Geostatistical modeling of groundwater properties and assessment of their uncertainties

Yamamoto, Shinya*; Honda, Makoto*; Sakurai, Hideyuki*; Suzuki, Makoto*; Sanada, Hiroyuki; Sugita, Yutaka ; Matsui, Hiroya 

The distribution of groundwater properties is important for understanding of the deep underground hydrogeological environments in the design of the HLW disposal facilities. This study proposes a geostatistical system for modeling the chemical properties of groundwater which have a correlation with the resistivity data obtained from widespread and exhaustive survey. The proposed system is consists of two methodologies (resistivity data integration using Simple Kriging with varying local mean (SKlm) and 3D distribution modeling of hydro-chemical properties of groundwater estimated by SKlm algorithm). 2 types of uncertainty are considered: (1) Aleatory uncertainty; (2) Epistemic uncertainty. As a result, the uncertainties of the proposal model have been estimated lower than other traditional model's.