Geologic storage of carbon dioxide (CO2) is a possible strategy to combat rising CO2 concentrations and thereby mitigate climate change. Deep saline aquifers (DSAs) have been identified as repositories for CO2 sequestration. To assess the suitability of DSAs for CO2 sequestration, computer simulations are required to predict how a candidate repository would respond to proposed injection of supercritical CO2. Numerical simulators require the modeler to make estimates of a large number of physical and chemical parameters and multi-phase flow relationships, and many of these estimates are highly uncertain. Hence, the objective of this research is to quantify how estimates of CO2 sequestration feasibility depend on uncertainties in key physico-chemical parameters and multi-phase relationships. We used TOUGHREACT, a reactive-transport simulator, to perform a set of simulations in which estimated input parameters were varied over realistic ranges. For each simulation, we assessed how the change in input parameter(s) affected the feasibility of CO2 sequestration in a hypothetical carbonate-based DSA based on five metrics: pressure-induced fracturing, extent of CO2 plume, storage efficiency, fraction of injected CO2 dissolved in the aqueous phase, and change in formation porosity (induced by mineral dissolution and precipitation). For the conditions we tested, the most important sensitivities were: 1) pressure build-up, which leads to pressure-induced fracturing, increases as formation permeability decreases or as residual CO2 fraction increases 2) plume extent decreases if the residual CO2 saturation increases, if the residual brine saturation decreases, or if the van Genuchten "m" parameter decreases 3) the fraction of injected CO2 dissolved in the aqueous phase (solubility trapping) decreases if the residual brine saturation decreases. In conclusion, formation permeability and multi-phase flow parameters were found to strongly affect the feasibility of CO2 sequestration in a hypothetical carbonate-based DSA. Although multi-phase flow parameters are difficult to characterize, modelers should avoid the temptation to over-rely on "default" values provided in the literature.