Introducing the OCI⁺

There is no standard oil and gas. A hodgepodge of hydrocarbons is produced, processed, and transformed into usable petroleum products that flow through today’s economy. Oil and gas characteristics, production methods, and operational stewardship vary widely. And, by extension, oil and gas climate impacts vary widely too. Significant opportunities exist to reduce greenhouse gas (GHG) emissions that start by differentiating oil and gas resources’ climate impacts.

Numerous approaches—bottom-up, top-down, and hybrid methods—are used to evaluate emissions from petroleum systems. Top-down methods typically record emissions via tower-based measuring stations, drive-by detection, and fly-over techniques, including satellites, aircraft, sensors, and drones. Bottom-up methods assess emissions from the ground-up, often accounting for emissions based on equipment counts and component-specific leak factors. No single method is entirely decisive. Looking ahead, a tiered ecosystem of both simple and advanced systems modeling, remote sensing, and— importantly—direct measurements by a suite of technologies will be necessary to continually improve our grasp on GHGs and enable widespread reductions across the sector.

The Oil Climate Index plus Gas (OCI+) is a bottom-up systems tool that also uses input top-down measurements (hybrid approach) to quantify emissions from oil and gas production, processing, refining, shipping, and end uses

The OCI+ uses three underlying models to assess GHG emissions from the upstream, midstream, and downstream segments of the oil and gas supply chain. The upstream model, Oil Production Greenhouse Gas Emissions Estimator (OPGEE) was first developed to analyze oil by researchers at Stanford University and the California Air Resources Board, and then expanded to include gas by researchers at Stanford University. The midstream model, Petroleum Refinery Life-Cycle Inventory Model (PRELIM) was developed and is being expanded by researchers at the University of Calgary. And the downstream model, OPEM (Oil and Gas Products Emissions Module) was developed by researchers at the Carnegie Endowment for International Peace, Koomey Analytics, and the Watson Institute at Brown University and, in 2022 for this OCI+ release, was expanded and coded in Python by RMI and Koomey Analytics

The OCI+ and its underlying models have been peer-reviewed and internationally cited and applied in energy policy decision making for over a decade, as documented in the OCI+ web tool “Studies” tab. The OCI+ is also deployed for analyzing the oil and gas sector as part of the Climate TRACE coalition¹.i A schematic of the OCI+ follows.

Modeling Overview

The OCI+ tracks equivalent barrels of extracted hydrocarbons through the oil and gas supply chain using a singular functional unit so that lifecycle emissions can be summed. This results in comparable GHG emissions estimates per barrel of oil equivalent (BOE) produced—barrel-forward total emissions—for all oils and gases. Thus, the functional unit of a BOE produced refers to the processed oil and gas that leaves a field’s production site and moves on to further refining and processing and then onto marketing and consumption.

The underlying OCI+ models are mass-balance, process engineering models that depend on various input data, including reservoir characteristics, operating characteristics, and downstream transport and end use specifications. Where available, the OCI+ uses open-access data in consistent, comparable, and verifiable formats. Smart defaults generated by statistical analyses from published studies, industry literature, proxy assays, engineering judgements, and other analytic methods fill in missing data. Examples of OCI+ input data are oil and gas field production volumes (proprietary at present), field depths and pressures, gas compositions, oil assays (analyses of predetermined data measuring a crude oil’s chemical and physical characteristics) reported out in a specified format, and shipping specifications. Academic sources, technical documents, and industry reports are used to gather OCI+ input data.

Modeled oil and gas resources are selected based on their volumetric, geographic, geologic, chemical, and physical diversity. But the final arbiter of the global resources selected is data availability. Numerous additional oil and gas resources would have been modeled if greater data transparency were voluntarily provided or mandated. As such, the lack of publicly available data currently limits the reach and granularity of the OCI+, and increases uncertainty bands of modeled results.

OPGEE models emissions associated with the extraction and processing of oil, gas, and other compounds stemming from the reservoir. Gas flows through OPGEE all the way to distribution, with user-defined options to also analyze gas at the field or transmission line. OPGEE models the oil transported from surface processing to the refinery gate. From here, refinery oil flows are modeled in PRELIM. Finally, OPEM uses the product slates modeled in PRELIM and the gas flows modeled in OPGEE to estimate product transport and end use consumption emissions. OPEM also includes the transport and combustion of petroleum coke (petcoke) that is removed upstream when upgrading extra-heavy oils and natural gas liquids that are separated from light oils and wet gases at the field. Petrochemical emissions are also estimated assuming in OCI+ v.2.1.0 that petrochemical feedstocks are turned into plastics and end-of-life emissions of plastics are included.Thus, the OCI+ estimates the sum of the results from these three models, which equals the total lifecycle emission intensity of a barrel oil equivalent through the entire supply chain.

In this version of the OCI+ (v2.1.0) web tool, the Total Emissions, Supply Chain, and Analysis tabs present the modeled emissions for over 500 upstream fields for the year 2022. Time series emissions data were also published for individual field from 2015-2022. Fields are defined as individual major named oil and gas fields and sub-basins. However, fields can have different names and aliases and can be grouped in different ways, depending on the data source

Top-producing global fields were selected for OCI+ modeling, representing about 70% of the 2022 global production.

