Spatio-Temporal Modeling of Groundwater Levels and Hydraulic Conductivity with Relation to Land Cover Changes

摘要

Groundwater table elevations are well known to be influenced by land cover changes and have seasonal fluctuations. The fluctuation pattern varies locally associated with hydraulic properties such as hydraulic conductivity, permeability and specific discharge of the aquifer and the distance from the water recharge to discharge areas. In order to understand groundwater table elevations and hydraulic conductivity decrease, Linear Spectral Mixture, Multivariate Cokriging, and inversion method based on Levenberg-Marquardt algorithm are examined for Kumamoto plain, southwest Japan where all the water supplies depend on the groundwaler. Linear Spectral Mixture was performed using the 10 Landsat TM/ETM+ satellites images during 16 years (1987-2003) to detect land cover changes.Multivariate Cokriging that includes variogram modeling between groundwater table elevations and, rainfall, and Linear Spectral Mixture data were applied to clarify the relationship between the decreases in groundwater table elevations and land cover changes. Levenberg-Marquardt method was used to estimate the hydraulic conductivity in the study area. Cokriging contributes to determine the dependency of groundwater table elevations on decrease of recharge zones generally covered by vegetations, trees, and rice fields. By combining these three methods, the relations on temporal changes among groundwater table elevations, precipitation, land uses, and hydraulic conductivity were clarified.