*Masanori KATSUYAMA1, Ken'ichiro KOSUGI2, Makoto TANI2
(1.Center for the Promotion of Interdisciplinary Education and Research, Kyoto University, 2.Graduate School of Agriculture, Kyoto University)
Keywords:Bedrock groundwater, Tracer, Mean residence time, granite catchment
Bedrock groundwater dynamics is one of the latest frontier of hillslope- and catchment hydrology. Although it relate to water resources as well as sediment disasters, only few studies have accessed directly with boreholes to bedrock groundwater aquifer because of, for example, high costs. In this context, tracer approach is effective to clarify the bedrock groundwater dynamics and water pathways within deeper layers of mountains. We have been keeping on monitoring of the chemical and isotopic compositions of bedrock groundwater and streamwater in Kiryu Experimental Watershed (KEW), Japan since 2003. We set up a nested observation system; a hillslope plot (AP, 0.024ha), a subcatchment (A catchment, 0.086ha), and whole of KEW (K catchment, 5.99ha), and monthly sampled the streamwater of K and A, the outflow from AP, which occurs as saturated throughflow on the soil-bedrock interface during rainstorms, and groundwater in the soil sediment. Moreover, we excavated the bedrock and installed some tension lysimeters at 0.1, 0.2, 0.4, and 0.8 m deep and boreholes at 12, 15, and 20 m deep below bedrock surface, and sampled them. The stream flow from K and A were perennial. The SiO2 and Na+ concentrations increased along with the infiltration process. On the other hand, the NO3- concentration was highest at the surface soil water, and removed along with the infiltration process. The concentrations of both solutes in the streamwater from A and K were intermediate between the concentrations in the surface soil water and bedrock groundwater. These facts mean that the streamwater is the mixture of shallow soil water and deep groundwater. The mean residence times calculated by delta 18O variations were about 4 or 5 months in the groundwater in the soil sediment and in the shallow (<0.8m) bedrock groundwaters, about 50 months in 12- and 15 m deep, and about 120 months in 20 m deep, respectively. That in the streamwater in A was estimated as about 30 months. Thus, the MRT in 20 m deep groundwater is quite different from the others. The relationship between the MRTs and the solute concentrations were different in each solute; for SiO2, the concentration increased as a saturation curve, and it increased as linearly for Na+. It exponentially decreased for NO3-. The streamwater chemistries in A were on these curves. Therefore, the solute concentrations can be described as functions of MRTs. These results suggest that a part of the bedrock groundwater can contribute to the stream from the shallower layers. The fact that the stream flow is perennial in this subcatchment A means that plentiful supply of groundwater from the relatively shallow bedrock layers exist. On the other hand, other part of the bedrock groundwater infiltrate deeply and less contribute to the stream in this small subcatchment; we have to consider whether the deeply infiltrated groundwater may contribute at the outlet of K catchment. Moreover, as the deeper bedrock groundwater have especially long residence time, we have to keep long-term monitoring to understand the dynamics and roles of this groundwater to hydrological and hydrochemical processes, because it will be a key of spatio-temporal scaling of these processes, as well as the water yield function of forests.