Paleolimnological analyses can be used to evaluate limnological responses to changing climate over decadal to centennial timescales, especially in regions with sparse lake monitoring data. We used a training set with 90 lakes to develop a diatom-based conductivity transfer function and address climatedriven changes in lakes on the Qinghai-Xizang Plateau, Tibet. This new training set is an expanded version of a previous model (Yang et al. in J Paleolimnol 30:1–19, 2003) and shows improved performance statistics for the conductivity model. The expanded training set also contains diatom species not previously identified from the region, such as Stephanodiscus sp. and Cyclotella sp., which are common eutrophic indicator species in other regions, but can also be influenced by water column conductivity. The new conductivity transfer function was applied to Lakes Nam Co and Chen Co in Tibet. Recent conductivity inferences were compared with climate data from the Dangxiong weather station and water level records from Yangzhuyong Co, which show increasing temperature and lower water levels, respectively, sinceAD1960. Other studies showed that the water balance for many lakes on the Qinghai-Xizang Plateau is complex, affected by both evaporation and glacial melting. Our paleolimnological reconstructions, which include sediment particle size data, indicate that over relatively short timescales glacial meltwater can influence lake hydrology, but over decadal timescales, increases in evaporation, driven by rising temperatures, dominate. Our findings suggest that regional warming is lowering water levels at these sites and will continue to do so given predicted future climate warming