic techniques, 3D vertical seismic profiles or cross well seismic data) to reveal the subsurface structure such as the location, orientation, and properties of faults in the area—are not generally available, total volume can be a tool to draw inferences about various technologies. However, a pure causal relationship between the largest induced magnitudes and fluid volume should not be assumed. Important also, exceptions occur in those cases where fluids are injected into sites such as depleted oil, gas, or geothermal reservoirs, or at sites where the volume of extracted fluids essentially equals or exceeds the volume injected. In those cases pore pressures may not reach the original levels, or in some cases may not increase at all due to the relative volumes of injection and extraction. These data (specifically for oil and gas withdrawal and geothermal energy) are included in Figure 3.16, but it is noted that these specific data points do not necessarily represent the total (net) fluid (injected and withdrawn) that may be related to the maximum magnitude event.
2. The volumes indicated in Figure 3.16 include both volumes for individual wells in single projects and volumes for fields. The data cannot be used to predict earthquake magnitudes for an entire region or industry, but rather only to infer what magnitudes might be possible for individual wells or fields.
3. The data in Figure 3.16 are maximum magnitudes associated with fluid injection or extraction and support the requirement, outlined in Chapter 2 and elsewhere in this chapter, that a certain net volume of fluid has to be injected to cause a seismic event of a certain magnitude (or in a similar sense for net fluid withdrawal). The graph does not represent causality, but a condition for an induced seismic event of a certain magnitude to occur. Importantly, the correlation in the figure does not predict what earthquake magnitude will be induced by a specific project, but it reports instead the observed limits (to date) of what earthquake magnitudes have
been observed and can be used to infer what might be the size of the largest induced seismic events, if the volume of injected or extracted fluid is known. However, the correlation cannot be used to directly infer hazard or risk associated with various energy technologies.
4. These data and the limitations described point toward the great value in collecting information about well projects and characteristics, including the size of earthquakes produced (if any). Data are critical to making progress in estimating hazard and risk (see Chapter 5).
Another important factor to consider in evaluating the potential for an energy project to induce felt seismic events is the variation in volume from technology to technology, and the variation in net volume over time (Figure 3.17). For example, although CCS does not have the highest daily injection volumes among the technologies investigated, it does have the highest annual injected volumes because the projects are designed to run continuously with relatively large injection volumes. Also, CCS, similar to waste and wastewater disposal, involves only net addition of fluid to a reservoir rather than both injection and extraction that occur with oil and gas production and geothermal energy development. This characteristic is represented in the bottom graph in Figure 3.17 by the high net volumes of fluid injected for both technologies. Comparatively, the two geothermal cases (The Geysers and the EGS project at Basel) and hydraulic fracturing for shale gas production have negative or low net injection volumes on an annual basis. In the case of The Geysers, the negative net fluid volume is due to the high volumes of fluid extracted; annually, the fluid volume in The Geysers reservoir has actually been declining yearly, despite the high injection volumes.
The tens of thousands of Class II water disposal wells located across the United States have proven to be mostly benign with respect to induced seismicity. However, there are clearly troublesome areas that have induced events as large as M 4.7 (Arkansas, 2011; see Horton, 2012) that warrant a closer examination. The dramatic increase in hydraulic fracturing over the past 5 years means an increased volume of wastewater from hydraulic fracturing requiring disposal. If the number of available Class II wastewater disposal wells remains the same, the volume of injected fluid in each well must increase to accommodate the increased wastewater. The long-term effect of this increased volume on the potential to induce felt seismic events is unknown but could be of concern.
The implication for subsurface storage demonstration sites, for instance for CO2, is that pilot plants that inject small volumes of fluid cannot be expected to represent or bound the induced seismicity that might occur for production plants that will inject much larger volumes. Evaluation of production facilities for large-scale CCS thus requires a complete presentation of the risk of induced seismicity and a comprehensive monitoring plan including bottom-hole pressures and time response to different injection regimes.
