3D Reconstruction and Scientific Visualization of Soil-Landscapes (WI and OH)
S. Grunwald, P. Barak, K. McSweeney, and B. Lowery Department of Soil Science, University of Wisconsin-Madison.
The objective was to investigate the use of Virtual Reality Modeling Language (VRML) to create 3-D soil landscape models at different scales ranging from pedon (1 m2), catena (2.73 ha), catchment (23.7 ha) to soil region scale (>100.0 ha). These soil-landscape models describe and visualize the three-dimensional distribution of soil and landscape attributes.
Scientific visualization (SciVis) transforms numerical symbolic data into geometric computer generated images. According to Barraclough and Guymer (1998) it is one of the most powerful communicators of spatial information. Advanced visualization techniques better communicate spatial information between people of different backgrounds such as scientists, administrators, educators and the public. Just as maps can visually enhance the spatial and temporal understanding of phenomena, 3D representations can enhance our understanding of soil patterns. Interactivity enhances the perception and interpretation of soil-landscapes. According to Stibbard (1997) information is absorbed best when using more than one human sense:
Virtual reality (VR) has different meanings. Full or immersive virtual reality requires the participant to be subject to stimuli affecting many senses, including vision, hearing, balance and touch. Such systems require head-mounted displays, audio speakers, moving platforms, and tactile gloves. Immersive VR systems are expensive and access is limited. Problems associated with immersive VR encompass cybersickness (problems with orientation and motion sickness symptoms) and psychological disturbance when the user returns from the synthetic world to the real world. Desktop virtual reality is the most commonly used form of VR systems, due to the fact that it can be presented on standard computer monitors. Here, conventional PC software is used to create and view artificial worlds in the office and over the internet. The World Wide Web (www) provides a desktop-based virtual environment (VE) where users can interactively navigate though VEs, they can interact in real time with objects, and have feelings of presence. Desktop VR is useful for representations of environmental systems, because it provides 3D capabilities, interactivity, and assists making extremely complex system transparent and supporting scientific interpretation and analysis of the natural environments. >
Virtual Reality Modeling Language (VRML) is a 3D object-oriented graphics language, suitable for stand-alone or browser-based interactive viewing. It is portable across platforms >and has an open standard (current ISO standard 'VRML97' accepted by International Standards Organization's (ISO) JTC1/SC24 committee). VRML worlds are accessible via the www. In VRML, 3-D objects are models extending in three dimensions. A VRML object has a form or geometry that defines its 3-D structure and it has an appearance based on the material and color from which it is made and its surface texture, like wood or brick. Coordinates of objects are defined using a 3-D coordinate system with x, y, z axis. The origin is defined by the triplet 0 0 0 representing x, y, and z-axis. The shape of every object has to be defined using triplets in relation to the origin.
Stibbard A., 1997, Warwick University Forum, No. 6.
Commonly, 2D maps are used to visualize the spatial distribution of soil and landscape patterns. 2D digital elevation model (DEM) and soils map (SSURGO data, NRCS) for a site in southern Wisconsin Geographic information systems (GIS) are still the most common tools to store, analyze, and visualize digital soil and landscape data. The most widely used digital soil data in the U.S. provided by the Natural Resources Conservation Service (NRCS) in the National Soil Geographic Database (NATSGO), State Soil Geographic Database (STATSGO), and Soil Survey Geographic Database (SSURGO) are attribute tables and 2D ArcView GIS shape- files. Other soil-landscape representations use a 2½D design, where soil or land use data are draped over a digital elevation model (DEM) to produce a 3D view. Since this technique describes patterns on 2D landscape surfaces rather than the spatial distribution of subsurface attributes (e.g., soil texture, soil horizons) it fails to address three-dimensional soil-landscape reality. Numerous 3D sketches of soil-landscapes can be found in Soil Survey Manuals. However, these mental models do not utilize field data nor do they utilize a geostatistical method. The relatively few 3D representations of soils at landscape-scale currently available are striking. For example, the Cooperative Research Center for Landscape Evolution and Mineral Exploration constructed a 3-D regolith model of the Temora study area in Central New South Wales, Australia (http://leme.anu.edu.au), and a 3D soil horizon model in a Swiss floodplain was created by Mendonça Santos et al. (2000) using a quadratic finite-element method. Sirakow and Muge (2001) developed a 3D Subsurface Objects Reconstruction and Visualization System (3D SORS) in which 2D planes are used to assemble 3D subsurface objects.
