Natural gas flaring is a common practice employed in many United States (U.S.) oil and gas regions to dispose of gas associated with oil production. Combustion of predominantly hydrocarbon gas ...results in the production of nitrogen oxides (NOx). Here, we present a large field data set of in situ sampling of real world flares, quantifying flaring NOx production in major U.S. oil production regions: the Bakken, Eagle Ford, and Permian. We find that a single emission factor does not capture the range of the observed NOx emission factors within these regions. For all three regions, the median emission factors fall within the range of four emission factors used by the Texas Commission for Environmental Quality. In the Bakken and Permian, the distribution of emission factors exhibits a heavy tail such that basin-average emission factors are 2–3 times larger than the value employed by the U.S. Environmental Protection Agency. Extrapolation to basin scale emissions using auxiliary satellite assessments of flare volumes indicates that NOx emissions from flares are skewed, with 20%–30% of the flares responsible for 80% of basin-wide flaring NOx emissions. Efforts to reduce flaring volume through alternative gas capture methods would have a larger impact on the NOx oil and gas budget than current inventories indicate.
Tropical wetlands contribute ∼30% of the global methane (CH4) budget. Limited observational constraints on tropical wetland CH4 emissions lead to large uncertainties and disparities in representing ...emissions. In this work, we combine remote sensing observations with atmospheric and wetland models to investigate dry season wetland CH4 emissions from the Pantanal region of South America. We incorporate inundation maps generated from the Cyclone Global Navigation Satellite System (CYGNSS) satellite constellation together with traditional inundation maps to generate an ensemble of wetland CH4 emission realizations. We challenge these realizations with daily satellite observations for May–July when wetland CH4 emission predictions diverge. We find that the CYGNSS inundation products predict larger emissions in May, in better agreement with observations. We use the model ensemble to generate an empirical observational constraint on CH4 emissions independent of choice of inundation map, finding large dry season wetland CH4 emissions (31.7 ± 13.6 and 32.0 ± 20.2 mg CH4/m2/day in May and June/July during 2018/2019, respectively). These May/June/July emissions are 2–3 times higher than current models, suggesting that annual wetland emissions may be higher than traditionally simulated. Observed trends in the early dry season indicate that dynamics during this period are of importance in representing tropical wetland CH4 behaviors.
We present estimates of regional methane (CH4) emissions from oil and natural gas operations in the Barnett Shale, Texas, using airborne atmospheric measurements. Using a mass balance approach on ...eight different flight days in March and October 2013, the total CH4 emissions for the region are estimated to be 76 ± 13 × 103 kg hr–1 (equivalent to 0.66 ± 0.11 Tg CH4 yr–1; 95% confidence interval (CI)). We estimate that 60 ± 11 × 103 kg CH4 hr–1 (95% CI) are emitted by natural gas and oil operations, including production, processing, and distribution in the urban areas of Dallas and Fort Worth. This estimate agrees with the U.S. Environmental Protection Agency (EPA) estimate for nationwide CH4 emissions from the natural gas sector when scaled by natural gas production, but it is higher than emissions reported by the EDGAR inventory or by industry to EPA’s Greenhouse Gas Reporting Program. This study is the first to show consistency between mass balance results on so many different days and in two different seasons, enabling better quantification of the related uncertainty. The Barnett is one of the largest production basins in the United States, with 8% of total U.S. natural gas production, and thus, our results represent a crucial step toward determining the greenhouse gas footprint of U.S. onshore natural gas production.
We use atmospheric observations to quantify methane (CH4) emissions from Mexico's most important onshore and offshore oil and gas production regions which account for 95% of oil production and 78% of ...gas production. We use aircraft-based top-down measurements at the regional and facility-levels to determine emissions. Satellite data (TROPOMI CH4 data and VIIRS night-time flare data) provide independent estimates of emissions over 2 years. Our airborne estimate of the offshore region's emissions is 2800 kg CH4 h−1 (95% confidence interval (CI): 1700-3900 kg CH4 h−1), more than an order of magnitude lower than the Mexican national greenhouse gas inventory estimate. In contrast, emissions from the onshore study region are 29 000 kg CH4 h−1 (95% CI: 19 000-39 000 kg CH4 h−1), more than an order of magnitude higher than the inventory. One single facility-a gas processing complex that receives offshore associated gas-emits 5700 kg CH4 h−1 (CI: 3500-7900 kg CH4 h−1), with the majority of those emissions related to inefficient flaring and representing as much as half of Mexico's residential gas consumption. This facility was responsible for greater emissions than the entirety of the largest offshore production region, suggesting that offshore-produced associated gas is being transported onshore where it is burned and in the process some released to the atmosphere. The satellite-based data suggest even higher emissions for the onshore region than did the temporally constrained aircraft data (>20 times higher than the inventory). If the onshore production region examined is representative of Mexican production generally, then total CH4 emissions from Mexico's oil and gas production would be similar to, or higher than, the official inventory, despite the large overestimate of offshore emissions. The main driver of inaccuracies in the inventory is the use of generic, non-Mexican specific emission factors. Our work highlights the need for local empirical characterization of emissions if effective emissions mitigation is to be undertaken.
