Geospatial variation in carbon accounting of hydrogen production and implications for the US Inflation Reduction Act | Nature Energy

Nature Energy (2024)Cite this article

Metrics details

Low-carbon hydrogen is considered a key component of global energy system decarbonization strategy. The US Inflation Reduction Act incentivizes low-carbon hydrogen production through tax credits that vary based on life-cycle greenhouse gas emissions intensity of hydrogen. Blue hydrogen or hydrogen produced from natural gas coupled with carbon capture and sequestration is one such pathway. Here we develop a geospatial, measurement-informed model to estimate supply-chain specific life-cycle greenhouse gas emissions intensity of blue hydrogen produced with natural gas sourced from the Marcellus and Permian shale basins. We find that blue hydrogen production using Permian gas has a life-cycle emissions intensity of 7.4 kg carbon dioxide equivalent per kg hydrogen (kgCO2e kg−1 H2), more than twice that of hydrogen produced using Marcellus gas of 3.3 kgCO2e kg−1 H2. Eligibility for tax credits should therefore be based on life-cycle assessments that are supply-chain specific and measurement informed to ensure blue hydrogen projects are truly low carbon.

Hydrogen is at the peak of technology hype cycles1. Recent pronouncements have declared hydrogen to be the decarbonization solution in heavy industry, agriculture, the power sector, shipping and aviation2,3,4,5,6,7. Governments around the world, led by the European Union and now the United States, have invested billions of dollars in jump-starting a hydrogen economy8,9,10. Even as potential applications of clean hydrogen continue to expand, demand for hydrogen in traditional applications such as fertilizer manufacturing and industry is expected to increase11. In parallel, energy system models of a net-zero economy by 2050 continue to suggest hydrogen as a key component of the global energy system12,13,14,15. Much of the hydrogen produced in the world is from fossil fuels such as coal or natural gas through steam methane reforming (SMR), with a greenhouse gas (GHG) emissions intensity of 20–25 kg carbon dioxide equivalent per kg hydrogen (kgCO2e kg−1 H2) or 10–30 kgCO2e kg−1 H2, respectively16,17,18. A large role for hydrogen in a decarbonized world requires urgent development and scale-up of low- and zero-carbon hydrogen production systems. In practice, the most likely candidate technologies include gas-based hydrogen production with carbon capture and sequestration (CCS), also called blue hydrogen, or electrolysis using clean electricity, also called green hydrogen18,19.

The US Department of Energy recently announced the creation of seven hydrogen hubs, each receiving approximately US$1 billion to accelerate commercial-scale deployment of clean hydrogen8. At least three of the announced hubs plan to produce low-carbon hydrogen from natural gas using CCS. In addition to federal investments, the Inflation Reduction Act (IRA) provided tax credits for clean hydrogen production20. The production tax credit (hereafter referred to as the 45V PTC) is available on a graded scale of the life-cycle GHG emissions intensity, with a minimum credit of US$0.60 per kg of H2 up to a maximum of US$3 per kg of H2. For comparison, recent estimates of costs of hydrogen production using clean-energy-based electrolysis and gas-based SMR were in the range of US$6–10 kg−1 H2 and US$1–4 kg−1 H2, respectively18,21. Whether a project is eligible for the full 45V PTC depends on the well-to-gate life-cycle GHG emissions intensity as estimated by Argonne National Lab’s greenhouse gases, regulated emissions and energy use in technologies (GREET) model22.

Several recent studies have estimated the life-cycle GHG emissions intensity of hydrogen production pathways16,17,18,23,24. Estimates of blue hydrogen emissions intensity range from 1 to 24 kgCO2e kg−1 H2, with the variability attributed to assumptions around supply-chain methane emissions, capture efficiency and global warming potential (GWP) values. A common theme is the use of nationally representative estimates for key model parameters such as average electricity grid carbon intensity or national average methane emissions rate. For example, the GREET model uses a base-case assumption of a nationally representative 1% methane leakage across US natural gas supply chains to estimate GHG intensity of blue hydrogen production. Recent work on the design of the 45V PTC for low-carbon electrolytic hydrogen production showed that verification and assurance of low-carbon attributes require stringent conditions to be imposed on time matching and additionality of clean electricity21.

The use of nationally representative parameter estimates for supply-chain methane emissions and other variables in a life-cycle framework is insufficient to estimate the GHG emissions intensity of hydrogen production. Measurements of methane emissions over the past decade have identified substantial spatio-temporal variability in emissions at different scales25,26,27,28,29,30. Field campaigns in the Permian Basin in Texas and New Mexico alone have reported methane emissions ranging from a production-normalized estimate of 3.7% to 9.4% (refs. 31,32,33). By contrast, recent aerial-based measurements in the Marcellus Shale Basin in Pennsylvania showed a methane emissions rate of less than 1% (refs. 34,35,36). In addition to spatial variation in emissions, analysis of recent measurements demonstrates that official methane emissions inventories such as the US Environmental Protection Agency (EPA) GHG Inventory underestimate methane emissions by 60% (refs. 25,37). Yet, most life-cycle assessment (LCA) studies of natural gas and blue hydrogen typically use EPA GHG inventory estimates of emissions38,39,40,41. To accurately estimate life-cycle emissions of blue hydrogen production in the United States, it is critical to use measurement-informed emissions inventories42,43,44.

Differences in supply-chain methane emissions, although important on their own, are just one of several key differences in emissions associated with the natural gas supply chain. For example, the degree of electrification affects the amount of grid-based electricity that is consumed across the supply chain at processing facilities and transmission compressor stations. The Marcellus Shale Basin, with a relatively high degree of electrification compared with the Permian Basin, uses more electricity. However, because of the coal-heavy electricity grids in Ohio and West Virginia where much of the natural gas processing facilities are located, embodied emissions associated with the use of grid electricity is higher for the Marcellus Basin compared with the Permian Basin. Thus, modelling the GHG emissions intensity of hydrogen production using national average estimates is unlikely to be representative of any real-world hydrogen production facility in the United States.

In this work, we develop a geospatial life-cycle assessment model to create supply-chain-specific estimates of the well-to-gate GHG emissions intensity of blue hydrogen production. We assume hydrogen is produced using SMR techniques and include emissions associated with capture, transport and sequestration of CO2 from the hydrogen production facility within life-cycle system boundaries. Our work represents two key advances: it creates a spatially explicit framework that helps assess the environmental impacts of supply-chain-specific hydrogen production and it incorporates a measurement-informed GHG emissions inventory within the LCA framework that addresses systemic emissions underestimation in official inventory estimates. We demonstrate the impact of these two advances using case studies of blue hydrogen production in Ohio and Texas using natural gas derived from the Marcellus and Permian basins, respectively.

Figure 1 shows the LCA boundary and relevant material flows through various life-cycle stages. The functional unit for this analysis is 1 kilogram of hydrogen production, thus the results are given in units of kgCO2e kg−1 H2. Emissions are aggregated using 100- and 20-year GWP for methane from the Sixth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC AR6). In this analysis, we do not consider the GHG emissions impact of hydrogen leakage because our supply-chain boundary ends at the hydrogen production facility. This choice enables our results to serve as a benchmark for the 45V PTC because it is based on a well-to-gate LCA that excludes emissions from hydrogen transportation and use. Recent work on modelling GWP of hydrogen has typically resulted in values between 1 and 1045,46.

Material flows of crude oil (brown), natural gas (blue) and natural gas liquids (red) are shown. Carbon capture and sequestration and electricity use are shown in purple and green, respectively.

