We carried out a multidata geophysics study in southern Colorado to explore for CO2 reservoirs in an area where seismic imaging is very limited due to the mountainous terrain, the presence of high-velocity volcanic rocks, and difficulty in obtaining land access permits. We have developed a modeling/interpretation methodology using ground magnetotelluric data as well as airborne magnetic and electromagnetic data combined with public domain gravity data and existing well and seismic data. We used the integration of these data sets to produce a series of 2D and 3D geophysical models that reveal basin architecture previously poorly defined through the analysis of limited seismic and well data alone. We found that this type of analysis aids in decreasing uncertainty in the interpreted geologic cross sections and a better understanding of the structural complexities of the region. Through the application of machine learning methods, we are also able to integrate several data sets into a mathematical framework resulting in a predictive model of spatial CO2 distribution. The integration of the interpretations from all data sets, predictive analytics results, and knowledge of CO2 production, allows us to delineate areas of interest for further exploration.