FIGURE 3.17 A comparison showing estimated injected fluid volumes for (1) shale gas hydraulic fracturing, (2) CCS, (3) Class II waste and wastewater disposal wells, (4) The Geysers geothermal steam field for an average injection well, and (5) the Basel EGS project per day (upper graph). The lower graph shows the same information over a 1-year period for each project, with the exception of the Basel EGS project (which operated in total for just 6 days before termination). Data are presented in Appendix L. The committee could not find reliable data per well or per field for hydrocarbon extraction (withdrawal) or for secondary recovery (waterflooding). Hydraulic fracture volumes for shale gas assume a six-stage-per-day program, with a 4.64 million gallon average per well (the “average freshwater volume for fracturing” listed for five shale projects in King, 2012), estimating six hydraulic fracture treatments per day. For the hydraulic yearly volume calculation, an estimate of 15 wells drilled over a project area in the course of a year is made with a 20 percent recovery rate of injected fluid used. The CCS volume shown assumes 1 million tons (~0.9 million metric tonnes) of CO2 injection per year, similar to the Sleipner field offshore Norway. Class II disposal well data assume 9,000 barrels per day of wastewater injected. The Basel injection volumes averaged 0.5 million gallons per day for 6 days.
Adushki, V.V., V.N. Rodionov, S. Turuntnev, and A.E. Yodin. 2000. Seismicity in the oil field. Oilfield Review Summer:2-17.
Al-Mutairi, S.M., and S.L. Kokal. 2011. EOR Potential in the Middle East: Current and Future Trends. Society of Petroleum Engineers (SPE) conference paper presented at the SPE EUROPEC/EAGE Annual Conference and Exhibition, Vienna, Austria, May 23-26.
Arts, R., A. Chadwick, O. Eiken, S. Thibeau, and S. Nooner. 2008. Ten years’ experience of monitoring CO2 injection in the Utsira Sand at Sleipner, offshore Norway. First Break 26:65-72.
Bachmann, C., J. Wössner, and S. Wiemer. 2009. A new probability-based monitoring system for induced seismicity: Insights from the 2006–2007 Basel earthquake sequence. Seismological Research Letters 80(2):327.
Barker, B.J., M.S. Gulati, M.A. Bryan, and K.L. Riedel. 1992. Geysers reservoir performance. Pp. 167-177 in Monograph on the Geothermal Field, edited by C. Stone, Special Report No. 17, Geothermal Resources Council.
Beall, J.J., M.A. Stark, J.L. Bill Smith, and A. Kirkpatrick. 1999. Microearthquakes in the SE Geysers before and after SEGEP injection. Geothermal Resources Council Transactions 23:253-257.
Bhattacharyya, J., and J.M. Lees. 2002. Seismicity and seismic stress in the Coso Range, Coso geothermal field, and Indian Wells valley region, southeast-central California. Geological Society of America Memoir 195:243-257.
Blackwell, D.D., and M. Richards. 2004. Geothermal Map of North America. Tulsa, OK: American Association of Petroleum Geologists.
Block, L. 2011. Paradox Valley Deep Disposal Well and Induced Seismicity. Presentation to the National Research Council Committee on Induced Seismicity Potential in Energy Technologies, Dallas, TX, September 14.
Boden, T., and T.J. Blasing. 2011. Record high 2010 global carbon dioxide emissions from fossil-fuel combustion and cement manufacture posted on CDIAC site. Carbon Dioxide Information Analysis Center. Available at cdiac.ornl.gov/trends/emis/prelim_2009_2010_estimates.html (accessed April 2012).
Buruss, R.C., S.T. Brennan, P.A. Freeman, M.D. Merrill, L.F. Ruppert, M.F. Becker, W.N. Herkelrath, Y.K. Kharaka, C.E. Neuzil, S.M. Swanson, T.A. Cook, T.R. Klett, P.H. Nelson, and C.J. Schenk. 2009. Development of a probabilistic assessment methodology for evaluation of carbon dioxide storage. U.S. Geological Survey Open-File Report 2009-1035, 81 p. Available at http://pubs.usgs.gov/of/2009/1035/ (accessed May 2012).
CDOGGR (California Division of Oil, Gas and Geothermal Resources). 2011. Oil, Gas & Geothermal—Geothermal Resources. State of California Department of Conservation. Available at www.conservation.ca.gov/dog/geothermal (accessed February 2012).
CGS (Colorado Geological Survey). 2012. Earthquakes Triggered by Humans in Colorado—a background paper by the Colorado Geological Survey. http://geosurvey.state.co.us/hazards/Earthquakes/Documents/Earthquakes%20Triggered.pdf (accessed April 2012).