Figure 2 shows the energy flows along the natural gas supply chain to produce 1 kg H2, accounting for all losses from methane emissions and flaring and process fuel consumption. To produce 1 kg of H2, the input energy that must be extracted in the Permian basin is nearly 2.5 times that required in the Marcellus Basin. One kg H2 produced using natural gas sourced from the Marcellus Basin requires 317 MJ of energy extracted, corresponding to a supply-chain energy conversion efficiency of 45%. By contrast, 1 kg of H2 produced using natural gas sourced from the Permian Basin requires 788 MJ of energy extracted, corresponding to a supply-chain energy conversion efficiency of 18%. The lower conversion efficiency in the Permian Basin compared with the Marcellus Basin can be attributed to two key factors: (1) difference in resource composition resulting from co-production of crude oil and natural gas liquids across the two basins and (2) differences in supply-chain methane emissions. In the southwest Marcellus Basin, co-products include dry natural gas and natural gas liquids (NGLs). In the Permian Basin, co-products include crude oil, dry natural gas and NGLs. Because of the higher energy density of crude oil, much of the extracted energy from the Permian Basin is embedded in crude oil, resulting in a lower energy conversion efficiency.

a,b, Energy flows of dry natural gas and co-products of crude oil and natural gas liquids across blue hydrogen production pathways using natural gas sourced from the Marcellus Basin (a) and Permian Basin (b). Losses correspond to methane emissions across the natural gas supply chain including leaks, venting and flaring, and fuel consumption corresponding to gas use across various process stages. SMR refers to steam methane reforming.

Figure 3 shows the difference in the life-cycle GHG emissions intensity of blue hydrogen production from the Marcellus and Permian shale basins. The base case considers CO2 capture from both the SMR by-product and fuel combustion required to operate the SMR plant. Carbon capture plants can theoretically achieve high capture rates of up to 96.2%. However, data from recent demonstration facilities show that observed capture rates can be lower than 96.2%—publicly available data suggest a pre-combustion capture rate of 78.2% and a post-combustion capture rate of 90% at full-load operation ('Key assumptions' in Methods)47,48. Results estimated using both optimistic and observed capture rates and two GWP values (GWP-20 and GWP-100) are shown in Fig. 3. Additional scenarios based on SMR-only CCS and zero-carbon energy inputs to hydrogen production are described in the Supplementary Information (Supplementary Figs. 3 and 4).

Life-cycle GHG emissions intensity in kgCO2e kg−1 H2 associated with blue hydrogen production using natural gas from the Marcellus and Permian shale basins considering theoretical and observed CO2 capture rates. Dark and light colours represent GWP-100 and GWP-20 results, respectively (IPCC AR 6 values). Overall, the Permian blue H2 supply chain has a base-case gas-production-normalized methane leakage of 5.2% whereas the Marcellus blue H2 supply chain has a base-case methane leakage of 1.25% before co-product allocation. The dotted line represents the threshold emissions intensity to qualify for the clean hydrogen production tax credits in the IRA.

Results from the base case with a 96.2% capture rate indicate that H2 produced from Marcellus gas has a life-cycle emissions intensity (EI) of 3.3 kgCO2e kg−1 H2. By contrast, H2 produced from Permian gas has a life-cycle GHG EI of approximately 7.4 kgCO2e kg−1 H2, over twice that of hydrogen produced from Marcellus gas (Supplementary Tables 7 and 8 for 20-year and 100-year GWP-based EI). Under this scenario, only blue hydrogen produced using gas from the Marcellus Basin will be eligible for the 45V PTC—such geospatial differences will not be apparent using GREET modeling as required by the Treasury. Using observed capture rates, life-cycle GHG EI of blue hydrogen increases to 5.1 and 9.2 kgCO2e kg−1 H2H2 for the Marcellus and Permian supply chains, respectively (Supplementary Tables 9 and 10). Thus, without high capture efficiencies, blue hydrogen production even in low methane leakage basins is unlikely to qualify for the 45V PTC. Furthermore, existing carbon capture projects have not demonstrated sustained operations at full load, thereby resulting in multi-year average capture efficiencies much lower than 90%. By contrast, high capture rates and use of hourly matched, zero-carbon clean electricity results in a life-cycle GHG EI of 2.6 and 6.9 kgCO2e kg−1 H2 for the Marcellus and Permian supply chains, respectively (Supplementary Fig. 3). These zero-carbon clean electricity results do not include embodied emissions from solar or wind in the life-cycle analysis, following Treasury guidelines. Furthermore, Treasury rules to procure clean power for hydrogen production, along with hourly matching requirements, increases the likelihood of additionality but does not guarantee it.

There are two key root causes for the difference in emissions intensity between the two H2 production supply chains: (1) higher upstream and midstream methane emissions in the Permian supply chain compared with the Marcellus supply chain and (2) differences in product streams from the two basins. Recent top-down methane measurements in the Permian Basin indicate a higher average production-normalized emissions rate compared with the Marcellus Basin. Midstream emissions associated with transmission compressor stations are larger for the Permian Basin because of the distance between the gas processing facilities in West Texas and demand centres in Southeast Texas, requiring more compressor stations to transport the gas (‘Natural gas transmission’ in Methods). By contrast, the distance between the gas processing facility in Southwest Pennsylvania for the Marcellus Basin is relatively closer to demand centres in Ohio, requiring fewer compressor stations for gas transport. Additionally, emissions from the natural gas supply chain are apportioned on an energy-based and product-assigned allocation basis, which varies by basin because of differences in co-products (‘Emissions allocation’ in Methods).

Recent measurement campaigns across the United States have demonstrated that methane emissions vary by facility, region and basin, often by an order of magnitude. Because of large uncertainties in measurement-based methane emissions estimates and efficiency of carbon capture processes, it is instructive to quantify the impact of these two key variables on the life-cycle EI of blue H2. This sensitivity analysis addresses two key challenges. First, it provides a quantitative tool to evaluate the eligibility of blue hydrogen for 45V PTC to determine benchmarks for methane emissions rate and carbon capture efficiency. Second, it identifies trade-offs between reducing methane leakage or improving capture efficiency in reducing emissions intensity of blue H2, relative to green hydrogen. This addresses principal agent issues—the entity developing the blue hydrogen project may only have direct control over capture technology (and therefore capture efficiency) and will be unable to reduce supply-chain methane emissions in the absence of a robust, domestic differentiated gas market43.

Figure 4 shows the effect of both methane emissions rate and carbon capture efficiency on the total life-cycle GHG EI of blue hydrogen production. Four scenarios are analysed: Marcellus and Permian gas-based hydrogen using average grid intensity in Ohio and Texas, respectively, and Marcellus and Permian gas-based hydrogen using hourly matched, zero-carbon electricity from additional wind capacity. The use of clean electricity to run the blue hydrogen plants in Ohio and Texas increases the allowable design space of supply-chain methane leakage and capture rate by reducing overall emissions intensity of blue H2. However, this requires that clean electricity is additional and is solely built to supply the hydrogen facility, as required by proposed Treasury guidelines. Thus, eligibility for the 45V PTC in the IRA for blue hydrogen depends on three key variables: supply-chain methane leakage, carbon capture rate and carbon intensity of electricity.

Impact of carbon capture rate (%, x axis) and supply-chain methane emissions rate (%, y axis) on the life-cycle GHG emissions intensity of blue hydrogen production from the Marcellus (a,b) and Permian (c,d) basins on a GWP-100 basis. The dotted line represents the threshold of 4 kgCO2e kg−1 H2 to qualify for 45V PTC in the IRA. The stars represent the base case with pre- and post-combustion capture rates (theoretical in blue and measured in black). The triangles represent scenarios with only pre-combustion capture. a,c, Assume electricity emissions based on average grid emissions factors. b,d, Assume the availability of hourly matched, zero emissions clean electricity.

