The Groningen natural gas field in the Netherlands – one of Europe’s major gas fields – deploys a “production cluster” infrastructure with extraction, some processing and storage in a single facility. This region is also the site of intensive agriculture and cattle operations. We present results from a multi-scale measurement campaign of methane emissions, including ground and airborne-based estimates. Results are compared with inventory at both the facility and regional level. Investigation of production cluster emissions in the Groningen gas field shows that production volume alone is not a good indicator of whether, and how much, a site is emitting methane. Sites that are nominally shut down may still be emitting, and vice-versa. As a result, the inventory emission factors applied to these sites (i.e. weighted by production) do a poor job of reproducing individual site emissions. Additional facility-level case studies are presented, including a plume at 150 ± 50 kg CH4 hr–1 with an unidentified off-shore emission source, a natural gas storage facility and landfills. Methane emissions in a study region covering 6000 km2 and including the majority of the Groningen field are dominated by biogenic sources (e.g. agriculture, wetlands, cattle). Total methane emissions (8 ± 2 Mg hr–1) are lower than inventory predictions (14 Mg hr–1) but the proportion of fossil fuel sources is higher than indicated by the inventory. Apportionment of methane emissions between thermogenic and biogenic source types used ethane/methane ratios in aircraft flasks and ground-based source characterization. We find that emissions from the oil and gas sector account for 20% of regional methane, with 95% confidence limits of (0%, 51%). The experimental uncertainties bound the inventory apportionment of 1.9%, though the central estimate of 20% exceeds this result by nearly 10 times. This study’s uncertainties demonstrate the need for additional research focusing on emissions apportionment, inventory refinement and offshore platforms.