Abrupt and large-scale climate changes have occurred repeatedly and within decades during the last glaciation. These events, where dramatic warming occurs over decades, are well represented in both Greenland ice core mineral dust and temperature records, suggesting a causal link. However, the feedbacks between atmospheric dust and climate change during these Dansgaard-Oeschger events are poorly known and the processes driving changes in atmospheric dust emission and transport remain elusive. Constraining dust provenance is key to resolving these gaps. Here, we present a multi-technique analysis of Greenland dust provenance using novel and established, source diagnostic isotopic tracers as well as results from a regional climate model including dust cycle simulations. We show that the existing dominant model for the provenance of Greenland dust as sourced from combined East Asian dust and Pacific volcanics is not supported. Rather, our clay mineralogical and Hf-Sr-Nd and D/H isotopic analyses from last glacial Greenland dust and an extensive range of Northern Hemisphere potential dust sources reveal three most likely scenarios (in order of probability): direct dust sourcing from the Taklimakan Desert in western China, direct sourcing from European glacial sources, or a mix of dust originating from Europe and North Africa. Furthermore, our regional climate modeling demonstrates the plausibility of European or mixed European/North African sources for the first time. We suggest that the origin of dust to Greenland is potentially more complex than previously recognized, demonstrating more uncertainty in our understanding dust climate feedbacks during abrupt events than previously understood.