The views and opinions expressed in guest editorials published in JPT are those of the individual authors and do not necessarily reflect the official policy, position, or views of SPE, its members, or its affiliates.
The hardest part of petroleum engineering is not the calculations or the inherent uncertainty. It is guarding against cognitive biases.
Avoiding bias begins with reliable measurements of appropriate questions, and therein lies the first of two problems with understanding our industry’s methane emissions:
The most common measure of methane emissions obscures poor gas-handling performance.
The second-biggest problem—availability bias—builds on this misstep to compound our misunderstanding of our industry’s performance by ignoring the bulk of the independent evidence about our performance.
Actionable insight—or misdirection—begins with the choice of measurement. Jim Collins’ bestseller, Good to Great, enshrined the now ubiquitous concept of key performance indicators. The measurement must be useful for discriminating better from worse in a way that can affect choices to achieve the goals. In this case, the objective is to reduce significantly methane emissions below historical levels. In Collins’ words, an appropriate measure must show the “brutal truth.”
Of course, natural gas is a valuable and useful product with strong, inherent, long-standing incentives against waste. But still, both venting and flaring have persisted at levels detrimental to near-term climate deformation, regulatory risk, and industry credibility. An appropriate measure would highlight where improvements can be made on top of the existing economic and regulatory structures. A measurement of emissions should differentiate good operations from bad operations when comparing like to like, and it should lead to improvements over the baseline.
Ironically, our industry of technical professionals often relies on a unit of measurement which obscures differences and conflates economic drivers with operational excellence: methane intensity. Although often expressed as a percentage, this unit compares energy vented (without regard to flaring) against energy captured, including crude oil and condensate.
Adding liquids to the denominator always drives the measurement value down without saying anything about the excellence of gas handling. It may measure the overall loss of energy to society, but it does not highlight shortcomings and does not help guide improvement.
Of course, what matters most directly for the climate is the mass of methane wasted.
At a small scale, the mass of methane in 1 Mcf of natural gas is calculated by multiplying 19.2 kg/Mcf by the methane concentration. Working in the other direction, 1 kg/hr of methane converts to about 1.5 Mcf/D of natural gas at an 84% methane content. Because methane is so much more potent in the short term than CO2, that 1.5 Mcf of natural gas has the same pollution effect as 4,500 lb of CO2. That is, pollution equivalent to the mass of a light-duty truck can be released each day from a single, small oil well.
At scale, emissions are measured by the benign-sounding unit of “terragram,” each of which can cause as much climate deformation over the next 20 years as 187 billion lb of CO2. The upstream industry in Texas alone emits six or more of these colossus units every year.
Reducing the mass of emissions is the goal; other units are means to an end. Methane intensity is useful for buyers to compare among crude oils, but the figure is dominated by oil production instead of gas handling.
Releasing all associated gas from a low gas-oil-ratio (GOR) crude can be the same intensity as a much lower percentage release of a high-GOR crude. Worse, high-productivity wells with inherently low venting dominate basin averages. Stripper wells releasing 100% of associated gas are averaged in with flowbacks of 3-mile laterals. Over time, methane intensity can further decline as new wells are drilled, even without improvements to operating practices.
For comparing natural gas suppliers and for most other purposes, the more obvious and intuitive measure of “loss ratio” or “leak rate” makes more sense: it’s simply the fraction of gas that is produced but wasted by loss to the atmosphere.
The best available research shows that for an average natural gas to have less climate effect than an average coal when both are used for electricity, the loss ratio needs to be less than 1%. The breakeven point is even lower when compared to low-methane coals. For example, the research found that coal mined and used in the Liaoning province of China produces less greenhouse effect than natural gas exported from the Permian Basin of Texas unless the methane leak rate in Texas is lower than 0.2%. Being less harmful than coal seems like a reasonable standard for excellence.
Yet crude oils can leak 5 or 10% of their associated gas and still have calculated methane intensities below 1%. Our industry is using a target performance metric that does not identify poor performance.
