Fundamental trapping mechanisms that occur during carbon sequestration at deep geo- logical conditions have been studied from different scientific perspectives throughout the last two decades. Fundamental capillary phenomena that determine the spatial distribution of non-wetting fluid migrating in saturated porous media need to be properly incorporated into numerical models in order to predict CO 2 plume evolution through the reservoir formation. The goal of our investiga- tion is to develop experimental methods and obtain datasets that will be used for model validation as well as a benchmark for more complex experiments conducted under different heterogeneous packing configurations in a 16 ft long tank. This paper presents the results from a set of experi- ments that were conducted in an intermediate scale test tank. Conducting these types of experi- ments is highly challenging, as methods have to be developed to extrapolate the data from experi- ments that are conducted under ambient laboratory conditions to high temperatures and pressures settings in deep geologic formations. We explored the use of a combination of surrogate fluids that have similar density, viscosity contrasts and analogous solubility and interfacial tension as super- critical CO 2 -brine in deep formations. The extrapolation approach involves the use of dimension- less numbers such as Capillary number (Ca) and the Bond number (Bo). The modeling analysis to verify whether existing models can capture the observed processes is carried out using TOUGH2/T2VOC codes developed by the Lawrence Berkeley National Laboratory. The results from the tank experiments and these model analyses are presented.