This Appendix includes the information, background, methods and results for the first four iterations of this thesis project. The project began as a modest idea. I wanted to try something I had learned in a GIS course given at Mesa College: a least-cost path analysis using one DEM block. I decided to use the DEM that included the area of Lost Valley, where I had worked through two summers of field school, and two semesters of laboratory work on the artifacts from that excavation. This became the Pilot Project and connected my graduate studies at SDSU with my course work in GIS at Mesa College. From this project I had the thrill of seeing the powerful GIS program in action and wanted to continue using it in research on my thesis. My thesis would most certainly be focused on the Lost Valley area. But how to incorporate GIS with the findings from an archaeological field school became the challenge.  I begin with the story of the pilot project.


A small-scale test to model a least-cost path was completed, the results of which gave me a fairly good idea of what could be expected in a larger project. One DEM for Hot Springs Mountain was used for the model. The GIS showed the path of least resistance between two points chosen non-randomly on the landscape: these two points were the locations of Cupa, a proto-historic village of the Cupeño, near Warner Springs for the beginning of the path, and CA-SDI-2506, our excavation site in Lost Valley, for the end of the path (Fig. 26, please note that I have offset the least-cost path to show the stream network beneath). The initial test of least resistance was for slope and distance only. The results of other studies that have been done, such as Pletka’s (2006), mirrored my expectations where the aim of the project was to show contact nodes using the spatial capabilities of ArcGIS software between several known site locations on the Santa Barbara coast. I wanted to make sure that least-cost path would give a relevant result and thus conducted this pilot study.

The pilot was limited to the construction of a least cost path or most effective travel route between two occupation site locations dating from the Late Period.  The end point for the pilot study was Lost Valley site CA-SDI-2506, in the north-central portion of San Diego County. Ethnographic (Piniolo et al. 1998:21; Strong 1929), and archaeological (Fleming 1999; Kline 2008) evidence indicates that Lost Valley was occupied during the Late Period. The start point was near the modern town of Warner Springs, which was the site of a Late Period village of the Cupeño Indian tribe, Kupa (or Cupa) until about 1902 (Bean and Smith 1978:589; Hill and Nolasquez 1973:vii; Strong 1929:183).

Figure 26. Pilot screenshot of ArcGIS 9.2 and least-cost path analysis (dark red line)
from Cupa (green circle) to Lost Valley and site CA-SDI-2506 (red circle).


A least-cost path was created in ArcView 9.2 (Fig. 26), showing a probable travel path from Cupa, in the lower San Jose Valley, to the higher elevation site at Lost Valley (at approximately 5,000 feet). Water runoff follows the path of least resistance as in the location of the waterways on this map. However, steeper slopes that would hinder the movement of water the least, would not be chosen by a foot traveling human as the effort required to surmount or descend the steeper terrain tends to cause more difficulty than taking an alternate, longer route. The algorithm engaged in ArcView GIS least-cost path analysis has a maximum slope that it recognizes as passable. If the slope is greater than this angle, the least-cost path cannot proceed through the area of too steep of a slope. Even with this exception, the least-cost path does seem to primarily follow the drainages.


With further readings and study, I decided to use Lost Valley as a place on the landscape and attempt to learn how Lost Valley might fit into a network of least-cost paths. Questions from previous study and research (Kidder and Leach 1981) asked 1) if Lost Valley was occupied year-round, and 2) what influence topography had on travel. I thought that these two questions could be addressed through least-cost path analysis. If Lost Valley were located where multiple least-cost paths came together, I could infer that it was possible that the area was used on a year-round basis. Lost Valley is at approximately 5,000 feet, but is sheltered by surrounding mountains and has a good water supply (see Chapter 2).

The second project involved the mosaicking of 27 DEMs, all of which were necessary to cover the San Luis Rey watershed. The San Luis Rey watershed was chosen because of its proximity to Cupeño territory, which is at the head of the San Luis Rey watershed (the bulbous portion on the east side of the watershed) (Fig. 27).

