Mineral dust is sourced from arid regions around the globe, and its composition impacts regional and global environments by affecting radiation balance, cloud formation, nutrient deposition, and snowmelt. As such, mineral dust plays a key role in Earth's energy and biogeochemical budgets. Global remote sensing of dust source regions have the potential to significantly improve our understanding of mineral dust composition and its environmental impacts. While, visible to shortwave infrared data in remote measurements effectively map surface mineralogy for large swaths of mineral groups, they struggle to detect quartz and feldspar, the two most abundant minerals on Earth's surface. Broadband thermal infrared remotely sensed data are capable of remote measurements of quartz and feldspar, but on their own are limited in their ability to resolve confounding factors including other minerals and vegetation. Here, we demonstrate the utility of joint visible to thermal infrared (VTIR) data by leveraging existing remote sensing data products and spectral libraries in a Monte Carlo spectral mixing analysis to remotely retrieve global quartz and feldspar areal distributions with grain size information. Resulting latitudinal and regional-scale distributions are consistent with expectations based on geological settings. Using a case study, we demonstrate our VTIR-based unmixing strategy is theoretically capable of further separating grain sizes and chemistries. These results highlight the utility of VTIR data, the value in further collection of these data as part of future space-based Earth observations, and the need for improved VTIR spectral libraries for interpretation.