Lidar Investigation of Infiltration Water Heterogeneity in the Tamala Limestone, SW WA

Abstract:

To better manage groundwater resources in carbonate areas and improve our understanding of speleothem archives, it is important to understand and predict unsaturated zone hydrology in karst. The high level of complexity and spatial heterogeneity of such systems is challenging and requires knowledge of the typical geometry of karstic features.

We present an exhaustive characterization of Golgotha Cave, SW Western Australia, based on an extensive LIDAR measurement campaign. The cave is developed in Quaternary age aeolianite (dune limestone) and contains speleothem records. We collect 30 representative 3D scan images from this site using FARO Focus^3D, a high-speed 3D laser scanner, to visualize, study and extract 2D and 3D information from various points of view and at different scales. In addition to LIDAR data, 32 automatic drip loggers are installed at this site to measure the distribution and volume of water flow.

We perform mathematical morphological analyses on the cave ceiling, to determine statistical information regarding the stalactites widths, lengths and spatial distribution. We determine a relationship between stalactites diameter and length. We perform tests for randomness to investigate the relationship between stalactite distribution and ceiling features such as fractures and apply this to identify different types of possible flow patterns such as fracture flow, solution pipe flow, primary matrix flow etc. We also relate stalactites density variation with topography of the cave ceiling which shows hydraulic gradient deviations. Finally we use Image Quilting, one of the recently developed multiple-point geostatistics methods, with the training images derived from LIDAR data to create a larger cave system to represent not only the caves that are visible, but the entire system which is inaccessible. As a result, an integral geological model is generated which may allow other scientists, geologist, to work on two different levels, integrating different speleothem datasets: (1) a basic level based on the accurate and metric support provided by the laser scanner; and (2) an advanced level using the image-based modelling.