Picturing the Subsurface With GIS:
Data Visualization Techniques for Geophysical Images Using ArcGIS
By Lucas Donny Setijadji
Geoscientists try to understand the earth's crust using geophysical methods such as gravity and magnetic surveys. Mining geologists also use geophysics to search new mineral deposits. In areas such as the Nevada Great Basin, geoscientists rely more on geophysical images than on geologic maps.
The Great Basin area holds one of the biggest mineral concentrations in the world.
Economically, gold is the most prominent commodity. The most important gold deposits are hosted by carbonate rocks, called the Carlin-type and Carlin-like gold deposits. Carlin-type and
Carlin-like gold deposits are unique geologically, are abundant in Nevada, and are hardly found elsewhere. Although much research has been done, a lot of the geology is still unclear.
This paper discusses how ArcGIS provides a tool for research on mineral resources. Emphasis is on the functionality of ArcGIS to handle multiple data sets and maps as well as data visualization techniques using ArcGIS.

Great Basin Geology
Figure 1 shows the distribution of known gold deposits combined with topography, young geologic rocks, and county lines. Data sets were downloaded and compiled from several United
States Geological Survey (USGS) Web sites (e.g., Raines et al., 1996). Data processing was done using ArcGIS 8.3 and its 3D Analyst extension.
The Great Basin is complicated by its physiography and surface geology. The area is one of the largest Cenozoic continental rift systems in the world. Its topography is characterized by a patchwork of mountain ranges and intervening valleys—it is thus also called the Basin-andRange geologic province. Most mountain ranges are bounded by young, normal faults that raised and tilted the blocks. The intermountain basins are filled with Quaternary period sediments derived from the erosion of the mountains. Together with upper Tertiary period rock units younger than the period of gold mineralization, such cover rocks occupy about 80 percent of the area.
Gold deposits are found through windows of older rocks (Mesozoic to Paleozoic in age) exposed along the narrow ridges between the basins. The mineralization occurred before the last geologic event that created the current basin-and-range phenomenon. This means that most geologic evidence related to the formation of gold deposits has been overprinted and buried by younger events. Any geographic information system (GIS) spatial analysis that ignores the concept of geologic time and the subsurface phenomena that are genetically related with mineral deposits will be misleading.

Figure 1—Great Basin Area

In the northern Nevada Great Basin area, the young extension faults created a patchwork of mountain ranges and intervening valleys, mostly in the northeast–southwest direction. Most areas are covered by upper Tertiary and Quaternary period rocks. Gold deposits are discovered within windows of older rocks that usually occupy the ridges.

Geophysical Images and Subsurface GIS
Gravity anomalies are produced by density variations within the rocks of the earth's crust and upper mantle. In the Great Basin, the gravity data allows the examination of large-scale changes in density within the basement. The gravity data comes from the isostatic residual gravity anomaly map of the western USA after the removal of the effect of Cenozoic basins in the Great
Basin area, provided by Dr. Thomas Hildenbrand from the USGS.
Magnetic anomalies are produced by variations in the distribution of iron minerals, usually magnetite, in the rocks of the earth's crust. Igneous and metamorphic rocks can be very magnetic. By comparison, sedimentary rocks are usually nonmagnetic. Magnetic anomalies, therefore, provide a way of mapping exposed and buried crystalline rocks. In this study, the magnetic data was downloaded from the Western U.S. Composite Magnetic Grid (Kucks, 2001).
To extract gravity and magnetic anomalies directly related to subsurface phenomena, digital filtering is applied to the original data. The goal is to separate the regional trends and residual anomalies. The digital separation processes sometimes involve the transformation of data from the space domain into the wave–number (Fourier) domain. This filtering process requires specific geophysical software such as the MAGMAP product from Oasis Montaj©.

The filtering algorithm applied to the gravity map was a band-pass filter to find intermediatewavelength gravity anomalies. These anomalies represent deep-seated geologic units and structural configurations (Figure 2). Meanwhile, the upward continuation filter for 1 km was applied on the aeromagnetic map to eliminate the noise from shallow magnetic sources. The result represents magnetic bodies at a significant depth below the surface (Figure 3).
Figure 2—The IntermediateWavelength Gravity Anomaly Map After
Application of Band-Pass Filtering

Figure 3—The Magnetic Anomaly
Map After the Application of 1 km
Upward Continuation Filtering

Data Integration With GIS and Some Concerns
Geophysical processing software usually lacks GIS functionality. Spatial analysis is much easier when done in a GIS environment. Transformation of geophysical images to a GIS format can be as simple as exporting the images that contain spatial information into geoTIFF format. In a GIS analysis, geophysical images are often treated as a type of evidence map, the same as other thematic maps such as mineral point, geologic, and structure maps.
For a vector-based GIS operation, GIS analysts usually create polygons by heads-up digitizing on raster images and adding attributes (e.g., high and low anomalies). For raster-based operations, polygons are converted into raster format. All evidence maps are then combined through map algebra (e.g., Boolean, Bayesian, fuzzy logic) to create a prediction map with high potential for new deposits. Such typical GIS operations can be found in the GIS textbooks such as Geographic Information System for Geoscientists by Bonham–Carter (1994). This multimap modeling has been widely accepted and is taught in many graduate institutions. The author experimented in his own graduate course with how to model potential maps, applying mathematical algebra on many evidence maps (including geophysical maps) using a GIS.
The author is, however, always concerned with the nature of the geophysical images, which differ from other data sets. It is important to stress that geophysical images are continuous phenomena in which values have a wide range and change gradually. It is different with discrete objects, such as mineral deposits or geologic units, that have entities with sharp boundaries.
Delineating the boundary of geophysical anomalies is a transformation process from continuous to discrete objects. Although several image algorithms have been developed to enhance such