Results from the base case indicate that only Marcellus gas-based blue hydrogen with a high 96.2% capture rate will qualify for the 45V PTC, corresponding to a subsidy equivalent to US$0.6 per kg H2. Furthermore, despite the relatively low supply-chain methane leakage rate, Marcellus-based blue hydrogen does not qualify for any tax credits if capture rates are less than 90% or the emissions from fuel combustion for the SMR are not captured. This increase in GHG emissions emphasizes the need for capturing not only the CO2 by-product from SMR but also the CO2 resulting from fuel combustion. The use of hourly matched clean electricity further reduces the emissions intensity of Marcellus gas-based blue H2 to 2.6 kgCO2e kg−1 H2, making it eligible for at least 20% of the maximum PTC. Efforts to reduce supply-chain methane leakage to about 0.5%—as some recent studies have shown—could further reduce EI to between 0.45 and 2.6 kgCO2e kg−1 H2. This would allow Marcellus gas-based blue H2 projects to be eligible for up to 33% of the maximum PTC.

By contrast, Permian gas-based H2 does not qualify for the 45V PTC in any of the three scenarios: base case, no post-combustion CO2 capture and clean electricity use. The only scenario where Permian gas-based blue hydrogen could qualify for tax credit provisions is if the total production-normalized supply-chain methane emissions rate is reduced to 2% or less, depending on overall CO2 capture rate. Furthermore, there is no scenario that only increases capture efficiency without addressing supply-chain methane emissions that would make Permian gas-based blue H2 eligible for 45V PTC. Thus, entities that want to develop blue H2 facilities in the Gulf Coast using gas sourced from the Permian Basin will need comprehensive strategies to address supply-chain methane emissions.

Although a wide range of capture rates are modelled here, they all assume long-term permanence of geologic sequestration49. This assumption is justified given that CCS in the context of the 45V PTC can only include permanent storage (‘CO2 transportation and injection’ in Methods). However, the capture rate in Fig. 4 can be used as a proxy for overall effectiveness of the CCS sub-system. In this context, the overall effectiveness of CCS represents the product of two efficiencies: the plant capture efficiency and the subsurface sequestration efficiency.

In both the Marcellus and Permian gas-based blue hydrogen, it is unlikely that any project will be eligible for the maximum PTC in the IRA of US$3 kg−1 H2, which requires an emissions intensity below 0.45 kgCO2e kg−1 H2. Achieving this in the context of blue H2 production requires advanced monitoring, widespread electrification and large-scale equipment replacement programs in developing near-zero methane leakage across natural gas supply chains.

Figure 5 shows a comparison of life-cycle GHG EI of blue hydrogen from this study with other hydrogen production pathways in the literature16,23,24,50,51,52,53,54,55. Green or zero-carbon hydrogen has some of the lowest GHG emissions intensity due to the use of electrolysis powered by renewable energy to produce hydrogen. However, these low emissions are a function of the electricity generation mix and hourly matching constraints powering the electrolysis. For example, if the electrolysers are connected to the local electric grid, GHG emissions will depend entirely on the average emissions intensity of the location-specific grid mix. Kleijne et al. argued that electrolysers obtain power from the average power mix of the grid that it is connected to unless additional renewable energy capacity is guaranteed51. Recent analyses have argued for stringent US Treasury guidelines for clean hydrogen including hourly matching of electrolyser energy needs with additional clean energy deployed for hydrogen production21. We estimate the emissions of grid-connected, electrolyser-based hydrogen production in both Ohio and Texas to compare it to our case studies. Emissions from grid-connected hydrogen production are 30.3 kgCO2e kg−1 H2 in Ohio and 21.5 kgCO2e kg−1 H2 in Texas, which are substantially higher than blue hydrogen produced in these states.

The dotted line represents the 4 kgCO2e kg−1 H2 threshold to qualify for clean hydrogen production tax credits in the IRA. Data from refs. 16,23,50,51,53,54,55. NREL, National Renewable Energy Laboratory; ETC, Energy Transitions Commission.

Emissions of green hydrogen produced from 100% renewables can also depend on the type of renewable source; for instance, hydrogen produced with wind power and photovoltaic cells have life-cycle GHG emissions of 0.05 and 0.5 kgCO2e kg−1 H2, respectively. These life-cycle emissions exclude the emissions associated with upstream and downstream processes related to the manufacturing and decommissioning of wind turbines or solar cells. The only hydrogen production technology that is comparable to green hydrogen produced with wind power is nuclear-based hydrogen production50.

Emissions associated with brown and grey hydrogen are substantially higher than those from clean energy and nuclear-based hydrogen. Hydrogen produced from fossil fuels without carbon capture have emissions intensity between 10 and 30 kgCO2e kg−1 H2 due to the release of CO2 from SMR into the atmosphere. Similarly, coal-based hydrogen has one of the highest emissions intensities in the literature of 25 kgCO2e kg−1 H2, similar to the highest gas-based hydrogen production54. For gas-based hydrogen production to be eligible for tax credits, it requires either effective methane emissions mitigation and high capture efficiency (blue hydrogen) or the use of methane pyrolysis24. Indeed, recent studies have shown that low-carbon blue hydrogen and green hydrogen production can be cost-competitive with carbon-intensive grey hydrogen56.

In this work, we developed a geospatial LCA of blue hydrogen production pathways using natural gas sourced from the Marcellus and Permian basins. We demonstrate the impact of spatial differences in GHG emissions across US natural gas supply chains on life-cycle EI of blue hydrogen production. Thus, eligibility for the hydrogen 45V PTC in the IRA should be project specific and location dependent to ensure tax credits go to truly low-carbon projects. We draw three primary conclusions from our analysis.

Location matters. Recent evidence from direct measurements across the oil and gas supply chain demonstrates the large spatio-temporal variation in methane emissions. These emissions directly affect the life-cycle GHG EI of blue hydrogen—results from our study indicate at least a 2× difference between the Permian and Marcellus basins. Crucially, using a nationally averaged methane emissions rate to determine average GHG emissions intensity of blue H2 production is unlikely to be representative of any blue H2 production facility. Guidelines to determine eligibility for 45V PTC should be based on regional or basin-specific modelling of the supply chain for natural gas. Major expansion in methane measurement campaigns across the United States over the past decade has made data on basin-specific emissions estimates readily available. Furthermore, an extensive body of work in the oil and gas methane emissions literature have concluded that measurement-informed emissions inventories are more accurate compared with conventional, bottom-up inventory estimates (Supplementary Sections 4–6). Updating conventional LCA models such as GREET using basin-specific, measurement-informed emissions inventory can enable analyses of blue H2 production pathways that are representative of US operations.

Levers available to blue hydrogen project developers to reduce emissions vary by location. In theory, there are three levers to reduce emissions intensity of H2 production—supply-chain methane emissions, capture efficiency and electricity source for hydrogen production. However, the importance of these levers varies by location. In the Marcellus Basin with relatively low methane emissions, all three levers can meaningfully reduce emissions with the most impact attributable to the use of clean electricity instead of grid electricity to power the SMR. However, in the Permian Basin, no amount of clean electricity or improved capture efficiency can mitigate the effects of high supply-chain methane emissions. Thus, without substantial reductions in supply-chain methane emissions in the Permian, blue hydrogen will not be a low-carbon hydrogen production pathway. This is further underscored by the potential need to avoid lock-in of fossil-based infrastructure with long lifetimes when alternative low-carbon pathways are likely to see rapid cost reductions in the near term57.