Cipolla, C.L., and C.A. Wright. 2002 (February). Diagnostic techniques to understand hydraulic fracturing: What? Why? and How? SPE Production & Facilities 17(1):23-35.
Clark, C.E., and J.A. Veil. 2009. Produced Water Volumes and Management Practices in the United States. ANL/EVS/R-09/1. Prepared by the Environmental Science Division, Argonne National Laboratory, for the U.S. Department of Energy, Office of Fossil Energy, National Energy Technology Laboratory, September.
Davis, S.D., and W.D. Pennington. 1989. Induced seismic deformation in the Cogdell oil field of West Texas. Bulletin of the Seismological Society of America 79(5):1477-1494.
De Pater, C.J., and S. Baisch. 2011. Geomechanical Study of Bowland Shale Seismicity—Synthesis Report. Prepared for Cuadrilla Resources. November 2. Available at http://www.cuadrillaresources.com/wp-content/uploads/2012/02/Geomechanical-Study-of-Bowland-Shale-Seismicity_02-11-11.pdf (accessed June 2012).
Deichmann, N., and D. Giardini. 2009. Earthquakes induced by the stimulation of an enhanced geothermal system below Basel (Switzerland). Seismological Research Letters 80(5):784-798.
DOE (U.S. Department of Energy). 2009. Modern Shale Gas Development in the United States: A Primer. Prepared by GWPC and ALL for the U.S. Department of Energy, National Energy Technology Laboratory, 116 pp. Available at http://www.netl.doe.gov/technologies/oil-gas/publications/epreports/shale_gas_primer_2009.pdf (accessed January 2012).
DOE. 2011. Enhanced Oil Recovery/CO2 Injection. DOE Fossil Energy Office. Available at fossil.energy.gov/programs/oilgas/eor/ (accessed March 2012).
Duncan, D.W., and E.A. Morrissey. 2011. The concept of geologic carbon sequestration. U.S. Geological Survey Fact Sheet 2010-3122. Available at http://pubs.usgs.gov/fs/2010/3122/ (accessed June 2012).
EIA (Energy Information Administration). 1993. Drilling Sideways—A Review of Horizontal Well Technology and Its Domestic Application. DOE/EIA-TR-0565. Available at ftp://tonto.eia.doe.gov/petroleum/tr0565.pdf (accessed January 2012).
EIA. 2011. Review of Emerging Resources: U.S. Shale Gas and Oil Shale Plays. Available at ftp://ftp.eia.doe.gov/natgas/usshaleplays.pdf (accessed November 2011).
EIA. 2012. AEO2012 Early Release Overview. Available at www.eia.gov/forecasts/aeo/er/early_carbonemiss.cfm (accessed February 2012).
Eiken, O., and P. Ringrose. 2011. Statoil experiences on storage and monitoring of CO2. Presentation (via teleconference) to the National Research Council Committee on Induced Seismicity Potential in Energy Technologies, Irvine, CA, August 18.
Eiken, O., P. Ringrose, C. Hermanrud, B. Nazarian, T.A. Torp, and L. Høier. 2011. Lessons learned from 14 years of CCS operations: Sleipner, in Salah and Snøhvit. Energy Procedia 4:5541-5548.
Eisner, L. 2011. Seismicity of DFW, Texas, USA. Presentation to the National Research Council Committee on Induced Seismicity Potential in Energy Technologies, Dallas, TX, September 14.
Ellsworth, W.L., S.H. Hickman, A.L. Llenos, A. McGarr, A.J. Michael, and J.L. Rubinstein. 2012. Are seismicity rate changes in the midcontinent natural or manmade? 2012 Seismological Society of America Annual Meeting abstracts.
Engelder, T. 1993. Stress Regimes in the Lithosphere. Princeton, NJ: Princeton University Press, 451 pp.
EPA (Environmental Protection Agency). 2011. Draft Plan to Study the Potential Impacts of Hydraulic Fracturing on Drinking Water Resources. EPA/600/E-11/001/February 2011/www.epa.gov/research. Available at water.epa.gov/type/groundwater/uic/class2/hydraulicfracturing/upload/hf_study_plan_110211_final_508.pdf (accessed January 2012).
