<p>Correlations of night-time atmospheric methane (CH<sub>4</sub>) and <sup>222</sup>Radon (<sup>222</sup>Rn) observations in Heidelberg, Germany, were evaluated with the Radon Tracer Method (RTM) to estimate the trend of annual CH<sub>4</sub> emissions from 1996–2020 in the catchment area of the station. After an initial 30 % decrease of emissions from 1996 to 2004, no further systematic trend but small inter-annual variations were observed thereafter. This is in accordance with the trend of emissions until 2010 reported by the EDGARv6.0 inventory for the surroundings of Heidelberg. We show that the reliability of total CH<sub>4</sub> emission estimates with the RTM critically depends on the accuracy and representativeness of the <sup>222</sup>Rn exhalation rate from soils in the catchment area of the site. Simply using <sup>222</sup>Rn fluxes as estimated by Karstens et al. (2015) could lead to biases in the estimated greenhouse gases (GHG) fluxes as large as a factor of two. RTM-based GHG flux estimates also depend on the parameters chosen for the night-time correlations of CH<sub>4</sub> and <sup>222</sup>Rn, such as the night-time period for regressions as well as the R<sup>2</sup> cut-off value for the goodness of the fit. Quantitative comparison of total RTM-based top-down with bottom-up emission inventories requires representative high-resolution footprint modelling, particularly in polluted areas where CH<sub>4</sub> emissions show large heterogeneity. Even then, RTM-based estimates are likely biased low if point sources play a significant role in the station/observation footprint as their emissions are not captured by the RTM method. Long-term representative <sup>222</sup>Rn flux observations in the catchment area of a station are indispensable in order to apply the RTM method for reliable quantitative flux estimations of GHG emissions from atmospheric observations.</p>