Leak rate reveals the brutal truth, which is that we need to escape cognitive biases so that we can identify and improve our handling of gas. (These other analysts have also written about the choice of emissions units: S&P Global, RMI, and academic researchers.)
At a larger scale is the second-biggest hurdle: we collectively suffer from availability bias and confirmation bias. The companies with the lowest emissions trumpet their accomplishments, though usually they do so as measured by their methane intensity.
Ironically, the claimed success may not reflect operational improvements, but often instead reflect pre-existing results based on inherent economic incentives, selling high-emitting fields, or drilling more oil wells to increase the denominator.
But the kinds of companies conscientious of the public view of their emissions are also those operating with high-production wells with inherently low emissions. They do not generally include the much larger population of low-producing wells with systematically higher emissions.
My recent analysis of gas-disposition data self-reported by Texas operators shows that 38% of the state’s unplugged oil wells are associated with leases reporting either zero gas production or explicitly reporting all gas was vented or flared. Said differently, 4.2% of the oil production in the state comes from leases for which all gas is not sold and evidently vented, or, in the case of higher rates, flared.
The Permian is sometimes highlighted as a case study in success, but this too seems to be an issue of confirmation bias. It has been studied by a wide range of analysts and analytical methods to give a better sense of emissions rates, uncertainties, and implied progress. One study gained a great deal of attention in the industry because it showed a significant decline in methane intensity, but other studies of the region have reached different conclusions.
Multiple independent top-down studies, using towers, aircraft, satellites, and basin-scale atmospheric inversions, continue to indicate Permian methane loss rates near 3%, despite some reported improvements in methane intensity.
Fig. 1 shows the meta-analysis of methane loss ratio in the Permian from a recent research study. The graph shows the results of six previous analyses plus the seventh presented in the paper using the full range of atmospheric measurement tools, from towers to aerial surveys by industry contractors to national satellite information.
An eighth analysis conducted by MethaneSAT confirmed a Delaware Basin average loss ratio of 2.8%. Still another—a ninth top-down study and the most recent—provides the most continuous history of releases using a single methodology. It shows little or no decline in methane intensity on the Texas side of the Delaware but some reduction on the New Mexico side that is coincident with changing regulations. More importantly, it showed no meaningful decline in overall mass of vented volume since 2019.
While one important study documented gains at the basin level and earned widespread attention inside the industry, these nine other studies show that this centerpiece of the American oil industry has held loss ratios steady around 3% or slightly down since 2020. Thus, accounting only for in-basin releases, natural gas from the Permian used for electricity causes three times as much warming as an average coal and 15 times as much warming as the least impactful coals.
The Permian is not alone.
Another study looked across the globe using similar data and methodologies to examine excess atmospheric methane. Among the world’s top 40 gas-producing nations, 35 exceeded the target methane intensity of oil and gas operations (including oil production), and most exceeded the targets by five times or more. The US ranked 25th out of 40, with about 2% of the energy produced being wasted as methane emissions. And that is without considering flaring practices, which in 2025 were the third-worst globally in 30 years of data.
The worldwide industry affects the worldwide atmosphere. Methane from the Permian pollutes the planet just as much as methane from Venezuela, and methane from an integrated major pollutes just as much as methane from a stripper well operator.
In some places, our industry always had low emissions, and by some measures, some have improved performance. But in general, our industry has a long way to go to better serve the needs of the world.
Dwayne Purvis, SPE, is a petroleum engineering consultant, writer, speaker, and trainer based in Texas. He assists operators, investors, NGOs, and others in making thorough analyses and circumspect decisions, especially in hard-to-understand reservoirs and complex situations. With nearly 30 years of practice in multiple leadership roles and projects around the world, Purvis has studied scores of fields and advised clients of all sizes. He holds a degree in petroleum engineering from Texas A&M University and earned a master's degree in sustainable energy from Johns Hopkins University in 2013.