Few technical terms carry as much ambiguity or provoke as much confusion as “fracking,” otherwise known as hydraulic fracturing. The word can be used (often incorrectly) to refer to the stimulation stage of a shale well, produced-water disposal, enhanced recovery, salt-cavern operations and myriad other subsurface work that uses wells but alters the ground in different ways. Consequently, which activity changed pressure, stress, fluid pathways or ground support, and whether local geology provided a failure path, is obscured, as is what is safe and effective versus which processes and where carry serious risks for human safety and the environment.

This distinction quickly surfaces in discussions about fracking and the risks of causing earthquakes or sinkholes, as they are different hazards with different relationships to hydraulic fracturing. Induced seismicity involves fault slip after pressure or stress changes, while sinkholes and subsidence involve loss of near-surface support through dissolution, void collapse, groundwater change or cavern failure. In both cases, fracking can be a cause, but the question is how often, how controllable and how large?

Hydraulic fracturing can trigger seismicity

Hydraulic fracturing is a short-duration stimulation process in a production well in which operators pump fluid and proppant at high pressure to create or extend fractures in a target reservoir. The process routinely generates microseismicity (small events used to map fracture growth), but microseismic events are not the same as felt earthquakes and are usually too small to be noticed at the surface.

Therefore, technically speaking, hydraulic fracturing itself can induce earthquakes, but most directly linked events are small, and felt events are uncommon. Reviews of hydraulic fracturing-induced seismicity describe a typical pattern of events clustering near the stimulated well, occurring during or soon after treatment, and reflecting interaction with a preexisting fault; the triggering pathway does not require the creation of a new large fault, and can reactivate an existing fault that is close to failure.

Felt events are uncommon indeed, but they do still occur and require several conditions to occur simultaneously. A fault must be large enough to host the earthquake, and the regional stress field must orient that fault favorably for slip. Fracturing fluid or a pressure disturbance must also reach the fault through direct intersection or hydraulic connectivity, and the pressure change must reduce effective normal stress enough to overcome frictional resistance. Basin studies show that risk can depend on well depth, formation, fluid type, pad development and overpressure, so hazard cannot be reduced to a simple count of wells, stages or pumped volume; regardless, the concurrent occurrence of these conditions is infrequent, so hydraulic fracturing is not considered a major cause of felt earthquakes.

Disposal wells pose the larger seismic risk

Wastewater or produced-water disposal differs from hydraulic fracturing in purpose, duration and pressure history. Disposal wells commonly inject water for months or years rather than hours or days, and they may inject larger volumes into formations selected for disposal capacity rather than hydrocarbon production. Pressure disturbances can migrate away from the wellbore where permeable formations, fractures or faults connect the injection interval to deeper structures.

For these reasons, the strongest evidence for felt oil and gas-related induced seismicity points to wastewater disposal rather than stimulation. Most induced earthquakes are not directly caused by hydraulic fracturing, and regional studies linked the central U.S. earthquake increase specifically to high-rate injection. For example, Oklahoma case studies connected earthquake swarms to pressure migration from high-volume disposal, including events away from the disposal well. Injection raises pore pressure in a reservoir or connected fault zone, and higher pore pressure lowers effective normal stress, or the clamping stress that helps keep a fault from slipping. If shear stress on the fault is already near failure, then an increase in pressure can trigger slip.

Depth is a significant consideration because many higher-consequence induced events occur on deeper faults, including faults in or near crystalline basement. Similarly, rate and pressure govern how quickly and how far pressure changes migrate, and hydraulic connectivity matters because a disposal zone that appears isolated may still communicate with faults through fractures, permeable beds or damaged wellbores.

However, disposal still does not make earthquakes inevitable, or even likely. Only a small fraction of disposal wells generate earthquakes of public concern, and causality requires evidence from timing, location, depth, focal mechanisms, injection history and geologic connectivity, making it difficult to confirm in many instances. Long-duration, high-volume disposal in a susceptible stress-and-fault setting generally poses a greater felt-earthquake risk than the short-duration hydraulic fracturing stage of a shale well, meaning that fracking is not directly the chief concern, but the supporting processes.

