Deepwater pre-salt lacustrine carbonate reservoirs in the Santos Basin hold hydrocarbon potential but present exploration challenges due to supercritical CO2. The dual physical nature of supercritical CO2 and reservoir heterogeneity make conventional single-domain fluid identification ambiguous. This study presents an integrated engineering workflow coupling log-scale petrophysical characterization with field-scale frequency-dependent AVO (FAVO) inversion to delineate these CO2 accumulations. We establish semi-quantitative fluid identification cross-plots based on the zero hydrogen index of CO2 using density-neutron and differential NMR logging data. To spatially extend this characterization, we implement a FAVO inversion via regularized least-squares optimization on time-frequency decomposed pre-stack angle gathers to extract the P-wave dispersion gradient. While conventional static AVO attributes are ambiguous due to the acoustic impedance overlap between crude oil and supercritical CO2, this dynamic gradient isolates the CO2 phase by exploiting its high fluid mobility and squirt-flow dispersion. Validation against recent well data confirms that the petrophysical models achieve up to 75% identification accuracy. Field application in Block P uses the P-wave dispersion gradient to delineate the spatial boundary between a structural CO2 gas cap and the adjacent flank oil ring. This framework provides a target-specific predictive application to mitigate exploration risks and optimize field development in CO2-rich geological settings.