Uncertainty reduction of hydrologic models using data from surface-based investigation

Geohydrologic model uncertainties include permeability, boundary, and initial conditions as well as the conceptual model it is based on. We present some examples of using information other than pressure data to constrain a geohydrologic model of the Horonobe area in Hokkaido, Japan. The initial model was constructed using information from surface geology and a few boreholes. Inversion analysis of pressure data implied the existence of a low-permeability cap rock. We then used river flow data and temperature data from a hot spring as a basis for estimating the recharge flux, which suggested that the overall permeability of the modeled area could be one order of magnitude larger than that of the base model. Next, we simulated a saltwater washout process and compared the simulated salinity distribution with the salinity data from a borehole. We found that a better match to the salinity data is obtained if the increase in permeability is taken up by a localized fault zone rather than uniformly by the entire model. A smaller-scale match to the temperature, pressure, and density profiles from two boreholes indicated that there was a low-permeability fault in between the two boreholes. The present study demonstrates that pressure data alone are insufficient to calibrate a model, and that additional observations are needed to accurately represent a site.