Sinkholes follow different failure mechanisms

Earthquake mechanisms should not be conflated with sinkhole mechanisms, although they frequently are. A sinkhole is a surface depression or collapse feature, commonly associated with soluble rock, voids or loss of near-surface support. Sinkholes often form in karst and other soluble rock settings, where groundwater dissolves limestone, gypsum, salt or similar materials, leaving unstable cavities. However, surface failure can also result from piping into voids, groundwater decline, cavern failure, mining-related collapse or gradual land subsidence.

Routine shale hydraulic fracturing does not provide a general mechanism for forming sinkholes. The stimulation target typically lies thousands of feet below the surface, and the created fractures are designed to remain within or near the reservoir. A deep pressure treatment does not automatically remove shallow ground support, dissolve near-surface rock or create a surface-connected void, so claims that treat "fractures underground" as equivalent to sinkhole formation skip the controlling hydrogeology and rock mechanics.

Oil and gas activities can contribute to sinkhole or subsidence hazards under specific conditions, but those conditions differ from those in ordinary shale stimulation, and other activities are more likely to cause sinkholes. Human activity can accelerate sinkholes through groundwater withdrawal, dewatering, construction and changes in subsurface drainage. In oilfield areas with evaporites, defective, abandoned or poorly sealed wells can create pathways for undersaturated water to dissolve salt. Salt-cavern storage and solution-mining operations pose another distinct hazard: a cavern wall or roof can fail, leading to major surface collapse, as shown by the collapse of an underground storage cavity in Louisiana in 2012.

Sinkholes are thus not a general consequence of hydraulic fracturing. Oil and gas-related sinkhole and subsidence problems can occur where operations alter groundwater, dissolve evaporites, destabilize caverns, leave mining voids or create defective fluid pathways, but those risks are site-specific rather than general consequences of hydraulic fracturing.

Risk controls should match the mechanism

Risk management strategies depend on the target mechanism. For induced seismicity, operators can screen for mapped and unmapped faults using 3D seismic data, well control, basement-depth maps, regional stress information, and historical seismicity, and geomechanical models can estimate whether faults are critically stressed and whether pressure changes could reach them. Baseline monitoring likewise helps distinguish natural background seismicity from operational changes, while microseismic and regional seismic monitoring can detect escalation.

In terms of operational controls, injection-rate limits, pressure limits, staged rate changes, avoidance of basement-connected intervals and disposal-zone selection minimize hydraulic communication with susceptible faults. Traffic-light protocols may require review, rate reductions or suspensions when seismicity exceeds defined thresholds; they reduce risk but do not eliminate it because pressure diffusion and fault response can lag operational changes. Evidence from Oklahoma indicates that reduced injection rates and shallower injection intervals helped mitigate induced seismicity in a high-risk region, and recycling produced water, moving volumes to lower-risk formations or using alternative disposal methods can also reduce pressure loading where disposal drives the hazard.

Sinkhole controls require a different set of tools and techniques. Regulators and operators need geohazard maps for karst, evaporites, salt domes, brine fields, caverns and abandoned wells. Well integrity programs verify casing, cement, plugging and mechanical isolation. Cavern operations require sonar surveys, pressure monitoring, conservative operating limits and groundwater monitoring and surface-deformation surveys to identify changing support conditions before collapse becomes evident. Response plans should address road closures, pipeline isolation, gas migration and public safety when surface failure begins.

Yes, hydraulic fracturing can cause earthquakes, but its typical direct effect is microseismicity or small induced events, and felt earthquakes require an unfavorable combination of faults, stress, pressure and connectivity. Wastewater disposal is more strongly linked to felt and damaging induced earthquakes because it can impose larger, longer pressure changes on connected fault systems. Fracking is also not a general cause of sinkholes. Most oil and gas-related sinkhole or subsidence hazards arise from evaporite dissolution, cavern collapse, brine extraction, groundwater change, mining voids or well-integrity failures. The practical concern is not fracking in general, but fracking processes in the wrong areas or without the right tools and infrastructure.