Long-term in situ atmospheric observations are essential for estimating surface fluxes and assessing temporal and spatial variability of greenhouse gases (GHGs). Achieving the World Meteorological Organization (WMO) compatibility goals for CO₂ and CH₄ dry mole fractions is critical for ensuring the accuracy of such measurements. However, compliance with these goals can be challenging even at nearby sites due to environmental and operational factors. This study presents a multi-year comparison of hourly atmospheric CO₂ and CH₄ dry mole fractions recorded by two independent monitoring programmes at the Mt. Cimone WMO/GAW Global Station in Italy. The air inlets are positioned approximately 90 m horizontally and 50 m vertically apart, and have different technical designs. Despite strong correlations and consistent seasonal and multiannual trends, the two datasets exhibit significant deviations, often exceeding the WMO compatibility goals for CO₂ throughout the year and for CH₄ primarily in winter and spring. Besides differences in data processing which led to large deviations during episodes with high GHG dry mole fractions, the discrepancies appear to be possibly driven by different factors. The different inlet positioning plays an important role. This is linked to measurement periods characterised by high temporal variability and different sensitivity to local biogenic influences during the active vegetative season. Additional contributors include the possible influence of human presence at the facilities. Accounting for these factors reduces the seasonal CO₂ bias by roughly 40 %, depending on the season, resulting in a residual bias of 0.20 – 0.35 ppm. For CH₄, mean seasonal deviations within the WMO compatibility goal occur from spring to autumn when periods characterised by low atmospheric variability and the absence of human presence are selected. The installation of a drying system at the CAMM laboratory was associated with a significant improvement in the agreement between the CAMM and CNR measurements. Finally, a preliminary comparison conducted after the relocation of the CAMM inlet in early 2026 showed a further reduction in the systematic differences, supporting the role of inlet configuration in the previously observed discrepancies. Our findings emphasise the complexity of interpreting discrepancies in GHG observations at closely located stations in mountainous environments and highlight the importance of standardised and centralised protocols, such as those implemented within the ICOS Research Infrastructure, to improve data harmonisation and ensure measurement compatibility.