Counterfactuals dictate relative advantages of blue hydrogen over other methods of producing hydrogen. In the Marcellus Basin, blue hydrogen has a much lower life-cycle GHG emissions intensity compared with grid-based electrolysis given that coal is a major source of electricity. In a world where hydrogen demand is high, the choice of production pathway should be based on a localized and relative analysis of available deployment options. In some regions such as the Marcellus where abundant and low-emissions natural gas is readily available, blue hydrogen could be a viable pathway, whereas regions in the Southwest with abundant solar potential would be better served with hourly matched electrolysis-based hydrogen production. Finally, blue hydrogen could be a viable alternative only in natural-gas-producing regions where infrastructure is already in place and if the life-cycle GHG emissions are lower than other technologies. Given long infrastructure timelines associated with pipelines and other natural gas infrastructure, pathways that rely on clean electricity may present a faster option for low-carbon hydrogen production in regions that do not have existing natural gas infrastructure.

Limitations of this study have been discussed throughout this analysis. In addition, we emphasize the need for continued updates to analyses of the GHG EI of blue hydrogen projects. Methane emissions from the oil and gas supply chain will rapidly decline because of several factors over the next decade—state and federal regulations, methane fees in the IRA and voluntary initiatives. With rapid developments in emissions monitoring technology, it is imperative to develop timely, transparent and measurement-based estimates of supply-chain methane emissions to continually evaluate the GHG intensity of blue hydrogen relative to other production pathways. Blue hydrogen facilities that may not be eligible for the PTC today may become eligible tomorrow if reductions in supply-chain methane emissions can be credibly verified.

The analysis of blue hydrogen GHG life-cycle assessment (LCA) presented here incorporates measurement-informed methane emissions across the natural gas supply chain. Aggregating data across all publicly available emissions estimates, we calculate a production through transmission methane emissions rate of 1.25% for the Marcellus Basin and 5.2% for the Permian Basin. Analysis presented in this study assumes that the marginal gas supply to hydrogen production in Houston and Ohio come from the Permian and Marcellus basins, respectively. Whereas a detailed gas delivery pathway analysis would require access to proprietary data on daily flows and contracts, a recent peer-reviewed studied developed an approximate pathway analysis for US natural gas transmission network. Their findings support our assumptions in this study58. Furthermore, the base-case blue hydrogen supply chain considers two carbon capture rates: (1) a theoretical capture rate of 96.2% with two solvent units from Lewis et al.18 pre-combustion and post-combustion capture and (2) an observed capture rate of 78.2% for pre-combustion capture from Shell’s Quest CCS facility in Canada and a post-combustion capture rate of 90% from Cansolv’s first commercial plant47,48. The pre-combustion carbon capture achieved by methyl diethanolamine captures CO2 from the by-product generated by the SMR reaction, whereas the post-combustion capture accomplished using Shell’s Cansolv solvent absorbs the CO2 generated by fuel combustion. As a result, the overall carbon capture efficiency increases by implementing these two absorption processes. Transmission compressor stations are assumed to be operated using reciprocating engines because approximately 78% of the compressors in the United States are the reciprocating type58.

Conventional LCA protocols as defined by International Standards Organization (ISO) 14040 require emissions allocation among different co-products when there is more than one output flow59. We use an energy-based and product-assigned allocation method, which allows for consideration of the different streams produced in each facility type60,61,62. In the Permian Basin, emissions are allocated between crude oil and produced gas in the production stage—these values are typically reported at the facility level to relevant state and federal regulatory agencies. The produced gas is a high-pressure fluid mixture that contains dry natural gas and natural gas liquids (NGLs). However, NGLs are only separated at the processing stage. Thus, a second emissions allocation is done at the processing stage between dry natural gas used for hydrogen production and NGLs. Similarly, although no crude oil is produced in the Marcellus Shale Basin, emissions are allocated between dry natural gas and NGLs at the processing stage. Supplementary Table 3 summarizes the proportion of GHG emissions allocated to different products in each stage of the hydrogen supply chain.

Methane emissions across the natural gas supply chain are estimated based on data from publicly available peer-reviewed studies. Across all stages, top-down methane emissions measurements are used to estimate measurement-informed emissions inventory for each stage. For the production and gathering and boosting stage of the supply chain, data from top-down aerial field campaigns and satellites are aggregated and averaged to generate a production-normalized methane emissions rate for each basin. Data for processing and transmission stages are obtained from a combination of facility-specific operational data and top-down measurement field campaigns. In all measurement-informed inventory estimates, emissions below the detection threshold of the measurement instrument were included when the original studies only report on measured emissions (Supplementary Tables 3 and 4).

We developed a model for a well-to-gate LCA for blue hydrogen with system boundaries that include drilling and completion of a well through the capture, compression and injection of CO2 in the subsurface. This assessment excludes emissions associated with hydrogen transportation to the end user and assumes that the hydrogen production facility is located near demand centres and will not require extensive hydrogen transportation networks. Thus, potential GHG impacts from hydrogen leakage are neglected in this study. The functional unit of our LCA is 1 kg H2—all mass units are converted to energy units using higher heating values (HHV) to track various co-products across the supply chain. LCAs are informed by a life-cycle inventory (LCI) conducted in accordance with standard ISO guidelines for life-cycle assessments61 (Supplementary Section 11).

Emissions across different stages of the supply chain are estimated through a detailed LCI that considers all material and energy inputs to each stage. Both the flow of fuels and emissions are tracked throughout the supply chain including CO2 emissions resulting from the combustion of fuels including natural gas and diesel, CH4 emissions from the natural gas supply chain and emissions associated with the electric grid used to power SMR and carbon capture and storage. The following sections describe key data sources and assumptions associated with estimating CO2 and CH4 emissions at each stage of the LCA. Differences in assumptions and data sources between the Marcellus and Permian supply chains are also included.

Methane emissions associated with drilling were estimated from a top-down aerial survey in southwestern Pennsylvania, which determined a methane mass flow rate of 34 g s−1 per well from a well pad in the drilling phase, occurring for an average of 22 days of drilling of typical unconventional wells63. During well completion, where natural gas is sent to a flare with an assumed 98% destruction efficiency, and to open tanks, where it is vented. These volumes of vented gas are considered in our analysis based on direct source methane measurements for well completion flowbacks in the Appalachian and Gulf Coast regions by Allen et al.64

Diesel is used for drilling and hydraulic fracturing, which releases CO2 into the atmosphere during combustion. The LCI for this process is informed by the operational parameters presented in Mallapragada et al. for the Marcellus Basin65. No public data were found for fuel use to drill and fracture a well in the Permian Basin, so parameters for fuel use from the Bakken and the same operator were used to account for these emissions66. The use of data from the Bakken Shale Basin as a proxy for the Permian Basin drilling is justified because the fuel for drilling is a function of the well’s total depth. The average total depth of 14,628 ft for unconventional wells in the Bakken Shale (Williston Basin) is comparable to the average total depth for unconventional wells in the Permian Basin of 12,177 ft (ref. 67). On the other hand, the average total well depth in the Appalachian Basin, where the Marcellus is located, is 7,947 ft (ref. 67).