thresholds and help with delineation, picking the threshold lines between low and high anomalies is an approximate and subjective process, which therefore, introduces a certain degree of error.
The second consideration is the nonunique relationship between geologic phenomena and geophysical anomalies. In a complex earth, there is no one-to-one relationship between geologic formations and their geophysical signature. In the case of mineral deposits, it is even worse.
Although there are several observed relationships between these two entities, there are many exceptions that prevent the establishment of definite rules about these relationships. The author suggests that previously assumed rules and relationships used in GIS modeling were not always appropriate. The third consideration is that most geophysical images capture geologic phenomena from all levels of depth below the earth's surface. Meanwhile, the GIS spatial analysis is done on a 2D plane of the earth's surface. Subsurface features can shift considerably from their vertical surface projection, and many deep features do not continue up to the earth's surface.

Experiment on the Display of Geophysical Images
In researching how to handle geophysical images in GIS, the author found that ArcGIS allowed transparency in displaying the raster maps. In an introductory course on ArcGIS at the ESRI
Learning Center, how to create transparencies was shown. This is typically used to reduce the apparent intensity of a backdrop. This option is available from the display option within the
Layer Properties, and opens the possibility of displaying several raster maps simultaneously.
An experiment was done as follows in ArcGIS 8.3. First, the final gravity and aeromagnetic anomaly maps were exported into RGB geoTIFF file formats. Second, both images were displayed together in one data frame using transparency display. The magnetic map was placed at the bottom with no transparency (default 0%). The gravity map was placed above with a transparency of 50 percent. A third raster TIN map was placed at the top with transparency set to 65 percent. The TIN map overlay is in grey-scale, rather than in color as in Figure 1. Other vector data sets, such as the mineral deposits and county polygons, were added at the very top
(Figure 4).

Figure 4—Final Map

Final map as a result of combining three types of raster maps (magnetic, gravity, and TIN) using
ArcGIS 8.3. The distribution of gold deposits is now shown to be controlled by deep-seated and old structures depicted by the merging of gravity and magnetic anomalies.

Results
The experiment of merging raster images in ArcGIS 8.3 was successful. The new map merges different geophysical features into one, while the original gravity and magnetic anomalies stay recognizable. The new map is superior to vector map operations in its information content.
The new map shows how the subsurface (basement) geology is different from the surface. While the latter is dominated by northeast–southwest trending young faults, the subsurface is controlled by older blocks with boomerang shapes trending northwest–southeast to northeast–southwest.
These deep blocks are considered to represent crustal phenomena of very old geologic age (such as Proterozoic). Subsurface magnetic anomalies are present within these blocks, especially along their margins. The relationship between the gold deposits and deep-seated structures is much more evident compared to surficial phenomena. Many major deposits are aligned in the northwest–southeast direction, parallel to basement trends. Other deposits are scattered along the margins of basement blocks. Only a few minor deposits show no clear spatial relationship with basement structures.

This experiment shows the importance of what data visualization technique is used. Data visualization using ArcGIS has revealed another way to deal with geophysical data. Rather than creating discrete representations of geophysical images, why not just display the full wealth of information contained in those images?
For additional information, contact
Lucas Donny Setijadji
Department of Geology
Faculty of Engineering, Gadjah Mada University
Grafika 2 Bulaksumur, Yogyakarta 55281 Indonesia
Tel. and Fax No.: +62-274-513668
E-Mail: lucas_donny@yahoo.com

About the Author
The author holds a bachelor's degree in geology from Gadjah Mada University, Indonesia, and a
Master of Science degree in geoinformation science with a specialization in mineral resources exploration and evaluation from the International Institute for Geo-Information Science and
Earth Observation (ITC), the Netherlands. He has been at ESRI in Redlands as an international intern since December 2002 and will be staying until June 2003. During this period, the author has worked with the industry solutions manager for petroleum and pipeline. Mr. Setijadji is a lecturer and researcher at the Department of Geology, Gadjah Mada University, Indonesia.

References
Bonham–Carter, G.F., Geographic Information System for Geoscientists: Modelling with GIS,
Delta Printing Ltd., Ontario, 1994, 398 pp.
Hildenbrand, T.G., B. Berger, R.C. Jachens, and S. Ludington, "Utility of Magnetic and Gravity
Data in Evaluating Regional Controls on Mineralization: Examples From the Western United
States," Structural Controls on Ore Genesis, Reviews in Economic Geology, Vol. 14, Society of
Economic Geologists, Inc., Littleton, Colorado, 2001, pp. 75–109.
Kucks, R.P., "Composite Magnetic Anomaly Grid of the Western U.S.," Individually Published
State Compilations at Simulated Flight Altitude of 1,000 Feet Above Ground, n.d., (2001).
Raines, G.L., D.L. Sawatzky, and K.A. Connors, "Great Basin Geoscience Database," U.S.
Geological Survey Digital Data Series DDS-41, n.d., (1996).