Espinoza, D.N., S.H. Kim, and J.C. Santamarina. 2011. CO2 geological storage—geotechnical implications. KSCE Journal of Civil Engineering 15(4):707-719.
Feng, Q., and J.M. Lees. 1998. Microseismicity, stress, and fracture in the Coso geothermal field, California. Tectonophysics 289:221-238.
Fialko, Y., and M. Simons. 2000. Deformation and seismicity in the Coso geothermal area, Inyo County, California: Observations and modeling using satellite radar interferometry. Journal of Geophysical Research 105:21,781-21,794.
Fisher, K. 2010. Data Confirm Safety of Well Fracturing. The American Oil and Gas Reporter. July 10. Available at www.fidelityepco.com/Documents/OilGasRept_072010.pdf (accessed December 2011).
Fisher, K., and N. Warpinski. 2011. Hydraulic Fracture Height-Growth: Real Data. SPE 145949 presented to the Society of Petroleum Engineers (SPE) Annual Technical Conference and Exhibition, Denver, CO, October 30-November 2.
Friedlingstein, P., R.A. Houghton, G. Marland, J. Hacker, and T.A. Boden. 2010. Update on CO2 emissions. Nature Geoscience 3:811-812.
Frohlich, C., E. Potter, C. Hayward, and B. Stump. 2010. Dallas-Fort Worth earthquakes coincident with activity associated with natural gas production. The Leading Edge 29(3):270-275.
Grasso, J.-R. 1992. Mechanics of seismic instabilities induced by the recovery of hydrocarbons. Pure and Applied Geophysics 139(3/4):506-534.
Hamilton, R.M., and L.J.P. Muffler. 1972. Microearthquakes at The Geysers geothermal area, California. Journal of Geophysical Research 77:2081-2086.
Häring, M.O., U. Schanz, F. Ladner, and B.C. Dyer. 2008. Characterization of the Basel 1 enhanced geothermal system. Geothermics 37:469-495.
Harper, T.R. 2011. Well Preese Hall-1: The Mechanism of Induced Seismicity. Report prepared by Geosphere Ltd. for Cuadrilla Resources Ltd. Lancashire, UK, October 10.
Helgesen, O.K. 2010. CO2-problemer på Snøhvit. TekniskUkeblad. March 18. Available at www.tu.no/olje-gass/article240108.ece (accessed September 2011).
Holland, A. 2011. Examination of Possibly Induced Seismicity from Hydraulic Fracturing in the Eola Field, Garvin County, Oklahoma. Oklahoma Geological Survey Open-File Report OF1-2011. Available at www.ogs.ou.edu/pubsscanned/openfile/OF1_2011.pdf (accessed February 2012).
Horton, S. 2012. Disposal of hydrofracking-waste fluid by injection into subsurface aquifers triggers earthquake swarm in central Arkansas with potential for damaging earthquake. Seismological Research Letters 83(2):250-260.
Kaldi, J.G., C.M. Gibson-Poole, and T.H.D. Payenberg. 2009. Geological input to selection and evaluation of CO2 geosequestration sites. Pp. 5-16 in The American Association of Petroleum Geologists Special Volumes, 1st ed. Tulsa, OK: American Association of Petroleum Geologists.
Kaven, J.O., S.H. Hickman, and N.C. Davatzes. 2011. Micro-seismicity, fault structure and hydraulic compartmentalization within the Coso geothermal field, California. Proceedings of the Thirty-Sixth Workshop on Geothermal Reservoir Engineering, Stanford University, CA, January 31-February 2.
King, G.E. 2010. Thirty years of gas shale fracturing: What have we learned? Presented at the Society of Petroleum Engineers Annual Technical Conference and Exhibition, Florence, Italy, September 19-22.
King, G.E. 2012. Hydraulic fracturing 101: What every representative, environmentalist, regulator, reporter, investor, university researcher, neighbor, and engineer should know about estimating frac risk and improving frac performance in unconventional gas and oil wells. SPE 152596 presented to the Society of Petroleum Engineers (SPE) Hydraulic Fracturing Technology Conference, The Woodlands, TX, February 6-8.
