Measurement-based carbon intensity of US offshore oil and gas production

The United States (US) produces oil and gas from six offshore regions: the North Slope of Alaska, Cook Inlet in Alaska, offshore California, and three Gulf of Mexico (GOM) sub-regions: state shallow, federal shallow, and deep waters. Measurement-based assessment of direct greenhouse gas emissions from this production can provide real-world information on carbon emissions to inform decisions on current and future production. In evaluating the climate impact of production, the carbon intensity (CI, the ratio of greenhouse gases emitted compared to the energy of fuels produced) is often used, though it is rarely quantified with measurements. Here, we complete an observational evaluation of the US offshore sector and present the largest current set of measurement-based CIs. We collected airborne measurements of methane, carbon dioxide, and nitrogen oxides from the North Slope, Cook Inlet, and California and combined with prior GOM results. For Alaska and California, we found emissions agree with facility-level inventories, however, the inventories miss some facilities. The US offshore CI, on a 100 year GWP basis, is 5.7 g CO2e/MJ[4.5, 6.8, 95% confidence interval]. This is greater than double the CI based on the national US inventory, with the discrepancy attributed primarily to methane emissions from GOM shallow waters, with a methane dominated CI of 16[12, 22] for GOM federal shallow waters and 43[25–65] for state shallow waters. Regional intensities vary, with carbon dioxide emissions largely responsible for CI on the North Slope 11[7.5, 15], in Cook Inlet 22[13, 34], offshore California 7.2[3.2, 13], and in GOM deep waters 1.1[1.0, 1.1]. These observations indicate offshore operations outside of the GOM in the US have modest methane emissions, but the CI can still be elevated due to direct carbon dioxide emissions. Accurate assessment of different offshore basins, with differing characteristics and practices, is important for the climate considerations of expanded production.


Introduction 1.The need for measurements of emissions from offshore oil and gas production
The production of oil and gas emits carbon dioxide (CO 2 ) and methane (CH 4 ), the two most prevalent pollutants driving global warming [1].Emissions from production represent a large fraction of the total greenhouse gas footprint of the fossil fuel sector [2][3][4].While future energy use is projected to rely considerably on fossil fuels well into 2100 [5], near term emission reductions can be achieved through reducing the overall climate impact of fossil fuels [6].In part, this can be realized through emission reductions at the production phase of the oil and gas supply chain.To this end, national programs seek to identify emission sources [7] and investors seek knowledge of oil and gas company carbon emissions [8,9].Identifying effective mitigation strategies, however, requires an accurate understanding of realworld emissions.Alvarez et al synthesized a multitude of measurement studies to update our understanding of CH 4 emissions from the United States (US) oil and gas supply chain, finding that 2015 CH 4 emissions were 60% higher than the Environmental Protection Agency value, largely due to elevated emissions from production operations [4].
The CH 4 synthesis by Alvarez et al did not update emission estimates for the offshore segment of the US oil and gas supply chain [4], due to a lack of measurements.Since then, new measurements of CH 4 from offshore facilities have become available for various offshore provinces, including the Gulf of Mexico (GOM) [10][11][12][13][14][15][16][17], offshore Canada [18], the North Sea [19][20][21], the Norwegian Sea [22], and Southeast Asia [23].The GOM is the largest offshore oil and gas basin in the US and has recently been the subject of multiple field studies measuring CH 4 emissions [10][11][12][13][14]17]. Gorchov Negron et al synthesized most of these recent studies and combined them with additional CH 4 and CO 2 measurements to develop a measurement-based carbon intensity (CI) of GOM oil and gas production [10].CI (in grams of CO 2 equivalent of CH 4 and CO 2 emissions divided by joules of oil and gas produced) is an indicator of the relative climate impact of an oil and gas field.It is a more complete metric compared to a gas loss rate since it includes both major greenhouse gas pollutants from oil and gas activities and attributes them to both oil and gas production, which are usually co-located in the production phase.This synthesis found that the CI differs across the three sub-basins within the US GOM: state shallow waters (Louisiana, Alabama, and Texas), federal shallow waters, and deep waters [10].In deep waters, the CI was relatively low at large high throughput facilities [10].In contrast, the CI in shallow waters was higher and attributed to underestimated CH 4 emissions from older infrastructure that was composed of multi-platform central hubs servicing small satellite facilities [10].Differences between sub-basins in the GOM raises the question as to whether these findings are representative of other US offshore basins.

