Tropospheric ozone is an important component of the Earth system as it can affect both climate and air quality. In this work, we use observed tropospheric column ozone derived from the Ozone Monitoring Instrument (OMI) and Microwave Limb Sounder (MLS) OMI-MLS, in addition to OMI ozone retrieved in discrete vertical layers, and compare it to tropospheric ozone from UM-UKCA simulations (which utilize the Unified Model, UM, coupled to UK Chemistry and Aerosol, UKCA). Our aim is to investigate recent changes (2005–2018) in tropospheric ozone in the North Atlantic region, specifically its seasonal, interannual and decadal variability, and to understand what factors are driving such changes. The model exhibits a large positive bias (greater than 5 DU or ∼ 50 %) in the tropical upper troposphere: through sensitivity experiments, time series correlation, and comparison with the Lightning Imaging Sensor and Optical Transient Detector lightning flash dataset, the model positive bias in the tropics is attributed to shortcomings in the convection and lightning parameterizations, which overestimate lightning flashes in the tropics relative to mid-latitudes. Use of OMI data, for which vertical averaging kernels and a priori information are available, suggests that the model negative bias (6–10 DU or ∼ 20 %) at mid-latitudes, relative to OMI-MLS tropospheric column, could be the result of vertical sampling. Ozone in the North Atlantic peaks in spring and early summer, with generally good agreement between the modelled and observed seasonal cycle. Recent trends in tropospheric ozone were investigated: whilst both observational datasets indicate positive trends of ∼ 5 % and ∼ 10 % in North Atlantic ozone, the modelled ozone trends are much closer to zero and have large uncertainties. North Atlantic ozone interannual variability (IAV) in the model was found to be correlated to the IAV of ozone transported to the North Atlantic from the stratosphere (R=0.77) and emission of NOx from lightning in the tropics (R=0.72). The discrepancy between modelled and observed trends for 2005–2018 could be linked to the model underestimating lower stratospheric ozone trends and associated stratosphere to troposphere transport. Modelled tropospheric ozone IAV is driven by IAV of tropical emissions of NOx from lightning and IAV of ozone transport from the stratosphere; however, the modelled and observed IAV differ. To understand the IAV discrepancy we investigated how modelled ozone and its drivers respond to large-scale modes of variability. Using OMI height-resolved data and model idealized tracers, we were able to identify stratospheric transport of ozone into the troposphere as the main driver of the dynamical response of North Atlantic ozone to the Arctic Oscillation (AO) and the North Atlantic Oscillation (NAO). Finally, we found that the modelled ozone IAV is too strongly correlated to the El Niño–Southern Oscillation (ENSO) compared to observed ozone IAV. This is again linked to shortcomings in the lightning flashes parameterization, which underestimates (overestimates) lightning flash production in the tropics during positive (negative) ENSO events.