Carbonate reservoirs are of significant importance as they host a high percentage of the world’s oil and gas reserves. They are however subject to dissolution and may be heavily karstified often resulting in large cavities which may impose major hazards in drilling and production. In many cases, difficulty remains due to poor understanding of karst geometry and how it is distributed. This research presents a geometry and stability analysis using LIDAR (Light Detection and Ranging) data on cave systems in Brazil that can serve as direct analogues to subsurface carbonate reservoirs and therefore enhance our ability to predict them. Our study area lies in the state of Bahia in Brazil and consists of five caves: Diva De Maura, Torrinha, Ioiô, Lapinha, and Paixáo. These cave systems occur in Neoproterozoic carbonates of the Salitre Formation in the Irecê basin of the São Francisco Craton. Analysis of the geometry shows existing preferential karstic developments and tunnel shape patterns mainly classified in two, horizontal ellipse and vertical ellipse. Analysis of geologic structures and orientations of cave passages show that these caves exhibit marked structural guidance with two main orientations N-S and E-W. This aligns with the regional phases of deformation corresponding to collisional events that occurred on the margin of the São Francisco Craton. Stability analysis shows that tunnels aspect ratio is a major parameter controlling the stability, whereas the dimension of a tunnel has a much less effect on stability. Tunnels with aspect ratios less than 1 show a collapse depth greater than 1000 m and are more stable than tunnels with aspect ratios greater than 1. Results show that almost all tunnels with a horizontal ellipse geometry are supported with rock that is already undergoing severe deformation at a depth of 500m. Sensitivity tests show that overburden and rock mass strength parameters namely cohesion have the largest effect on the stability results.