Analysis of time-lapse GPR data to visualize preferential flow paths

Time-lapse GPR data are commonly interpreted by analyzing picked travel-time differences between individual reflectors or sample-to-sample differences of the data. In this study, we employ a novel, structural similarity attribute for the interpretation of time-lapse GPR data, to image preferential flow paths. We test this attribute using synthetic GPR data calculated on a heterogeneous subsurface model and compare its performance to the classical employed sample-to-sample difference. After evaluating the attribute definition, we apply it to field data, collected at a field site, which is known for its potential to develop preferential flow path. Our results indicate subsurface areas of increased differences, which, due to their temporal evolution are interpreted as preferential lateral preferential flow path.