The timing and pattern of Tibetan Plateau rise provide a critical test of possible mechanisms for the development and support of high topography, yet views range widely on the history of surface uplift to modern elevations of ~4.5 km. To address this issue we present clumped isotope thermometry data from two well-studied basins in central and southwestern Tibet, for which previous carbonate δ 18 O data have been used to reconstruct high paleoelevations from late Oligocene to Pliocene time. Clumped isotope thermometry uses measurements of the 13 C-18 O bond ordering in carbonates to constrain the temperature [T(Δ 47 )] and δ 18 O value of the water from which the carbonate grew. These data can be used to infer paleoelevation by exploiting the systematic decrease of surface temperature and the δ 18 O value of meteoric water with elevation, provided samples record original depositional conditions and appropriate context exists for interpreting T(Δ 47 ) and δ 18 O values. Previous calcite δ 18 O and δ 13 C values for Oligocene-age marls from the Nima Basin in central Tibet are thought to refl ect original depositional conditions; however, T(Δ 47 ) values exceed Earth-surface temperatures, indicating that the samples have been diagenetically altered. Maximum burial temperatures were not high enough to cause solid-state C-O bond reordering. Instead, the elevated T(Δ 47 ) and water δ 18 O values are consistent with recrystallization of the samples in a rock-buffered system.Miocene-Pliocene aragonitic gastropod shells from the Zhada Basin in southwestern Tibet record primary environmental temperatures, which we interpret in the context of modern shell and tufa T(Δ 47 ) values and lake water temperatures. Modern shell and tufa T(Δ 47 ) values are similar to warmseason water temperatures. The ca. 9-4 Ma shell temperatures are signifi cantly colder, suggesting a 9 ± 3 °C (2σ) average increase in warm-season lake water temperatures since the late Miocene. If the temperature increase is due entirely to elevation change and the modern June-July-August (JJA) surface water lapse rate of 6.1 °C/km applies, it implies >1 km of elevation loss since the late Miocene-Pliocene-corresponding to an average basin fl oor paleoelevation of 5.4 ± 0.5 km (2σ). A warmer mid-Pliocene climate would make this a minimum estimate. Our fi nding of cold paleotemperatures contrasts with previous conclusions based on Pliocene snail shell T(Δ 47 ) data interpreted in the absence of modern shell and water temperature data, but is consistent with δ 18 O-based paleoaltimetry and with paleontological and isotopic data indicating the presence of coldadapted mammals living in a cold, high-elevation climate. We suggest that late Neogene elevation loss across the Zhada Basin catchment probably related to local expression of east-west extension across much of the southern Tibetan Plateau at this time.