Abstract Haddam Meadows State Park in Haddam, Connecticut, serves as a test site for the Branch of Geophysical Support of the USGS in Connecticut. At this site, various geophysical techniques are tested for detection of structures that could influence dense non-aqueous phase liquids (DNAPL) migration. DNAPLs sink through the water column until they reach an impermeable boundary. The geology of Haddam Meadows consists of Ordovician metamorphic bedrock, approximately 43 m deep, overlain by glaciofluvial and alluvial deposits. If no aquitards exist above bedrock, any DNAPLs present at this location should collect in topographic lows in the bedrock. A 3-D seismic dataset was collected at the site to determine the feasibility of using this methodology to constrain bedrock topography. Data were recorded from an evenly spaced shot grid along five parallel receiver lines. Although acquisition was rapid (two days), the offset, azimuth, and full-fold ranges of the dataset are restricted by the nonoptimized shot-receiver pattern, limiting our ability to examine bedrock structure. Data are reduced and interpreted using conventional oil-exploration methods. Two prominent reflections are observed. Both of these tie to a vertical seismic profile (VSP) dataset collected in the same area. The upper reflector corresponds to the refusal depth of the well from which the VSP was acquired. This reflector can be interpreted as either the top of fractured bedrock or the top of an overlying till layer. On the basis of previous refraction studies, we prefer the former interpretation. The lower reflector shows less topographic relief and is approximately 61 m deep. It is distinguished by larger amplitudes and is spatially more continuous. We interpret this reflection as originating from the top of competent bedrock. Amplitude anomalies along each reflection are examined. We interpret the observed amplitude anomalies as being indicative of mineralogic changes within the bedrock, which could be related to remineralization of open fractures or to the presence of pegmatites. Compared to drilling shallow test wells or acquiring less costly geophysical datasets, 3-D seismic surveys at this site have limited economic viability. Investigations that might be deemed suitable for the application of 3-D near-surface seismic include those that (1) contain aquicludes of sufficient thickness, lateral extent, and velocity contrast to be resolved by seismic methods, (2) have targets that are relatively deep (>15 m), (3) display significant spatial variability in target depth, target properties, or target continuity, and (4) require the use of noninvasive methods.