The twenty-seven DEMs for this portion of the project were downloaded individually, extracted with the free zip tool, ZipGenius 6, and then individually converted with the Conversion Tool: SDTS Raster to Grid. After conversion, the DEMs were mosaicked into one large rectangular block. The same process was applied to the block in preparation for creating the least-cost paths as had been applied to the single DEM for the pilot project. Start and end point shape files were created with the Editor Tool, then the distance and direction surfaces were derived from the derived slope layer from the mosaicked DEMs. This method does not always work, and some adjustments in the map mosaic process were necessary until the block of DEMs lined up properly and did not cause the path to halt at a gap. The gap meant no data to the computer program, and was therefore not passable. Finally, it was necessary to break the 27 DEMs into blocks of twelve, and create the start and end points and then the least cost path analysis on each of the blocks. The blocks of 12 shared the same DEMs at adjacent edges. Because of this overlap, I was able to assemble the paths together to cover the whole area.

Figure 27. The location of the San Luis Rey watershed within San Diego County.

One example of a least-cost path from a possible habitation location along the San Luis Rey River to Lost Valley is shown in Figure 28. When choosing the start and end points for least-cost path analysis within the San Luis Rey watershed, I used Lost Valley as an end point, and then other locations along the San Luis Rey River and its tributaries that had the environmental conditions for good locations of possible village sites in Luiseño territory. The San Luis Rey watershed, with the exception of the head of the watershed (which was Cupeño territory), was completely within Luiseño territory.

Figure 28. ArcView 9.2 map of San Luis Rey Watershed with elevation shown in relief,  waterways (blue lines) within
the watershed shown as well as roads (red lines). The least-cost path analysis (neon green line) was made from a
point near the coast along the San Luis Rey River to CA-SDI-2506 in Lost Valley.


A problem with the results of the least-cost path analysis using the San Luis Rey watershed as the boundary of the study area is that the least-cost paths produced tended to follow the San Luis Rey River if start and end points were any where near the river, and that in order for any of the least-cost paths to cross into Lost Valley, an end point of the path had to be located at Lost Valley. There was no way to know if the path would make its way through Lost Valley if Lost Valley were not the destination. How was I to determine how Lost Valley was positioned in relation to the surrounding settlement system if I had to use Lost Valley as the end of the path (or the start)? I needed some way to model least-cost paths where the paths would cross through Lost Valley between points that were outside of the area of Lost Valley.

With the San Luis Rey watershed least-cost paths as inspiration, I decided that I needed to use a different set of criteria for the start and end of paths used for the least-cost path analyses. I determined to use a block of DEMs in the same way that was used by Whitley and Hicks (2003). The block of DEMs that was chosen were those that included and surrounded the head of the San Luis Rey watershed, that of the territory of the protohistoric Cupeño, and adjacent territory.

Large Block Around Cupeño Territory

The head of the San Luis Rey watershed became the next area of focus in an attempt to narrow down the area where I could concentrate my efforts. A block of 16 DEMs that created a buffer around the head of the San Luis Rey watershed was mosaicked together (Fig. 29).


For this block, the method of marking the start and end paths changed as I incorporated the methods of Whitley and Hicks (2003) into the least-cost path analysis. Whitley and Hicks’ work in the eastern United States (see Chapter 3) utilized a block of DEMs for least-cost path analysis, creating the start and end points for multiple least-cost paths at the boundary of the mosaicked DEMs and running the analysis from north to south and from east to west across the mosaic.


The results of the new technique with start and end points around the edge of the mosaicked DEM block was more promising than the previously used placement of the points at Lost Valley, then at other possible locations of sites. By changing the start and end points to random (but spaced) locations along the edge of the mosaicked DEM, I was able to see a reasonable model of where paths would possibly be located if sojourners traveled across an expanse of ground from point A to point B. The only problem with this approach is that it did not lead to any paths making their way through or even near Lost Valley, the focus of the project. This was a dilemma that was remedied by bringing in the west and south boundaries of the study area so that it would be probable that a least-cost path would traverse Lost Valley. There was no point in going any further with the project if none of the least-cost paths entered and exited or came close to Lost Valley. The results of this project were the creation of the smaller bounded portion of the San Luis Rey watershed, and subsequent least-cost path analyses using the same technique of equally spacing start and end points around the perimeter of the new boundary and running the least-cost path analysis from north to south, and from west to east.

