Importing and separating a Digital Elevation Model (DEM) for near surface geological models in EarthVision.

The top surface of 3D Geological Models that represent the near surface geology in catchments is the surface topography. Data that define the topographic surface can be obtained from many sources including satellite, land-based and airborne geophysical surveys, geological field mapping and borehole logging. The most common form of topographic data used in spatial data analysis are the freely available Digital Elevation Models (DEMs). Often sediment filled valleys are surrounded by mountains with exposed rock outcrops. In this situation the DEM defines the top of the rock surface and the top of the sediments. When building a 3D EarthVision geological model of a catchment the DEM needs to be separated into the two domains. This manual details how this can be achieved in EarthVision. If the DEM data are in a GIS raster format then the data need to be preprocessed in a GIS package before being imported into EarthVision. For this manual the DEM is firstly imported into ArcGIS, cleaned, trimmed and then exported into a format compatible for EarthVision. The data used for this example are from the Maules Creek catchment located in the state of New South Wales, in Australia. The methodology developed for this case study can be successfully applied on any area of steep rocky mountains surrounded by flat alluvial plains. In these locations the gradient of the DEM data is sufficient for sorting the data into regions of rock outcrop and alluvial plains. High gradient (the rock outcrop) are sorted from the flat areas (the alluvial sediments) by calculating the partial differential in X and Y directions. In EarthVision there is a Slope Grid program. Using this program to separate low gradient alluvial areas from step rocky areas may be appropriate in some catchments. The process detailed below although based on examining the slope of the surface, has some subtle variations compared to using the Slope Grid program and yields a slightly different degree of separation. The major differences are adding the X and Y partial differential grids and then filtering this output.