Generating Viewsheds:
Using Digital Elevation Data to Create Soundscapes for Archaeology

The use of viewsheds in archaeology has become increasingly common in recent years. The ways in which archaeologists employ this ability of GIS is diverse, but each use begins with the same simple concepts and methods. This tutorial will equip you with the ability to generate viewsheds.

Programs Used
ArcMap
ArcScene

Extensions Required
3D Analyst or Spatial Analyst

A viewshed, derived from the term watershed, computes a raster image of all points viewable from a single point on a landscape. Therefore, in order to create a viewshed, you will need some kind of raster that has elevation data (you can generate your own elevation models with this tutorial). Also, you will need to create a point shapefile to represent the viewpoint from which the viewshed will be cast.

In this tutorial, I am creating a viewshed in the hopes that it will also function as a soundscape, but more on this later. First, we want to add the digital elevation model (DEM), in this case a tingrid created from countour lines) to an empty ArcScene document.

I prefer to actually have the DEM conform to its elevations (the above image shows a DEM still in 2D), lets make this raster into a 3D image. In order to do this, you will want to right-click on the tingrid in the table of contents and access its properties.

Once the Properties dialogue box opens, click on the base heights tab and select 'Obtain heights for layer from surface' (the tingrid should be automatically selected in the drop-down box as it is the only DEM in the ArcScene document).

Once you click okay, the tingrid will be re-drawn in ArcScene in 3D.

The next step is to add the point shapefile to the above ArcScene document. In order to make sure that your point shapefile is occupying a 3D point in space, you should repeat step above to set the points base height just like you did the DEM. You may also want to offset the point a few meters/feet by entering a value in the 'Offset' box at the bottom of the Properties box.

Once you click okay, we are ready to generate the viewshed with the 3D Analyst or Spatial Analyst extension (both have Surface Analysis capabilities). In order to create the viewshed, simply select 'Viewshed' underneath 'Surface Analysis' on the '3D Analyst' or 'Spatial Analyst' toolbar.

This will launch a new window where you can set the viewshed's parameters. The Input Surface will provide the elevation data neccessary for determing which points are viewable on the landscape from the Observer points. You can also save the output raster (which will have two values, visible and non-visible).

Once you hit OK, the program will computer the viewshed and add it to the ArcScene document.

That's it! Its actually not very hard to generate a viewshed, what is sometimes challenging is finding ways of using this information in an archaeological context. While in New Zealand, I had a chance to work on a site called Nenthorn, which was goldmining site. What follows is a brief excerpt from an upcoming article of mine that discusses one way of using viewshed analysis for answering a question which occured to me and others while working at the site of Nenthorn.

The use of viewsheds for quantifying a perception on the ground presents interesting possibilities. The sites of Nenthorn includes hardrock mining operations, and their subsequent stamping batteries. A stamping battery, also known as a stamp-mill, is very loud when operating. While visiting these two sites with Peter Bristow, he commented that sound may have guided the placement of structures in relation to these loud machines. In essence, we considered that the natural landscape may have been utilized by the workers in order to shield them from the noises of these mills, and locations which provided geographical buffers from them might have been sought out.

Location of town and stamp mills at Nenthorn, New Zealand

The organization of features on the landscape at Nenthorn was specifically intriguing in relation to this idea. Specifically, the location of the town itself was situated above and away from the stamp mills. Of course, the placement of stamp mills is partially one of economics, in response to factors such as the availability of water, location of mine entrances, and suitable land features for construction. The location of the town is on a portion of land with a gentle slope. However, other areas closer to the stamp mills offer the same gentle slope, but the town and individual miner huts are not placed there. The figure below shows two other possible locations for the town or individual miner huts, but where no remains currently exist.

Possible habitation areas at Nenthorn exhibiting no structural remains.

Using the viewshed to compute soundscapes from the two mills provided at least one possible reason why no habitation structures were placed in the other two possible locations. Figures fourteen and fifteen show how line of site from the stamp mills include portions of the other possible habitation areas. Of course, soundscapes do not directly correspond to line-of-sight, but the viewsheds do suggest that sound might have been a determining factor in discouraging the placement of huts in these areas. While the placement of the town itself was most likely the result of numerous factors, there is not apparent reason that individual miners wouldn’t have placed their own huts closer to the sites were they worked.

Viewshed from Stamp Mill 1 clearly impacting the second possible habitation location.

Viewshed from Stamp Mill 2 clearly impacting the first possible habitation location.

Of course, there are certainly other ways to employ viewshed analysis within archaeological studies. Here is a link to an article by David Wheatley that combines viewshed and statistical analysis in an attempt to determine the motives behind mound placement in Great Britain.