Figure 29. The head of the San Luis Rey watershed and adjacent area.

Bringing in the Boundary for Cupeño and Adjacent Territory

A new smaller block for the boundary of the study area was created using a new polygon shapefile, and a new set of start and end points spaced evenly on the sides of the rectangular area. A new set of least-cost paths were run from east to west, then from north to south as in the previous larger block (Figs. 30 and 31).


This smaller block and the least-cost paths that were created across it were the most promising in the way of answers that I had been able to produce. At least three of the paths that were produced crossed through Lost Valley. Additionally, the paths crossed the watershed boundary at two locations upon leaving Lost Valley. These were two locations that could be located on aerial maps and inspected for possible trails. Also of interest, other watershed boundary crossings at the head of the San Luis Rey watershed were marked by least-cost paths that did not traverse Lost Valley, but are nevertheless in the vicinity of Lost Valley and therefore of interest to this study.

It is the previous four projects that formed the foundation of my thesis. I provide this appendix to show the evolution of my thinking for the choices that were made at each interval. Least-cost paths were collected from these initial projects and merged together to form the network of paths used in the results of this thesis.

Figure 30. Smaller block around the head of the San Luis Rey watershed.

Figure 31. Close-up of smaller block in Figure 30.

Appendix References

Bean, Lowell John, and Charles R. Smith
1978   Cupeño. In Handbook of North American Indians, Vol. 8, California, Robert F. Heizer, ed., pp. 588-591. Washington, D.C.: Smithsonian Institution.

Kidder, Fred, and Larry L. Leach
1981   Archaeological Reconnaissance of Sky Oaks, Chihuahua Valley, San Diego County, California. Report on file at SCIC, San Diego State University.

Fleming, Kaylene Patricia
1999   Life at 5000 Feet: An Archaeological Investigation of CA-SDI-2508 (Leaning Pines), Lost Valley, San Diego County, California. Master’s thesis, San Diego State University.

Kline, George E.
2008   Metates to Merit Badges: The Contrasting Occupational Sequences of Lost Valley. Master’s thesis, San Diego State University.

Hill, Jane H, and Rosalinda Nolasquez
1974   Mulu’wetam: The First People. Banning, California: Malki Museum Press.

Pigniolo, Andrew R., Michael Baksh, and Stephen R. Van Wormer
1999   Wiatava: An Evaluation of Cultural Resources within Lost Valley, County of San Diego, California. A Cultural Resource Evaluation, prepared by Tierra Environmental Services, San Diego, CA. (Prepared for the County of San Diego Planning and Land Use). On file at the South Coastal Information Center, San Diego, California.

Pletka, Nicole
2006   Late Holocene Gabrielino Settlement Patterns in the Newport Coast Area of California: A Geographic Systems Based Approach. Master’s thesis, California State University, Long Beach.

Strong, William Duncan
1930   Aboriginal Society in Southern California. University of California Publications in American Archaeology and Ethnology 26. Berkeley: University of California Press.

Whitley, Thomas G., and Lacey M. Hicks
2002   Using a Geographic Information Systems (GIS) Approach to Extract Potential Prehistoric and Historic Travel Period Corridors Across a Portion of North Georgia. Electronic document,, accessed November 5, 2006.



Regional Network Analysis Situating Lost Valley in the Inter-Site Landscape
Victoria Lynn Kline
Master of Arts in Anthropology
San Diego State University, 2009

The purpose of this study is to analyze possible contact pathways through Lost Valley, San Diego County, California, using the methods of least-cost path analysis for both real and modeled travel corridors through the San Luis Rey Watershed. I conducted this study using ArcView 9.2 GIS (geographic information systems), digital elevation models, and aerial photography, in order to show the most likely corridors of travel and trade using least-cost path modeling, and to compare the modeled paths to real trail networks, i.e., those that show up in the aerial photographs and/or those written about in the literature. This study shows how prehistoric and protohistoric peoples traveled through the Lost Valley area and how contact networks were likely established and maintained. The modeled travel corridors are compared to the ethnographic and historic knowledge of the Cupeño who seasonally occupied Lost Valley.

Back to Opening Page -----------------------------------------------------

Go to Chapter 1 Introduction