The GEOS-Chem 3D global chemical transport model is used to demonstrate the sensitivity of the global ozone budget to spatial and temporal variability in the parameterization of lightning emissions of reactive nitrogen oxides (NO_x). Global atmospheric chemistry models must parameterize lightning formation as part of their convective simulations and on the basis of large-scale meteorological variables, attenuating sub-grid processes that drive lightning variability. The High Resolution Monthly Climatology (HRMC) combined product of the Optical Transient Detector (OTD) and the Lightning Imaging Sensor (LIS) covering the period 1995-2005 available from the NASA Global Hydrology and Climate Center (GHCC), as well as the individual Science Data orbits of the LIS (1998-present) are used here as top-down constraints on the lightning parameterization in GEOS-Chem, driven by assimilated meteorological products of the NASA Global Modeling and Assimilation Office (GMAO). Different techniques are explored for spatially redistributing the lightning density via local scaling factors determined by comparing the unscaled parameterized and observed LIS/OTD climatology product for the 11-year observation window. A series of simulations compares the impact of this redistribution performed at different scales: at native model resolution (“local”), increasingly larger-scale regions as identified by a hierarchal clustering algorithm (“regional”), and no redistribution at all. The local technique allows for sharper spatial variability and stronger correlation to the climatology, but regional methods stay truer to the physically-based parameterization and model physics with scaling factors closer to unity, enable more robust statistics, while yielding similar impacts on the ozone budget. The regional methods also enable additional constraint of temporal variability within GEOS-Chem by permitting much more robust sample sizes of the LIS Science Data orbits for determination of interannual monthly scaling factors for each tropical and subtropical region. The impact of constraining temporal variability in the model is studied over simulations over the first 9-years with available LIS data. Validation is performed against various in situ and satellite observations of upper tropospheric photochemical species, in particular, the long-term ozonesonde observations available through the World Ozone and Ultraviolet Radiation Data Centre (WOUDC). The ability of GEOS-Chem to capture large-scale features of the interannual ozone budget in space and time (e.g., the South Atlantic maximum) is explored.