Once the well has been completed, it enters the production stage. Here oil is gathered within a pipeline network to be sent to refineries, and natural gas is compressed and transported to gas processing plants. The analysis for the Marcellus Shale is based on the southwestern region, which is known for producing wet gas. In contrast, most of the Permian Basin produces oil and associated gas with a gas-oil-ratio less than 4,000 standard cubic feet per barrel (ref. 68). This means that for the production stage, all the emissions in the Marcellus Shale correspond to produced natural gas, whereas the emissions in the Permian Basin must be allocated between crude oil and produced natural gas. The energy flows at this stage include fuel consumption and gas loss or emissions due to methane leakage, liquids unloading and flaring.

For production and gathering and boosting facilities in the Marcellus, fuel use is estimated to be 2.32% unit volume of fuel gas per unit volume of throughput based on field data65. For the Permian basin, this ratio is estimated to be 5.02%, according to ref. 38. The higher fraction of natural gas used in operations in the Permian is a result of the lesser degree of electrification compared with the Marcellus Basin.

CO2 emissions from fuel combustion is given by:

where Fuel use is percentage of gas throughput used as fuel (unit volume fuel gas per unit volume throughput in %), mgas throughput is mass of gas throughput (kg) and CO2 emissions factor is kgCO2 emitted per kg natural gas burned (2.69 kgCO2 kg−1 NG).

Measurement-informed methane emissions were calculated by aggregating data from peer-reviewed studies that conducted top-down methane measurements. These estimates were then converted into production-normalized emissions rates in the Marcellus and Permian basins (Supplementary Tables 3 and 4). Methane emissions for each LCA stage are given by:

where Avg CH4 emissions rate is average production-normalized methane emissions rate (%), mgas is mass of gas required per stage to produce 1 kg H2 (kg), \({\chi }_{\mathrm{CH}_4}\) is mass fraction of methane in raw natural gas (for production and processing) and pipeline quality gas (for transmission and SMR) and GWP is global warming potential (kgCO2e kg−1 CH4).

Gas composition varies depending on the supply-chain stage as noted in Supplementary Tables 6 and 7 (refs. 65,69).

Liquids unloading was also considered as an intermittent source of methane emissions. Only 13% of gas wells in the United States in 2012 vented gas resulting in emissions from liquids unloading70. For this study, one liquid unloading event with emissions per well is considered. Data from liquid unloading in the Appalachian and Permian basins from Zaimes et al.67 was used to estimate emissions associated with liquid unloading.

In the production stage, we assume a flare destruction efficiency of 98%, following EPA guidelines for well-operated flares. Direct measurements of flare destruction efficiency by Caulton et al.71 concluded that all the flares functioned with an efficiency of >99.8%. These emissions are embedded in the top-down methane measurements and are not considered separately in the LCI to avoid double counting. However, CO2 emissions from flares are included separately as most field campaigns only measure methane emissions. Data from the satellite-informed tool SkyTruth shows that the volume of gas flared in the Permian is higher than that being flared in the Marcellus72. In addition, no publicly available data were found regarding flare volumes from wells in routine production—recent studies indicate that flare volumes depend on several basin-specific factors such as pipeline takeaway capacity, price of natural gas and crude oil and demand73. Only Marcellus data for flaring during well completions was found, and these emissions were included. Furthermore, a recent analysis by Rystad Energy and the Environmental Defense Fund concluded that flaring intensity in the Permian is about 1% (ref. 74).

Gas processing allows for the separation of produced gas into two product streams NGLs and dry natural gas. The energy allocation for this stage is achieved based on the NGL yield for each region, which is the ratio of barrels of NGL to the volume of produced natural gas. The Energy Information Administration (EIA) reported that the Northern Appalachian Basin has a NGL yield of 72 bbl MMscf−1 on average, whereas the Permian Basin generates a higher yield of 95 bbl MMscf−1 due to the larger percentage of heavier hydrocarbons75. During processing, energy from NG flows as produced gas condensed to NGL, fuel consumption and gas loss or emissions due to methane leakage.

The fuel consumption ratio for the Marcellus and Permian basins is estimated to be 0.19% and 2.3%, respectively, based on available data from the literature65,69,76. The only source of methane emissions in this stage is from the methane leakage from equipment (Supplementary Tables 3 and 4). Gas compositions are obtained from Mallapragada et al. and Contreras et al. for the Marcellus and Permian basins, respectively65,69.

The compressors in this study are assumed to be gas-driven reciprocating engines. Approximately 78% of compressors in the United States are reciprocating engines, 19% are centrifugal engines and 3% are electrical centrifugal engines58. Fuel consumption is directly proportional to the number of compression stations as volumes are assumed to be constant for a steady state blue hydrogen production. Furthermore, the number of compressor stations is a function of the transmission distance from the gas-producing region to the demand point. For the Marcellus Shale, the demand centre is assumed be in Ohio77. For the Permian Basin, the demand centre is Houston. For this analysis, an average of one compressor station is considered to transport gas from the southwest Marcellus Shale to Ohio and five compressor stations to transport gas from the Permian Basin to Houston (Supplementary Section 11). The number of compressor stations was estimated based on the assumption that the average distance between transmission compressor stations in the United States is about 75 miles and direct mapping of individual stations from the Homeland Infrastructure Foundation-Level Data. The energy associated with fuel required for operating compressor stations is given by:

where Efuel is energy required from fuel for each compressor station (British Thermal Units (BTU) yr−1), p is reciprocating engine power (horsepower, hp), 𝜂 is reciprocating engine efficiency (dimensionless), top is operating hours (h yr−1), nunits is number of compressor units per compressor station and nstations is number of compressor stations.

Compressor stations also emit methane emissions through methane slip from the engines and other fugitive and vented emissions. The same top-down studies that conducted methane measurements for the production and processing stages also included measurements from select compressor stations. In addition, Zimmerle et al. estimated a mean production-normalized gas leakage of 0.22% for transmission compression stations with reciprocating engines78. We aggregate data from all publicly available studies on methane emissions from compressor stations and account for emissions from pipeline leaks and venting, which is 0.03% of gas throughput.

The technology for hydrogen production included in this study is SMR due to its efficiency, economic viability and commercial readiness. There are two chemical reactions that govern this process:

SMR reaction: \(\mathrm{C{H}}_{4}+\mathrm{H}_{2}\mathrm{O}\to 3\mathrm{H}_{2}+\mathrm{CO}\)

Water–gas shift reaction: \(\mathrm{CO}+\mathrm{H}_{2}\mathrm{O}\to{\mathrm{H}_{2}}+\mathrm{C{O}}_{2}\)

This LCA includes the capture and compression of the CO2 by-product generated by the SMR reaction and the CO2 generated from fuel combustion. Lewis et al. modelled the flows of feedstocks and fuels for the SMR process with CO2 capture at each stage18. The plant configuration utilized in our work is an SMR plant with both pre-combustion and post-combustion CO2 capture. HHV of hydrogen is used to calculate the energy required to produce 1 kg H2 because the efficiency for SMR is 72.1% on an HHV basis. An overall carbon capture efficiency of 96.2% following both pre-combustion and post-combustion capture is used in the base-case analysis.

