Mineral dust is an essential element of the Earth’s natural biogeochemical cycle, but has become a global concern due to its profound transnational impact on the environment, human health, agriculture, livelihoods, and socioeconomic well-being. Quantifying the diverse impacts of dust necessitates a good understanding and model representation of the dust emission process, particularly its response to climate change and land use management. Past studies have shown substantial inter-model discrepancies in the global dust emissions in terms of the total flux, annual cycle, geographic distribution, and the sensitivity to wind and land surface conditions. In this study, we aim to better understand the model disagreement on the interannual variability of dust emission. Similar to the view of Koster et al. (2009) on soil moisture, we consider dust emission as a model-specific quantity with minimal observational constraint, and believe that the true information content of dust emission lies in its temporal variability. Dominance analysis and multiple linear regression techniques will be used to examine the model consistency in representing the dust emission variability and the relative importance of wind and hydroclimate variables within each model. Results will be presented to show the model agreement and disagreement by geographic regions and climate zones.