In this study, we assessed induced polarization as potential non-destructive method for root detection in dike embankments. We used both laboratory and field experiment to describe the electrical signal with a focus on soil water content. Our objective was to determine in which hydric state of the soil, and related electrical properties, roots could be accurately discriminated. We hypothesized that preferential water zone absorption near the roots could, in some conditions, contribute to locate them. During the laboratory experiments, we compared the response of containers filled with the same homogeneous silty clay bare material, and without (A) or with freshly cut root (B) at different levels of soil water content. Resistivity and phase variations with soil water content indicated that it was preferable to work in dry conditions since the contrast was higher. Interactions and overlapping between polarization effects of both root and soil made it difficult to interpret first chargeability maps. This led us to study temporal-spatial variations by considering the dynamics of water absorption during a field experiment. High resolution time lapses images showed a correlation between root location and complex resistivity anoma- lies. Although these first results have to be confirmed by further measurements, induced polarization seems to add useful information to interpret anomalies produced by woody roots. .