On the basis of the following equations from McCollum and Ogden, the power required to compress CO2 in its gaseous form and pump it when it reaches a supercritical state at approximately 7.38 MPa, is estimated79. Supplementary Table 37 (Supplementary Section 11) shows typical CO2 compressibility and ratio of specific heats for each compression state.

where Ws,i is compression power requirement for each individual stage (kW), m is CO2 mass flow rate to be transported to injection site (tonnes per day), Zs is average CO2 compressibility for each individual stage (dimensionless), R is gas constant (kJ kmol-K−1), Tin is CO2 temperature at compressor inlet (K), M is molecular weight of CO2 (kg kmol−1), 𝜂is is Isentropic efficiency of compressor (dimensionless), ks is average ratio of specific heat at constant pressure (Cp) to the specific heat at constant volume (Cv) of CO2 for each individual stage, Cp/Cv (dimensionless), CR is compression ratio of each stage (dimensionless), Ws,total is total combined compression power requirement for all stages (kW), Pfinal is final pressure of CO2 for pipeline transport (MPa), Pcut-off is pressure at which compression changes to pumping (MPa), 𝜌 is density of CO2 during pumping (kg m−3) and 𝜂p is efficiency of pump (dimensionless).

The total power requirement for a plant configuration of SMR with CO2 capture and compression integrated as described above is 26,893.6 kW or 1.34 kWh kg−1 H2. The power required for SMR and CO2 capture was obtained from Lewis et al.18, and the power requirement for CO2 compression and pumping for supercritical CO2 was calculated using equations (4)–(6). The normalized SMR power is 0.35 kWh kg−1 H2, which is 26% of the total power and for CO2 handling (that is, both capture and compression) is 0.98 kWh kg−1 H2, equal to 74% of the total power.

Emissions from the electric grid vary for different supply chains based on the location of the SMR facility. According to the latest data from the Emissions and Generation Resource Integrated Database (eGRID) from the Environmental Protection Agency (EPA), the state of Texas has an average electricity emissions factor of 0.39 kgCO2e kWh−1, whereas Ohio has electric grid emissions factor of 0.55 kgCO2e kWh−1.

We also estimated the pipeline distance from the CO2 capture and compression facility to the closest CO2 injection well with ArcGIS. The power requirement to pump CO2 through this distance and the power required to inject CO2 based on injection pressure needed are calculated based on equation (6). An estimate of the injection pressure is based on the normal pressure gradient of 0.433 pounds per square inch (psi) ft−1 and the depth of the candidate formations for CO2 storage for each case80,81.

The supercritical CO2 must be pumped to the CO2 injection site. The estimated pressure drop along CO2 pipelines is 0.13 MPa per mile, according to the National Energy Technology Laboratory (NETL). The CO2 pressure at the injection site is given by:

where \({P}_{{\mathrm{CO}}_2}\) is CO2 pressure at injection site (MPa), Pfinal is final CO2 pressure after pumping (MPa), Δ𝑃drop is pressure drop along CO2 transportation pipeline (MPa per mile) and d is distance between the hydrogen plant and the CO2 injection site (miles).

An estimate of the injection pressure assuming a normal hydrostatic pressure gradient of 0.433 psi ft−1 and an average depth of formations for CO2 sequestration based on the literature can be determined by80,82:

where Pinj is injection pressure (MPa), Δ𝑃g is normal hydrostatic pressure gradient (0.433 psi ft−1) and D is depth of formation for CO2 sequestration (ft).

Whereas there are ongoing concerns regarding the long-term efficacy of CO2 subsurface storage, we assume the selection of adequate geologic formations for CO2 injection and well-regulated storage sites. This is a reasonable assumption because CCS in the context of hydrogen PTC is only applicable for Class II or Class VI injection well permits issued by the EPA that assume permanent sequestration. On the basis of these assumptions, the annual CO2 leakage rate is negligible81.

All data necessary to replicate the results presented in this study are provided along with the paper. Additional data tables and figures are available in Supplementary Information. The development of measurement-informed emissions inventories is based on peer-reviewed and publicly available data on methane emissions. Corrections to published estimates to be incorporated in this work are described in Supplementary Information, whereas the original measurement data can be found in the peer-reviewed literature.

The Excel spreadsheet tool associated with the model in this study is provided as Supplementary Information.

van Renssen, S. The hydrogen solution? Nat. Clim. Change 10, 799–801 (2020).

Article Google Scholar

Global Hydrogen Review 2023 (IEA, 2023); https://www.iea.org/reports/global-hydrogen-review-2023

The Role of Hydrogen in Meeting Our 2030 Climate and Energy Targets (DG COMM, 2021); https://data.europa.eu/doi/10.2775/833

Adler, E. J. & Martins, J. R. R. A. Hydrogen-powered aircraft: fundamental concepts, key technologies, and environmental impacts. Prog. Aerosp. Sci. 141, 100922 (2023).

Article Google Scholar

Hoang, A. T. et al. Technological solutions for boosting hydrogen role in decarbonization strategies and net-zero goals of world shipping: challenges and perspectives. Renew. Sustain. Energy Rev. 188, 113790 (2023).

Article Google Scholar

Griffiths, S., Sovacool, B. K., Kim, J., Bazilian, M. & Uratani, J. M. Industrial decarbonization via hydrogen: a critical and systematic review of developments, socio-technical systems and policy options. Energy Res. Social Sci. 80, 102208 (2021).

Article Google Scholar

Sepulveda, N. A., Jenkins, J. D., Edington, A., Mallapragada, D. S. & Lester, R. K. The design space for long-duration energy storage in decarbonized power systems. Nat. Energy 6, 506–516 (2021).

Article Google Scholar

White House Biden-Harris Administration Announces $7 Billion for America’s First Clean Hydrogen Hubs, Driving Clean Manufacturing and Delivering New Economic Opportunities Nationwide (US DOE, 2023).

Hydrogen Net Zero Investment Roadmap (DESNZ, 2023); https://www.gov.uk/government/publications/hydrogen-net-zero-investment-roadmap

Commission Approves up to €5.2 Billion of Public Support by Thirteen Member States for the Second Important Project of Common European Interest in the Hydrogen Value Chain (European Commission, 2022).

Hydrogen (IEA, 2023); https://www.iea.org/reports/global-hydrogen-review-2023

van der Spek, M. et al. Perspective on the hydrogen economy as a pathway to reach net-zero CO2 emissions in Europe. Energy Environ. Sci. 15, 1034–1077 (2022).

Article Google Scholar

Odenweller, A., Ueckerdt, F., Nemet, G. F., Jensterle, M. & Luderer, G. Probabilistic feasibility space of scaling up green hydrogen supply. Nat. Energy 7, 854–865 (2022).

Article Google Scholar

Yang, X., Nielsen, C. P., Song, S. & McElroy, M. B. Breaking the hard-to-abate bottleneck in China’s path to carbon neutrality with clean hydrogen. Nat. Energy 7, 955–965 (2022).

Article Google Scholar

Davis, S. J. et al. Net-zero emissions energy systems. Science 360, eaas9793 (2018).

Article Google Scholar

Bauer, C. et al. On the climate impacts of blue hydrogen production. Sustain. Energy Fuels 6, 66–75 (2022).

Article Google Scholar

Oni, A. O., Anaya, K., Giwa, T., Di Lullo, G. & Kumar, A. Comparative assessment of blue hydrogen from steam methane reforming, autothermal reforming, and natural gas decomposition technologies for natural gas-producing regions. Energy Convers. Manag. 254, 115245 (2022).

Lewis, E. et al. Comparison of Commercial, State-of-the-Art, Fossil-Based Hydrogen Production Technologies Report No. DOE/NETL-2022/3241 (OSTI, 2022); https://doi.org/10.2172/1862910

Terlouw, T., Bauer, C., McKenna, R. & Mazzotti, M. Large-scale hydrogen production via water electrolysis: a techno-economic and environmental assessment. Energy Environ. Sci. 15, 3583–3602 (2022).

