We present cross-hole P-and S-wave seismic experiments that have been performed along a ~100 m long transect for the detailed characterization of a contaminated sedimentary site (Bitterfeld research test site, Germany). We invert the corresponding first break arrival times for the P-and S-wave velocity structure and compare two different strategies to interpret these models in terms of pertinent lithological and geotechnical parameter variations. The first (common) approach is based on directly translating the tomographic velocity models into the parameters of interest (e.g., elastic moduli). The second (zonal) approach first reduces the tomographic parameter information to a limited number of characteristic velocity combinations via k-means cluster analysis. Then, for each zone (cluster) further parameters including uncertainties can be estimated. In the presented case study, our results indicate that the zonal approach provides an effective means for the integrated interpretation of different co-located data.rial at a site located in the New Madrid Seismic Zone, Missouri, USA. In another recent example, Daley et al. (2004) performed cross-hole P-and S-wave tomography to characterize fractured basalt at the Idaho National Engineering and Environmental Laboratory, USA. However, the number of published case studies illustrating the benefit of combining P-and S-wave cross-hole data is limited and further experience is needed to establish this combined approach for a variety of near-surface applications.Commonly, tomographic parameter fields (such as P-and S-wave velocity models) are interpreted manually by delineating regions characterized by similar parameter ranges. Such an interpretation is usually focused on outlining subsurface structures or zones which, then, can eventually be further interpreted, e.g., in terms of lithology by comparing the tomographic images to corresponding borehole records. In addition, we usually can estimate further petrophysical parameter models by directly translating the tomographic models using adequate petrophysical translations or by establishing statistical relations between borehole data and the tomographic parameters in the vicinity of the borehole (e.g., Yamamoto 1994;Angioni et al. 2003). However, the outcome of these common approaches often depends on the experience and preconceptions of the interpreter and the validity and applicability of the petrophysical model used.In this paper, we present the results of P-and S-wave crosshole seismic experiments performed to characterize a contaminated sedimentary site in Bitterfeld, Germany. After providing background information on the field site and the experiments, we present the inverted P-and S-wave velocity models and compare