Although quite speculative at this point, there are some suggestive relationships among the viewsheds of Cavanaugh Mound and mounds at the Spiro and Skidgel sites. I offer this analysis as one way to view the possible relationships of the mounds on the landscape and in relation to one another. GIS viewshed analysis is necessary for testing the potential intervisibility of the sites because of historic changes to the landscape. Dense vegetation on the mounds, houses and other buildings, and high road embankments now block much of the view, making it impossible to determine what would have been visible in prehistoric times simply by going to the mounds and looking.
A "viewshed" is the sum of what can be seen from any particular point on the landscape. From a digital elevation model (DEM) in GIS we can specify a location, the height of the viewer, a maximum potential distance to which a person can see, and derive an output showing us which portions of the landscape would be in view. The viewsheds in this analysis are based on U.S. Geological Survey DEMs with 30 meter resolution. There are several considerations which much be kept in mind when employing such viewshed studies, however. Atmospheric haze can often limit the maximum distance from which features can be seen, for example (termed distance decay), and modern landscape changes expressed in DEMs can significantly alter viewsheds from what they would have been in prehistoric times.
Modern landscape changes in the Cavanaugh area that have the potential to alter the viewshed include tall road embankments and changes in the course of the Arkansas and Poteau Rivers due to natural meandering and intentional modification. The road embankments are easily "erased" from the DEM and leveled to the elevation of the surrounding ground. The river channels may be similarly altered in the DEM to reflect previous positions, but I do not consider this necessary for the current study. The exact course of the rivers will affect the viewsheds within the bottomlands in possibly significant ways, but because the mounds are above the floodplain itself the exact river location will not alter the potential intervisibility of the sites, or the overall shape and size of the viewsheds in general.
I also do not consider distance decay to be a significant factor for this study. The sites in question are relatively close to one another (Cavanaugh and Skidgel are only about 16 km apart), and historic precedence shows us that atmospheric conditions in the region were often clear enough to allow for far greater lines of site. Thomas Nuttall's journal of his travels through the area in 1819, for example, includes the following passage for April 24:
This morning we passed the hills of Lee's Creek, which, for a short distance, borders the Arkansas, and about noon arrived at the garrison, which comes into view at the distance of about four miles, agreeably terminating a stretch of the river. Rising, as it were, out of the alluvial forest, is seen from hence at a distance of 35 miles, a conic mountain nearly as blue as the sky, and known by the French hunters under the name of Point de Sucre, or Sugar Loaf. (Nuttall 1821:197).
Tall vegetation can also obstruct views, but is not accounted for in the topography of DEMs. For the bottomlands I do not consider this to be a hindrance in determining the overall viewsheds or intervisibility of the sites. Cavanaugh, Brown, and Skidgel Mounds are all located on terraces above the bottomlands and several meters tall. Even relatively high vegetation in the bottoms would not obstruct the views from one site to another. Trees or other vegetation on the terraces near the mounds could easily block the views, however, and there is no way of reconstructing exactly what vegetation was present at specific times in the occupation of the sites. I therefore consider the following examples to be the maximum potential viewshed for each specific location, and not necessarily the exact viewsheds actually existing prehistorically. From what we know of the nature of the mounds, however, particularly the public nature of their construction and the use of colorful and contrasting colors for fill, it seems unlikely that they would have been intentionally screened from view. Possibly the ability to see the mounds from a great distance was even a consideration in deciding where to locate them.
For this analysis I used a somewhat conservative viewing height of 150 cm as the approximate eye-level for a prehistoric person of average height. The maximum viewing distance was set at 25 km. This is large enough to encompass all three sites, but small enough that edge effects from the curvature of the earth are negligible.
Figure 19 shows the viewshed from the ground level at the location of Cavanaugh Mound. Cavanaugh Mound is substantial enough on the landscape that it is expressed as a topographic high even on the U.S. Geological Survey 30 m DEM. For this analysis, the DEM was altered to produce a level surface in the location of Cavanaugh Mound, effectively "erasing" it from the landscape. The light areas in this figure thus represent what a person would be able to see, if no tall vegetation were present, standing at the location of Cavanaugh, before the mound was constructed. The light areas also represent portions of the landscape from which that person would be visible. The viewshed is primarily directed to the west and northwest, overlooking the Poteau/Arkansas River bottoms. The Arkansas River off to the east is out of sight. A low rise in the landscape about 150 m west of Cavanaugh blocks the view down into the southern part of the bottoms. A few of the hills above the bluff line to the north of the Arkansas River are visible as well. The viewshed from this analysis covers about half of the Spiro site, and comes within about 90 m of Brown Mound. 90 m in a DEM with 30 m resolution is of course only equal to 3 raster cells, and it would be unwarranted to assume accuracy of the model at this level of detail. What we can say is that from the ground surface, the Cavanaugh and Spiro areas may have been just barely visible to one another. If there were even low vegetation on the ground, however, the view would have been totally obscured. Skidgel Mound, even though it sits on a local topographic prominence, is obscured by the landscape alone in this analysis, even without the added obstruction of vegetation.
