A set of lake surface sediment samples and aquatic macrophytes from the eastern and central Tibetan Plateau and a sediment core from Koucha Lake (northeastern Tibetan Plateau) were analysed for hydrogen isotope signals of aliphatic lipid biomarkers. The δD values of n-alkanes were correlated with those of meteoric water in the summer months and showed increasing Pearson correlation coefficients with increasing n-alkane chain length (r, 0.80 for n-C31). While in samples from the most humid regions (southeastern Tibetan Plateau), the apparent enrichment factors between meteoric water and lipids are close to those reported for other humid sites (ε≈-135‰ for mid-chain and ≈-125‰ for long chain n-alkanes), they become smaller (ε≈-118‰ to -41‰; mean value: -86‰) at sites with a negative moisture balance (northeastern and central Tibetan Plateau and Qaidam Basin). This might be caused by enrichment in deuterium of the source water used for lipid synthesis as a result of lake water evaporation and soil and leaf water evapotranspiration. In the core from Lake Koucha, decreasing δD values of all n-alkanes between 15 and 10calka BP give evidence for an intensification of the Asian summer monsoon. At ca. 9calka BP, a ca. 15‰ shift towards higher δD values of mid-chain n-alkanes suggests enhanced evaporation of the lake water, which outweighs higher precipitation amounts because of a temperature increase during the early Holocene monsoonal maximum. Only a few surface sediment samples show a significant offset between δD values of mid-chain n-alkanes and n-C31 (ΔδDmid-n-C31) and no systematic dependence between this offset and environmental parameters, such as mean annual precipitation, was found. In the core, available data for most samples also do not show a significant offset, only a slight increase (≈+15‰) of ΔδDmid-n-C31 at ca 10.0calka BP and a decrease (≈-17‰) at ca 6.8calka BP might indicate wetter and drier conditions, respectively. However, interpretation of ΔδDmid-n-C31 throughout the record is difficult because of a lack of reliable data for n-C31 in key sections of the core. So, more records are necessary to further test the potential applicability of ΔδDmid-n-C31 as a proxy for palaeo-hydrological conditions. © 2010 Elsevier Ltd.