Time-lapse cross-well electromagnetic (EM) surveys are used to monitor two types of fluid injection (Water Injection and Water Alternating Gas) in a giant field in the Middle East. Cross-well EM data will help optimize sweep efficiency, identify bypassed pay, and predict fluid-related issues such as water breakthrough by providing an image of the resistivity distribution between boreholes in time lapse. This paper explores the influence of a high quality background geologic model in constraining the interwell results and providing a higher resolution image of the ongoing flooding processes. The classic EM inversion process determines a coarse (3 to 5 m resolution) resistivity distribution from a basic initial static reservoir model built from logs. This study refines the model by adding variable resolutions to encompass the small-scale heterogeneities common to carbonate reservoirs. Incorporating geological data derived from seismic attributes, core descriptions, and detailed log analyses into the static model helps optimize the EM inversion and increases the resolution of the resulting inverted model.