Field-scale lithologic applications of complex conductivity (sigma*) imaging have been hindered by the challenges of (1) acquiring reliable induced polarization (IP) measurements and (2) obtaining reliable sigma* images from the measurements. We performed a series of 2D time domain resistivity/IP surveys at the Hanford 300 Area (Richland, Washington) where the challenge was to image the spatial distribution of two lithologic units that control the exchange between groundwater and surface water of the Columbia River. Exploiting the equivalence between time domain and frequency domain measurements of polarization, a 2D sigma* inversion (real conductivity sigma', imaginary conductivity sigma '', and phase angle phi) was used to image the spatial distribution of sigma* across the site. Synthetic studies were carried out to investigate the effects of noise on the resolution of sigma* images and to add confidence on the interpretation of possible paleochannels observed in the field data sets. The synthetic studies show that, with increasing representative noise levels, degradation of the resolution of lithologic structures in the parameters most controlled by the IP measurements (phi and sigma '') is significantly greater than degradation of resolution of sigma' images. However, the acquisition of IP measurements, and the analysis of changes in sigma' and sigma '' constrains the lithological interpretation of the geoelectrical data set due to the strong dependency of sigma '' on lithological properties. A threshold based on sigma '' measurements from cores at the site was used to estimate the elevation of the contact between the two key units, which is consistent with boreholes at the site. Variation in the elevation of this contact provides evidence of a depression in the Hanford-Ringold contact connecting the aquifer and the Columbia River; this depression likely represents a paleochannel regulating flow and transport at the site.