The model parameters used to generate the OCI+ dataset are described in further detail in the next section, Model Deployment. Users can download data using the “Data Download” tab to explore inputs and outputs of the OCI+ models. We recommend that users fill out the OCI+ user input form on the web tool to share how they are using OCI+ data so RMI can improve this tool and plan for future user-based expansions. Please also feel free to contact us if you find anomalous results, can provide updated data to input, or have suggestions.

The model parameters used to generate the OCI⁺ dataset are described in further detail in the next section, Model Deployment. The Scenarios tab includes data from a subset of fields to study sensitivity of the emission intensities to changes in data inputs for select emissions mitigation measures. This is described in further detail in the Scenarios section.

Model Deployment

OPGEE 3.0a

Model Summary

The OPGEE model covers:Upstream oil and gas production, processing, and transport to oil refineries and/or various points of gas distribution prior to end use.

Link to OPGEE version 3.0b developed by and hosted at Stanford University (users please note the copyright provisions for this model): https://github.com/arbrandt/OPGEE.

Model Boundaries

OPGEE covers emissions from all upstream operations required to produce and transport crude hydrocarbons to the refinery gate--this includes direct (Scope 1) emissions (i.e., flaring, venting, fugitives, and onsite combustion), as well as indirect (Scope 2) emissions embodied in imported electricity and materials acquired. OPGEE also accounts for emissions affiliated with oil and gas site exploration and land use. The model is divided into modules, broken up by major production stages. Several activities and processes occur in each stage. Over 100 emission sources are classified across all upstream process stages.

In OPGEE, the user has the option to define the boundary of emissions assessment for each hydrocarbon (oil and gas). For oil-dominant fields, the oil boundary was set to “Refinery” and the gas boundary was set to “Field”; for gas-dominant fields, the oil boundary was set to “Field” and the gas boundary was set to ”Distribution” for OCI+ v2.1.0 runs.

Data Inputs

OPGEE can accept hundreds of user data inputs and relies on public data wherever possible (see Appendix for full list of main OPGEE inputs). However, where input data are lacking, smart defaults allow the model to assign reasonable estimates based on fewer than a dozen key oil and gas characteristics. Key inputs include: field name, field age, field depth, oil, gas, and water production volumes, gas composition, number of injecting wells, satellite-derived flaring volumes, crude API gravity, production methods, means and distance of crude transport. Fracked and LNG resources receive special designation found on the “Secondary Inputs” tab of the OPGEE model. Grid electricity mix is also adjusted for certain fields that source renewable electricity for their operations. For key upstream data inputs for OCI+, download the data through “Data Download” tab.

The field and resource data specified are used to calculate base-run GHG emissions outputs for each oil and gas run. Under California regulations, OPGEE reports GHG-emissions outputs in units of grams CO2 equivalent per megajoule of petroleum products generated. These outputs are converted into kilograms emitted per barrel of oil equivalent by multiplying them by the weighted lower heating value of all processed oil and gas products (in megajoule per barrel of oil equivalent), which is output in OPGEE’s Energy Summary tab.

Field Locations

The location and areal extent for each oil and gas field was collected from publicly available data where possible (e.g., U.S. government datasets) and also through multiple proprietary datasets that provides latitude and longitude values. Where multiple fields are aggregated under a single name, a production-weighted average of the coordinates was taken.

Flaring Emissions

To estimate emissions from flares, OPGEE uses a Flaring-Oil-Ratio (volume of gas flared per barrel of oil produced) as an input. Flaring volumes were estimated by aligning satellite measurements of flares with oil and gas field boundaries, where available. The flares and volumes were identified with data from NOAA’s Visible Infrared Imaging Radiometer Suite (VIIRS) satellite instrumentation, which, when combined with the VIIRS Nightfire (VNF) algorithm from the Earth Observation Group, Payne Institute for Public Policy, Colorado School of Mines, detects flare location and estimates flared volume². For fields where geographical polygon boundaries are not available, but latitude and longitude coordinates are (e.g., most fields in the United States and Canada), a 10-mile radius was used to encompass surrounding flares and assign them to that field. The exception is for the Permian Basin in the U.S., for which a 50-mile radius was used due to the size of the field.

To estimate emissions from flares, OPGEE takes a Flaring-Oil-Ratio (volume of gas flared per barrel of oil produced) as an input. Flaring volumes were estimated by aligning satellite measurements of flares with outlines of the oil and gas fields. The flares and volumes were identified with data from the Visible Infrared Imaging Radiometer Suite (VIIRS) sensor, which, when combined with the Nightfire Algorithm from Colorado School of Mines, can detect and provide estimates to quantify flare location and volume. Not all fields contained detectable flares. In these cases, a country-wide flaring rate was assumed. Sky Truth and other public tools were used to aggregate flare volumes at the field, basin, and country level. The flaring rates reflect 2020 upstream flaring volumes.²For fields where geographical polygon boundaries are not available, but latitude and longitude coordinates are (e.g., most fields in the United States and Canada), a 10-mile radius was used to encompass surrounding flares and assign them to that field. The exception is for the Permian Basin in the U.S., for which a 50-mile radius was used due to the size of the field.