Abstract 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 CO 2 e/MJ4.5, 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 1612, 22 for GOM federal shallow waters and 4325–65 for state shallow waters. Regional intensities vary, with carbon dioxide emissions largely responsible for CI on the North Slope 117.5, 15, in Cook Inlet 2213, 34, offshore California 7.23.2, 13, and in GOM deep waters 1.11.0, 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.
Incomplete combustion during flaring can lead to production of black carbon (BC) and loss of methane and other pollutants to the atmosphere, impacting climate and air quality. However, few studies ...have measured flare efficiency in a real-world setting. We use airborne data of plume samples from 37 unique flares in the Bakken region of North Dakota in May 2014 to calculate emission factors for BC, methane, ethane, and combustion efficiency for methane and ethane. We find no clear relationship between emission factors and aircraft-level wind speed or between methane and BC emission factors. Observed median combustion efficiencies for methane and ethane are close to expected values for typical flares according to the US EPA (98%). However, we find that the efficiency distribution is skewed, exhibiting log-normal behavior. This suggests incomplete combustion from flares contributes almost 1/5 of the total field emissions of methane and ethane measured in the Bakken shale, more than double the expected value if 98% efficiency was representative. BC emission factors also have a skewed distribution, but we find lower emission values than previous studies. The direct observation for the first time of a heavy-tail emissions distribution from flares suggests the need to consider skewed distributions when assessing flare impacts globally.
We have compared a suite of recent global CO2 atmospheric inversion results to independent airborne observations and to each other, to assess their dependence on differences in northern extratropical ...(NET) vertical transport and to identify some of the drivers of model spread. We evaluate posterior CO2 concentration profiles against observations from the High-Performance Instrumented Airborne Platform for Environmental Research (HIAPER) Pole-to-Pole Observations (HIPPO) aircraft campaigns over the mid-Pacific in 2009–2011. Although the models differ in inverse approaches, assimilated observations, prior fluxes, and transport models, their broad latitudinal separation of land fluxes has converged significantly since the Atmospheric Carbon Cycle Inversion Intercomparison (TransCom 3) and the REgional Carbon Cycle Assessment and Processes (RECCAP) projects, with model spread reduced by 80 % since TransCom 3 and 70 % since RECCAP. Most modeled CO2 fields agree reasonably well with the HIPPO observations, specifically for the annual mean vertical gradients in the Northern Hemisphere. Northern Hemisphere vertical mixing no longer appears to be a dominant driver of northern versus tropical (T) annual flux differences. Our newer suite of models still gives northern extratropical land uptake that is modest relative to previous estimates (Gurney et al., 2002; Peylin et al., 2013) and near-neutral tropical land uptake for 2009–2011. Given estimates of emissions from deforestation, this implies a continued uptake in intact tropical forests that is strong relative to historical estimates (Gurney et al., 2002; Peylin et al., 2013). The results from these models for other time periods (2004–2014, 2001–2004, 1992–1996) and re-evaluation of the TransCom 3 Level 2 and RECCAP results confirm that tropical land carbon fluxes including deforestation have been near neutral for several decades. However, models still have large disagreements on ocean–land partitioning. The fossil fuel (FF) and the atmospheric growth rate terms have been thought to be the best-known terms in the global carbon budget, but we show that they currently limit our ability to assess regional-scale terrestrial fluxes and ocean–land partitioning from the model ensemble.
