How NOAA measures atmospheric CO₂ at Mauna Loa

Measuring a proportion, not counting emissions

The amount of atmospheric CO₂ is commonly expressed as a dry-air mole fraction, in parts per million or ppm. It describes the proportion of carbon dioxide molecules relative to the total dry air. It is not equivalent to tons emitted in a year and does not allow emissions to be directly attributed to a facility.

NOAA’s measurement documentation explains how this proportion is observed at Mauna Loa. The record requires a calibrated instrument and criteria for selecting representative measurements. The observatory does not operate like a chamber that counts molecules across the entire planet: it obtains observations at one location, using a documented procedure and in relation to other measurement networks.

What the instrument does

NOAA describes the cavity ring-down spectroscopy technique, known as CRDS, used at Mauna Loa since April 2019. The method uses the way absorption changes the behavior of light inside a cavity. Reference standards and instrument checks are needed to relate that signal to a mole fraction.

Calibration is not a formality added at the end. It links the instrument’s response to a measurement scale and makes it possible to detect changes that could be confused with variation in the observed air. If an instrument changes, a responsible interpretation needs to know how continuity of the record is maintained. The methods page identifies techniques and procedures; it does not replace the metadata for each published data point.

From observations to averages

NOAA’s CO₂ trends page distinguishes between time-based products. A daily value, a monthly value, and an annual average do not describe the same thing. The annual average summarizes a complete period and does not preserve the full seasonal oscillation. Comparing a monthly maximum with an annual average would introduce a difference in definition, in addition to any actual change over time.

The Mauna Loa annual averages file specifies units, provenance, and quality notices. The version created on 9 October 2026 contains, among others, these values: 316,91 ppm in 1960, 369,71 ppm in 2000, and 427,35 ppm in 2025. These are values from that product and version, not a global average or an instantaneous reading from today.

What the graph shows

The accompanying resource plots all annual averages from 1959 to 2025 included in that file. The line connects annual observations to make them easier to read; it does not turn the record into a continuous measurement or show monthly values. The vertical axis is fitted to the range of the series and does not start at zero: read its values, not just the visual slope.

The file itself warns that the data may be revised through recalibration and quality control. The most recent figure in this graph corresponds to a complete year; no assumed annual average for 2026 is added before the year ends. Keeping the file date makes it possible to identify which version is being cited.

An interruption that must remain visible

NOAA records an interruption in measurements at Mauna Loa following the November 2022 eruption and their resumption in July 2023. Observations from Maunakea were used for part of that interval. This notice matters because a long series should not be presented as if all its observations always came from the same place and procedure without incident.

The Scripps Keeling Curve provides historical context for CO₂ observations. NOAA’s file acknowledges Scripps data for an initial part of the record. Therefore, linking to both institutions does not make all those years two independent measurements: the shared provenance must be examined.

What we can conclude and what question remains open

The series documents a sustained increase in the annual mole fraction in this product over the period shown. Studying causes, global distribution, or emissions flows requires additional observations and analyses. A concentration graph is not, by itself, an emissions inventory.

The term dry air maintains a comparable definition by removing variation due to water vapor. The table summarizes other decisions worth preserving when sharing a figure. Understanding them makes it possible to use the record as quantitative evidence without attributing a different precision or geographic scope to it than its source publishes.