A localized bedrock aquifer distribution explains discharge from a headwater catchment

Understanding a discharge hydrograph is one of the leading interests in catchment hydrology. Recent researches have provided credible information on the importance of bedrock groundwater for formations of discharge hydrograph from headwater catchments. Some recent studies have monitored bedrock groundwater in nested wells to analyze its recharge process, flow path, and interaction with streams and lakes. However, intensive monitoring of bedrock groundwater is rare in mountains with steep topography. Consequently, how bedrock groundwater controls discharge from a steep headwater catchment is in dispute. In this study, we conducted long-term hydrological observations using densely nested bedrock wells along with monitoring of discharge hydrograph in a headwater catchment underlain by granitic bedrock. Bedrock wells with depths of 21-70 m were drilled at eight points within the 2.10-ha catchment. The observed discharge hydrograph was notably characterized by gentle and significant variations in base-flow, and exhibited anomalous triple-peak responses. That is, whereas the flashy first peaks occurred just after rainfall peaks, the second peaks lagged behind the rainfall peaks by about 72 h. Broad peaks in the base-flow discharge corresponded to the third peaks which occurred once or twice in each hydrological year. The observations using the bedrock wells indicated a fairly regionalized distribution of bedrock groundwater; that is, upper, middle, and lower aquifers were present. We observed large differences in water level between the upper and middle aquifers, and between the middle and lower aquifers. The triple-peak discharge responses can be explained by three types of water pathways: the first peak was caused by the peak in soil mantle groundwater around the outlet of the watershed; the second peak was caused by the first peak in the lower aquifer, which was fed by vertical rainwater infiltration; and the third peak was caused by the second peak in the lower aquifer, resulting from an increased lateral water supply from the middle aquifer. We developed a simple model to simulate discharge hydrograph, which evaluated each of contributions from the soil mantle groundwater, from the lower aquifer, and from the middle aquifer to the discharge. The model employed the power low relationship between water storage and discharge from the soil mantle. The contributions from the lower and middle aquifers were evaluated by the expanded Dupuit-Forchheimer formula. The modeling result generally succeeded in reproducing the observed hydrograph, implying that our suggested processes for the formation of the triple-peak discharge responses were reasonable. Thus, this study demonstrated that grasping localized bedrock aquifer distribution is essential for understanding discharge hydrograph from the headwater catchment. Future studies in this watershed should conduct physically-based numerical simulations which explain developments of the bedrock aquifers as well as their effects on formations of the discharge hydrograph. For this purpose, we additionally excavated 20 bedrock wells in the watershed in 2010. Results of monitoring these bedrock wells will be also presented.