Un-sampled US offshore fields
The carbon emissions from US offshore fields outside of the GOM have not been previously surveyed.There are three additional US offshore fields: offshore the North Slope of Alaska, in Cook Inlet in southern Alaska, and offshore of California.Most US offshore fields are mature: the largest ramp up in production began in ∼1955 in Louisiana state shallow waters [24], ∼1960 for GOM federal shallow waters [25], ∼1967 in Cook Inlet [26], before 1977 in state waters offshore California [27], ∼1980 in federal waters offshore California [28], ∼1988 offshore the North Slope [26], and ∼1994 for GOM deep waters [25].However, these fields differ in characteristics to the GOM and therefore present an opportunity to explore how CH 4 emissions and CIs vary in fields with different infrastructure, practices, and history.

Offshore the North Slope, Alaska
Offshore North Slope production is concentrated in offshore and coastal state water facilities, most of which have had declining production for decades.Facilities are built to endure the harsh high latitude environment and encompass a diverse range of infrastructure: artificial islands, a causeway, and coastal facilities servicing offshore wells.There is substantial gas production on the North Slope, but gas is used either on-site or re-injected because there is no infrastructure to transport gas south; only oil is transported south by the Trans-Alaska pipeline (figure 1).

Cook Inlet, Alaska
Platforms in Cook Inlet are located in state waters with moderate production rates of oil and gas and are more traditional infrastructure compared to the North Slope.While production has declined for years, a number of platforms have been replaced and/or ramped up production as recent efforts re-opened shut-in wells and drilled new wells [29,30].Five nearby onshore facilities either transport gas fuel to these platforms, process, and/or treat the hydrocarbons.Three of these are directly connected to offshore platforms by pipeline.

Offshore California
Offshore of California, platforms are located on the northern Santa Barbara Channel and the southern San Pedro Shelf.The regulatory environment for oil and gas within California is notably stringent [31].The offshore industry has experienced long periods of moratoriums on new production in response to oil spills for both state [32] and federal waters [33] with no new leases since 1982 [34].The platforms are old facilities with moderate production rates and many have ceased production or are on track for decommissioning [35,36].Given this and current trends [37], it is unlikely there will be new offshore wells in the near-term.Most active facilities are located in federal waters, while most currently active state water facilities are comprised of three artificial islands on the San Pedro Shelf.

The role of CI in current and future expansions in production
Updating our understanding of US offshore emissions, including how emissions differ between regions, is important at this time.US offshore production is under consideration for expanded development given recent and planned lease sales in Cook Inlet [38,39] and the GOM [38,40,41], and a proposed liquefied natural gas (LNG) pipeline to transport stranded gas out of the North Slope [42][43][44].Expanded development is partly contingent on environmental impact statements (EISs), which for these offshore regions contain an evaluation of the CI of new production.Specifically, the EISs compare the CI of fossil fuel use from two scenarios: one with expanded production and another where production demands are met from fuels extracted elsewhere [45,46].These CIs carry legal weight in court proceedings [47], however, these EISs have yet to incorporate a measurement-based CI, partly for a lack of observations.

This work
In this study, we expand upon the work of Gorchov Negron et al in the GOM with new field measurements in the remaining US offshore fields to complete our survey of US offshore production.Based on new airborne measurements of CH 4 , CO 2 , and nitrogen oxides (NOx = NO + NO 2 ), offshore the North Slope of Alaska, in Cook Inlet in southern Alaska, and offshore of California, we estimated emissions at the facility-level and regional-level and compared with inventories.We further calculated measurement-based CIs of production activities for each field and the total US offshore sector, which could be incorporated into future synthesis studies of the oil and gas supply chain at the national scale.Here, we present the first measurement-based assessment of the CI of US offshore oil and gas production and describe in detail how and why emissions vary.This is currently the largest set of measurementbased CIs and we discuss what these imply about how CIs may vary for oil and gas production in the real world.Finally, we summarize the key implications of this assessment for the state of emissions from US offshore production and speculate how the climate impact may change in response to expanded development.

