A study of synthetic 13CH4 retrievals from TROPOMI and Sentinel-5/UVNS

Retrievals of methane isotopologues have the potential to differentiate between natural and anthropogenic methane sources types, which can provide much needed information about the current global methane budget. We investigate the feasibility of retrieving the second most abundant isotopologue of atmospheric methane ((CH4)-C-13, roughly 1.1% of total atmospheric methane) from the shortwave infrared (SWIR) channels of the future Sentinel-5/ultra-violet, visible, near-infrared, shortwave infrared (UVNS) and current Copernicus Sentinel-5 Precursor TROPOspheric Monitoring Instrument (TROPOMI) instruments. With the intended goal of calculating the delta C-13 value, we assume that a delta C-13 uncertainty of better than 1% is sufficient to differentiate between source types, which corresponds to a (CH4)-C-13 uncertainty of < 0.02 ppb. Using the well-established information content analysis techniques and assuming clear-sky, non-scattering conditions, we find that the SWIR3 (2305-2385 nm) channel on the TROPOMI instrument can achieve a mean uncertainty of < 1 ppb, while the SWIR1 channel (1590-1675 nm) on the Sentinel-5 UVNS instrument can achieve < 0:68 ppb or < 0:2 ppb in high signal-to-noise ratio (SNR) cases. These uncertainties combined with significant spatial and/or temporal averaging techniques can reduce delta C-13 uncertainty to the target magnitude or better. However, we find that (CH4)-C-13 retrievals are highly sensitive to errors in a priori knowledge of temperature and pressure, and accurate knowledge of these profiles is required before (CH4)-C-13 retrievals can be performed on TROPOMI and future Sentinel-5/UVNS data. In addition, we assess the assumption that scattering-induced light path errors are cancelled out by comparing the delta C-13 values calculated for non-scattering and scattering scenarios. We find that there is a minor bias in delta C-13 values from scattering and non-scattering retrievals, but this is unrelated to scattering-induced errors.