IN043-0004
Automated Subsampling of Geological Maps: A step towards 3D Geological Multiscale Modelling
Wednesday, 16 December 2020
Poster
Ranee Joshi, University of Western Australia, Centre for Exploration Targeting (School of Earth Sciences), Crawley, WA, Australia, Mark Jessell, The University of Western Australia, Centre for Exploration Targeting (School of Earth Sciences), Crawley, WA, Australia and Mark Lindsay, University of Western Australia, Centre of Exploration Targeting (School of Earth Sciences), Crawley, WA, Australia
Abstract:
Three-dimensional modelling is a powerful visualization and communication tool that proves useful in solving geological problems. Currently, the geologic modelling process produces a
single static model created with geological data upscaled and/or downscaled to a predetermined scale appropriate to answer specific geological questions. However, this original scale is not always optimal for other purposes sought after the model is finalized. Moreover, these scaling methods do not mimic the way geology occurs and may involve loss of critical information or data imputation. T
his results into models that limits the meaningful geological information provided and increases the inherent risk involved in relying on the model to make informed geological assessments. We propose a provide a new approach to refining geological understanding by producing multiscale geological models that allows data to be dynamically subsampled and visualized at a scale optimal to better answer specific geological questions.
This contribution presents a workflow on automatically subsampling geological maps and its initial findings on upscaling geological vector maps based on stratigraphy hierarchy and by means of vector simplification. The case study for this work uses the 500K Geological Survey of Western Australia (GSWA) geological vector maps of the lithologically diverse and structurally complex Yalgoo-Singleton Greenstone Belt located in the Murchison Domain, western Youanmi Terrane, Yilgarn Craton. As a proof of concept, the three scales considered for the multiscale model are: camp scale, regional scale and large scale. The workflow addresses the selection and emphasis of relevant geological features of interest in a given scale. It also aggregates and simplifies these features using a modified Visvalingam-Whyatt algorithm that at a vertex reduction of up to 95% in a 1:500,000 map can preserve the features’ shape and topological relationships readable up to a camp scale level (1:100,000). The code to this workflow will be compiled and be available as a part of the Python package map2loop (https://github.com/Loop3D/map2loop).
