What happened
Earthview Corporation makes the BluBird Continuous Monitoring System. Each fixed node draws air over three metal oxide sensors and records wind, humidity and temperature. Cloud software judges whether an emission is present and estimates its location and size. Earthview's approved method says that a machine learning-based model for each sensor predicts its clean-air resistance. EPA's memo describes a "digital twin" that predicts the same baseline. Earthview applied on 3 February 2025, citing controlled releases at Colorado State University's METEC facility from February to April 2024. Earthview's analysis put the 90% probability of detection at about 2.4 kg/hr for single releases. EPA notes that this analysis weighted leak duration and gave a much lower figure than the facility's public reports. EPA found that the sensors would identify emissions of about 5 kg/hr within a reasonable time, subject to environmental factors. It approved the system for periodic screening on 30 September 2025. EPA accepted that field use will place fewer sensors than the tests did. Earthview's application cited field detections under 5 kg/hr at one upstream site. No emission cuts are reported.
Sources
-
Regulatory document, by authors with no tie to the system.
EPA has determined that The BluBird CMS meets the periodic screening requirements for the 5 kg/hr, 10 kg/hr, and 15 kg/hr detection thresholds. Additionally, EPA is approving Earthview's solution for use by an owner or operator, on an affected facility, for the alternative periodic screening process
EPA found Earthview's analysis to be reflective of their potential application in the field and that subject to environmental factors, their Blubird sensors would identify methane emissions of approximate 5 kg/hr coming from affected equipment within a reasonable time frame.
-
Report, by the makers. Context or contrary evidence.
This is combined with a machine learning-based model constructed for each sensor based on measured resistance and atmospheric conditions.
Details
AI system
- Workflow
-
Each BluBird node drew air over three metal oxide sensors and measured humidity, temperature and wind. A model for each sensor predicted its clean-air resistance so that readings could be corrected for drift and weather. Cloud software used inverse plume models to estimate emission rates and notified the operator when a plume passed a set level.
Follow-up
- Conditions for replication
-
Wind speed, relative humidity and temperature must stay within the method's limits. The number and placement of nodes must follow the method's siting requirements.
Entry details
- Sector
-
Climate change mitigation / methane emissions monitoring.
- Purpose
-
Earthview built the BluBird Continuous Monitoring System to detect, locate and size methane emissions at oil and gas sites with a network of fixed sensors.
31 more questions not yet checked
model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, AI contribution, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.
Cite this record
Peter H. Vartanian. (2026). BluBird sensors meet EPA's 5 kg per hour methane screening threshold (AIBD-1133). Artificial Intelligence Benefits Database, edition of 1 November 2026. https://petervartanian.xyz/aibd/records/AIBD-1133.html (archived at https://doi.org/10.7910/DVN/AADSII, version 7.0)