Methane (CH4) emissions from oil and gas activities are large and poorly quantified, with onshore studies showing systematic inventory underestimates. We present aircraft measurements of CH4 ...emissions from offshore oil and gas platforms collected over the U.S. Gulf of Mexico in January 2018. Flights sampled individual facilities as well as regions of 5–70 facilities. We combine facility-level samples, production data, and inventory estimates to generate an aerial measurement-based inventory of CH4 emissions for the U.S. Gulf of Mexico. We compare our inventory and the Environmental Protection Agency Greenhouse Gas Inventory (GHGI) with regional airborne estimates. The new inventory and regional airborne estimates are consistent with the GHGI in deep water but appear higher for shallow water. For the full U.S. Gulf of Mexico our inventory estimates total emissions of 0.53 Tg CH4/yr 0.40–0.71 Tg CH4/yr, 95% CI and corresponds to a loss rate of 2.9% 2.2–3.8% of natural gas production. Our estimate is a factor of 2 higher than the GHGI updated with 2018 platform counts. We attribute this disagreement to incomplete platform counts and emission factors that both underestimate emissions for shallow water platforms and do not account for disproportionately high emissions from large shallow water facilities.
Airborne estimates of greenhouse gas emissions are becoming more prevalent with the advent of rapid commercial development of trace gas instrumentation featuring increased measurement accuracy, ...precision, and frequency, and the swelling interest in the verification of current emission inventories. Multiple airborne studies have indicated that emission inventories may underestimate some hydrocarbon emission sources in US oil- and gas-producing basins. Consequently, a proper assessment of the accuracy of these airborne methods is crucial to interpreting the meaning of such discrepancies. We present a new method of sampling surface sources of any trace gas for which fast and precise measurements can be made and apply it to methane, ethane, and carbon dioxide on spatial scales of ∼ 1000 m, where consecutive loops are flown around a targeted source region at multiple altitudes. Using Reynolds decomposition for the scalar concentrations, along with Gauss's theorem, we show that the method accurately accounts for the smaller-scale turbulent dispersion of the local plume, which is often ignored in other average mass balance methods. With the help of large eddy simulations (LES) we further show how the circling radius can be optimized for the micrometeorological conditions encountered during any flight. Furthermore, by sampling controlled releases of methane and ethane on the ground we can ascertain that the accuracy of the method, in appropriate meteorological conditions, is often better than 10 %, with limits of detection below 5 kg h−1 for both methane and ethane. Because of the FAA-mandated minimum flight safe altitude of 150 m, placement of the aircraft is critical to preventing a large portion of the emission plume from flowing underneath the lowest aircraft sampling altitude, which is generally the leading source of uncertainty in these measurements. Finally, we show how the accuracy of the method is strongly dependent on the number of sampling loops and/or time spent sampling the source plume.
Wetlands are the single largest source of methane to the atmosphere and their emissions are expected to respond to a changing climate. Inaccuracy and uncertainty in inundation extent drives ...differences in modeled wetland emissions and impacts representation of wetland emissions on inter‐annual and seasonal time frames. Existing wetland maps are based on optical or NIR satellite data obscured by clouds and vegetation, often leading to underestimates in wetlands extent, especially in the Tropics. Here, we present new inundation maps based on the Cyclone Global Navigation Satellite System (CYGNSS) satellite constellation, operating in L‐band that is not impacted by clouds or vegetation, providing reliable observations through canopy and cloudy periods. We map the temporal and spatial dynamics of the Pantanal and Sudd wetlands, two of the largest wetlands in the world, using CYGNSS data and a computer vision algorithm. We link these inundation maps to methane fluxes via WetCHARTs, a global wetland methane emissions model ensemble. We contrast CYGNSS‐modeled methane emissions with WetCHARTs standard runs that use monthly rainfall data from ECMWF re‐analysis (ERA5), as well as the commonly used SWAMPS wetland maps. We find that the CYGNSS‐based inundation maps modify the methane emissions in multiple ways. The seasonality of inundation and methane emissions is shifted by two months because of the lag in wetland recharge following peak rainfall. Both inundation and methane emissions also respond non‐linearly to wet‐season precipitation totals, leading to large interannual variability in emissions. Finally, the annual magnitude of emissions is found to be greater than previously estimated.
Key Points
Cyclone Global Navigation Satellite System (CYGNSS) data is used to produce monthly maps of tropical wetlands. The maps are used to drive the WetCHARTs methane emission model
The seasonality of inundation‐based model results lags two months behind the rainfall‐based models and shows larger dry‐season emission
CYGNSS‐based estimates, consistent with independent observations, show higher emissions with larger variability than rainfall‐driven ones