Paper No. 5
Presentation Time: 10:00 AM
3-D VISUALIZATION OF SISTEMA ZACATóN: INTEGRATING MULTIPLE SPATIAL DATASETS TO CHARACTERIZE EXTREME KARST
Deep karst aquifers represent some of the most complex hydrogeological features known, and characterizing these three dimensional voids and interpreting the various geological and hydrological relationships between sub-surface and surficial features presents unique challenges. Three technologies: lidar, electrical resistivity geophysics, and phased array sonar provide the ability to model such features in detailed 3-D and link metadata to the model through the use of immersive interactive graphics. Zacatón, the deepest underwater sinkhole in the world, has a sub-aquatic void space exceeding 1.5 x 106 cubic meters, and was the focus of this study in which a detailed 3-D map of the entire system was created. The interactive map includes data from above the ground surface, beneath the water table, and in the rock matrix itself. It is used to gain more accurate knowledge of the extent of these immense karst features and interpret the geologic processes that formed them. Phase 1 of the research used high resolution lidar scanning of surficial features, including four of the largest cenotes in the system. A ground-based, static mount equivalent to aerial platform based LIDAR imagery collected 3-D point cloud data under tree canopies, manmade structures, within overhung rock structures and in air-filled caves. Phase 2 included multiple electrical resistivity surveys of sealed sinkholes, and revealed likely massive water-filled voids exist beneath thin skins of travertine formed at the surface. Phase 3 was completed by the DEPTHX project, which mapped all accessible underwater environments in detail using sonar aboard a fully autonomous underwater robot. A large suite of aqueous geochemical data were also collected by the DEPTHX robot. Data from all three phases were processed and integrated into a variety of formats, including Adobe 3D pdf, allowing for broad distribution of attributed models.