How reliable are process-based ²²² radon emission maps? Results from an atmospheric ²²² radon inversion in Europe

Abstract. The radioactive noble gas radon (222Rn) is a suitable tracer for atmospheric transport and mixing processes that can be used to evaluate and calibrate atmospheric transport models or to estimate greenhouse gas (GHG) emissions using the so-called radon tracer method (RTM). However, these applications require reliable estimates of the 222Rn fluxes from the soil. This study evaluates two process-based 222Rn flux maps in central Europe in 2021 using the flux results from a 1-year 222Rn inversion. The maps are based on different soil moisture reanalysis products (GLDAS-Noah and ERA5-Land), which are used to describe the diffusive 222Rn transport in the soil. The 222Rn inversion was conducted using the CarboScope-Regional inversion system and observational data from 17 atmospheric sites in central Europe in 2021. We observe that, in particular, the ERA5-Land-based 222Rn flux map underestimates the data-driven fluxes from the inversion. Our inversion yields ca. 20 % (GLDAS-Noah) to almost 100 % (ERA5-Land) larger 222Rn fluxes than the respective process-based prior fluxes within a domain covering Germany. Also, the temporal variability seems to be underestimated by the process-based flux maps. Using a flat (uniform) prior inversion, we found a significant anti-correlation of −0.6 (and −0.8) between the posterior 222Rn flux and the GLDAS-Noah (and ERA5-Land) soil moisture time series, indicating that soil moisture is an important driver for the temporal variability in the 222Rn fluxes. To investigate the robustness of our flux estimates, we run the inversion with three different transport models (STILT, FLEXPART, NAME). The respective annual mean flux results agree within ca. 10 %.