Denitrification in soils is a very complex phenomenon due to the multiple factors it depends on. Therefore, making accurate predictions has been an elusive task. In this study we measured N₂O emissions daily during 7 days from a set of 20 undisturbed small soil cores that were subsequently scanned using X-ray micro-tomography. Macropores, particulate organic matter (POM) and the mineral matrix were detected based on a locally-adaptive segmentation method. We proposed an indicator based on the morphology of the soil micro-structure to predict the N₂O emissions. The indicator, Id_POM-air, relies on the hypothesis that more N₂O will be emitted when POM is occluded in the soil matrix, i.e. is located at large distances from the next air-filled pore, most likely leading to anoxic conditions, favorable to the production of N₂O. We computed Id_POM-air as the average value of the geodesic distances from the surface of every POM to the closest air-filled pore. For each of the 7 days of measurements Id_POM-air showed a linear trend (each day with an r₂ > 0.75) with respect to the N₂O emissions, indicating that the spatial distributions of the POM and air-filled pores were key factors to determine the N₂O emissions in our soil cores.