Analysis and Integration
From www.archaeogeophysics.org
Integrating Geophysical Data in ArcGIS
One of the most powerful and dynamic tools for interpreting processed geophysical data is GIS programs such as ArcGIS. Geophysical data may be integrated into GIS by directly importing processed data as image files or interpolating raw data within ArcGIS itself. Either way different datasets may be overlain on top of one another as layers. These layers may then be compared using statistical processes located in the programs toolbox.
The tutorial will use the data sets collected in Zzyzx, CA.
Methodology for Integrating Geophysical Data
The first step after collecting our data in the field was to process the data and import it all into ArcGIS. All of the data from the GPS unit and the geophysical instruments must be extracted and imported into ArcGIS. The data can be directly imported into ArcGIS or imported as Raster or Shape files via a program called Surfer. Regardless of how the data is imported from the geophysical instrument it must be georeferenced to the points collected on the GPS. Because we recorded the locations of the corners of our survey grid we can then georeference the corners of the geophysical data sets so that the data is overlain on the base data (topographic maps, satellite images, aerial photos, etc.). Georeferencing can be easily performed by using the georeference toolbar. The corners of the images imported into ArcGIS including the images that we captured of the area with the aerial camera rig and raster datasets including the GPR data must be referenced to the known coordinates of our grid points. After the corners are tied to the correct coordinates then you must update the georeferencing or rectify the image. If you rectify the image you risk loosing some clarity in the image. The image can also become distorted. Rectifying the image creates a new file and layer, whereas simply updating the georeferencing orients and size the original image properly and a world file is created that is associated with the image. After georeferencing the image, a projection should be defined for the image that corresponds to the projection associated with the reference points. The next step that is necessary in order to perform our spatial analysis is to convert the raster layer associated with our GPR data to a vector layer (Figure 1). This can be done by using the Conversions Toolbox. The “From Raster to Polygon” function allows you to simply change the color values associated with the pixels in the raster to individual polygons. We were interested in the areas of the GPR raster that were represented by red because they represented areas of high dielectric values within the subsurface of our grid. These high values are probably associated with subsurface features that are much harder and compact than the surrounding soils suggesting that they might have previously been used for sedentary activities.
After generating a shapefile, the areas of high dielectric values were selected by the numerical value associated with them and a layer file was created from the selection (Figure 2). This isolated the desired features into their own shapefile. Then the artifacts layer was queried and a selection of all artifacts within 2 meters of the subsurface features was highlighted. The results of this query selected a large percentage of artifacts within our grid. The majority of the artifacts and artifact clusters that we had recorded locations for in our GPS were located within at least 2 meters of subsurface features with high dielectric values. This tells us that there is a possible non-random relationship between the spatial distribution of the artifacts on the surface and characteristics of the underlying subsurface.
Because the relationship between the spatial distributions of the artifacts on the surface seemed to be non-random with respect to the location of the subsurface features, we sought to explain the distribution of the artifacts in relation to the subsurface feature with respect to topography.
Elevation
In order to prove that the local topography plays a role in how the artifacts were distributed we need to show that the artifacts align themselves along key topographical gradients. Since we wanted to examine the relationship between the elevational differences and the artifact distribution we decided to import the topographical GPS points we took with our Trimble unit and bring it into ArcGIS. Since the Trimble Pathfinder software does not import directly export into a .dbf file (at least not in a straightforward way that we were aware of), we exported the data under the setting “Sample Configurable ASCII Setup” as a tab delimited text file (Figure 3). In this menu of Pathfinder were able to select the projection and coordinate system we were going to export our data under as well as any critical details of the features we wanted to examine.
From here we opened up the text file under excel and then saved it as a .dbf (IV) file to import in ArcGIS. I then added the flat file into ArcView. From here we went into the “Tools” menu function and clicked “Add X/Y Data”. From here I got to choose which column was the X and the Y as well as define the spatial reference that the points were going to be under (Figre 4). The spatial reference we chose was the Projected Coordinate System UTM Zone 11N with the WGS 1984 datum since the original data was also exported under the same constraints.
The geographic data was then imported as a series of points as shown in Figure 5.
From here we added the “Spatial Analyst” extension in ArcGIS in order to create a contour map of the area (Figure 6). First I clicked “Spatial Analyst” and then selected “Interpolate to Raster.” We then selected the Kriging method since I preferred this interpolation method over that of inverse distance. We then set the Z field heading in the attribute table as my “Z value field” and hit “OK”. The resulting map presented below was okay but did not sufficiently highlight the differences we wanted to see between gradients. We then right clicked this layer in the table of contents and opened up its “properties”. Under the Symbology and Classified tabs I then hit the “Classify” button and set the intervals under the “Quantiles” method since the “Equal Distance” intervals did not visually highlight what I was trying to see. We then switched the GPS point layer on and off to see where the interpolations were going to be the most precise. Naturally the contour map was going to be the most accurate at the areas with the most points. Once that was done we turned off the GPS point layer
While the raster data helped me see the overall gradients in elevation we also wanted a more definitive and structured look at the elevation. For this we wanted a line contour map. Again we went into the “Spatial Analyst” extension and went into the “Surface Analyst” function and selected the “Contour Option” (Figure 7). For the “Z factor” we chose 0.25 since 0.5 and 1 drew too many lines to see anything clearly.
Slope
After viewing the elevational data it also occurred to us to try and examine the slope of the elevational gradient as well (Figre 8). To do this we once again went into the “Spatial Analyst” extension and selected the “Surface Analyst” function. This time chose “Slope” instead of “Contour”.
We then added in the artifact distribution data and examined the relationship between the artifact distribution, the geophysical data, the elevational differences, as well as the slope gradient. Once we positioned the data we wanted displayed and messed with the color settings to both our satisfaction we then went into the “View” menu and chose to display our data in the layout view in the form of a map and added in the symbols, legend, text, orientation, and scale. We then went into “File” and clicked “Export Map” to move our analysis into a readily transferable format.
See Also: