Abstract. For the first time, a plume-in-grid approach is implemented in a chemical transport model (CTM) to parameterize the effects of the nonlinear reactions occurring within high concentrated NO x plumes from lightning NO x emissions (LNO x ) in the upper troposphere. It is characterized by a set of parameters including the plume lifetime, the effective reaction rate constant related to NO x -O 3 chemical interactions, and the fractions of NO x conversion into HNO 3 within the plume. Parameter estimates were made using the Dynamical Simple Model of Atmospheric Chemical Complexity (DSMACC) box model, simple plume dispersion simulations, and the 3-D Meso-NH (non-hydrostatic mesoscale atmospheric model). In order to assess the impact of the LNO x plume approach on the NO x and O 3 distributions on a large scale, simulations for the year 2006 were performed using the GEOS-Chem global model with a horizontal resolution of 2 • × 2.5 • . The implementation of the LNO x parameterization implies an NO x and O 3 decrease on a large scale over the region characterized by a strong lightning activity (up to 25 and 8 %, respectively, over central Africa in July) and a relative increase downwind of LNO x emissions (up to 18 and 2 % for NO x and O 3 , respectively, in July). The calculated variability in NO x and O 3 mixing ratios around the mean value according to the known uncertainties in the parameter estimates is at a maximum over continental tropical regions with NO [−1.18, +1.93] ppb, in July, mainly depending on the determination of the diffusion properties of the atmosphere and the initial NO mixing ratio injected by lightning. This approach allows us (i) to reproduce a more realistic lightning NO x chemistry leading to better NO x and O 3 distributions on the large scale and (ii) to focus on other improvements to reduce remaining uncertainties from processes related to NO x chemistry in CTM.