Field campaign and sites sampled
We collected new data from almost all infrastructure related to offshore production in Alaska and California through an airborne campaign.This project is under the Flaring and Fossil Fuels: Uncovering Emissions and Losses (F 3 UEL) effort, which has produced multiple studies [10,17,48,49] intended to improve knowledge of emissions from offshore Photos are chosen to represent the diversity of infrastructure sampled, with photos of facilities of a similar nature excluded.On the North Slope, we sampled artificial islands (A), coastal facilities with nearby offshore subsea wells (B), a gathering and gas re-injection facility that processed both offshore and onshore hydrocarbons (C), and a causeway with an artificial island connected by road to the mainland (D).Most if not all facilities had active flares.The map of facilities in figure 1(C) differentiates facilities of each category of North Slope infrastructure as shown here.In Cook Inlet, we sampled multiple offshore platforms (E), which appeared to have either a vent or flare boom (F).We also sampled two of the three onshore facilities connected by pipeline to offshore platforms that served a processing and/or fuel transport service (G).In California, we sampled both offshore platforms on the southern San Pedro Shelf, that included a multiplatform facility with a processing capability (H), and in the northern Santa Barbara Channel (I).For photos of platform categories in the Gulf of Mexico, see Gorchov Negron et al [10].Reproduced from [10].CC BY 40. and flaring activities (see https://graham.umich.edu/f3uel).
We measured CH 4 , CO 2 , and NO x atmospheric concentrations and ambient weather conditions (temperature, relative humidity, wind speed and direction) from a Mooney aircraft operated by Scientific Aviation.Included onboard was a FLIR infrared camera for visualization of hydrocarbon sources.For details on a similar payload, see Gorchov Negron et al [10].We used these measurements to estimate 70 fluxes (see next section for method) from 37 sites (with some repeat sampling), encompassing 41 unique facilities (with those in close proximity sampled together).We flew during July 2021 in Alaska and during August 2021 in California.We chose these months to minimize weather interference (e.g.fog and storms) and attempt to have optimal atmospheric mixing conditions where surface waters were expected to be warmer than the atmosphere.
A map and photos of facilities sampled are shown in figures 1 and 2. In the North Slope, we sampled all facilities tied to offshore production at least once over four sample days (10 July 2021-15 July 2021).We sampled nearly all facilities two to three times to discern daily variability.The easternmost facility sampled is the recently constructed Point Thomson, which, with a few expansions, is projected to produce 25% of future natural gas for the proposed LNG pipeline [50].We also sampled a central gas facility and compression plant, which is connected by pipeline to offshore facilities.At this site, operations are focused on gathering and separating oil and natural gas liquids and reinjecting gas [51].
In Cook Inlet, we sampled all active platforms over the course of three sample days (17 July 2021-22 July 2021).In addition, we sampled two of three onshore facilities that had pipelines tied to these platforms.The first of these is the Trading Bay Production Facility, which processes hydrocarbons from multiple platforms and sends gas fuel back to at least one [52].The second is the Granite Point Production Facility and Tank Farm, which appears less active and still reportedly handles waste water and transports gas back at least one facility [48].Low wind speeds and weather limited totals days flown, limiting repeat sampling to three sites.
In California, we sampled platforms over the course of six sample days (10 August 2021-15 August 2021) in mostly federal waters.We targeted all producing federal water facilities and incidentally sampled some facilities that appear to have ceased production operations.Our samples include the only official processing facility, which services two nearby production facilities.We did not sample state water artificial islands on the San Pedro Shelf near Long Beach.These locations were not safe for sampling as they were too close to either the mainland or boats.Also, we were unable to sample nearby onshore facilities that we suspected had a processing function for platforms due to steep terrain in close proximity to the site.

Quantification of facility and regional emissions
We quantified facility-level emissions using gauss's theorem applied to a cylindrical flight pattern [53] (SI appendix S1).Next, we combined facilitylevel emissions for only sites directly involved in offshore production to estimate regional totals.These totals represent upstream production and onsite processing emissions and exclude mid-stream onshore facilities and facilities not sampled.
We summed emissions using a boot-strapping approach that considers both the uncertainty for a given facility-level emission flux and any repeat measurements made for same facility on a separate day.For each iteration, the approach randomly chose a daily flux mean and error for each facility, with replacement, and assigned a flux rate from the normal distribution created from that mean and error (treated as a standard deviation).We summed all facilities in a region and repeated the simulation 1000 times to generate a distribution of total emissions.From this distribution, we calculated a mean and 95% confidence interval, reported in table S1.