Methane Venting and Fugitives

In 2021, the Global Methane Pledge was signed at COP26, heightening the importance of quantifying, tracking, and reducing methane emissions. The OCI+ provides field-level analysis of methane from oil and gas, from production through distribution. A specific and extensive recent effort was undertaken by researchers at Stanford university to build out the methane fugitives module in OPGEE.³Estimates are based on advanced statistical methods and derived from published academic surveys, thousands of measured emissions data points, and other public data sources

When using OPGEE, the model requires “component” or “site” level analysis to be selected. Users can also input equipment component data and adjust methane loss rates to fine tune these results. For most OCI+ fields, the model was run using “component” analysis and default leakage rates. Offshore Norwegian fields were run on “site” analysis to better reconcile with externally published results. Upstream equipment and gathering system loss rates were adjusted for a handful of fields including Permian, Hassi R’Mel, and Tin Fouye, to better align total methane results with recent scientific studies on methane super emitters (Cusworth et al. and Lauvaux et al). Similarly, gas transmission pipeline fugitive loss rates were adjusted upwards for fields in Russia and Turkmenistan.

Flare slip is another source of methane—default OCI+ runs were set to flare efficiency of 90%. Flare efficiencies were adjusted for countries with regulations on flaring efficiency, such as Norway, which was set to 98%.

Since gas resources were run on the “Distribution” setting, OPGEE naturally estimates methane leakage from gas transport and LNG segments of the supply chain. In the OCI+, these emissions were reallocated to the “Gathering, Storage, and Transport” segment on the Supply Chain tab to reflect the downstream nature of pipeline gas and LNG transport.

Transport Distance

OPGEE uses methods and data from CA-GREET to calculate the emissions from transport of crude oil and gas. While detailed transport methods and destinations of crude and gas from each field are largely unavailable, regional assumptions of destinations were made using import/export data from the BP Statistical Review of World Energy, theIEA, and theEIA. Distances were then estimated using available pipeline and LNG infrastructure data for continental transport, and port-to-port shipping distances for exported products. Where pipeline length data was not available, continental transport distance was assumed to be the distance from the field to the largest city in the destination country. Where liquefaction and regasification facility locations were not available, other ports were assumed for LNG shipment distances.

• Crude oil transport – to refinery
  • Assumption for domestic oil was to calculate major pipeline distance (based onGlobal Energy Monitor pipeline trackeror distance to nearest city if pipeline data not available for region) and assume 100% of produced oil travels by pipeline.
  • Crude export fraction and main crude destinations were used to inform a percentage of oil traveling by ship and shipping distance.
  • Example: Approximately 40% of Mexico crude oil consumed domestically and piped an assumed 350 miles viamain pipelineto Mexico City. Additional 60% exported via ocean tanker to U.S., Asia, and Europe.
• Gas pipeline – to point of distribution
  • Assumption is all gas stays in country and travels distance to largest city, except countries where gas production is predominantly exported, in which case publicly available gas pipeline length was used.
  • Example: Algerian gas assumed to be piped 1010 miles to Europe viaMaghreb-Europe Gas pipeline.
• LNG – to liquefaction and shipment to regasification
  • Assumed a pipeline distance from gas field to nearest liquefaction terminal. LNG shipping distances calculated based on country-level gas trade data patterns
  • Example: Qatari gas travels about 50 miles to Ras Laffen liquefaction terminal and about 8300 miles via tanker to Dusseldorf, Germany.

PRELIM Version 1.6

Model summary

The PRELIM model covers: Midstream refining

Link to the model hosted by the University of Calgary (users please note the copyright provisions for this model): https://www.ucalgary.ca/energy-technology-assessment/open-source-models/prelim.

PRELIM is crude oil refining process model that provides an estimate of the emissions and product slates volumes from a refinery per barrel of crude. It has been developed by researchers at the University of Calgary. For the purposes of the OCI+, PRELIM is used to estimate the oil refining emissions associated with one barrel of crude output from a given upstream field.

These model results do not trace individual barrel supply chains from field to specific refinery. Rather, provide an estimate of the likely refinery complexity and associated emissions based on the crude properties.

Data Inputs

PRELIM takes several inputs to specifically adapt process units and is highly configurable for the user. For the OCI+, each crude oil was run using PRELIM’s default conditions, with the following input settings:

PRELIM Main Input Number	InputName	Value
1A	Process Unit Configuration	Default
1B	Naphtha Catalytic reformer	Default
1C	FCC hydrotreater	Default
1D	Electricity generation	NG Fired power plant
1E	SMR Hydrogen Purification	Default
2	Allocation	Energy Basis
3	Allocation Products	Turnon LPG and Petrochem allocation
4	Heating Value	LHV
5	GWP	AR5 20 year and 100 year
6	Upstream Releases	Included**
7	Off Site waste	Default (on)
8	Asphalt Production	0%
9	Off gasproduct production	LPG, Petrochem feedstock on
10	Cogeneration Unit	None

Refinery Configuration:

PRELIM contains a set of refinery configurations. Default refinery configuration is set automatically based on the API gravity and sulfur-content of the input crude:

• Deep conversion refinery—heavy crude (15 to 22 API) with any sulfur level
• Medium conversion refinery—medium sweet crude (22 to 32 API, with less than 0.5 percent sulfur content by weight); medium sour crude (22 to 32 API, with more than 0.5 percent sulfur content by weight); and light sour crude (over 32 API, with more than 0.5 percent sulfur content by weight)
• Hydroskimming refinery—light sweet crude (over 32 API, with less than 0.5 percent sulfur content by weight)

Given the lack of data on crude-refinery mapping, the default configurations represent the OCI+ team’s best engineering judgments at present. Under the Scenarios tab of the OCI+ tool, users can simulate running oils through different refinery configurations other than the default to estimate how a suboptimal refinery affects GHG emissions.

Upstream Releases:

The upstream releases setting corresponds to emissions from offsite electricity and emissions from production, including methane leakage and upstream energy use. To obtain the offsite electricity related emissions without double counting the production emissions from OPGEE, PRELIM was run with upstream releases enabled, and production emissions were manually removed.

Offgas Product Production:

This setting enables offgases, or gases that are released as part of other processes, to be converted to value products. While refinery fuel gas (RFG) offtake is not often of high value for refineries compared to onsite use, there is increasing value given to natural gas liquids (NGLs) and liquid petroleum gas (LPG) in the market as fuels or chemical feedstocks.

Cogen:

Cogen is assumed to be off for all cases due to lack of data.

Assays

One of the important inputs to the PRELIM model is crude assay properties corresponding to fixed temperature cuts. An example of the required data input is below. Please refer to PRELIMdocumentationon distillation curve formatting.

PRELIM currently contains 686 pre-loaded oil assays in its inventory.

Assay matching

Each upstream field run was assigned one or more assays based on the best available information in the public domain. A confidence score was assigned to each match on a scale of 1-5 (1-least confidence, 5-highest confidence) based on the following criteria:

1. Exact assay name match
2. Fuzzy assay name match
3. Country match with matching sour designation and assay API within 20% of field API
4. Region match with matching sour designation and assay API within 20% of field API
5. Global match with matching sour designation and assay API within 20% of field API

Matches were reviewed by internal experts for quality control. ivThe assay used to model refining emissions for each field is noted in the OCI+ Oil and Gas Details page. It is also available in the downloadable csv dataset.

Outputs

The output of PRELIM includes a product slate and emissions by process step and source, all of which can be found in the “Main Inputs & Outputs” tab of the PRELIM model. PRELIM outputs are based on an input crude volume of approximately 100 kbd, (volume specified in the assay), so the total emissions and product volumes are scaled to the volume of crude entering the refinery as taken from OPGEE before being used in the final data outputs. Emissions intensity values are normalized to the total BOE of oil and gas leaving the field boundary as calculated using OPGEE outputs.

OPEM Version 3.0

Model Summary

The OPEM model covers: Downstream Oil Product Transport and Oil and Gas End Use Emissions

Link to the model (users please note the copyright provisions for this model): https://github.com/RMI-Climate-Intelligence/OPEM

OPEM estimates two emissions sources: the transport of petroleum products by shipping entities and the end use of all petroleum products by various consumers. The combustion of natural gas and petroleum products by consumers (Scope 3 emissions) is increasingly being considered in this sector’s climate impacts.

OPEM considers all associated oil and gas products that are consumed. Historically, petroleum end use centered only on transport fuels, including gasoline and diesel, and ignored or incompletely and inconsistently reported GHG emissions from petroleum co-products like petcoke, fuel oil, residual fuels, asphalt, and petrochemical feedstocks. OPEM v.3.0 includes these co-products, as well as gas and natural gas liquids that were produced along with each barrel of crude.

OPEM estimates emissions using combustion, transport, and process emissions factors reported by the U.S. Environmental Protection Agency for GHG inventories and Argonne National Laboratory’s GREET model. Emissions factors are documented here. Note that the EPA assumes high fuel quality and near-complete fuel combustion in the calculation of published combustion emissions factors. Depending on the quality of the engine in which a fuel is burned, EPA emissions factors may result in a best-case (lowest emissions) estimate. The model also assumes some percentage of ethane is converted to ethylene for petrochemical use, and assumes a conversion process emissions intensity, but does not include emissions associated with further processing and use of the ethylene product, nor does it include other petrochemical processes.

Given current limitations posed by the reporting of global petroleum product transport data, the OCI+ assumes that lighter petroleum liquid products (petrochemical feedstocks, gasoline, diesel, jet fuel, NGL, LPG) are transported via pipeline 2,414 kilometers or 1,500 miles (the equivalent distance from Houston to New York) and then by heavy-duty tanker truck 380 kilometers or 236 miles (the equivalent distance to either the Washington, DC or Boston metropolitan areas). Heavy liquid products (fuel Oil, residual fuels) are assumed to travel 1,200 km by rail, 1,200 km by tanker, and 500 miles by truck. Solid fuels (petroleum coke and sulfur) are assumed to travel 3,352 km by tanker and 750 miles by rail. Users can input different distances, shipping modes, and shipping fuels into OPEM to model different transport scenarios.

Natural gas processing and transport emissions are computed through the OPGEE model, and therefore are not part of OPEM. Details for calculation of emissions from natural gas pipelining and LNG shipping can be found in the OPGEE documentation.

Data Inputs

OPEM data inputs require a detailed product slate (measured in barrels per day for liquid products, kilograms per day for solid products, or gas mass for gaseous products), which are reported out in PRELIM, and gross volumes of crude, gas, NGLs, and upgrader coke produced, which are reported in OPGEE. Petroleum coke, LPG, and petrochemical feedstock densities are used to convert solid and gaseous refinery products to barrels equivalent. LPG and petrochemical feedstock are assumed to have a 270:1 gas volume to liquid volume ratio⁴. LPG and petrochemical feedstock gas densities are sourced from PRELIM, under 20°C and 1 atm conditions. It is assumed that petrochemical feedstock ethane, while LPG is 3/4 propane and 1/4 butane. The product slates for OPEM inputs are generated from PRELIM, by modeling a standard crude input volume of 100,000 barrels per day and then normalized per barrel crude input to the refinery.

The total BOED (barrels of oil per day) produced at field are calculated as follows:

• Natural gas (btu) is extracted from the ‘Energy Summary’ tab in OPGEE and converted to volume using a natural gas LHV of 983 btu/ft³
• Crude (kg) is extracted from the ‘Energy Summary’ tab in OPGEE and converted to volume using the API gravity of the crude, using the following formula to calculate density:
• Density [kg/bbl] = (158.9873*141.5/(131.5+°API))
• NGL masses (kg) are extracted from individual ethane, propane, and butane mass flows in the ‘Flowsheet’ tab in OPGEE (columns W and CP). Conversion constants, taken from OPGEE, are found in Appendix Table A2.
• Upstream coke mass from upgrading heavy oils is extracted (kg) from the ‘Flowsheet’ and ‘Energy Summary’ tabs.

All other model input parameters are left as OPEM defaults.

The model outputs include emission intensities of transportation, product combustion, and non-combusted products, per BOE produced at the upstream field. This downstream emission intensity is then added to the PRELIM and OPGEE emissions intensities to create whole lifecycle emission intensity of a BOE produced at a field.

Global Warming Potentials

OCI+ GHG emissions estimates are based on equivalent GHG emissions for carbon dioxide, methane, nitrous oxide, (and volatile organic compounds in the case of OPGEE), combining these into one CO2eq result using different global warming potentials (GWPs) that compare the other GHGs to carbon dioxide. For the OCI+, we used data reported by the IPCC in AR6. OCI+ default emissions estimates on the Supply Chain and Total Emissions pages use 20-year GWPs, with a toggle for 100-year GWPs, while estimates on the Analysis page use 20-year GWPs. The IPCC’s 2021 assessment report (AR6) states that GWP for carbon dioxide (CO2) is referenced at 1. The GWP for fossil methane (CH4) is 30 times greater than CO2 on a 100-year timescale and 82.5 times greater on a 20-year timescale (without including climate carbon-cycle feedbacks). For nitrous oxide (N2O), the value is consistently near 270 times greater than CO2. Download the United Nations Intergovernmental Panel on Climate Change (IPCC) global warming potentials here.

Web Tool Functionality

Map page

The OCI+ flaring map reflects VIIRS NightFire flare data that serves as a direct OPGEE input. However, the methane map does not reflect data used in the estimation of GHG emissions for the OCI+ oils and gases. The methane map was obtained from independent analysis by Harvard professor and NASA affiliate, Dr. Daniel Jacob, and Dr. Tia Scarpelli. Drs. Jacob and Scarpelli and their team used global methane data from UNFCCC inventories and infrastructure datasets to spatially allocate emissions for 2010-2019 in a Global Fuel Exploitation Inventory v2 (Scarpelli, Tia R.; Jacob, Daniel J., 2019, “Global Fuel Exploitation Inventory (GFEI)” , https://doi.org/10.7910/DVN/HH4EUM, Harvard Dataverse, V2). They evaluate GFEI v2 against global inversions of atmospheric methane observations in this analysis. The GFEI data is mapped and overlaid on the Map page with OCI⁺ fields and GHG footprints.

Converting Emissions Outputs to Other Metrics

The default functional unit or metric in the OCI+ web tool is an emissions intensity--GHG emissions per barrel of oil equivalent of processed oil and gas. Both CO2e and methane-only emissions are estimated. These conversions are calculated by multiplying default results by reported lower heating values (in megajoule per BOE of total processed oil, NGLs and gas).

In the Oil and Gas Details page, several metrics related to methane intensity are also provided: 1) “Upstream Methane Intensity” in the unit of kilograms CH₄ per BOE; 2) “Upstream Methane Leakage” based on NGSI standard in the unit of kilograms CH₄ per kilogram methane produced(only emissions from the production segment are included); This unit can be used to benchmark with RMI’s MIQ standards; 3) “Upstream Methane Loss” in the unit of grams CH₄ per total megajoule produced.

Resource Classification

The classification bounds used to sort resources in the OCI⁺ are detailed in the tables below.

Crude oil types	Natural Gas Types	Other Resources
Ultra-light	Acid	Condensate
Light	Wet	Natural Gas Liquids
Medium	Dry
Heavy	Coal-bed
Extra-heavy

Gravity Bounds	Classification
<15	Extra-Heavy Oil
15 - 23	Heavy Oil
23 - 38	Medium Oil
38 - 45	Light Oil
45 - 50	Ultra-Light Oil
>50	Condensate

Lower Bound C2-C4 %Vol	Classification
0	Dry Gas
15	Wet Gas

Condition	Classification
If Max(Gas BOE, Oil BOE, Condensate BOE)==Gas BOE	Gas
If Max(Gas BOE, Oil BOE, Condensate BOE)==Oil BOE	Oil
If Max(Gas BOE, Oil BOE, Condensate BOE)==Condensate BOE	Condensate

Lower Bound CO₂ mol%	LowerBound H₂S mol%	Classification
0	0	Non Acid
10	2	Acid

Sulfur range (%Wt)	Classification
0-.42	Sweet
.42-.5	Uncategorized
0.50	Sour

Flare Range	Classification
0-1	Very Low
1-5	Low
5-50	Medium
50-500	High Flare
500-1000	Very High
>1000	Ultra High

Label	Crude_production_bbld: Cutpoints	Gas_production_Mscfd: Cutpoints	NGLcon_production_bbld: Cutpoints	Total_production_bbld: Cutpoints
Very Low	<5,000	<1,000	<1,000	<5,000
Low	5,000 : 50,000	1,000 : 50,000	1,000 : 25,000	5,000 : 50,000
Medium	50,000 : 250,000	50,000 : 500,000	25,000 : 50,000	50,000 : 250,000
High	250,000 : 500,000	500,000 : 2,000,000	50,000 : 250,000	250,000 : 500,000

Flaring Risk Map

Map Layers

Flare Layer

Flare stack locations and flared gas volumes were identified with data from NOAA’s Visible Infrared Imaging Radiometer Suite (VIIRS) processed through the VIIRS Nightfire (VNF) algorithm from the Earth Observation Group, Payne Institute for Public Policy and Colorado School of Mines. The flaring activity represented in the webtool is an annual dataset for 2022 and has been filtered to remove detections that are not oil and gas related such as landfill gas flares, and those that occur offshore or at downstream facilities such as refineries and shipping terminals. A small number of flare points that have been recorded in prior years but had no detected flaring in 2022 remain in the dataset.

Block Group Layer

Block groups are geographic units defined by the US Census Bureau that represent contiguous areas with populations between 600 and 3,000 people. Block group demographic and geographic data was obtained from EPA’s EJScreen Environmental Justice Screening and Mapping tool for the year 2022⁶. The block groups within a 5-kilometer distance to a flare in the flare dataset comprise the study area shown in the webtool of areas directly impacted by upstream flaring.

Contextual Data

The Impacted Native American Lands Layer was derived from US Census American Indian/Alaska Native/Native Hawaiian (AIANNH) Areas representing federal and state recognized reservations and off-reservation trust land areas⁷.

The Production Areas Layer was derived through well data from FracTracker Alliance in combination with resource classification data⁸. This layer displays 2-kilometer pixels of the predominant type of resource extraction (Oil, Gas, or Both) in the underlying area.

Index Methodology

Flare Index

The Flare Index quantifies and ranks the negative impact of upstream flaring, based on flare characteristics alone. When applied to the Flare Layer, this index is derived from the flared gas volume and detection frequency of an individual flare. When applied to the Block Group Layer, the Flare Index is based on cumulative statistics of flares within a 5-kilometer buffer zone of a block group. These metrics include the sum of flared gas volume, the average detection frequency, and the density of flares in the buffered block group. In both layers, the variables are normalized, aggregated into the Flare Index, and displayed as ranks and percentiles to show the relative impact within a state or across the nation.

Environmental Justice Index

The Environmental Justice Index quantifies and ranks the negative impact of upstream flaring, based on flare characteristics and social vulnerability. Indicators of social vulnerability are chosen from EPA’s EJScreen demographic variables that specify percentages of a block group’s population that identify as people of color ('PEOPCOLORP'), low income ('LOWINCPCT'), unemployed (‘UNEMPPCT'),speaking English as a second language ('LINGISOPCT'), having less than a high school degree (LESSHSPCT), being over 64 years old (OVER64PCT), or being under 5 years old('UNDER5PCT’). These demographic characteristics combine into a weighted average representing a block group’s vulnerability that places emphasis on the first two indicators as shown below.

Block group vulnerability score = 0.25 * ('PEOPCOLORPCT'+ 'LOWINCPCT') + 0.1 * (‘UNEMPPCT'+'LINGISOPCT'+'LESSHSPCT'+'OVER64PCT'+'UNDER5PCT’)

For the Block Group Layer, block group vulnerability scores are normalized into an index and multiplied with the block group Flare Index. This creates the block group Environmental Justice Index that quantifies the cumulative impact of flaring and social vulnerability in a block group.

The Environmental Justice Index is also derived for the Flare Layer using GIS analysis on the block group vulnerability scores to quantify the environmental justice impact of individual flares. Block groups within a 5-kilometer radius of a flare contribute to that flare’s vulnerability score based on the degree of intersection. From there, a flare’s vulnerability score is normalized and multiplied by its Flare Index to produce the Flare Environmental Justice Index. In both layers, the Environmental Justice Index is displayed through ranks and percentiles to show the relative impact within a state or across the nation.

Data Quality and Uncertainty

Modeling uncertainty

The OCI+ is built upon dynamic, complex engineering models. As such, the results of each model (OPGEE, PRELIM, and OPEM) carry inherent uncertainty, as does the overall result. For OPGEE, user inputs affect the way the model functions (i.e., how statistical simulations are performed). This can be read about further in the OPGEE methodology available on the Github site.

For the base runs in OCI+, OPGEE was run with 1 uncertainty simulation. This means that for any inputs whose values occur along a distribution, the mean was selected as the smart default. As an additional exercise, we perform 100 uncertainty runs for a select handful of fields to demonstrate the impact on the upstream emissions intensities (Figure 2). In the next phase of OCI+, with a python version of OPGEE model, we will be able to run sufficient uncertainty iterations for every field modeled so that an uncertainty band is included for every field.

The coefficient of variation for OPGEE, calculated by dividing the standard deviation in emissions intensity by the mean, reflects uncertainty of the modeled results. Uncertainty is reduced as the number of provided model inputs increases (Figure 3). This emphasizes the need for greater data availability to improve confident understanding of the climate impacts of every oil and gas resource.

For the midstream model PRELIM, the most influential parameters are processing unit energy use (including electricity, gas, and steam), hydrogen production emission factor (via steam methane reforming), natural gas combustion emission factor, and electricity emission factor. Normal distributions are assumed for these parameters by 1) using their baseline values in PRELIM as means if they are close (±5%) to the midpoint of their actual ranges; 2) using one sixth of their actual ranges as standard deviation to cover 99.7% of the data (99.7% of the data is within 3 standard deviations of the mean; 3) if their baseline values in PRELIM are not close to the midpoint but are rather close to either end of their actual ranges, add skewness to normal distribution by using MATLAB Pearson system random numbers, where skewness and kurtosis coefficients measure the symmetry of a distribution and the thickness of the tail ends, respectively

For each Monte Carlo iteration, a randomly sampled dataset is generated using above normal distributions. If the value of any parameter generated is beyond its actual range, then use the closer range end value instead. Results show that the refining CI of all five crudes vary within a 10-15 kg CO2eq bbl-1 range, which is around 30% of the baseline case results.

At the process unit-level, uncertainty associated with PRELIM parameters and modeling structure can also be propagated to the modeling results (e.g., refining climate intensity and energy use). Modeling structure, such as fluid catalytic cracking (FCC) yields and gas oil hydrocracker yields in PRELIM are fixed for each configuration. However, refineries with the same configuration may have different yield patterns (e.g., more gasoline less diesel or more diesel less gasoline) driven by refining margin and market dynamics. Due to the lack of data in such refinery-wise yield patterns, this type of uncertainty is not considered in this study.

Archived versions of the OCI⁺

The OCI+ has undergone numerous iterations since its introduction in 2015. As new, open access data is provided or uncovered through research, the research team has updated the OCI+ models and revised the estimates for lifecycle GHG emissions. Moreover, the OPGEE, PRELIM, and OPEM models are continuously undergoing improvements and expansions. Previously, a date-stamped log was used to detail changes made to the OCI+. This format continues below. For archived Change Log dating back to 2015, refer to the original Oil Climate Index site hosted by the Carnegie Endowment for International Peace. For the v.1.0.0 of the OCI+ webtool, go to archive-ociplus.rmi.org.

Date	Change Description
Summer 2022	API gravities For some OCI+ oils, the OPGEE input API gravity and ensuing classification may not always agree with the optimal refinery configuration selected for that field. This can happen because not all crudes have a one-to-one match to an assay. These fields were matched with a blend of known assays that can reflect different APIs gravities. For example, an OPGEE-classified heavy oil may be matched with an assay blend that results in selection of a medium refinery configuration. Additionally, small differences in OPGEE input gravity and the assay gravity can result in different refinery configurations.
June 23, 2022	v1.0.0 of the OCI+ webtool was launched, covering 135 oil and gas resources (~50% of world’s oil and gas production). The input data for all the oil and gas models are from public sources.
April 5, 2023	V2.0.0 of the OCI+ webtool was launched, covering 500+ oil and gas resources (~70% of world's oil and gas production). The changes from the v.1.0.0 of the Webtool includes: Renamed “Upstream” supply chain segment to “Production”, renamed “Midstream” to “Refinery and Petrochemical Processing”, renamed “Downstream” to “End Uses”, and added transport and storage as a separate supply segment. Added plastics associated petrochemical processing emissions into the “Refinery and Petrochemical Processing segment.” Added plastics end-of-life emissions into the “End Uses” segment. Added more units of emissions intensity, such as gCO2eq/MJ O&G Produced and gCO2eq/MJ Petroleum Products. Added more variables, such as, Yield of Petchem Feeds Per Barrel. Added time series of emission profiles from 2015-2022 Some of the model inputs, such as Production Volume, are sourced from a variety of proprietary data providers.
April 24, 2024	V2.1.0 of the OCI+ webtool was launched, with several new webtool functions: Flaring Risk Map and Benchmark Crude. The changes from the v.2.0.0 of the Webtool includes: Added Flaring Risk Map Added Benchmark Crude Removed certain data attributes and oil and gas fields due to a data provider contract change.

