Carbon footprint of global natural gas supplies to China

China has historically relied on coal as its largest energy source, but recent air pollution concerns have driven strong policies to promote coal-to-gas switching, making natural gas the fastest-growing fossil fuel in China. Domestic conventional gas production is insufficient, prompting development of domestic unconventional resources (especially shale gas) and rapid growth in imports via international pipelines and LNG. However, the climate benefits of switching from coal to natural gas depend on natural gas life-cycle emissions, which vary substantially, particularly in upstream stages (well-to-city-gate). Emissions differ across countries, extraction techniques, and even among fields due to geology, gas composition, and transport distances. The study aims to quantify well-to-city-gate GHG intensities of natural gas supplies to China at the field level for 104 fields in 15 countries, construct supply curves for 2016 and 2030, and identify key drivers and implications for emissions reduction and energy policy.

Prior work indicates that 20–50% of natural gas life-cycle emissions arise upstream (well-to-city-gate), with combustion emissions being relatively constant. Studies by NETL and others have modeled extraction, processing, and transmission, while GREET has been used widely for fuel life-cycle assessment, including LNG stages. Literature highlights variability in methane leakage, the influence of gas composition (e.g., CO₂ and H₂S) on processing emissions, and the importance of transport distances and liquefaction/shipping conditions. Earlier comparisons of coal vs. gas emphasize uncertainty in upstream emissions. There is a gap in field-level, engineering-based assessments tailored to China’s diversified supply mix; this work addresses that gap by using field-specific parameters across domestic and imported supplies to China.

The study compiles a well-to-city-gate life-cycle assessment (LCA) model for natural gas supplies to China by integrating: (1) NETL models for extraction, processing, and transmission; (2) Argonne’s GREET model for LNG storage, LNG shipping, and offsite fuel and electricity generation; and (3) literature-based models for liquefaction and regasification. The system boundary includes energy inputs production and the following processes: extraction (well drilling, completion, liquids unloading, workovers, gathering, flaring/venting), gas processing (amine treatment, dehydration, NGL separation, compression, flaring/venting), pipeline transmission (compression energy and leakage), and for LNG: liquefaction (refrigeration, turbines, leakage, heating/boilers, flaring/venting), LNG storage and ocean shipping (boil-off, marine fuel), regasification, and downstream transmission to the city-gate. Local distribution beyond city-gate is excluded. Field-specific inputs include country origin, extraction characteristics (EUR, well depth, production rates, drilling/completion type including horizontal drilling and hydraulic fracturing), raw gas composition (e.g., CO₂, H₂S), pipeline age and conditions, transmission distances from fields to domestic consumers or to LNG plants and from Chinese LNG terminals to city-gate (assumed common across LNG supplies once in China), liquefaction plant data, ambient temperatures (affecting liquefaction efficiency), and shipping distances to China. Transmission distances for pipeline gas were estimated by mapping fields to provincial demand centers and weighting distances by demand shares; for Xinjiang, distinct north/south destinations were used. Electricity grid mixes for producer countries were modeled in GREET using IEA shares of generation sources. Emissions are reported primarily in 100-year GWP (GWP100) with GWP20 presented in the Supplementary Information. Uncertainties in key inputs (e.g., pipeline leakage rates and other process parameters) were propagated via Monte Carlo simulation to produce 90% confidence intervals. Supply-energy-weighted averages and empirical distribution functions were constructed for 2016 and projected to 2030 using documented contracts, production statistics, and projections for field-level supply shares.

