INCORPORATING AUXILIARY GEOPHYSICAL-DATA INTO GROUNDWATER ESTIMATION

The incorporation of auxiliary data into ground-water flow parameter e stimation is a challenging task which can ultimately result in a bette r site characterization. In this study a maximum likelihood estimation procedure has been applied to the joint identification of the paramet ers of the aquifer transmissivity random field, and the parameters of the linear regression between the logarithm of transmissivity and the logarithm of the electrical transverse formation factor (TF), determin ed from surface geoelectrical methods (Vertical Electrical Sounding or V.E.S.). This approach is basically a co-kriging technique applied to the transmissivity and transverse formation factor random fields, but it avoids the independent estimation of the cross-covariances and the secondary variable covariance. The procedure needs some direct well d ata for transmissivity and a (usually larger) number of V.E.S. measure ments which have to be in part at a distance from the well locations i n order to provide useful information. The algorithm determines the ch aracteristics of the local (site dependent) transmissivity-transverse formation factor relationship and utilizes this auxiliary information for a geostatistical transmissivity field estimation. The methodology is tested on a real field scenario: a fractured aquifer impacted by la ndfill leachate contamination. The use of the formation factor in plac e of the raw resistivity of the subsoil layers accounts for possible e ffects of clay and contaminant concentration on pore-water resistivity . The information provided by the V.E.S. can add, to some extent, to t he understanding of the aquifer characteristics and vulnerability. How ever, tbe specificity of each site has to be fully understood for an e ffective application of the present procedure. It seems unlikely that geoelectric data can differentiate between transmissivity values diffe ring by less than two or three orders of magnitude.