Colorado Ongoing Basins Emissions (COBE): Measurements and Source Attribution

Ethan Rimelman3, Jenna Brown¹, Michael Moy¹, Arthur Santos¹, Olga Khaliukova², William Daniels², Callan Okenberg², Dorit Hammerling², Daniel Zimmerle¹, Anna Hodshire3

¹CSU Energy Institute, Colorado State University, ²Dept. of Applied Mathematics and Statistics, Colorado School of Mines, 3Systems Engineering Dept., Colorado State University

Download Poster PDF

Background

COBE is a joint project partnering with three aerial instrument vendors (Bridger Photonics, GHGSat, Insight M) to measure statewide methane emissions from upstream oil & gas facilities. The project will inform the 2026 Colorado Greenhouse Gas (GHG) Intensity Verification Rule through two independent modeling approaches.

Each vendor conducted three different campaigns (spring, summer and fall) in 2024 over upstream facilities throughout the state, which were grouped by the modeling team into 3 basin groups and 4 Prototypical Site (PS) classes.

Primary Objectives:​​

  • Collect representative methane measurements via aerial campaigns (this poster)​
  • Develop measurement-informed inventories (MIIs)​ (Michael Moy, Jenna Brown)
  • Compare MIIs to operator-reported emissions to provide ratios (Jenna Brown)

 


Methodology

Figure 1: Simplified data pipeline for collecting and implementing measurements into model

Aircraft flight occurs (operator not pre-informed)→CSU receives data, sorts to individual facilities/operators→Operators receive sorted data, respond with event data→CSU/study team anonymizes data; creates scaling model separating fugitives from activities →CDPHE receives emissions report (annon. aggregated data)
←
Logic Tree. Has Gas Lift? If "Yes," then COBE-PS1. If "No," Has Tank? If "No," then COBE-PS6. If "Yes" then if "No Flares or VRUs) then COBE-PS2 or if "Has Flares or VRU/s (engines
Figure 2: Logic tree for classifying sites into Prototypical Sites (PS)

Based on publicly available reporting data, facilities of interest are delegated into Prototypical Site (PS) Configurations


CSU MII Process:

  1. Collect representative measurements of CO’s upstream O&G facilities
  2. Attribute a likely source cause for each significant emission
  • This process involves using operator provided cause analysis (when available), along with aerial imagery and metadata
  • 12 operators in the state participated, representing approximately 77% of facilities in the state
Aerial view of a oil and gas site with a sample of what a plume looks like when captured by aerial instrument
Figure 3: (Generic) Example of methane plume image captured by aerial instrument

3. Compare emissions to bottom-up inventory (simulated using MAES, Mechanistic Air Emissions Simulator)

  • This process involves using operator provided cause analysis (when available), along with aerial imagery and metadata
  • 12 operators in the state participated, representing approximately 77% of facilities in the state

4. Incorporate emissions that are not already captured in inventory -> MII

Top chart: MAES Simulation (14 days, 300 MC) Instantaneous Emissions at Site Level (Normal Emissions Only) Probability Density Function (PDF). Probability on y-axis and Aggregated Emissions (kg/h) on x-axis. Bottom chart: Cumulative Distribution Function (CDF) with Cumulative Probability on y-axis and aggregated emissions (kg/h) on x-axis. Text on graph reads: "We expect emissions ~0-8 kg/hr. Aerial value: 15 kg/hr."
Figure 4: Example of MAES predicted emission profile. Top: Probability distribution developed from inventory (e.g. ONGAIER) (Colorado’s reporting database) Bottom: Cumulative distribution function (CDF) of inventory

Results

Map of Colorado with Aerian Scans marked. Aerial Scans include Bridger, InsightM, GHGSat, COBE_DJ_Basin, COBE_Piceance_basic, ONGAEIR Sites OpenStreetMap
Figure 5: Map of measured facilities across Colorado. The partitioning of each basin (Denver-Julesberg)  (DJ) [blue], Piceance [grey], and 'Other' were defined by study partners at CDPHE

Table 1: Counts of scanned facilities across all measurement campaigns. ‘Repeat Facilities’ is applied when a facility is measured more than once in 24 hours.

Aerial CompanyTotal ScansUnique FacilitiesRepeat Facilities
Bridger7,0433,7081,836
GHGSat10,9157,2093,057
Insight M15,1277,7494,296
Campaign Total33,08510,7717,732
Facilities scanned in all basins by PS class. List includes "At least one vendor, "Bridger," "GHGSat," and "Insight M."
Figure 6: Aerial coverage: 91.4% of facilities in the DJ basin, 96.8% in the Piceance basin, and 92.4% in other basins

Emissions that are determined to be from midstream activities (96), preproduction activities (44), maintenance actions (47), or off of the reported site (40) are excluded from the MII model (but included in the public final dataset).

