<p><strong>Abstract.</strong> Dry deposition is the second largest sink of tropospheric ozone. Increasing evidence has shown that ozone dry deposition actively links meteorology and hydrology with ozone air quality. However, there is little systematic investigation on the performance of different ozone dry deposition parameterizations at the global scale, and how parameterization choice can impact surface ozone simulations. Here we present the results of the first global, multi-decade modelling and evaluation of ozone dry deposition velocity (<i>v<sub>d</sub></i>) using multiple ozone dry deposition parameterizations. We use consistent assimilated meteorology and satellite-derived leaf area index (LAI) to simulate <i>v<sub>d</sub></i> over 1982&#8211;2011 driven by four sets of ozone dry deposition parametrization that are representative of the current approaches of global ozone dry deposition modelling, such that the differences in simulated <i>v<sub>d</sub></i> are entirely due to differences in deposition model structures. In addition, we use the surface ozone sensitivity to <i>v<sub>d</sub></i> predicted by a chemical transport model to estimate the impact of mean and variability of ozone dry deposition velocity on surface ozone. Our estimated <i>v<sub>d</sub></i> from four different parameterizations are evaluated against field observations, and while performance varies considerably by land cover types, our results suggest that none of the parameterizations are universally better than the others. Discrepancy in simulated mean <i>v<sub>d</sub></i> among the parameterizations is estimated to cause 2 to 5&#8201;ppbv of discrepancy in surface ozone in the Northern Hemisphere (NH) and up to 8&#8201;ppbv in tropical rainforest in July, and up to 8&#8201;ppbv in tropical rainforests and seasonally dry tropical forests in Indochina in December. Parameterization-specific biases based on individual land cover type and hydroclimate are found to be the two main drivers of such discrepancies. We find statistically significant trends in the multiannual time series of simulated July daytime <i>v<sub>d</sub></i> in all parameterizations, driven by warming and drying (southern Amazonia, southern African savannah and Mongolia) or greening (high latitudes). The trends in July daytime <i>v<sub>d</sub></i> is estimated to be 1&#8201;%&#8201;yr<sup>&#8722;1</sup> and leads to up to 3&#8201;ppbv of surface ozone changes over 1982&#8211;2011. The interannual coefficient of variation (CV) of July daytime mean <i>v<sub>d</sub></i> in NH is found to be 5&#8201;%&#8211;15&#8201;%, with spatial distribution that varies with the dry deposition parameterization. Our sensitivity simulations suggest this can contribute between 0.5 to 2&#8201;ppbv to interannual variability (IAV) in surface ozone, but all models tend to underestimate interannual CV when compared to long-term ozone flux observations. We also find that IAV in some dry deposition parameterizations are more sensitive to LAI while others are more sensitive to climate. Comparisons with other published estimates of the IAV of background ozone confirm that ozone dry deposition can be an important part of natural surface ozone variability. Our results demonstrate the importance of ozone dry deposition parameterization choice on surface ozone modelling, and the impact of IAV of <i>v<sub>d</sub></i> on surface ozone, thus making a strong case for further measurement, evaluation and model-data integration of ozone dry deposition on different spatiotemporal scales.</p>