A newly developed broadband digital-based seismic landstreamer system was employed for the planning of a double-train-track tunnel in the city of Varberg, southwest Sweden. Twenty-five seismic profiles, totalling more than 7.5 km of data, were acquired using a 2-to 4-m receiver and source spacing. At places where it was not possible to move the streamer such as road crossings, wireless recorders connected to 28-Hz geophones were used. In addition to the earlier refraction data analysis and first-break traveltime tomographic modelling, reflection processing of the data was considered in this study, given the realisation of reflections in raw shot gathers and their good quality. Bedrock is strongly reflective in most cases but is not evident in the sections when it gets near the surface. Bedrock undulation is noticeable in most reflection sections, and at one occasion, strong diffraction is observed in the bedrock or near to it. The diffraction is originated, not known during the survey, from a 400-m 3 cylindrical (of about 3-m-height and 13-m-diameter) concrete-made fireprotection water tank situated in the bedrock and used in emergency situations. Reflection seismic data greatly complement the tomographic models and support deep bedrock where the excavation of the tunnel is planned in downtown Varberg. This interpretation implies different reinforcements and tunnel construction methods (e.g., roofed concrete) at this section of the tunnel. In addition, weakness zones associated with fracture systems are inferred from the reflection characteristics and in conjunction with the velocity models requiring verification by additional boreholes. Malehmir 2013, 2014;Wang, Malehmir and Bastani 2016;Place and Malehmir 2016). Most of these cases, however, deal with either deep bedrock (>30 m) or use shearwave (horizontal component) surveys for these purposes (Pugin et al. 2004;Benjumea et al. 2008;Krawczyk et al. 2012;Krawczyk, Polom and Beilecke 2013;Pugin et al. 2013). When bedrock is shallow (10-20 m) and the cover sediments are of high velocity (e.g., >1500 m/s), reflection from bedrock is difficult to identify because of strong source-generated noise at these shallow depths and lack of high seismic fold. Fine spacing between source and receivers (0.5-1 m) and using high-frequency sources (>300 Hz) can tackle this (e.g., Miller et al. 1986;Steeples and Miller 1998;van der Veen et al. 2000;Schmelzbach, Green and Horstmeyer 2005;Sloan et al. 2007), but then, refraction data analysis becomes an issue since long-offset first arrivals are not recorded if the receiver spread is not sufficiently long enough and since high frequencies will not travel far. A compro-