Koenig, J.B. 1992. History of development at The Geysers geothermal field, California. Pp. 7-18 in Monograph on The Geothermal Field, Special Report No. 17, edited by C. Stone. Geothermal Resources Council.
Koottungal, L. 2010. Special report: 2010 worldwide EOR survey. Oil and Gas Journal 108(14):41-53.
Kraft, T., P.M. Mai, S. Wiemer, N. Deichmann, J. Ripperger, P. Kastli, C. Bachmann, D. Fah, J. Wossner, and D. Giardini. 2009. Enhanced geothermal systems: Mitigating risk in urban areas. Eos 90(32):273-280.
Kuuskraa, V.A. 2010. Challenges of implementing large-scale CO2 enhanced oil recovery with CO2 capture and storage. White paper prepared for the Symposium on the Role of Enhanced Oil Recovery in Accelerating the Deployment of Carbon Capture and Storage, Cambridge, MA, July 23. Available at web.mit.edu/mitei/docs/reports/eor-css/kuuskraa.pdf (accessed April 2012).
Luza, K.V., and J.E. Lawson, Jr. 1980. Seismicity and tectonic relationships of the Nernaha Uplift in Oklahoma. U.S. Nuclear Regulatory Commission NUREG/CR-1500, Part 3, 70 pp.
Majer, E.L., R. Baria, M. Stark, S. Oates, J. Bonner, B. Smith, and H. Asanuma 2007. Induced seismicity associated with enhanced geothermal systems. Geothermics 36:185-222.
McGarr, A., D. Simpson, and L. Seeber. 2002. Case histories of induced and triggered seismicity. Pp. 647-661 in International Handbook of Earthquake and Engineering Seismology, Part A, edited by W.H.K. Lee et al. New York: Academic Press.
Melzer, S. 2011. CO2 Enhanced Oil Recovery. Presentation to the National Research Council Committee on Induced Seismicity Potential in Energy Technologies, Dallas, TX, September 14.
Metz, B., O. Davidson, H. de Coninck, M. Loos, and L. Meyer (eds). 2005. Carbon Dioxide Capture and Storage. Intergovernmental Panel on Climate Change (IPCC) Special Report. Cambridge, UK: Cambridge University Press. 431 pp.
Meyer, J.P. 2007. Summary of Carbon Dioxide Enhanced Oil Recovery (CO2 EOR) Injection Well Technology. Prepared for the American Petroleum Institute. Plano, TX: Contek Solutions. Available at www.gwpc.org/e-library/documents/co2/APIpercent20CO2percent20Report.pdf (accessed February 2012).
MIT (Massachusetts Institute of Technology). 2006. The Future of Geothermal Energy: Impact of Enhanced Geothermal Systems (EGS) on the United States in the 21st Century. Boston: MIT Press.
Montgomery, C.T., and M.B. Smith. 2010. Hydraulic fracturing: History of an enduring technology. Journal of Petroluem Technology December:26-32. Available at http://www.spe.org/jpt/print/archives/2010/12/10Hydraulic.pdf (accessed January 2012).
NETL (National Energy Technology Laboratory). 2007. DOE’s Unconventional Gas Research Programs 1976-1995. U.S. Department of Energy. Available at http://www.netl.doe.gov/kmd/cds/disk7/disk2/Final%20Report.pdf (accessed December 2011).
NETL. 2009. Best Practices for: Public Outreach and Education for Carbon Storage Projects. DOE/NETL-2009/1391. U.S. Department of Energy. Available at www.netl.doe.gov/technologies/carbon_seq/refshelf/BPM_PublicOutreach.pdf (accessed December 2011).
NETL. 2010. Carbon Dioxide Enhanced Oil Recovery: Untapped Domestic Energy Supply and Long Term Carbon Storage Solution. U.S. Department of Energy. Available at www.netl.doe.gov/technologies/oil-gas/publications/EP/CO2_EOR_Primer.pdf (accessed January 2012).
NETL. 2012. Carbon Sequestration FAQ Information Portal: Permanence and Safety of CCS. Available at www.netl.doe.gov/technologies/carbon_seq/FAQs/permanence2.html (accessed February 2012).
NRC (National Research Council). 1996. Rock Fractures and Fluid Flow: Contemporary Understanding and Applications. Washington, DC: National Academy Press.
