Abstract
This study proposes to estimate an averaged hydraulic conductivity as a probability-weighted average in our geoinformation system based on adjacent water-well data. In this study, individual hydraulic conductivity of each soil/rock type was determined to obtain the least square error against field measurements in 105 pumping test locations. A basic class of 8 soil/rock types and another class of 32 types with respect to topography and geologic periods were assumed for comparison. As a result, the proposed method yielded sufficient agreements in both classes, relative to the conventional method based on simply weighted averages of the drilling records. A stepwise cross validation also determined an optimal class of 19 soil/rock types for the least prediction square error. Finally, a relation of individual hydraulic conductivities with effective thermal conductivities was shown to discuss the potential applicability of ground source heat pump systems in each soil/rock type considering groundwater advection.