Mendonça Santos M.L., C. Guenat, M. Bouzelboudjen, and F. Golay. 2000. Three-dimensional GIS cartography applied to the study of the spatial variation of soil horizons in a Swiss floodplain. Geoderma 97: 351-366.
Sirakov N.M. and F.H. Muge. 2001. A system for reconstructing and visualizing three-dimensional objects. Computers & Geosciences 27: 59-69.
(1) Reconstruction and scientific visualization of soil-landscapes (geo-data modeling): The simplest geographic data model of reality is a basic spatial entity, which is further specified by attributes and geographical location (spatial coordinates or geometry) and relationships (topology). This can be further subdivided according to one of the three basic geographical data primitives, namely a ‘point’, a ‘line’, or an ‘area’. We used the geographic entities ‘pixels’ and ‘voxels’ (volume cells) to model real soil-landscapes. The input datasets to create soil-landscape models comprised x, y, and z (depth) coordinates, elevations, and subsurface attribute values (e.g. texture, bulk density, soil water content, etc.). Geostatistical methods were used to create continuous soil-landscape representations. We distinguished three different model types: (a) Models representing subsurface attributes as points: These are the simplest type of virtual soil-landscape models. The PointSet VRML class (node) was used to create point geometry. (b) Models representing subsurface attributes as polyhedrons or “volume objects” (stratigraphic models). Surfaces of models were created using 2D ordinary kriging and volumes (layers) were created with linear interpolation in the vertical direction between these surfaces. The IndexedFaceSet VRML class was employed to render polyhedrons (e.g., representing soil horizons) Steps (pdf file) (c) Models representing subsurface attributes as voxels (block models). Soil attributes were interpolated based on the spatial structure identified in 3D variograms, which plot semivariance on the z-axis, distance (h) between data pairs in the x-y plane (horizontal) on the x-axis, and distance (h) between data pairs in the z-plane (vertical) on the y-axis. Variograms are displayed in three-space as a surface. The observed values were interpolated horizontally and vertically using 3D ordinary kriging. The IndexedFaceSet VRML class was employed to render voxels. Steps (pdf file) We used Environmental Visualization System (EVS) (EVS-PRO; CTech Development Corporation, Huntington Beach, CA) to create soil-landscape models in VRML format. Our models visualize the 3D spatial distribution of subsurface and topographic attributes. Colors and surface textures were used to specify the appearance of objects. Subsurface attributes were portrayed using the red-green-blue (RGB) color specification system and topographic attributes were portrayed on the z-axis. The VRML capable browser automatically computes shading to give objects a 3D appearance.
3D point model showing the spatial distribution of soil cores which were analyzed for bulk density in 5-cm depth increments (site location: southern Wisconsin)
3D block model showing the spatial distribution of bulk density across a 2.73-ha site in southern Wisconsin
Grunwald S., V. Ramasundaram, N.B. Comerford and C.M. Bliss. 2006. Are current scientific visualization and virtual reality techniques capable to represent real soil-landscapes? pp. 571-580 (chapter 42). In Lagacherie P., A.B. McBratney and M. Voltz (eds.), Digital Soil Mapping - An Introductory Perspective. Developments in Soil Science Vol. 31, Elsevier, Berlin.
Grunwald S. 2006. Reconstruction and three-dimensional scientific visualization of soil-landscapes, pp. 373-392. In Grunwald S. (ed.), Environmental Soil-Landscape Modeling - Geographic Information Technologies and Pedometrics. CRC Press, New York.
Chen S.-S. and S. Grunwald. 2005. The spatial/temporal indexing and information visualization genre for environmental digital libraries. J. of Zhejiang University Science (JZUS) 6A(11): 1235-1248.
Grunwald S. and P. Barak. 2003. 3D Geographic reconstruction and visualization techniques applied to land resource management. Transactions in GIS 7(2): 231-241.
Grunwald S. and P. Barak. 2001. The use of VRML for virtual soil landscape modeling. Systems Analysis Modelling Simulation 41: 755-776.
Grunwald S., K. McSweeney, D.J. Rooney, and B. Lowery. 2001. Soil layer models created with profile cone penetrometer data. Geoderma 103: 181-201.
Grunwald S., P. Barak, K. McSweeney, and B. Lowery. 2000. Soil landscape models at different scales portrayed in Virtual Reality Modeling Language (VRML). Soil Sci. 165(8): 598-615.