NRC. 2001. Energy Research at DOE: Was It Worth It? Washington, DC: National Academy Press.
NRC. 2010. Management and Effects of Coalbed Methane Produced Water in the United States. Washington, DC: The National Academies Press.
NRC. 2011. America’s Climate Choices. Washington, DC: The National Academies Press.
NRC. 2012. Water Reuse: Potential for Expanding the Nation’s Water Supply through Reuse of Municipal Wastewater. Washington, DC: The National Academies Press.
Nicholson, C., and R.L. Wesson. 1990. Earthquake hazard associated with deep well injection: A report to the U.S. Environmental Protection Agency. U.S. Geological Survey Bulletin 1951, 74 pp.
Nicol, A., R. Carne, M. Gerstenberger, and A. Christopher. 2011. Induced seismicity and its implications for CO2 storage. Energy Procedia 4:3699-3706.
Nicot, P., and B.R. Scanlon. 2012. Water use for shale-gas production in Texas, U.S. Environmental Science and Technology 46:3580-3586.
NPD (Norwegian Petroleum Directorate). 2011. CO2 Storage Atlas Norwegian North Sea. Available at http://www.npd.no/Global/Norsk/3-Publikasjoner/Rapporter/PDF/CO2-ATLAS-lav.pdf (accessed June 2012).
Preiss, J.S., S.R. Walter, and D.H. Oppenheimer. 1996. Seismicity maps of the Santa Rosa 1deg by 2deg Quadrangle, California for the period 1969-1995. U.S. Geological Survey Miscellaneous Investigations Series.
Price, J., and B. Smith. 2008. Geologic Storage of Carbon Dioxide: Staying Safely Underground. Booklet of the International Energy Agency (IEA) Working Party on Fossil Fuels. Available at www.co2crc.com.au/dls/external/geostoragesafe-IEA.pdf (accessed January 2012).
Ringrose, P. and O. Eiken. 2011. Microseismic monitoring for CO2 storage. Presentation (via teleconference) to the National Research Council Committee on Induced Seismicity Potential in Energy Technologies, Irvine, CA, August 18.
Shepherd, M. 2009. Oil field production geology. AAPG Memoir 91. American Association of Petroleum Geologists, Tulsa, OK, 350 pp.
Simpson, D.W., and W. Leith. 1985. The 1976 and 1984 Gazli, USSR, earthquakes—Were they induced? Bulletin of the Seismological Society of America 75:1465-1468.
Sminchak, J., and N. Gupta. 2003. Aspects of induced seismic activity and deep-well sequestration of carbon dioxide. Environmental Geosciences 10(2):81-89.
Smith, J.L.B., J.J. Beall, and M.A. Stark. 2000. Induced seismicity in the SE Geysers Field, California, USA. Proceedings of the World Geothermal Congress, Kyushu-Tohoku, Japan, May 28-June 10.
Soeder, D.J., and W.M. Kappel. 2009. Water Resources and Natural Gas Production from the Marcellus Shale. U.S. Geological Survey Fact Sheet 2009-3032, 6 pp. Available at http://pubs.usgs.gov/fs/2009/3032/ (accessed January 2012).
Statoil. 2009. Carbon Dioxide Capture, Transport and Storage (CCS). Available at www04.abb.com/GLOBAL/SEITP/seitp202.nsf/c71c66c1f02e6575c125711f004660e6/9ebcbfaedaedfa36c12576f1004a49de/$FILE/StatoilHydro+CCS.pdf (accessed October 2011).
Szulczewski, M.L., C.W. MacMinn, H.J. Herzog, and R. Juanes. 2012. Lifetime of carbon capture and storage as a climate-change mitigation technology. Proceedings of the National Academy of Sciences. Available at www.pnas.org/content/early/2012/03/15/1115347109 (accessed April 2012).
Texas RRC (Railroad Commission). 2010. Saltwater Disposal Wells Frequently Asked Questions (FAQs). Updated February 1, 2010. Available at www.rrc.state.tx.us/about/faqs/saltwaterwells.php (accessed February 2012).
Williams, C.F., M.J. Reed, R.H. Mariner, J. DeAngelo, and S.P. Galanis, Jr. 2008. Assessment of Moderate- and High Temperature Geothermal Resources of the United States. USGS Fact Sheet 2008-3082.