Appendix

Data Inputs	Specifications			Units
Production Methods				0-1
	Downhole pump
	Water reinjection
	Natural gas reinjection
	Water flooding
	Gas lifting
	Gas flooding
	Steam flooding
	Oil sands mine (integrated with upgrader)
	Oil sands mine (non-integrated with upgrader)
Field properties
	Field location (Country)			name
	Field name			name
	Field age			years
	Oil production volume			bbl/d
	Number of producing wells			#
	Number of water injecting wells			#
	Production tubing diameter			inches
	Productivity index			bbl/psi-d
	Reservoir pressure			psi
	Reservoir temperature			◦F
	Offshore?			0-1
Fluid properties
	API gravity			deg. API
	Gas composition
		N₂		mol%
		CO₂		mol%
		C₁		mol%
		C₂		mol%
		C₃		mol%
		C₄+		mol%
		H₂S		mol%
Production practices
	Gas-to-oil ratio (GOR)			scf/bbl oil
	Water-to-oil ratio (WOR)			bbl water/bbl oil
	Water injection ratio			bbl water/bbl oil
	Gas lifting injection ratio			scf/bbl liquid
	Gas flooding injection ratio			scf/bbl oil
	Flood gas			NA
		1= Natural gas
		2= Nitrogen (N₂)
		3= Carbon Dioxide (CO₂)
		Fraction of CO₂ breaking through to producers		%
		Source of makeup CO₂		NA
			1= Natural subsurface reservoir
			2= Anthropogenic
		Percentage of sequestration credit assigned to the oilfield		%
	Steam-to-oil ratio (SOR)			bbl steam/bbl oil
	Fraction of required electricity generated onsite			-
	Fraction of remaining natural gas reinjected			-
	Fraction of produced water reinjected			-
	Fraction of steam generation via cogeneration			-
	Fraction of steam generation via solar thermal			-
Processing practices
	Heater/treater			NA
	Stabilizer column			NA
	Upgrader type
		0 = None
		1 = Delayed coking
		2 = Hydroconversion
		3 = Combined hydroconversion and fluid coking
	Associated Gas Processing Path			NA
		1= None
		2= Minimal: Dehydrator
	Flaring-to-oil ratio			scf/bbl oil
	Venting-to-oil ratio (purposeful)			scf/bbl oil
	Volume fraction of diluent			-
Land use impacts
	Crude ecosystem carbon richness
		Low carbon richness (semi-arid grasslands)		NA
		Moderate carbon richness (mixed)		NA
		High carbon richness (forested)		NA
	Field development intensity
		Low intensity development and low oxidation		NA
		Moderate intensity development and moderate oxidation		NA
		High intensity development and high oxidation		NA
Crude oil transport
	Fraction of oil transported by each mode
		Ocean tanker		-
		Barge		-
		Pipeline		-
		Rail		-
		Truck
	Transport distance (one way)
		Ocean tanker		Mile
		Barge		Mile
		Pipeline		Mile
		Rail		Mile
		Truck		Mile
	Ocean tanker size, if applicable			Ton
Small sources emissions				gCO₂eq/MJ

Ethane	20.98lb/ft³ liquid ethane; 2.2 lb/kg; 5.614 ft³/bbl
Propane	42gal/bbl; 1920 g/gal propane density; 1923 g/gal lpg
Butane	42gal/bbl; 2213 g/gal butane density; 1923 g/gal lpg
Pentanes+	84950btu/gal; 1923 g/gal; 42 gal/bbl; .110 mmbtu/gal

---

Footnotes

1. Climate TRACE is a coalition of NGOs and companies who seek to bring transparency to global GHGs with greater accuracy and speed through the use of satellite data and advancing modeling techniques.https://www.climatetrace.org/ ↑

2. Elvidge, C.D.; Zhizhin, M.; Baugh, K.; Hsu, F.-C.; Ghosh, T. Methods for Global Survey of Natural Gas Flaring from Visible Infrared Imaging Radiometer Suite Data. Energies 2016, 9, 14.https://doi.org/10.3390/en9010014 ↑

3. https://www.nature.com/articles/s41467-021-25017-4↑

4. https://www.elgas.com.au/blog/453-the-science-a-properties-of-lpg ↑

5. https://www.investopedia.com/terms/s/sourcrude.asp ↑

6. Source: EJScreen: Environmental Justice Screening and Mapping Tool.https://www.epa.gov/ejscreen ↑

7. Source:https://catalog.data.gov/dataset/tiger-line-shapefile-2019-nation-u-s-current-american-indian-alaska-native-native-hawaiian-area ↑

8. Source: FracTracker 2021 National Well File.https://www.fractracker.org/data/data-resources/ ↑