Inventories
We compared emissions at the facility-level and at the regional field-level with the most recent inventories available at the time of this investigation.CH 4 and CO 2 emissions are compared against the 2020 Greenhouse Gas Reporting Program (GHGRP) Facility Level Inventory GreenHouse gases Tool (FLIGHT) [54], which contains emitters considered to be at least 25 000 mt CO 2 eq/y.The GHGRP contains most of the facilities in Alaska, but does not contain reports for offshore California.NO x emissions are compared against the National Emissions Inventory (NEI) 2017 annual spatial emissions [55].
We also compared regional CH 4 emissions with the Environmental Protection Agency Inventory of Greenhouse Gas Emissions and Sinks for 1990-2019 (GHGI) [56][57][58].Offshore emissions for Alaska and California waters are relatively new categories and are developed using production-based emission factors [57,58].To ensure our comparison was fair, we recalculated the GHGI emission rates using our own estimated production data for the facilities sampled in this study multiplied by GHGI emissions factors (table S2).This also allowed us to calculate separate GHGI CH 4 emissions for offshore North Slope and Cook Inlet.We did not treat re-injected gas as production in this calculation.

Production
We gathered production data from Enverus [59], which is a proprietary database that reports monthly production at the well-level for oil (in barrels per month) and gas (in 1000 standard cubic feet per month).We used production data for the months we sampled: July in Alaska and August in California.We linked well-level production to platforms based on the proximity of wells to platforms, pipeline information where present, and maps of production blocks.Table S2 compares production data gathered from Enverus to the original GHGI.Next, we calculated produced energy content (Joules) of both oil and gas using the equations outlined by Gorchov Negron et al [10].We further separated gas into gas that is marketed or used on-site vs re-injected.See SI appendix S2 to see how we partitioned the re-injected gas fraction.

Emissions offshore Alaska and California
Measurements of facility-level emissions are shown in figure 3 and regional-level emissions are discussed in SI appendix S3, reported in table S1, and shown in figures S1-S3.Emissions of all gases are generally higher in the North Slope, closely followed by Cook Inlet, and relatively low offshore California.Broadly, emissions are comparable to rates detected in other offshore basins.Below, we contextualize these emissions against previous observations in other offshore fields and compare with inventories when available.The combination of these measurements from Alaska and California with previous work in the GOM, completes the first observational assessment of carbon emissions from US offshore oil and gas production.
Most rates of CH 4 emissions from producing facilities range from ∼0-100 kg hr −1 with most sites exhibiting relatively low inter-daily variability where measured.This is close to the range detected in the North Sea [19,21], Norwegian Sea [22], offshore Borneo [23], and deep water GOM [10,13,14], and lower than the 100s-1000s kg hr −1 detected offshore Malay Peninsula [23] and some shallow water facilities in the GOM [10][11][12].However, on the North Slope, the gathering site and the recently constructed Point Thomson central pad site had emissions that reached up to 300 kg hr −1 .Although we were unable to sample the three artificial islands offshore California, one of these facilities was sampled by Carbon Mapper for a total of 54 ± 15 kg CH 4 hr −1 [60].We are unable to identify CH 4 sources on the facilities with the FLIR camera, because either emission rates were below detection or temperature contrasts between the background and emissions were insufficient to be resolved.
Substantial CO 2 emissions occur in Alaskan waters, which are comparable to deep water GOM CO 2 emissions of ∼5000-25 000 kg hr −1 per facility [10].Two North Slope facilities with large natural gas production and re-injection rates emit considerable emissions close to 40 000 kg hr −1 (Endicott main production island) and 90 000 kg hr −1 (North Star artificial island).The GHGRP attributes most of these combustion emissions to stationary combustion as opposed to petroleum and natural gas systems, which may indicate that emissions are driven by on-site fuel combustion for energy instead of flaring of excess gas.California CO 2 emissions are comparatively low, but not negligible, and compare with rates observed in GOM shallow waters [10].NO x emissions from producing facilities were generally low, but ranged from 0-130 kg hr −1 .This range is similar to the range previously quantified in the GOM [10], but with higher maximum values.
Inventory emission rates, when reported, tend to be similar to observed emissions (figures 3 and S1-S3).Both the GHGRP CO 2 and CH 4 and NEI NO x emission rates are similar to at least one daily emission rate for each facility.There is at least one instance of a major emitter on both the North Slope and in Cook Inlet that is not reported in the GHGRP.These missing sites explain most of the gap in total emissions between observations and inventories.While there are no emissions reported in the GHGRP for California, these sites appear to be relatively small sources.The GHGI CH 4 updated with production data performs fairly well in Cook Inlet and California and overall agrees with observations in that it expects low emission rates (figure S1).Overall, we find relatively low CH 4 emissions, which is expected by inventories, and evidence of considerable combustion, which is generally consist with inventories for both CO 2 and NO x , suggesting that (1) facilities were operating under standard conditions and (2) CO 2 emissions will likely have a greater impact on climate than CH 4 emissions from these fields.

