Spectroscopy, SWIR-TIR, and isotopes from orbit.

This is the most under-told story in the project. It ran as a quiet technical work package and produced one of the clearest results.

The problem

Methane's thermal-infrared absorption is imperfectly described in the standard spectroscopic databases. That matters because the strongest part of the methane band — the Q-branch, around 1300–1310 cm⁻¹ — is exactly where the extra information lives. In the earlier Methane+ project, RAL tried to use it and had to give up: fit residuals there were too large. The retrieval was confined to 1232–1290 cm⁻¹. [TN1 §3.4]

What SMART-CH4 did

SPASCIA reviewed three CH₄ cross-section lookup tables and found lmtran — which includes a line-mixing model — reproduces IASI spectra best in the TIR. They reparameterised it onto RAL's spectral grid (polynomial in temperature, precision better than 1/1000) and delivered it. RAL rebuilt their RTTOV coefficients around it. [TN1 §3.3, §3.4]

What it bought

The Q-branch residuals largely disappeared, and the fit window could be opened up to 1232–1340 cm⁻¹ — the range that had been ruled out three years earlier. The consequences:

Plot of IASI brightness-temperature fit residuals in Kelvin against wavenumber from about 1230 to 1330 per centimetre, comparing the standard cross-section table (black) against the reparameterised SPASCIA lmtran table (red). The original spectral range used, 1232-1290 per centimetre, is shaded. Beyond that range, around the Q-branch near 1300-1310 per centimetre, the standard table's residuals spike sharply while the SPASCIA table stays comparatively flat. — IASI brightness-temperature fit residuals: the standard cross-section table (black) spikes sharply around the Q-branch (~1300–1310 cm⁻¹), just beyond the original 1232–1290 cm⁻¹ range (shaded); the reparameterised SPASCIA `lmtran` table (red) stays comparatively flat there, which is what justified widening the fit window.

More vertical information. Degrees of freedom for the methane profile rose by 0.3–0.4, from a range of 1.3–2.2 to 1.3–2.6 — meaning the retrieval can now genuinely separate upper tropospheric / lower stratospheric methane from the rest of the column.
Better agreement with independent models in the 6–12 km layer, compared against CAMS flux inversions with averaging kernels applied.
A step toward isotopes from space. The wider window contains stronger ¹³CH₄ features. Retrieved δ¹³C-CH₄ artefacts shrank from spatial variations of ±50‰ to about ±20‰, and from ±10‰ down to trends smaller than 0.33‰/decade in latitude-band averages. For context, the real atmospheric signal is roughly 1‰ of variation around −47.5‰ — so these are still artefacts, not measurements. But the gap closed substantially, and the direction of travel is clear.

Laboratory work continues in parallel: quantum-cascade-laser absorption spectra from the University of Namur were analysed against lmtran, an instrument response function derived, and a framework established for extracting improved line-by-line parameters (Lorentzian broadening, line-mixing coefficients) from the four remaining micro-windows.

Why it matters

Different methane sources carry different isotopic fingerprints — a cow and a gas leak are not the same carbon. Measuring δ¹³C-CH₄ from orbit would let inversions attribute emissions to sectors, not just locations. SMART-CH4 did not get there. It did establish, with numbers, how far away it is and what closes the gap — see Isotopes.

Read more:

Isotopes results — what the wider window buys for δ¹³C-CH₄ detectability
Technical Note 1 — full spectroscopy methodology (§3.3–3.4)
Phase 2, WP2 — TIR spectroscopy and IASI-NG remain an open priority