Methane is the primary component of natural gas, so every undetected release is product lost before the sales meter and a reportable event under EPA's methane rule. That combination of revenue leakage and compliance exposure is what makes detection a priority for operators.
Satellites, drones, and ground sensors get discussed constantly in oil & gas circles. But few operators actually understand how these systems work, where they fail, and what that means for facility-level decisions. This guide breaks down the mechanics, step by step.
Key Takeaways
- Methane molecules absorb radiation at specific wavelengths: 3.2-3.55 µm (MWIR), 7.2-8.3 µm (LWIR), and 2.31-2.39 µm (SWIR)
- Satellites (TROPOMI, GHGSat, PRISMA, EnMAP) operate at global, regional, and facility scales
- Ground-based sensing fills the gaps satellites miss, especially for continuous facility monitoring
- Detection follows four stages: capture, spectral analysis, false-positive filtering, quantification
- Defensible emissions data requires a layered multi-sensor approach
What Is Remote Sensing for Methane Detection?
Remote sensing uses satellite, aerial, or ground-based sensors to detect methane concentrations without physical contact. It relies entirely on light absorption physics: methane absorbs specific infrared wavelengths, and sensors measure that absorption pattern.
Methane is invisible and odorless, so manual detection at scale is essentially impossible without instrumentation.
It is not a substitute for walking LDAR surveys or continuous ground monitoring—it complements both with broader spatial coverage.
Even sophisticated missions face limits. MethaneSAT launched in March 2024, then lost contact with mission control on June 20, 2025.
The official anomaly investigation logged nearly 400 failed reconnection attempts over 21 days before declaring the mission unrecoverable; the root cause was never determined. Space-based instruments still carry real operational risk.
Those limits help explain why operators combine platforms rather than relying on one sensor class:
- Satellite: Hyperspectral and multispectral spectrometers for basin-scale screening
- Aerial / drone: Mobile, targeted surveys over facilities and pipeline corridors
- Ground-based: Optical gas imaging, laser, and acoustic sensors for continuous site coverage
The absorption physics stays the same across all three. What changes is spatial resolution and how often you get a fresh look.
How Does Remote Sensing Detect Methane?
Every detection method, space or ground, follows the same sequence: capture light, analyze the spectrum, filter noise, quantify the result.
Initiation: Data Capture
Detection starts when a sensor captures reflected or emitted radiation over a target area. Satellite systems work on scheduled orbital passes: passive, continuous global sweeps. Ground sensors, by contrast, run continuously and automatically.
The bottleneck: clouds, weather, and low reflectance surfaces like water or dense vegetation block usable readings. This is a documented limitation for TROPOMI and other SWIR-based systems. One peer-reviewed 2024 urban study found a retrieval success rate of just 3% in evaluated observations.
Core Operation: Spectral Absorption Analysis
Methane molecules absorb radiation at specific wavelengths: 3.2-3.55 µm (MWIR), 7.2-8.3 µm (LWIR), and 2.31-2.39 µm (SWIR). Sensors measure how much light goes "missing" at those bands compared to background radiation.
Algorithms then take over:
- Matched filtering compares observed spectra against known methane absorption profiles
- Attenuation analysis measures signal loss across the plume
- Inversion models convert that loss into a column concentration estimate
Resolution and revisit frequency determine what you can actually see. A sensor with kilometer-scale resolution catches a regional hotspot. One with meter-scale resolution can isolate a single facility.
Regulation/Control: Filtering False Positives
Raw plume detections aren't automatically real. They need correction before anyone calls them confirmed emissions:
- Wind speed and direction
- Surface albedo
- Water vapor interference
- Instrument noise
Researchers often cross-check plume orientation against wind data such as ERA5 reanalysis. A peer-reviewed landfill methane study rotated satellite observations into wind coordinates and looked for consistent downwind enhancement to rule out noise. That same study found ERA5 wind data produced a 9% low bias against GEOS-FP winds, a reminder that even the correction step carries uncertainty.
Without this stage, operators and regulators drown in false alarms instead of actionable data. At the facility level, that same filtering problem is what Well Checked's Zentinal Core™ addresses.
It fuses video, acoustic AI, and Long-Wave Infrared optical gas imaging, then uses a roughly two-day AI site-learning cycle to learn what "normal" looks like at a given wellsite. Once that baseline exists, the system flags only true fugitive anomalies, not routine process emissions, cutting through what the company calls "needle in stacks of needles."
Output/Result: Quantification and Reporting
The end product is an estimated emission rate, typically expressed in kg/hr or tons/year, tied to a location and timestamp. That number feeds into national inventories, carbon market verification, and operator LDAR programs.
