Can radon-222 help to improve methane emission estimates? Results from a dual-tracer inversion

Atmospheric transport models cause a large part of the uncertainty in top-down estimates of greenhouse gas fluxes derived by atmospheric inversions. In particular, deficits in transport models, such as inadequate description of vertical mixing in the planetary boundary layer (PBL), can lead to systematic biases in the flux estimates. While their quantification is critical for reliable flux estimation, such model biases and uncertainties are difficult to assess. One way of evaluating the performance of atmospheric transport models is to compare the modelled with the measured activity concentration of the radioactive noble gas radon-222 (Rn), provided that the Rn fluxes are sufficiently well known. Rn is produced by the decay of radium-226 in the soil and diffuses through the soil pores into the atmosphere. As the Rn lifetime (3.8 days) is comparable to the ventilation time scale of the PBL, atmospheric measurements of Rn activity concentrations provide sensitive information on vertical mixing.

By comparing the mismatch between the modelled (using the Stochastic Time-Inverted Lagrangian Transport model, STILT, and posterior flux estimates) and measured concentrations of methane (CH4) with that of Rn, we found significant correlations for many sites in Europe (the median correlation coefficient of all sites is r=0.6), indicating that a large part of the variability in the CH4 and Rn model-data mismatch can be explained by transport model errors. To exploit this information, we set up a joint inversion for (the targeted tracer) CH4 and Rn, taking into account realistic prior uncertainties and making use of the fact that the transport model error is correlated between the two gases. By comparing the results of the CH4-Rn inversion with those of a single-tracer CH4-only inversion, we assess the potential of Rn to improve CH4 emission estimates and highlight the importance of having accurate Rn flux maps.