The western Himalaya, Karakoram and Tibet are known to be heterogeneous with regard to Pb isotope compositions in K-feldspars, which allows this system to be used as a sediment provenance tool. We used secondary ion mass spectrometry to measure the isotopic character of silt and sand-sized grains from the modern Sutlej and Chenab Rivers, together with Thar Desert sands, in order to constrain their origin. The rivers show a clear Himalayan provenance, contrasting with grains from the Indus Suture Zone, but with overlap to known Karakoram compositions. The desert dunes commonly show 207Pb/ 204Pb and 206Pb/ 204Pb values that are much higher than those seen in the rivers, most consistent with erosion from Nanga Parbat. This implies at least some origin from the trunk Indus, probably reworked by summer monsoon winds from the SW, a hypothesis supported by bulk Nd and U–Pb zircon dating. Further data collected from Holocene and Pleistocene sands shows that filled and abandoned channels on the western edge of the Thar Desert were sourced from Himalayan rivers before and at 6–8 ka, but that after that time the proportion of high isotopic ratio grains rose, indicating increased contribution from the Thar Desert dunes prior to ∼4.5 ka when flow ceased entirely. This may be linked to climatic drying, northward expansion of the Thar Desert, or changes in drainage style including regional capture, channel abandonment, or active local Thar tributaries. Our data further show a Himalayan river channel east of the present Indus, close to the delta, in the Nara River valley during the middle Holocene. While this cannot be distinguished from the Indus it is not heavily contaminated by reworking from the desert. The Pb system shows some use as a provenance tool, but is not effective at demonstrating whether these Nara sediments represent a Ghaggar-Hakra stream independent from the Indus. Our study highlights an important role for eolian reworking of floodplain sediments in arid rivers such as the Indus.