Adding nine meters of height to the viewer's elevation and repeating
the analysis results in the viewshed shown in Figure 20. This represents
the landscape viewable from the top of Cavanaugh Mound, and, possibly
more importantly, the areas on the landscape from which Cavanaugh Mound
itself would be visible. The viewshed off to the west overlooking the
Poteau/Arkansas River bottoms is now much wider, and the Spiro and Skidgel
sites are clearly within view. Even relatively tall vegetation in the
bottoms would not obstruct this intervisibility. The narrow views off
to the east and north in this viewshed are somewhat suspect. The ground
along these paths is not much lower than the mound itself, and any vegetation
of even moderate height would completely obstruct these lines of sight.
Unless these corridors were completely cleared of trees and tall bushes
(for about 10 km to the east and about 6 km to the north), these views
would not have been possible. The view off to the west may be more realistic.
Cavanaugh Mound is just over 200 m from the sharp drop of the bluff line,
and only tall vegetation within this area would block the view into the
bottoms.
 Figure 19. Viewshed from ground level at the location of Cavanaugh Mound. The Spiro and Skidgel sites are possibly within view, but would have easily been obscured by vegetation. |
 Figure 20. Viewshed from the top of Cavanaugh Mound. The Spiro and Skidgel sites are easily within this view. |
 Figure 21. Viewshed from the top of Brown Mound at Spiro. |
 Figure 22. Viewshed from the top of Skidgel Mound. Note that although Skidgel is on the western edge of a meander bend of the Arkansas River, the view from the mound is still primarily to the east, overlooking the Poteau/Arkansas River bottoms. |
Figure 21 shows the viewshed from the top of Brown Mound, the largest platform mound at the Spiro site. The viewshed from Copple Mound, the other platform mound at the site, is substantially similar. This viewshed is primarily directed to the east, overlooking the Poteau/Arkansas River bottoms. Nearly all of the bottomland area between Cavanaugh and Spiro, in fact, is within this viewshed. Some of the hills behind the Arkansas River bluff line both to the north and to the south are within this viewshed. Although there is some lower-lying land between Spiro and Skidgel, the prominence on which the Skidgel Mound is located is well within view. As with Cavanaugh Mound, Brown would also have been visible from much of the wide bottomlands between the sites.
Figure 22 shows the viewshed from the top of Skidgel Mound. Because it is already on a local topographic high, its overall viewshed is larger than that for both Brown and Cavanaugh Mounds. Again, the largest portion of the viewshed looks east over the Poteau/Arkansas River bottoms. Similar to the viewshed from Brown Mound, several of the hills back from the bluff line of the Arkansas are within the viewshed, both to the south and to the north. The viewshed also encompasses much of the bottomlands west of Skidgel (the Redland Bottom). The high, steep northern bluff of the Arkansas River southwest of Skidgel blocks the view into the bottoms farther up river.
To travelers coming up the Arkansas River either on foot or by boat, the first mounds to come into sight would have been those at Spiro and Skidgel. Barring tall vegetation blocking either, they would have come into view at about the same time, just as the travelers came into the broad stretch of bottomlands a little east of what is now the Oklahoma/Arkansas state line. Although Skidgel would have been a little farther away, it is situated on a local prominence and may have come into view first. About halfway through the bottomlands (again barring any tall vegetation in the way) Cavanaugh Mound would have come into view, off to the left.
To travelers coming down the Arkansas River, Skidgel Mound would have been clearly visible as a mound on an already existing peak, right at the top of the steeply rising bluff line. This would have come into sight as the travelers entered the Redland Bottom. Cavanaugh Mound and mounds at Spiro would not have been visible until the travelers rounded the final bend into the Poteau/Arkansas River bottoms. From most locations within the Poteau/Arkansas River bottoms, a person would likely have been able to look up and see a mound at one or more of the sites.