Methane sensors are affected by two main types of errors. Some are “random,” meaning that readings are sometimes too high and sometimes too low. Other errors are “systematic,” which means that a sensor is biased toward routinely overestimating or underestimating emissions.

Systematic errors don’t go away

Random errors tend to go away when large numbers of measurements are made, with overestimates cancelling out the underestimates. But systematic errors reinforce rather than cancel each other out.

A continuous monitor that quantifies methane emissions once every hour makes 8,760 measurements per year, so systematic errors can add up and have a significant impact.

Bias is money

Many metrics are used to describe methane monitors, and systematic error is one of the most important. When large numbers of measurements are performed, systematic error is what controls the accuracy of the aggregated reports. And this accuracy is what matters for OOOOb compliance and OGMP reporting.

It’s why we’ve spent extensive time and effort eliminating systematic errors from our methane point instrument—a continuous monitoring device designed for mass deployment.

Measuring performance at METEC

How do you find out if a continuous monitor has systematic errors? By knowing exactly what’s been emitted and comparing it with the measured value. This is done during controlled release tests, where sensor measurements are compared with the “ground truth.”

At Colorado State University’s Methane Emissions Technology Evaluation Center (METEC), sensor readings are compared against the ground truth, and systematic error is quantified as “regression bias.” Results of the 2024 long-duration tests of continuous monitors show that the SLB methane point instrument (identified in the results as Sensor S) has a regression bias of 4%. This means that across the large number of measurements typically made by a continuous monitor, the SLB technology can report total emissions with an error potentially as low as 4%.

Our instrument has the lowest regression bias among all the sensors evaluated at METEC—as much as an order of magnitude lower than some of the other sensors tested. Cleary, knowing the bias of what you buy is of critical importance.

Technology is moving fast. And it needs to. A true picture of methane emissions is not just there for compliance reporting. It’s crucial to incentivizing and prioritizing action. A sensor with low regression bias reveals that picture more accurately, providing the confidence necessary to take decisive remedial action.