Find a gas in infrared light

MOPITT, short for Measurements of Pollution in the Troposphere, was Canada’s contribution to NASA’s Terra observatory. Terra launched on December 18, 1999. MOPITT began its science record in March 2000, measuring radiation that could be used to infer carbon monoxide in the lower atmosphere. Its instrument, host spacecraft and processed data products are related parts of a mission, not interchangeable names.

Carbon monoxide is written CO: one carbon atom and one oxygen atom. Carbon dioxide, CO₂, is a different molecule. Identifying which gas a product measures matters before using it to discuss emissions, atmospheric chemistry or environmental change. A map of CO is not automatically a map of carbon dioxide, all pollution or every greenhouse gas.

A satellite supplies a broad view that complements observations from aircraft and the ground. It also brings a particular footprint, viewing geometry and instrument response. MOPITT’s approximately 22-kilometre footprint makes it useful for regional atmospheric questions; it does not resolve the air in an individual room. The later lessons explain how to connect this broad view to evidence at other scales.

MOPITT used gas correlation spectroscopy. Radiation passed through an internal cell containing the target gas, and the amount of gas in that optical path was alternated. Comparing signals in the two states emphasized the target gas’s absorption pattern. The reference gas stayed inside the instrument; the instrument did not collect an air sample from every location it mapped.

The carbon monoxide channels used two spectral regions. Thermal infrared radiation near 4.7 micrometres is emitted by Earth and its atmosphere. Near-infrared radiation near 2.3 micrometres is mainly reflected sunlight. The two regions respond differently to the vertical distribution of CO, which creates opportunities to combine information rather than treating all channels as duplicates.

For an original signal example, one cell state produces 120 arbitrary detector units and the other produces 90. Their difference is 30 units. This subtraction illustrates a measurement step, not a concentration result: those units are not parts per billion. Calibration and a model of how radiation travels through the atmosphere are still needed to infer a gas distribution.

A detector difference is not yet a concentration

A retrieval is the calculation that estimates an atmospheric state from measured radiation. It uses a forward model to predict radiation for possible states, together with information such as temperature and water vapour. Different combinations of gas, surface properties and atmospheric conditions can affect the signal, so the inference must include uncertainty and additional constraints.

MOPITT’s design also included methane channels, but those did not yield an operational methane product. The presence of a channel in a diagram is therefore not evidence that a validated public product exists for that gas. Follow the documented product, not just the original instrument ambition.

The Version 9 study distinguishes thermal-only, near-infrared-only and combined retrievals. Thermal channels usually provide greatest sensitivity higher in the troposphere; the combined product can improve lower-tropospheric sensitivity under suitable daytime land conditions. Reflected sunlight is unavailable at night, and surface reflectance also matters. Combining channels does not remove these physical constraints.