Atmospheric Rivers (ARs) are huge narrow bands of enhanced water vapour. When they cross coastlines and interact with orography, ARs can cause intense rainfall and flooding. ARs only appeared in the literature in recent decades and there has been much debate about how to define ARs, and how to identify them. As a result, a wide range of identification algorithms have been produced often with variations in the conditions required for an object to be classified as an AR, and variations in the input data. One of the key conditions in all AR identification algorithms is a minimum threshold of water vapour or water vapour flux. In this study, we conduct a sensitivity analysis of a generic AR identification algorithm to different Integrated Water Vapour (IVT) thresholds, input data resolutions and regridding methods using the Year of Tropical Convection operational analysis. We found that the resolution and regridding method affects the magnitude of ARs identified but the shape of the distribution is well conserved. AR identification is highly sensitive to the choice of IVT threshold, which can impact both the magnitude of ARs identified and shift the AR spatial distribution. We show that the typical definition used (250 kgm^-1s^-1 ) fails to identify the most intense ARs, due to geometric reasons, especially in the North Pacific and North Atlantic – two basins where ARs are known to have strong impacts on highly populated regions i.e. the North American West Coast and Western Europe. We conclude by recommending a threshold between 350-500kgm^-1s^-1 for global studies in order to adequately capture the diversity of ARs. The next steps in this project will involve applying these results to understanding the role ARs play in Australian rainfall extremes.