Tracing the source of dust storm (DS) in mega-cities of northern China currently suffers ambiguities from different approaches including source-sink proxy comparison, air mass back trajectory modeling, and satellite image monitoring. By integrating advantages of all three methods, we present a multidisciplinary approach to trace the provenance of dust fall in Xi'an during the spring season (March to May) of 2012. We collected daily dust fall to calculate dust flux variation, and detected eight DS events with remarkable high flux values based on meteorological comparison and extreme detection algorithm. By combining MODIS images and accompanying real-time air mass back trajectories, we attribute four of them as natural DS events and the other four as anthropogenic DS events, suggesting the importance of natural and anthropogenic processes in supplying long-range transported dust. The primary sources of these DS events were constrained to three possible areas, including the northern Chinese deserts, Taldimakan desert, and Gurbantunggut desert. Proxy comparisons based upon the quartz crystallinity index and oxygen isotope further confirmed the source-to-sink linkage between the natural DS events in Xi'an and the dust emissions from the northern Chinese deserts. The integration of geochemical and meteorological tracing approaches favors the dominant contribution of short-distance transportation of modern dust fall on the Chinese Loess Plateau. Our study shows that the multidisciplinary approach could permit a better source identification of modern dust and should be applied properly for tracing the provenance fluctuations of geological dust deposits. (C) 2015 Elsevier Ltd. All rights reserved.