<p>Though they cover less than 3 % of the global land area, urban areas are responsible for over 70 % of the global greenhouse gas (GHG) emissions and contain 55 % of the global population. A quantitative tracking of GHG emissions in urban areas is therefore of great importance, with the aim of accurately assessing the amount of emissions and identifying the emission sources. The Weather Research and Forecasting model (WRF) coupled with GHG modules (WRF-GHG) developed for mesoscale atmospheric GHG transport, can predict column-averaged abundances of CO<sub>2</sub> and CH<sub>4</sub> (XCO<sub>2</sub> and XCH<sub>4</sub>). In this study, we use WRF-GHG to model the Berlin area at a high spatial resolution of 1 km. The simulated wind and concentration fields were compared with the measurements from a campaign performed around Berlin in 2014 (Hase et al., 2015). The measured and simulated wind fields mostly demonstrate good agreement and the simulated XCO<sub>2</sub> agrees well with the measurement. In contrast, a bias in the simulated XCH<sub>4</sub> of around 2.7 % is found, caused by relatively high initialization values for the background concentration field. We find that an analysis using differential column methodology (DCM) works well for the XCH<sub>4</sub> comparison, as corresponding background biases then cancel out. From the tracer analysis, we find that the enhancement of XCH<sub>4</sub> is highly dependent on human activities. The XCO<sub>2</sub> signal in the vicinity of Berlin is dominated by anthropogenic behavior rather than biogenic activities. We conclude that DCM is an effective method for comparing models to observations independently of biases caused, e.g., by initial conditions. It allows us to use our high resolution WRF-GHG model to detect and understand sources of GHG emissions quantitatively in urban areas.</p>