Chronic well leakage probability relative to basin and fluid characteristics

Regulation of wells is largely concerned with precluding leakage along them. Studies of the proportion of wells in large populations that leak provide one means of reviewing the effectiveness of these regulations. The article entitled “A portrait of wellbore leakage in northeastern British Columbia, Canada” in PNAS (1) provides the next such study.

Our civilization depends to a remarkable degree on wells. Groundwater extraction via wells provides one-half of the drinking water in the world and over two-fifths of irrigation water (2). Almost all oil and gas production occurs via wells. For more than a century, oil wells have been a key part of the global energy system providing transportation fuels and petrochemicals that have raised quality of life around the world. In the future, even as power for transportation becomes supplied more and more by electricity from renewable sources and oil production declines, an important new class of wells may emerge for the purpose of sequestering carbon dioxide from the continued combustion of coal and gas in centralized facilities in geologic reservoirs deep underground.

As oil and gas, and perhaps in the future carbon sequestration, wells extend down into the ground, their borings typically extend through multiple geologic units. These create potential pathways for unintended movement of fluids between different units, which can result in the degradation of fluid resources (groundwater, oil, natural gas) and air quality. For groundwater wells, the main hazard of concern has been movement of surface water of lower quality into groundwater. For oil and gas wells, there are a number of hazards of concern owing to the greater number of zones they cross and the nature of the hydrocarbons they access.

These hazards have been a driver of well design for over a century. For instance, in California where I live, this hazard was a key motivation for the founding in 1915 of an agency to regulate oil and gas production, as indicated by the following from the enabling legislation.

“It shall be the duty of the state oil and gas supervisor so to supervise the drilling, operation and maintenance and abandonment of petroleum or gas wells in the State of California, as to prevent, as far as possible, damage to underground petroleum and gas deposits from infiltrating water and other causes and loss of petroleum and natural gas” (ref. 3, chapter 718, section 3).

As indicated, at that time in California, the main concern was oil wells allowing the flow of water into the oil resource. This was because there was no standard practice of installing seals in the well to impede such flow. Well casing (pipe) was generally installed without any engineered material between it and the borehole wall. As result, regulation came to require construction of wells with engineered sealing material to prevent flow between zones, that material being almost exclusively based on Portland cement.

Over time, other hazards have emerged as the primary concern for oil and gas wells.

Concern for contamination of groundwater by fluids flowing up wells from deeper zones led to further advances in regulatory requirements of well construction: mainly, the addition of an outer casing cemented through the shallow groundwater zone through which the boring to the deeper hydrocarbon resource is advanced. While drinking water protection laws have been the main driver of engineering to prevent groundwater contamination, a new hazard of deep wells has been recognized: well leakage of natural gas to the atmosphere contributing to global warming (the global warming potential of methane is tens of times that of carbon dioxide by mass; the exact ratio is dependent on the time span of consideration owing to the different removal rates of each from the atmosphere).

The flow rate of a well leak is the main characteristic that determines society’s response to that leak. Most well-known because of their immediate impact to their environs, with well workers often being the most impacted, are leaks termed well blowouts. These are leaks of such magnitude or impact that society responds immediately upon discovery to stop them. Well-known instances, at least in the United States, are the Macondo oil and gas well blowout in the Gulf of Mexico in 2010 and the Aliso Canyon underground natural gas storage well blowout in 2015–2016, also known as the Deepwater Horizon and Porter Ranch blowouts, respectively.

Relatively unknown by the general public are leaks whose magnitude and impacts are such that society does not move immediately to stop them upon discovery or does not require efforts to discover them at all. Provinces in Canada such as Alberta lead the way in requiring efforts to discover these leaks. As a result, data existed to determine that a minimum of 5% of oil and gas wells in Alberta chronically leak gas (the study found this rate varies considerably based on various well features) (4). Since this result was published a decade ago, there has not been a similar result based on a sample population on the order of tens of thousands of wells until now to my knowledge.