Measurement-based CI of US offshore production
Figure 4 presents the 100 year CI, emissions, and production for regional and total US offshore production.Observations (O) of measurement-based CI are compared against an equivalent inventory-based (I) estimate.Table S3 reports values, with mean and 95% confidence intervals, and greenhouse gas percentage contributions over 100 yr and 20 yr horizons with CH 4 global warming potentials of 28 and 84 g CO 2 eq/g CH 4 respectfully [61].These values represent a snapshot of ongoing production operations for 2021.Onshore processing or gathering facilities are excluded.The inventory CI is calculated from the GHGRP for only production facilities where emission information is reported.In the sections below, we (1) present the full US offshore CI, compare with the inventory, and explain where the majority of emissions are coming from, (2) show the range of regional CIs and describe the role that CO 2 and CH 4 can play, including on a 20 year horizon, (3) clarify how to interpret gaps between measurement-based versus inventory-based CIs at the field-level, and (4) compare our results with EIS CIs.
The CI of 2021 offshore production is 5.7 g CO 2 eq/MJ, double the GHGRP inventory-based intensity of 2.4 g CO 2 e/MJ.The discrepancy is due to elevated CH 4 emissions in GOM shallow waters, which have been previously attributed to emissions from old central hub facilities and small satellite facilities [10].Various emission sources have been identified at these sites, with particular emphasis on direct venting of CH 4 from vent booms [10,11,17].For more details see Ayasse et al [11], Gorchov Negron et al [10] and Biener et al [17].The offshore CI is lower than the estimated average CI of US oil production of ∼11 g CO 2 eq/MJ [9-17, 95% confidence interval] [3].Our update estimates that CH 4 contributes to 66% of the total CI, placing CH 4 as the dominant greenhouse gas, which is in contrast to the inventory breakdown (table S3).But, CO 2 emissions from on-site fuel use and flaring are still important.Production and emissions are driven by different fields with 80% of production attributed to GOM deep waters, while 80% of emissions are attributed to outside of GOM deep waters.
Measurement-based offshore CIs vary widely across fields (figure 4 and table S3).The CI of GOM deep waters (1.1 g CO 2 eq/MJ), offshore California (7.2 g CO 2 eq/MJ), offshore the North Slope (11 g CO 2 eq/MJ), and Cook Inlet (22 g CO 2 eq/MJ) are all primarily attributed to CO 2 emissions.However, the GOM federal shallow waters (16 g CO 2 eq/MJ) and state shallow waters (43 g CO 2 eq/MJ) are primarily attributed to CH 4 emissions.The variable role of different greenhouse gases is exemplified by comparing the CO 2 :CH 4 percentage contribution to CI on the North Slope (96%:4%) versus GOM state shallow waters (12%:88%).Over 20 years, CH 4 has a higher warming potential.This triples the CI of fields dominated by CH 4 , while the intensity of fields dominated by CO 2 show little increase (table S2).However, the most impactful greenhouse gas species at the fieldlevel remains the same as over 100 years (table S3).
Differences with inventory-based intensities reflect a real gap in the GOM shallow waters [10], however for Alaskan fields they reflect the sensitivity to the exclusion of individual facilities.The GHGRP in Alaska does not report emissions at some facilities that either have high production or high emissions (figure 3), which drastically change the ratio of emissions to production, thereby driving apparent gaps.However, inventories still expect considerably high CIs from CO 2 alone, in agreement with our observations.
The measurement-based CIs of current operations are higher than the CI estimated by the EIS for expanded production (SI appendix S3).This is expected because the EISs estimate a full lifecycle CI, which should be lower due to the incorporation of high production rates in the early stages of development.The variation of CI from older fields in this study (7-43 g CO 2 eq/MJ, excluding the deep GOM) suggests that different practices, regulations, and maintenance over the course of a field's life can have a strong influence on the climate impact as a field matures.This is particularly true for CH 4 -dominated emissions profiles, where CH 4 super-emitters, which are hard to predict, can drive high emissions intensities.These EISs should expect increases in CI as any new developments mature, but management and infrastructure choices could reduce future emissions from aging infrastructure.