Table 2: Summary of facility-level detected emission rates measured in kg/hr by aerial measurement company and basin.

CompanyBasinMedianAverageMinMaxRange
BridgerDJ2.135.330.203189188
Piceance1.533.960.13581.981.7
Other2.095.390.20343.743.5
GHGSatDJ10511834248214
Piceance2457.310157147
Other2946.58285277
Insight MDJ361137353346
Piceance4349.43143140
Other17333114111

References

  • Bell, C., Ilonze, C., Duggan, A., & Zimmerle, D. (2023). Performance of continuous emission monitoring solutions under a single-blind controlled testing protocol. Environmental science & technology, 57(14), 5794-5805.
  • Ilonze, C., Emerson, E., Duggan, A., & Zimmerle, D. (2024). Assessing the Progress of the Performance of Continuous Monitoring Solutions under a Single-Blind Controlled Testing Protocol. Environmental Science & Technology, 58(25), 10941-10955.
  • Day, R. E., Emerson, E., Bell, C., & Zimmerle, D. (2024). Point Sensor Networks Struggle to Detect and Quantify Short Controlled Releases at Oil and Gas Sites. Sensors, 24(8), 2419.
  • Cheptonui, F., Emerson, E., Ilonze, C., Day, R., Levin, E., Fleischmann, D., … & Zimmerle, D. (2025). Assessing the Performance of Emerging and Existing Continuous Monitoring Solutions under a Single-blind Controlled Testing Protocol. Elementa Sci https://doi.org/10.1525/elementa.2025.00020
  • EPA. 2024d. 40 CFR Part 60 Subpart OOOOb—Standards of performance for crude oil and natural gas facilities for which construction, modification or reconstruction commenced after December 6, 2022. Available at https://www.ecfr.gov/current/title-40/chapter-I/subchapter-C/part-60/subpart-OOOOb

COBE Measurements Summary and Future Work


Average Emission Rate by Failure Type (Normal vs Abnormal)

Average Emission Rate by Failure Type (Normal vs Abnormal). Average Emission Rate (kg/h) on the y-axis. On the x-axis is Failure Type broken out by: Misc, Tank, Compressor, Flare, Heater, Separator, Other, and Well.
Figure 7: Average emission rate by measured equipment/source, from 999 measurements determined to be abnormal, and 892 normal emissions

After sorting emissions and labeling suspected cause, a combined uncertainty model is applied and measurements are incorporated into an MII (see other COBE posters by Jenna Brown & Michael Moy)

  • To incorporate measurements that are determined to be from a failure mode, pLeak (probability of detecting a leak) is used to input frequency into MAES MII.

Future work

  • What is the best way to leverage the different vendor’s detection capabilities (lower detection limit, number of facilities scanned, equipment identification, etc.)?
  • Can other measurement technologies, such as satellite instruments and continuous monitors, further improve this MII?

Table 3: Calculated likelihood of leak by associated equipment : pLeak = # failures measured / # units scanned

Sample sizepLeak
Compressors11,0150.0160
Miscellaneous emissions32,8650.0368
Flares23,9410.0038
Heaters118,7990.0026
Controlled Tanks74,0510.0028
Uncontrolled Tanks26,8540.0076

Acknowledgments and Contact Information

Final dataset is available to the public online, which has 2,102 emissions measured (and anonymized).

Funding for COBE was provided by the Colorado Department of Public Health and Environment Agreement #2024*3364.

Poster author: Ethan Rimelman | Master’s Student | Systems Engineering Department | Colorado State University | [email protected]

Project PI: Anna Hodshire | Assistant Professor | Systems Engineering Department | Colorado State University | [email protected]


References

COBE final report:
Brown, J. A.; Moy, M.; Santos, A.; Rimelman, E.; Okenberg, C.; Daniels, W. S.; Hammerling, D. M.; Zimmerle, D.; Hodshire, A. L. Colorado Ongoing Basin Emissions (COBE) Final Report.

COBE anonymized dataset:
Brown, J. A.; Hodshire, A. Colorado Ongoing Basin Emissions Study (COBE) Anonymized Final Data Set of Emissions Measurements, 2025. https://doi.org/10.5061/dryad.8kprr4z0p.

For other COBE-related posters, see:  

Equipment at the METEC Site