Fractures in carbonate strata are often associated with dissolution features such as karst cavities along the fracture plane. Integration of full resolution 3D GPR data and outcrop observation allows the visualization and quantification of fractures and karst cavites in three dimensions. Three 3D GPR cubes were acquired in the Solvay quarry, Provence region, SE France. The quarry exposes an intensly fractured and karstified shallow-water Barremian carbonates with matrix porosity less than 1.8%. The quarry walls provide the 2D information of the fractures (location, type, orientation, dip, length, aperture, spacing and stratigraphic position) and the karst (location, size, shape, stratigraphic position and relationship with fractures), that were documented using 1D scan line measurement. Several deformation features are observed in the quarry, dominated by meter-scale joints and tension gashes, and also minor faults, burial and tectonic-related stylolites. Three major fracture orientations (E-W, NW-SE and N-S) were found in the quarry with a non-uniform distribution influenced by structural position relative to major structural features (eg: large fractures (>10m) and faults). Interpretation of 3D GPR data reveals one major fracture orientation (NE-SW) that was not observed in the outcrop because it is oriented parallel with the quarry wall. Besides fractures, karst were also documented in outcrop as well as in 3D GPR data. Combination between 3D GPR and outcrop observation show that in the Solvay quarry karst occur along fracture plane and preferentially at fracture intersections. The karst are decimeter to meter-size karst which are below the resolution of 3D seismic reflection. Quantification of karst in the quarry show that karst size-frequency distribution follows power-law distribution, where small karst is more common and large karst is less common. The volume of karst in the quarry is up to 3.8% of the total rock volume. The power-law scaling relationship and the quantification of the sub-seismic karst volume will help predicting karst arrays in wide range of sizes in subsurface reservoirs, where visualization of karst is restricted to the resolution of 3D seismic data.