Article Google Scholar

117th Congress, Inflation Reduction Act, Public Law 117–169 136 Stat. 1818 (US Congress, 2022).

Ricks, W., Xu, Q. & Jenkins, J. D. Minimizing emissions from grid-based hydrogen production in the United States. Environ. Res. Lett. 18, 014025 (2023).

Article Google Scholar

The Greenhouse Gases, Regulated Emissions, and Energy Use in Technologies Model (Argonne National Lab, 2017); https://greet.anl.gov/

Howarth, R. W. & Jacobson, M. Z. How green is blue hydrogen? Energy Sci. Eng. 9, 1676–1687 (2021).

Article Google Scholar

Sánchez-Bastardo, N., Schlögl, R. & Ruland, H. Methane pyrolysis for zero-emission hydrogen production: a potential bridge technology from fossil fuels to a renewable and sustainable hydrogen economy. Ind. Eng. Chem. Res. 60, 11855–11881 (2021).

Article Google Scholar

Alvarez, R. A. et al. Assessment of methane emissions from the U.S. oil and gas supply chain. Science 361, 186–188 (2018).

Article Google Scholar

Wang, J. L. et al. Multiscale methane measurements at oil and gas facilities reveal necessary frameworks for improved emissions accounting. Environ. Sci. Technol. https://doi.org/10.1021/acs.est.2c06211 (2022).

Article Google Scholar

Chan, E. et al. Eight-year estimates of methane emissions from oil and gas operations in western Canada are nearly twice those reported in inventories. Environ. Sci. Technol. 54, 14899–14909 (2020).

Article Google Scholar

Daniels, W. S. et al. Toward multiscale measurement-informed methane inventories: reconciling bottom-up site-level inventories with top-down measurements using continuous monitoring systems. Environ. Sci. Technol. 57, 11823–11833 (2023).

Article Google Scholar

Lauvaux, T. et al. Global assessment of oil and gas methane ultra-emitters. Science 375, 557–561 (2022).

Article Google Scholar

de Gouw, J. A. et al. Daily satellite observations of methane from oil and gas production regions in the United States. Sci. Rep. 10, 1379 (2020).

Article Google Scholar

Chen, Y. et al. Quantifying regional methane emissions in the New Mexico permian basin with a comprehensive aerial survey. Environ. Sci. Technol. 56, 4317–4323 (2022).

Article Google Scholar

Cusworth, D. H. et al. Intermittency of large methane emitters in the permian basin. Environ. Sci. Technol. Lett. 8, 567–573 (2021).

Article Google Scholar

Robertson, A. M. et al. New Mexico Permian Basin measured well pad methane emissions are a factor of 5–9 times higher than U.S. EPA estimates. Environ. Sci. Technol. 54, 13926–13934 (2020).

Article Google Scholar

Omara, M. et al. Methane emissions from conventional and unconventional natural gas production sites in the Marcellus Shale Basin. Environ. Sci. Technol. 50, 2099–2107 (2016).

Article Google Scholar

Barkley, Z. et al. Quantification of oil and gas methane emissions in the Delaware and Marcellus basins using a network of continuous tower-based measurements. Atmos. Chem. Phys. 23, 6127–6144 (2023).

Article Google Scholar

Ren, X. et al. Methane emissions from the Marcellus Shale in Southwestern Pennsylvania and Northern West Virginia based on airborne measurements. J. Geophys. Res. Atmos. 124, 1862–1878 (2019).

Article Google Scholar

Rutherford, J. S. et al. Closing the methane gap in US oil and natural gas production emissions inventories. Nat. Commun. 12, 4715 (2021).

Article Google Scholar

Roman-White, S. A. et al. LNG supply chains: a supplier-specific life-cycle assessment for improved emission accounting. ACS Sustain. Chem. Eng. 9, 10857–10867 (2021).

Article Google Scholar

Gilbert, A. Q. & Sovacool, B. K. Carbon pathways in the global gas market: an attributional lifecycle assessment of the climate impacts of liquefied natural gas exports from the United States to Asia. Energy Policy 120, 635–643 (2018).

Article Google Scholar

Nie, Y. et al. Greenhouse-gas emissions of Canadian liquefied natural gas for use in China: comparison and synthesis of three independent life cycle assessments. J. Clean. Prod. 258, 120701 (2020).

Article Google Scholar

Jordaan, S. M. et al. Global mitigation opportunities for the life cycle of natural gas-fired power. Nat. Clim. Change 12, 1059–1067 (2022).

Article Google Scholar

Johnson, M. R., Conrad, B. M. & Tyner, D. R. Creating measurement-based oil and gas sector methane inventories using source-resolved aerial surveys. Commun. Earth Environ. 4, 139 (2023).

Ravikumar, A. P. et al. Measurement-based differentiation of low-emission global natural gas supply chains. Nat. Energy https://doi.org/10.1038/s41560-023-01381-x (2023).

Zhu, Y., Allen, D. & Ravikumar, A. Geospatial life cycle analysis of greenhouse gas emissions from US liquefied natural gas supply chains. Preprint at ChemRxiv https://doi.org/10.26434/chemrxiv-2024-9v8dw (2024).

Sand, M. et al. A multi-model assessment of the global warming potential of hydrogen. Commun. Earth Environ. 4, 203 (2023).

Ocko, I. B. & Hamburg, S. P. Climate consequences of hydrogen emissions. Atmos. Chem. Phys. 22, 9349–9368 (2022).

Article Google Scholar

Quest Carbon Capture and Storage Project: Annual Report, 2021 (Alberta Department of Energy, 2022); https://open.alberta.ca/publications/quest-carbon-capture-and-storage-project-annual-report-2021

Singh, A. & Stéphenne, K. Shell Cansolv CO2 capture technology: achievement from first commercial plant. Energy Procedia 63, 1678–1685 (2014).

Article Google Scholar

Snæbjörnsdóttir, S. Ó. et al. Carbon dioxide storage through mineral carbonation. Nat. Rev. Earth Environ. 1, 90–102 (2020).

Article Google Scholar

Ozbilen, A., Dincer, I. & Rosen, M. A. Comparative environmental impact and efficiency assessment of selected hydrogen production methods. Environ. Impact Assess. Rev. 42, 1–9 (2013).

Article Google Scholar

de Kleijne, K., de Coninck, H., van Zelm, R., Huijbregts, M. A. J. & Hanssen, S. V. The many greenhouse gas footprints of green hydrogen. Sustain. Energy Fuels 6, 4383–4387 (2022).

Article Google Scholar

Making the Hydrogen Economy Possible: Accelerating Clean Hydrogen in an Electrified Economy (Energy Transitions Commission, 2021); https://www.energy-transitions.org/publications/making-clean-hydrogen-possible/

Ajanovic, A., Sayer, M. & Haas, R. The economics and the environmental benignity of different colors of hydrogen. Int. J. Hydrog. Energy 47, 24136–24154 (2022).

Article Google Scholar

Machhammer, O., Bode, A. & Hormuth, W. Financial and ecological evaluation of hydrogen production processes on large scale. Chem. Eng. Technol. 39, 1185–1193 (2016).

Article Google Scholar

Pettersen, J. et al. Blue hydrogen must be done properly. Energy Sci. Eng. 10, 3220–3236 (2022).

Article Google Scholar

Cheng, F., Luo, H., Jenkins, J. D. & Larson, E. D. Impacts of the Inflation Reduction Act on the economics of clean hydrogen and synthetic liquid fuels. Environ. Sci. Technol. 57, 15336–15347 (2023).

