Detection and hydrologic modeling of aquifers in unconsolidated alluvial plains through combination of borehole data sets: a case study of the Arao area, Southwest Japan

In addition to spatial distribution of groundwater-flow parameters, aquifer properties of location and shape are also significant for assessing groundwater resources because they strongly affect water flow. We present a selection of geologic data suitable for aquifer analysis, a mathematical method of processing them, and a combination of several maps produced by it. The data used in the analysis are typically obtained by borehole investigation. Our targets are the areas underlain by geologic bodies with different ages and lithologies; the spatial correlation structures of geologic data over the areas tend to change locally. The processing method should be a versatile one that is applicable to areas where geostatistical stationarity is not satisfied. The aquifer analysis consistent with that requirement consists of two steps: the first is the transformation of screen locations, locations of sand and gravel layers, and resistivity by electric logging into indicator values, and the second is three-dimensional interpolation of these using the optimization principle method to produce three kinds of distribution models. A stochastic simulation is also used for modeling the resistivity distribution. The three distribution models are integrated to generate a value for evaluating the possibility of aquifer existence. A case study of an alluvial coastal plain, situated in southwest Japan, describes an aquifer model that contains three permeable layers. Each layer has about 10-m thickness and is lens shaped. To confirm the validity of the model, we have drilled two wells that reached one of the middle and bottom permeable layers, observing the water level change. Low correlation of the temporal changes of levels between the wells indicates that the two layers are hydrologically independent of each other. Additionally, groundwater-flow patterns have been estimated by transforming the simulation model parameter into hydraulic conductivity through a simple function and using a finite difference method for flow analysis. The procedure proposed by this study can be applied to other areas by changing the weights assigned to each geologic and geotechnical factor in the generation of the coefficient for aquifer existence, considering the reliability of each factor and hydrologic properties of study area.