Implications for CI
We present a comparison of measurement-based CIs for six unique offshore fields, which is the largest set currently available.This presents an opportunity to discuss how CIs vary in the real world.Most CIs are typically estimated from bottom-up methods [2,3,62], without field-specific observations, emphasize the role of CO 2 [3,62], and usually separate gas [2] or oil [2,3].There is almost no literature on CIs estimated from measurements, as defined here.Besides work in the GOM [10], there is only one somewhat similar study [63] with all other work estimating intensities for only CH 4 [18,64] or CO 2 [65].The current understanding expects that CO 2 emissions intensities increase with field maturity [62,66,67], because as fields age, field pressure declines and wells tend to produce more water.Therefore the per unit energy (i.e.combustion) used to support greater lift or injection methods of extraction and process hydrocarbons may increase [67].Gas handling is expected to be an important driver of CI with high intensities expected in regions with high flaring and high energy demands for gas reinjection [3].If gas is vented instead of flared, CIs could be further elevated by up to an additional 13 g CO 2 eq/MJ [3].
We find that measurement-based CI varies widely in magnitude (1-43 g CO 2 eq/MJ), comparable to the modeled range of ∼3-50 g CO 2 eq/MJ found by Masnadi et al [3] for oil production.Both CO 2 and CH 4 are important and each can individually drive high CI from offshore production.For example, CO 2 alone is responsible for 20 g CO 2 eq/MJ in Cook Inlet and CH 4 alone is responsible for 38 g CO 2 eq/MJ in GOM state waters.Observations onshore have also found high CO 2 intensities of ∼240 kg CO 2 eq/bbl or ∼39 g CO 2 eq/MJ of oil in Canadian oil sands [65] and high CH 4 intensities up to ∼32 g CO 2 eq/MJ in Canadian oil and gas fields [64].
It is possible that variation in CIs reflects differences in field maturity and gas handling.The range of modeled CIs is primarily driven by varying CO 2 emissions intensities [3].In this work, we observe the lowest CI in the recently developed GOM deep waters and higher intensities in the older CO 2 dominant fields despite visibly similar infrastructure in Cook Inlet and highly centralized weather-resilient infrastructure on the North Slope.Gas handling varies widely between extremes of re-injection on the North Slope, flaring in the deep GOM, and high venting rates in the shallow GOM.Re-injection is expected to be an energy-intensive process, and likely explains the moderately high CI on the North Slope.The elevated CI found in the state and federal shallow GOM supports the expectation that venting can further increase CI, as presented by Masnadi et al.