Article Google Scholar

Dillman, K. & Heinonen, J. Towards a safe hydrogen economy: an absolute climate sustainability assessment of hydrogen production. Climate 11, 25 (2023).

Article Google Scholar

Littlefield, J., Rai, S. & Skone, T. J. Life cycle GHG perspective on U.S. natural gas delivery pathways. Environ. Sci. Technol. 56, 16033–16042 (2022).

Article Google Scholar

Skone, T. J., Littlefield, J. & Marriott, J. Life Cycle Greenhouse Gas Inventory of Natural Gas Extraction, Delivery and Electricity Production Report No. NETL/DOE-2011/1522 (OSTI, 2011); https://doi.org/10.2172/1515238

ISO 14040:2006, Environmental Management—Life Cycle Assessment—Principles and Framework (International Organization for Standardization, 2006).

Zavala-Araiza, D., Allen, D. T., Harrison, M., George, F. C. & Jersey, G. R. Allocating methane emissions to natural gas and oil production from shale formations. ACS Sustain. Chem. Eng. 3, 492–498 (2015).

Article Google Scholar

Aldrich, R., Llauró, F. X., Puig, J., Mutjé, P. & Pèlach, M. À. Allocation of GHG emissions in combined heat and power systems: a new proposal for considering inefficiencies of the system. J. Clean. Prod. 19, 1072–1079 (2011).

Article Google Scholar

Rosselot, K. S., Allen, D. T. & Ku, A. Y. Comparing greenhouse gas impacts from domestic coal and imported natural gas electricity generation in China. ACS Sustain. Chem. Eng. 9, 8759–8769 (2021).

Article Google Scholar

Caulton, D. R. et al. Toward a better understanding and quantification of methane emissions from shale gas development. Proc. Natl Acad. Sci. USA 111, 6237–6242 (2014).

Article Google Scholar

Allen, D. T. et al. Measurements of methane emissions at natural gas production sites in the United States. Proc. Natl Acad. Sci. USA 110, 17768–17773 (2013).

Article Google Scholar

Mallapragada, D. S. et al. Life cycle greenhouse gas emissions and freshwater consumption of liquefied Marcellus Shale gas used for international power generation. J. Clean. Prod. 205, 672–680 (2018).

Article Google Scholar

Laurenzi, I. J., Bergerson, J. A. & Motazedi, K. Life cycle greenhouse gas emissions and freshwater consumption associated with Bakken tight oil. Proc. Natl Acad. Sci. USA 113, E7672–E7680 (2016).

Article Google Scholar

Zaimes, G. G. et al. Characterizing regional methane emissions from natural gas liquid unloading. Environ. Sci. Technol. 53, 4619–4629 (2019).

Article Google Scholar

Advances in Technology Led to Record New Well Productivity in the Permian Basin in 2021 (US EIA, 2022); https://www.eia.gov/todayinenergy/detail.php?id=54079

Contreras, W. et al. Life cycle greenhouse gas emissions of crude oil and natural gas from the Delaware Basin. J. Clean. Prod. 328, 129530 (2021).

Article Google Scholar

Allen, D. T. et al. Methane emissions from process equipment at natural gas production sites in the United States: liquid unloadings. Environ. Sci. Technol. 49, 641–648 (2015).

Article Google Scholar

Caulton, D. R. et al. Methane destruction efficiency of natural gas flares associated with shale formation wells. Environ. Sci. Technol. 48, 9548–9554 (2014).

Article Google Scholar

Allen, D. T., Roman-White, S. A., McCormick, M. & George, F. Moving toward zero routine flaring in the Permian Basin oil and gas production region: measuring progress and driving factors. ACS Sustain. Resour. Manag. 1, 1041–1046 (2024).

Lyon, D. R. et al. Concurrent variation in oil and gas methane emissions and oil price during the COVID-19 pandemic. Atmos. Chem. Phys. 21, 6605–6626 (2021).

Article Google Scholar

Goldstein, J. & Weatherall, G. New Rystad cost analysis makes case for EPA to end routine flaring in final methane rule. EDF Blogs https://blogs.edf.org/energyexchange/2022/02/28/new-rystad-cost-analysis-makes-case-for-epa-to-end-routine-flaring-in-final-methane-rule/ (2022).

Energy Information Administration. U.S. natural gas plant liquid production continues to hit record highs. Today in Energy https://www.eia.gov/todayinenergy/detail.php?id=38772 (2019).

Roman-White, S., Rai, S., Littlefield, J., Cooney, G. & Skone, T. J. Life Cycle Greenhouse Gas Perspective on Exporting Liquefied Natural Gas from the Unites States: 2019 Update (OSTI, 2019); https://doi.org/10.2172/1607677

Henning, M., Thomas, A., Psarras, P. & Triozzi, M. Developing a Hydrogen Economy in Ohio: Challenges and Opportunities (Cleveland State Univ., 2022); https://engagedscholarship.csuohio.edu/urban_facpub/1765

Zimmerle, D. J. et al. Methane emissions from the natural gas transmission and storage system in the United States. Environ. Sci. Technol. 49, 9374–9383 (2015).

Article Google Scholar

McCollum, D. L. & Ogden, J. M. Techno-Economic Models for Carbon Dioxide Compression, Transport, and Storage & Correlations for Estimating Carbon Dioxide Density and Viscosity (UC Davis, 2006); https://escholarship.org/uc/item/1zg00532

Cumming, L., Hawkins, J., Sminchak, J., Valluri, M. & Gupta, N. Researching candidate sites for a carbon storage complex in the Central Appalachian Basin, USA. Int. J. Greenh. Gas Control 88, 168–181 (2019).

Alcalde, J. et al. Estimating geological CO2 storage security to deliver on climate mitigation. Nat. Commun. 9, 2201 (2018).

Article Google Scholar

Download references

Department of Petroleum and Geosystems Engineering, The University of Texas at Austin, Austin, TX, USA

Valeria Vallejo, Quoc Nguyen & Arvind P. Ravikumar

Energy Emissions Modeling and Data Lab, The University of Texas at Austin, Austin, TX, USA

Arvind P. Ravikumar

Center for Energy and Environmental Systems Analysis, The University of Texas at Austin, Austin, TX, USA

Arvind P. Ravikumar

You can also search for this author in PubMed Google Scholar

You can also search for this author in PubMed Google Scholar

You can also search for this author in PubMed Google Scholar

A.P.R. conceived the study. A.P.R. and Q.N. designed and coordinated study elements. V.V. created the life-cycle assessment model, performed the analysis, created the figures and tables and wrote the paper. V.V., Q.N. and A.P.R. discussed the results and contributed to the writing.

Correspondence to Arvind P. Ravikumar.

A.P.R. is currently a member of the Gas Pipeline Advisory Committee of the US Department of Transportation; in this role, he is a Special Government Employee. A.P.R.’s research is currently supported by the US Department of Energy, Environmental Defense Fund and sponsors of the Energy Emissions Modeling and Data Lab (EEMDL). The other authors declare no competing interests.

Nature Energy thanks Kevin Dillman, Wilson Ricks and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Tables 1–43, Figs. 1–4 and Methods.

Workbook describing all calculations and modelling presented in the study.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

Vallejo, V., Nguyen, Q. & Ravikumar, A.P. Geospatial variation in carbon accounting of hydrogen production and implications for the US Inflation Reduction Act. Nat Energy (2024). https://doi.org/10.1038/s41560-024-01653-0

Download citation

Received: 25 November 2023

Accepted: 18 September 2024

Published: 14 October 2024

DOI: https://doi.org/10.1038/s41560-024-01653-0

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative