How an induced seismicity assessment is built

What "induced" means

Rock in the subsurface is under stress, and faults in it are held in place by friction. Some faults sit close to the point where the stress along them would overcome that friction; they're loaded, and have been for geological time, and nothing has tipped them. Raise the pore pressure in the rock around such a fault and the fluid pushes the two faces apart a little, reducing the friction that was holding them, and the fault can slip. The slip is the earthquake.

An induced earthquake, in this sense, isn't one that injection created. The stress was already there and the fault was already close to failure. Injection supplied the last increment, by raising the pressure on the fault, and that's why the question in an assessment isn't whether injection caused stress but whether injected pressure plausibly reached a fault that was ready to move. The mechanism by which pressure spreads from a disposal well is covered in How pressure moves out from a disposal well.

The earthquake catalog

The first input is the record of what earthquakes have occurred near the site, when, where, how deep, and how large. State seismic networks and the national network publish catalogs of located events, and the assessment compiles them for an area large enough to show the pattern, over a period long enough to show what the area looked like before injection began.

A catalog has two limits that shape everything built on it. It has a magnitude of completeness, below which the network doesn't reliably detect events, and that threshold changes over time as stations are added, so an apparent increase in small earthquakes can be an increase in detection. And every location carries an uncertainty, in map position and especially in depth, that depends on how many stations recorded the event and how close they were. Depth matters because injection-related events tend to occur in or below the injection interval, and a depth that could be two miles or five isn't much use for saying which. A careful assessment states both limits and works within them.

The injection record

The second input is what was injected, where, and when: monthly volumes and pressures for the disposal wells in the area, from the reports operators file with the regulator, along with each well's injection interval and depth. Together they say how much fluid went into which formation and how the rate changed over time.

The record is only as good as the reporting. Volumes are usually reliable; pressures are reported at the surface, not in the formation, and bottomhole pressure has to be estimated from them. Where several wells inject into the same formation, the assessment treats them as a group, because what reaches a fault is the combined pressure field, not one well's contribution.

Faults, stress, and slip potential

The third input is the structural setting: what faults are mapped near the site, how they're oriented, how deep they go, and whether they reach the basement rock where most injection-related earthquakes originate. Some of this is in published geological maps and papers, some in seismic survey interpretations where operators or agencies have released them, and some can be inferred from the alignment of the earthquakes themselves.

Orientation matters because a fault's tendency to slip depends on its angle to the regional stress. Fault slip potential studies, published for the major producing basins, combine mapped faults, the measured stress field, and pore pressure to estimate how much additional pressure each fault could take before failing. They're an estimate with wide uncertainty, and they're the best available answer to whether a particular fault was the kind that a modest pressure rise could move.

This is also where a disposal permit's own representations get tested. In some states the application attests that no faults transect the injection zone. Set that attestation beside the regional structure maps and the earthquake locations and it either holds or it doesn't.

The regulator's protocol

Most states with disposal-related seismicity have a published response procedure: when an earthquake above a certain magnitude occurs within a certain distance of a disposal well, the permit is reviewed, the rate is cut, or injection is suspended, in stages that escalate with magnitude and proximity. The thresholds differ by state and are revised as experience accumulates.

The protocol is a useful yardstick, because it says what the regulator itself treats as significant. An assessment reads the catalog against it: which events would have triggered which stage, whether they did, and what the operator's response was. It's not a substitute for the physical question, since a protocol is a management rule and not a finding of cause, but it anchors the discussion in something the regulator has already accepted.

Surface deformation

Where satellite radar data exist for the area, they add a fourth input. Interferometric synthetic aperture radar (InSAR) measures small changes in ground elevation over months and years, and injection at scale shows up as ground rising over the disposal zone, sometimes in a pattern that traces the faults the pressure is following. Deformation doesn't locate earthquakes, but it can show where the pressure has gone, which is otherwise the hardest part of the picture to see.

Putting it together

With the inputs assembled, the assessment asks a set of ordered questions. Was there seismicity in the area before injection began, and if so, how much and how large? Did the rate of earthquakes change after injection started or after the injected volume increased, allowing for the months it takes pressure to travel? Are the events located near the injection wells, at depths consistent with the injection interval or the basement beneath it, and along mapped or inferred faults? Do the faults involved have the orientation and the slip potential that a pressure rise of the estimated size could move? And does the sequence fit what's been documented elsewhere in the basin under similar conditions?

Each question is answered with the data and its limits stated, and none of them is decisive alone. Earthquakes near a disposal well in a region that was seismically quiet before, at depths that fit, along a fault the slip potential studies flag, following a rise in injected volume by a plausible lag, add up to a strong association. Earthquakes at a distance, at the wrong depth, on faults with no known link to the injection zone, in an area with a history of natural seismicity, don't. Most real cases sit between those, and the assessment says where.

What the finding says, and doesn't

The finding an assessment can support is that the seismicity near a site is, or isn't, consistent with injection: that the timing, location, depth, and structural setting fit the mechanism, and that alternative explanations have been considered and are less well supported. It can say whether the events crossed the thresholds the regulator would act on. It can say what the structural setting implies for a permit that attested to no faults. And it can say what the record doesn't contain and what would be needed to close the gap.

What it can't do, on its own, is prove that a particular well caused a particular earthquake. That requires modeling the pressure field from every well in the area and the response of the specific fault, with data that's rarely complete, and even then the answer is a probability. A careful assessment doesn't claim more than the data carry, because the other side will read it with the same catalog and the same maps, and the assumptions are declared so they can. The practice's induced seismicity assessment work is built on that principle.

This page describes a method. It isn't engineering certification, a regulatory determination, or legal advice, and it doesn't describe any engagement. Terms are defined in the glossary.