Implications for emissions from US offshore production
Overall, US offshore production is primarily driven by the deep GOM, but other fields with marginal production contribute substantial emissions.CO 2 emissions are an important contribution to the overall climate impact and well characterized where reported in inventories.CH 4 emissions are also important to the overall climate impact and are elevated due to high emissions associated with older infrastructure in the shallow water GOM.Old infrastructure is expected to be a more probable CH 4 source [68], but while much of the US offshore industry is old, CH 4 emissions tend to be low, except at a specific pattern of infrastructure: multi-platform central hubs serviced by small satellite platforms.This suggests that the US offshore industry is capable of maintaining low CH 4 emissions at old infrastructure, and high CH 4 emissions may be related to operator practices or maintenance.This trend appears consistent with the trends outside of the US, where (a) instances of high CH 4 emissions of order of 500-1000 kg CH 4 hr −1 associated with plumes in Southeast Asia [23], a site in the North Sea [21], and an ultra emission event in the Mexican GOM [16] were detected at facilities or in fields that contain central hubs and satellite facilities and (b) lower emissions of order 10-100 kg CH 4 hr −1 were detected at facilities with different infrastructure in the North Sea [19][20][21] and the Norwegian Sea [22].Despite generally low CH 4 emissions outside of the GOM, CO 2 emissions can still drive a relatively high CI.
Expanded production will likely change offshore CI.From a strictly CO 2 intensity perspective, we may presume, based on trends presented here and modeling in the literature [3,62,66,67], that the ratio of emissions to production will be low for new sites.However, the total CI may be impacted by how natural gas is handled.For the North Slope, if (a) production expands and gas is marketed instead of reinjected and (b) operations implement low fugitive and vented gas rates, then we may speculate that the increased production and lower energy needs for reinjection will result in a lower CI.Similarly, newly constructed sites in the GOM must avoid the patterns of considerable gas losses in shallow waters [10] to achieve a low CI.As the US enters what may be a major expansion in offshore production, consideration of these factors may help maintain low carbon intensities in the coming decades.

Figure 1 .
Figure 1.(A) The four offshore production regions in the US.The currently existing Trans-Alaska Crude Oil Pipeline and proposed Alaska LNG pipeline are shown.(B) All likely active platforms offshore and in wetlands located in the US Gulf of Mexico separated by water depth and state jurisdiction.Platforms sampled by atmospheric measurement campaigns before 2021, as reported by Gorchov Negron et al [10], are shown with black borders.(C)-(E) Oil and gas infrastructure related to offshore operations on the North Slope Alaska, Cook Inlet Alaska, and offshore California.Airborne flight tracks from the 2021 F 3 UEL campaign are shown in black lines.Red numbers in panels C-E identify individual sites sampled and correspond to site numbers in figure 3, which link to fluxes.

Figure 2 .
Figure 2. Photos of facility types sampled by the 2021 F3 UEL campaign.Photos are chosen to represent the diversity of infrastructure sampled, with photos of facilities of a similar nature excluded.On the North Slope, we sampled artificial islands (A), coastal facilities with nearby offshore subsea wells (B), a gathering and gas re-injection facility that processed both offshore and onshore hydrocarbons (C), and a causeway with an artificial island connected by road to the mainland (D).Most if not all facilities had active flares.The map of facilities in figure1(C) differentiates facilities of each category of North Slope infrastructure as shown here.In Cook Inlet, we sampled multiple offshore platforms (E), which appeared to have either a vent or flare boom (F).We also sampled two of the three onshore facilities connected by pipeline to offshore platforms that served a processing and/or fuel transport service (G).In California, we sampled both offshore platforms on the southern San Pedro Shelf, that included a multiplatform facility with a processing capability (H), and in the northern Santa Barbara Channel (I).For photos of platform categories in the Gulf of Mexico, see Gorchov Negron et al[10].Reproduced from[10].CC BY 40.

Figure 3 .
Figure 3. Facility-level emissions and production for offshore related infrastructure on the North Slope (left), Cook Inlet (middle), and off the coast of California (right).Aircraft quantifications are filled in black if the flux is statistically different from zero (the error bars do not include zero) and are not filled in and shown as empty if the error bars overlap with zero.Fluxes where the mean fell below zero, regardless of the size of error bars, are considered and shown as zero emissions.Fluxes are unique to each day and connected by a solid line to indicate repeat samples for the same site.Site names correspond to official facility names.The number in front of each site name corresponds to the red numbers mapped out in figure 1, thereby linking these fluxes to latitude and longitude locations.Inventories tend to agree with observations.

Figure 4 .
Figure 4. US offshore and regional 2021 carbon intensity over a 100 year horizon (top), emissions (middle), and production (bottom).Observations (O) of measurement-based CI are compared against the GHGRP inventory-based estimate (I).Gulf of Mexico values are taken from Gorchov Negron et al[10].Re-injected natural gas is not included in the North Slope carbon intensity.No carbon intensity is estimated from inventories in California because no CO2 information is available.Mean and 95% confidence intervals are shown.Carbon intensity varies widely with both CO2 or CH4 driving high intensities.