GEOSTATISTICAL ANALYSIS OF GROUND-PENETRATING RADAR DATA - A MEANS OFDESCRIBING SPATIAL VARIATION IN THE SUBSURFACE

We have investigated the use of ground-penetrating radar (GPR) as a me ans of characterizing the heterogeneity of the subsurface. Radar data were collected at several sites in southwestern British Columbia under lain by glaciodeltaic sediments. A cliff face study was conducted in w hich geostatistical analysis of a digitized photograph of the face and the radar image of the face showed excellent agreement in the maximum correlation direction and the correlation length determined from thes e two da-ta sets. Other two-dimensional (2-D) sections of radar data w ere divided into sedimentary architectural elements on the basis of th e distinct radar appearance of these sedimentary units. Examples of fo ur sedimentary units were used to obtain semivariograms from the radar data and resulted in maximum correlation lengths between 0.5 and 4.8 m. A 3-D radar survey, collected over a package of gravel and sand for esets, was analyzed to determine the paleoflow direction; a correlatio n length of 4 m was found in that direction.