• Wide range of well-to-city-gate GHG intensities across 104 fields supplying China: 6.2–43.3 g CO₂eq MJ⁻¹ (GWP100).
• Category averages (GWP100): domestic conventional gas 15.5 g CO₂eq MJ⁻¹ (largest within-category variability), domestic unconventional gas 21.4 g CO₂eq MJ⁻¹, international pipeline gas 35.9 g CO₂eq MJ⁻¹ (highest), overseas LNG 19.7 g CO₂eq MJ⁻¹.
• 2016 supply-energy-weighted average GHG intensity: 21.7 g CO₂eq MJ⁻¹; percentile values: 5th 6.6, 25th 17.2, median 19.4, 75th 28.7, 95th 41.5 g CO₂eq MJ⁻¹.
• 2030 projected supply-energy-weighted average increases to 23.3 g CO₂eq MJ⁻¹ due to higher shares of GHG-intensive supplies (e.g., Russia’s Urengoi and Nadym, Turkmenistan’s Galkynysh and Bagtiyarlyk, and domestic shale gas such as Fuling).
• Total well-to-city-gate emissions are estimated to grow by roughly threefold from 2016 to 2030 as demand rises.
• Drivers of variability: • Transmission distance strongly influences pipeline emissions; Western China fields serving eastern coastal demand incur high transmission emissions (e.g., West–East pipeline ~4000 km). • Processing emissions increase with higher impurity contents, especially high CO₂ and H₂S (e.g., Dongfang, Ledong >20 vol% CO₂; Gorgon ~15 vol% CO₂ with highest processing emissions ~12.5 g CO₂eq MJ⁻¹; CCS delays increased emissions). • Extraction emissions depend on technology (conventional vs. horizontal drilling and hydraulic fracturing), EUR, initial production rates, and well depth (e.g., Qatargas North Field low extraction ~3.9 g CO₂eq MJ⁻¹ due to high EUR; some US shale sources high extraction up to ~19.7 g CO₂eq MJ⁻¹ due to higher fugitive emissions and low EUR). • Liquefaction emissions vary with ambient temperature (e.g., Norway’s Snohvit has low liquefaction emissions 4.1–7.6 g CO₂eq MJ⁻¹); shipping emissions scale with voyage distance (e.g., US Sabine Pass ~18,000 km to China).
• Major 2016 contributors: Sulige (11% of supply; 17.3 g CO₂eq MJ⁻¹), Galkynysh (41.5 g CO₂eq MJ⁻¹) and Bagtiyarlyk (36.9 g CO₂eq MJ⁻¹) together 14%, Anyue-longwangmiao (6.6 g CO₂eq MJ⁻¹; 4%), Puguang (~median+; 4%), Jingbian (22.1 g CO₂eq MJ⁻¹; 4%), Kela (22.5 g CO₂eq MJ⁻¹; 3%), Fuling (28.7 g CO₂eq MJ⁻¹; 3%), Australia Pacific LNG (18.4 g CO₂eq MJ⁻¹; 4%), Qatargas LNG (17.2 g CO₂eq MJ⁻¹; 4%).
• Policy-relevant scenario results for coal-to-gas switching (2016–2030): assuming incremental gas displaces coal in power generation and total power output is held constant across scenarios: • Mean gas supply intensity yields estimated emission reduction benefits of 7.4 Gt CO₂eq (GWP100) and 7.8 Gt CO₂eq (GWP20). • High-intensity gas (80th percentile) reduces those benefits by 1.3 Gt CO₂eq (GWP100) and 2.8 Gt CO₂eq (GWP20). • Low-intensity gas (20th percentile) increases benefits by 12% (GWP100) and 34% (GWP20).

Field-level, engineering-based LCA reveals substantial variability in well-to-city-gate GHG intensities across China’s diverse natural gas supply portfolio. These differences materially affect the net climate benefits of coal-to-gas switching. The analysis highlights actionable levers: prioritize lower-intensity sources (e.g., certain LNG from Asia-Pacific and Middle East for coastal demand centers), optimize pipeline routing and network operation to reduce transmission distances, monitor and mitigate methane leakage in transmission systems, and deploy CCS where raw gas contains exceptionally high CO₂. Strategic allocation of supply within China (e.g., favoring LNG for coastal loads and avoiding very long pipeline transports from Western Russia and Central Asia to the east) can reduce overall intensity. The findings also raise questions about the climate efficacy of aggressive domestic shale gas expansion given its relatively higher upstream intensities compared with some LNG options. Incorporating GHG performance in supply contracts and domestic development plans can support China’s broader emissions reduction commitments.

This study provides the first comprehensive, field-level well-to-city-gate GHG intensity assessment of natural gas supplies to China, constructing supply curves for 2016 and 2030. It quantifies large variability (6.2–43.3 g CO₂eq MJ⁻¹) and shows a projected rise in the supply-weighted average intensity from 21.7 to 23.3 g CO₂eq MJ⁻¹ by 2030 driven by growth in higher-intensity sources. These results enable climate-informed procurement, infrastructure planning, and mitigation prioritization (e.g., leakage control, CCS at high-CO₂ fields, and regional allocation of supplies). Future research should refine cost and pricing analyses, improve regional demand resolution, enhance empirical data on leakage and field operations, and evaluate policy mechanisms to incentivize low-intensity gas supply chains.

• System boundary excludes local distribution beyond city-gate; results focus on well-to-city-gate stages only.
• Significant uncertainties remain in key parameters (e.g., pipeline leakage rates, field EURs, operational practices). Uncertainty is addressed via Monte Carlo simulations but may still affect precision.
• Transmission distance estimates rely on provincial demand aggregation and available pipeline disclosures; limited sub-provincial demand data may introduce approximation errors.
• Projections for 2020/2030 depend on evolving energy policies, trade contracts, and field development status; actual supply shares may differ.
• Some input data gaps were filled using commercial datasets and literature; CCS deployment timelines (e.g., Gorgon) and operational changes can materially alter future intensities.
• Assumed common post-regasification transmission within China for LNG supplies may mask regional differences in actual last-mile transmission.