A Multi-tracer Approach to Characterize Hydrological Process in Kherlen River Basin, Eastern Mongolia Maki Tsujimura Graduate School of Life and Environmental Sciences, University of Tsukuba 1. Introduction The purposes of this study focusing on hydrological processes in Kherlen River basin, eastern Mongolia, are to elucidate flow path and residence time of water using isotope tracers of 2 H, 18 O and 3 H, to determine the origin of precipitation, river water, groundwater and atmospheric water vapor, to investigate plant water use strategy using 18 O, and to investigate the soil erosion and overland flow processes using 137 Cs and 210 Pb. Modeling approach is also applied to support the results by tracer approach regarding atmospheric water vapor source and groundwater flow system. 2. Where does water vapor come from? The stable isotopes of 2 H (D) and 18 O in precipitation, atmospheric water vapor and subsurface water were investigated at forest and grassland sites. Atmospheric water vapor was sampled at heights of 0.5–1000 m from the ground surface (throughout atmospheric boundary layer: ABL) using an aircraft and aboveground observations, and soil water was sampled at depths of 0.1–1.5 m by digging a trench and using suction lysimeters from June to October 2003. The δD and δ 18 O of 230 water samples were determined (Tsujimura et al., 2006). The δ 18 O of precipitation in the forest and grassland sites showed clear seasonal variation from October 2002 to September 2003, with higher values in summer and lower values in winter. The δ 18 O values in the atmospheric water vapor decreased from June to October 2003, parallel to those of precipitation. The vertical profile of δ 18 O in the water vapor tended to show a gentle decrease with altitude in the atmospheric boundary layer at both the forest and grassland sites. This was caused by evapotranspiration and mixing with air in the free atmosphere over the atmospheric boundary layer. Using the δ 18 O of water vapor in the free atmosphere and evapotranspired water vapor on the 21st and 23rd August 2003, the contribution of free atmospheric water vapor in the water vapor in the ABL was evaluated as approximately 54–70% at forest site and 56–75% at grassland site. These estimated values suggest that evapotranspired water vapor is not dominant in the water vapor of the ABL especially above KBU. Thus, local recycling of water vapor does not play a significant role in the formation of water vapor in the ABL in the grassland region of the Kherlen River basin under a steady atmospheric condition. Yamanaka et al. (2006) showed that the isotopic composition of precipitation is primarily controlled by rainout history during transportation from the region outside of Mongolia as the ultimate origin to eastern Mongolia, and that continental recycling has a generally minor effect on isotopes in precipitation. Sato et al. also (in preparation) suggests inner Eurasia, Siberia and eastern China as source areas causing the atmospheric water vapor to eastern Mongolia by using regional climate model including isotope transport process. We separated evaporation and transpiration components by Keeling plots analysis using the δ 18 O of atmospheric water vapor and soil water. Accordingly, the ratio of transpiration rate to evapotranspiration rate was estimated to be 60–73% at the forest site and 35–59% at the grassland site. 3. Where are groundwater and river water recharged? Inorganic solute ion concentrations and the stable isotopes in groundwater, river water and precipitation were investigated to gain insight into the groundwater recharge process in the Kherlen River basin, a semi-arid region in eastern Mongolia (Tsujimura et al., 2006). The solute constituents in the river water (main stream) were of Ca–HCO 3 type, spatially invariant and low in concentration. Groundwater in the upstream region was also characterized by a Ca–HCO 3 type, though all ion concentrations were higher than in the river water. On the other hand, the chemical composition of the groundwater in the midstream region (southern and eastern) was spatially variable and the Na + , Mg 2+ , Cl - and HCO 3 - concentrations were considerably higher than in the river water and upstream groundwater. The stable isotopic compositions showed an evaporation effect on the groundwater and river water, as well as an altitude effect in the precipitation and river water. Preferential recharge by relatively large rainfall events is thought to