Greenhouse gas emissions are monitored through a combination of direct measurement, calculation-based estimates, satellite observation, and ground-level sensor networks. The right approach depends on what you’re monitoring: a single facility, a supply chain, a city, or an entire region. Most organizations use a mix of methods, pairing direct hardware measurements where possible with standardized calculations to fill in the gaps.
Direct Measurement at the Source
The most precise way to monitor emissions from a facility is a continuous emission monitoring system, or CEMS. This is a package of analyzers, sensors, and software installed directly at the point where gases leave a building, smokestack, or vent. The system measures pollutant concentrations in real time and converts them into emission rates that can be compared against regulatory limits. In the United States, the EPA requires CEMS at certain industrial facilities for ongoing compliance.
A related approach, called predictive emission monitoring, skips the direct gas analyzer and instead uses measurements of operating conditions (temperature, fuel flow, equipment load) to estimate emissions through equations calibrated against actual measurements. This can be cheaper to maintain, though it’s less widely accepted for regulatory purposes.
The sensor technology inside these systems matters. Traditional infrared absorption sensors (NDIR) are the workhorse of CO2 monitoring but drift by roughly 0.3 ppm per day, requiring hourly calibration. Newer laser-based instruments using cavity ring-down spectroscopy are far more stable, drifting less than 0.38 ppm over an entire 30-month deployment. After periodic calibration with known reference gases, laser-based systems achieve accuracy within 0.02 to 0.17 ppm. Their stability means less hands-on maintenance, which can offset the higher upfront cost.
Calculation-Based Inventories
Most organizations don’t measure every emission directly. Instead, they build a greenhouse gas inventory using a simple formula: multiply an activity (like fuel burned or kilometers driven) by a published emission factor, then convert the result to metric tons. The GHG Protocol, the most widely used corporate accounting standard, structures it this way:
- CO2 from fuel: Fuel consumed (in liters, gallons, or kilograms) × emission factor (kg CO2 per unit) ÷ 1,000 = metric tons CO2
- Other gases: The same approach applies to methane (CH4) and nitrous oxide (N2O), using gas-specific emission factors for each activity
- CO2 equivalent: Each gas is then multiplied by its global warming potential (GWP) to convert everything into a single unit. Methane has a GWP of 21, meaning one ton of methane traps as much heat as 21 tons of CO2. Nitrous oxide has a GWP of 310.
The EPA publishes a regularly updated Emission Factors Hub, most recently refreshed in January 2025. It covers purchased electricity, vehicle fleets, business travel, employee commuting, and product transport. The electricity factors are drawn from the EPA’s eGRID database, which breaks emission rates down by regional power grid, so a company in the Pacific Northwest (with more hydropower) will report lower electricity emissions than one drawing from a coal-heavy grid in the Midwest. The 2025 update added grid loss percentages for each subregion for the first time, letting organizations account for energy lost during transmission and distribution.
If you’re building an inventory that needs external credibility, ISO 14064-1:2018 lays out the formal requirements. It specifies how to design, develop, manage, report, and verify an organizational GHG inventory. It’s program-neutral, meaning it works alongside whatever reporting framework or regulation applies to you.
Satellite and Remote Sensing
Satellites have transformed large-scale emissions monitoring. Japan’s GOSAT, launched in 2009, was the first satellite built specifically to measure CO2 and methane from space. NASA’s Orbiting Carbon Observatory-2 (OCO-2), launched in 2014, pushed precision much further. OCO-2 collects measurements three times per second along a narrow ground track, with each reading covering less than 3 square kilometers. After bias correction, its CO2 measurements are accurate to within 0.5 ppm (median difference) when compared against ground stations, with root-mean-square differences typically below 1.5 ppm.
For context, ground-based stations achieve about 0.07 ppm precision, but there are only around 147 of them worldwide. Satellites trade some precision for global coverage, filling in the vast stretches of ocean, forest, and developing regions where no ground station exists. Earlier instruments like SCIAMACHY (2002 to 2012) had much coarser resolution, covering 30 by 60 kilometer patches with sensitivity of only 4 to 8 ppm. Each generation of satellite gets sharper.
Satellite data is particularly valuable for identifying large methane leaks from oil and gas infrastructure, landfills, and agricultural regions that would otherwise go undetected. These observations complement ground-level monitoring by flagging hotspots that warrant closer investigation.
Ground-Based Monitoring Networks
NOAA’s Global Monitoring Laboratory operates a network of surface stations that serve as the backbone of long-term atmospheric tracking. These stations follow strict quality control and standard operating procedures tied to common, traceable measurement standards, which makes their data reliable enough to validate satellite observations and improve climate models. All the data is publicly available.
In Europe, the Integrated Carbon Observation System (ICOS) fills a similar role. These networks don’t tell you who emitted what. They tell you what’s actually accumulating in the atmosphere, which is the ultimate check on whether reported emissions match reality.
Monitoring Methane From Agriculture
Livestock are a major methane source, but measuring their emissions is tricky because animals move freely and emit gas continuously. The gold standard is a respiration chamber: a sealed enclosure where exhaust air is sampled every 15 minutes by an infrared gas analyzer. These chambers are expensive to build and maintain, and they only measure a few animals at a time under artificial conditions.
Portable laser methane detectors offer a practical alternative. The device shoots an infrared laser beam at an animal’s nostril and measures the methane concentration along the beam’s path, taking readings every half second. The laser wavelength is fixed at 1,653 nanometers, which corresponds precisely to methane’s absorption signature. Researchers have validated these handheld detectors against respiration chambers and found them useful for estimating emissions without disturbing the animals’ normal behavior. They’re far cheaper than chambers, making it feasible to survey larger herds.
Low-Cost Sensor Networks
A newer approach uses networks of inexpensive sensors distributed across a city or region. NIST has been developing a low-cost sensor platform built from commercially available components and microcontrollers, costing a fraction of traditional monitoring equipment. The sensors are designed to be easily assembled and deployed by end users in varied locations, with data reported securely in real time through an internet-connected backend.
The trade-off is accuracy. These sensors are less precise than research-grade instruments, and they’re susceptible to drift over time and interference from changing weather. NIST is working to reduce uncertainty to the 1 ppm level through frequent calibration, real-time data corrections, and custom algorithms. Even with those limitations, dense networks of cheap sensors can map emission patterns across a city in ways a handful of expensive stations never could.
Third-Party Verification
Monitoring only matters if the numbers are trustworthy. Third-party verification involves an independent auditor reviewing an organization’s emissions data, methodology, and supporting records. The process typically includes on-site facility visits, identification of errors or inconsistencies, and a report that the organization can use to correct and finalize its inventory. For established verification programs, this process generally takes three to six months.
Setting up a credible verification system requires defined protocols, a process for accrediting qualified verifiers, and safeguards against conflicts of interest (since the facility being audited often hires the verifier). California’s cap-and-trade program and the EU Emissions Trading System both mandate third-party verification. For voluntary corporate reporting, frameworks like the GHG Protocol and ISO 14064 strongly encourage it as a way to build stakeholder confidence in disclosed numbers.