Progress towards net‑zero targets depends on independent verification of urban greenhouse gas emissions. Cities account for ~75% of energy‑related CO₂ emissions, yet their quantification remains challenging due to spatially complex sources, heterogeneous meteorology, and limited observation coverage. We deploy seven EM27/SUN ground‑based Fourier‑transform spectrometers, within weatherproof enclosures, across the City of Edinburgh to measure column‑integrated carbon dioxide_,methane and carbon monoxide (XCO₂, XCH₄, XCO), which are less sensitive than <em>in situ</em> sensors to highly localised surface signals. A comprehensive calibration and quality-control framework leads to&nbsp;inter‑instrument precision of 0.26&ndash;0.62 ppb for XCH₄ and 0.05&ndash;0.07 ppm for XCO₂. The influence of the weatherproof enclosure dome on XCO₂ and XCH₄ retrievals is negligible but introduces a correctable XCO bias. City-wide observations reveal&nbsp;clear seasonal cycles and detectable intra-city gradients, with XCH₄ showing consistent upwind-downwind behaviour, while XCO₂ and XCO are strongly influenced by background variability and atmospheric transport. Comparison with the ICON-ART mesoscale model (6.5 km resolution) shows that total column variability is dominated by background conditions, with anthropogenic contributions typically &lt; 0.2&ndash;0.3 ppm. These signals exceed model minus observation differences in only ~5% of cases, highlighting the challenge of isolating urban emissions from column measurements. Overall, our results demonstrate that dense column networks provide robust, scalable constraints on urban greenhouse gas emission estimates and valuable validation for atmospheric transport models and satellite retrievals, forming a foundation for inverse modelling with the UK-wide network of similar instruments.