Mineral aerosols play multiple roles in the global system, but the magnitude of their impact on climate and global biogeochemical cycles remains uncertain. The extent to which atmospheric dust concentrations warm or cool the climate is still poorly quantified, as it depends on factors such as the physical and radiative characteristics of dust. Dust-cycle models must be improved before dust impacts can be more reliably predicted. Progress in this direction is hindered by a shortage of global data sets to determine model input parameters on multiple spatial scales and data sets to comprehensively validate model simulations. A workshop, which was recently held in Germany, brought experts together to formulate strategies for using existing global data sets and creating new data sets to improve model parameterizations and model evaluation.