In the Permian Basin, a Science Advances study found top-down satellite-based emissions estimates of 2.9 Tg/yr, more than double the 1.2 Tg/yr EPA bottom-up inventory estimate for the same basin. That gap is why accurate quantification is essential, not optional.
On the facility side, Well Checked's Zentinal IQ™ activates only after Core validates an event. It estimates methane volume, event duration, and emissions rate with LWIR OGI-based volumetric estimation and AI-refined plume analysis.
The output is a regulatory-ready log, not a raw guess, aligned with the defensible records operators need for LDAR and inventory reporting.
Satellite vs. Ground-Based Methane Detection
Different systems serve different jobs. Here's how the major satellite platforms stack up:
| System | Resolution | Revisit | Published Detection Limit |
|---|---|---|---|
| TROPOMI | ~7×5.5 km | Daily (cloud-free) | Not universally stated |
| GHGSat | 25 m | Daily | 100 kg CH4/h (118 mcf/d) at 50% probability |
| Carbon Mapper Tanager | 30 m | Not stated | 90-180 kg/h (106-212 mcf/d) at 90% probability |
| PRISMA | 30 m | Not stated | Not stated |
| EnMAP | 30×30 m | 27 days (4 days off-nadir) | Not stated |
The coverage gap is real. Satellites miss diffuse, intermittent, or small emissions due to detection thresholds, weather interference, and orbital timing. A facility that leaks for six hours between passes may never register.
Ground-based sensing closes that gap with continuous, facility-level coverage:
- Optical gas imaging for day/night plume detection
- Acoustic monitoring for equipment and leak signatures
- AI-driven anomaly detection that filters false alarms
No orbital sensor matches 24/7 on-site monitoring. It isn't waiting for the next overpass.
Where Remote Sensing Fits Into Methane Management Workflows
The most effective methane programs layer detection types rather than relying on a single method:
- Broad satellite screening flags regions or basins showing elevated concentrations
- Facility-level investigation narrows down to specific sites and equipment
- Continuous monitoring watches those sites around the clock
- Repair verification confirms fixes actually worked Satellites excel at scanning large basins cheaply. Ground and edge sensors excel at continuous, weather-independent facility oversight — exactly where satellites fall short. Well Checked's Zensory.ai™ platform runs this continuous layer across remote oil & gas sites, including a continuous monitoring deployment in the Appalachian Basin. The platform combines video, LWIR optical gas imaging, and acoustic sensing. Data is structured to support EPA methane rule and OGMP 2.0-aligned reporting where satellite screening alone leaves blind spots.
Conclusion
Methane detection, whether from orbit or a wellsite fence line, works through the same layered process: capture, spectral analysis, filtering, quantification. That shared mechanism is what should shape how operators build their monitoring stack.
Operators who grasp these fundamentals can pair satellite screening for basin-wide awareness with continuous facility-level monitoring for the leaks satellites can't see. That combination cuts compliance risk and unplanned repair costs at the same time.
Frequently Asked Questions
What types of sensors detect methane for remote sensing?
Satellite spectrometers (TROPOMI, GHGSat, PRISMA, EnMAP), aerial and drone payloads, and ground-based optical gas imaging or acoustic sensors all detect methane. Each suits a different detection scale.
Was the methane tracking satellite lost?
Yes. MethaneSAT launched in March 2024 and lost contact with mission control in June 2025. It was declared unrecoverable, leaving a monitoring gap between TROPOMI's regional scans and GHGSat's facility-level tasking.
How accurate is satellite methane detection?
Accuracy depends heavily on sensor resolution and atmospheric conditions. GHGSat, for example, has a published detection limit around 100 kg CH4/hour, or 118 mcf/d, at 50% probability, and quantification carries real uncertainty tied to wind modeling.
Can remote sensing detect small or diffuse methane leaks?
Often not from space. Small, diffuse, or intermittent leaks frequently fall below satellite detection thresholds, which is why continuous ground-based or facility-level sensing matters for catching what satellites miss.
Why does methane detection matter for oil & gas operators?
Operators face EPA methane rule and OGMP 2.0 reporting obligations. Continuous detection catches leaks early, reducing both regulatory fines and the cost of delayed repairs.
How is methane emission data used for regulatory compliance?
Quantified emissions data feeds EPA, state agency, OGMP 2.0, SASB, and TCFD disclosure frameworks. These frameworks increasingly require defensible, continuous records rather than periodic snapshot reports.