At long last, the next such study, of oil and gas wells in the Canadian Province of British Columbia, is available in PNAS (1). It finds a minimum of 11% of wells chronically leak gas. Both this study and the Alberta study (4) found the share of wells leaking increased over time. Both studies found this is due to increasing efforts to detect those leaks rather than to the probability of leakage increasing with well age. This is a fascinating conjunction of results given the expected long-term degradation of well casing, which is almost exclusively composed of carbon steel. Rather than the age of a well, the Alberta study found the probability of leakage was significantly correlated to aspects of a well’s initial engineering and construction (4).

A potential process that could contribute to the lack of increased probability of leakage with well age consists of inward movement of the borehole wall. This occurs if the strength of the geologic materials surrounding the boring results in those materials creeping or sloughing inward in response to increased differential stress caused by the boring void. By analogy, think of digging a hole in sand at the beach. At some depth after water is encountered, the walls of the hole collapse.

A study of leakage to surface from abandoned wells in Texas found the probability of leakage varied between sedimentary basins (5). The probability of leakage was about 20 times higher in the basin with stronger sediments than in the basin with weaker sediments. These results of the Texas study suggest it would be useful to compare the strength of sediments between and within the basins in Alberta and British Columbia and/or other interbasin comparison studies to determine whether the correlation between the percentage of wells leaking and sediment strength is replicated.

If this result is replicated, it has implications for regulations based on the risk specific to a basin. The Alberta study also has implications for basing regulations on the risk specific to a well type (4). Currently, well construction regulations are more one-size-fits-all (prescriptive) within a political jurisdiction.

A particular instance of this is underground injection control regulation in the United States. A key aspect of this regulation is the requirement for injection well operators to review the construction of all legacy wells within the area of increased pressure due to the injection. Technically, the area of increased pressure is only the area over which the pressure increase is sufficient to cause fluids to move via a hypothetical conduit from the injection interval to the deepest zone with groundwater designated for protection. However, the pressure necessary to do this is so low that it results in almost the entire area with any pressure increase at all requiring review of the construction of preexisting wells (6). If well construction records suggest the wells in the area of increased pressure are not configured to prevent leaks according to current standards, the proponent of the new injection well must bring all of the legacy wells up to standard.

The Texas study of leakage from abandoned wells found that in basins with low-strength sediments, the hypothetical conduit-prescriptive approach to defining the area over which legacy wells need to be reviewed for potential upgrade may be a substantial overstatement of the risk, resulting in an unwarranted expenditure of effort and funds. For instance, the Texas study found only 0.0004% of oil and gas fields in the basin with low-strength sediments had evidence of a liquid leak from an abandoned well (5).

The British Columbia and Alberta studies suggest the phases available to leak are not considered in the regulation. In contrast to the proportion of wells that leak gas at the ground surface, the Canadian studies find that the proportion that leak liquid is orders of magnitude less (1, 4). Table 1 lists the proportion of wells leaking gas versus liquid in each study. The Alberta study mentioned consideration of liquid leakage but did not provide the proportion, suggesting it is so low as to not be noteworthy (4). The British Columbia study is uniquely useful in reporting the proportion of wells leaking each phase (1).

Table 1.

Proportion of wells leaking gas and liquid in different political jurisdictions

Jurisdiction	Leaking gas	Leaking liquid
Alberta (4)	5%	∼0%? (not reported)
British Columbia (1)	11%	0.002%

Numerical studies undertaken to assess the risk of geologic carbon sequestration provide a physics-based reason for the substantially lower proportion of wells leaking liquid than gas at the ground surface. One study regards leakage from a reservoir to groundwater in the presence of an intervening transmissive zone (7). This entailed hundreds of simulations across the values of various parameters, such as the permeability of each zone. The results indicate that for pathways along a well that contact the intervening transmissive zone, such as those involving the annulus (space) between the casing and the boring wall, water leakage will not propagate above the transmissive zone. Gas leakage is also substantially attenuated by the transmissive zone, but some leakage continues upward due to buoyancy.

The area of review definition and well construction and leakage testing requirements would benefit from moving from a prescriptive basis that does not consider basin geomechanics and hydrogeology or potential leaking fluid-phase characteristics, such as density, to a risk-based approach that considers each of these.