Publications and Papers

Pierce, C
1982 Stratigraphy, Features, and Site Formation. In Archaeological Investigations at Talepop (LAN-229), compiled by C. D. King, pp. 5.1-5.87. Office of Public Archaeology, University of California, Santa Barbara.

Pierce, C
1988 California's Milling Stone Horizon: Of Mice or Men? 53rd Annual Meeting of the Society for American Archaeology, Phoenix, Az.

Pierce, C
1992 Effects of Pocket Gopher Burrowing on Archaeological Deposits: A Simulation Approach. Geoarchaeology 7(3): 185-208.

View HTML - Download PDF

Ramenofsky, A, F Neiman and C Pierce
2008 Measuring Time, Population, and Residential Mobility from the Surface at San Marcos Pueblo, North Central New Mexico. American Antiquity Volume 74, Number 3, pp. 505-530

Download PDF preprint

Running the Simulation

Rodent Burrowing Disturbance Research

Any archaeologist who has excavated has encountered traces of burrowing animals, usually in the form of filled burrows visible as differences in the color or consistency of the sediment. While working in California during the latter half of the 1970s, I became curious about how all that burrowing may have affected the archaeological record. No one I asked seemed to know much about the critters or the nature of their impacts on the sites we were investigating. Burrowing rodents are quite common in California, and pocket gophers or ground squirrels were actively burrowing in virtually every site at which I worked. At the same time, I was taking classes in geology at San Jose State University and beginning to explore the interfaces between geology and archaeology, an area that would shortly come to be called geoarchaeology. I was mostly working on using soil chemistry to identify the uses or functions of different areas in archaeological sites and deposits. However, I took a course on soil ecology and learned that there was quite a bit known about the habits of animals that live in the soil including burrowing mammals.

This knowledge lay dormant until I was hired by UC Santa Barbara in December of 1980 to supervise surface collections and test excavations at a Late Period village site in Malibu Creek State Park located in the Santa Monica Mountains of southern California. This site turned out to be the historic Chumash village of Talepop, but that is another story. While completely removing the plants (mostly grasses) from the site to conduct the surface collection, I noticed that almost every inch of the exposed ground was covered with the small hills made by rodents when they expel dirt losened during burrowing out onto the surface. This was the first time I had noticed evidence of such prolific burrowing, but it was also the first time I had stripped a site in California of its vegetative cover in this way. When we dug into the site, active and filled burrows were visible, but they were not abnormally abundant. However, the soil had a crumb structure, which, as I had learned in my soil ecology course, can be indicative of substantial bioturbation.

At the end of the fieldwork in the Spring of 1981, I took on the responsibility of documenting and writing up the stratigraphic information from the site. During work on the the writing, I did a little research at the UCLA library on burrowing rodents, particularly pocket gophers and ground squirrels because they were both present at the site. I mostly examined general books on mammals from California, which gave me some general information on burrowing and other aspects of the behavior and of these animals. I also turned up an article published by Joseph Grinnell in the Journal of Mammalogy in 1923 titled "The Burrowing Rodents of California as Agents of Soil Formation" (vol. 4, no. 3, pp. 137-149). Not only did this article provide some wonderfully descriptive information on the burrowing habits of pocket gophers, it also had a significant impact on the future direction of my interest in this problem. The following paragraph, located on page 145 of Grinnell's article, caught my attention:

As in the case of Darwin's earthworms, there is plentiful evidence in California as to the function of burrowing rodents in burying large objects, such as rocks and logs. Ground squirrels and pocket gophers both show an inclination toward placing their nesting chambers beneath objects that will protect them from being dug out by burrowing enemies, such as badgers and coyotes. The earth is taken out from beneath a rock or a boulder by the rodents and deposited around the margin of the object, which thus, as the years go by, gradually disappears, as a result of the process of undermining and settling plus that of the building up of the ground round-about. The seeker need not go far to find good cases of this kind. Mr. Joseph Dixon cites the case of a rock pile on one corner of his ranch which was half buried in ten years. One certain rock settled six inches in comparison with the general land level, during a period of ten years.

At the time I read this, I was working with Chester King who was the project director for the work at Talepop. Chester had excavated many sites in California including sites assigned to a group known as the "Milling Stone Horizon." Milling Stone Horizon sites commonly occur on the tops of hills, terraces, or other prominent and geologically stable landforms. These sites often display a characteristic vertical distribution of artifacts with the upper levels dominated by flakes, projectile points and other small artifacts, which are underlain by a more-or-less dense zone or floor of large artifacts, commonly manos, metates, and other tools used to mill seeds into flour. The common interpretation of this vertical distribution was, and to some extent still is, that there was an early "Milling Stone Horizon" during which people subsisted mainly on plant products, and this period was followed by a "Hunting Horizon" when people subsisted mainly by hunting game. Upon reading Grinnell's description of the burying of rocks cited above, I was immediately struck by the thought that perhaps this same process was responsible for the peculiar vertical distribution of artifacts found at Milling Stone Horizon sites. I discussed this possibility with Chester King, and he concurred that this was certainly possible as he had often encountered evidence of burrowing by rodents in Milling Stone Horizon sites.

The site of Talepop where I was working did not display the features of a Milling Stone Horizon site, but did show clear evidence of vertical movement or mixing of small artifacts and sediment as well as deterioration of fragile artifacts such as bone, shell and charred plant remains. I noted these aspects of burrowing-related disturbance along with a brief description of pocket gopher burrowing behavior in the report I wrote on the stratigraphy, features and site formation at Talepop, which was completed in 1981 but never published.

In 1982, I took a job at the Desert Research Institute (DRI) in Reno, Nevada as a supervisor on the Archaic Oshara Project run by Cynthia Irwin-Williams. Between working on that project full-time and returning to school at the University of Nevada, Reno to complete my BA, I had little time initially to pursue my interests in the burrowing-Milling Stone Horizon connection. However, in the spring of 1983, I was able to complete a first draft of a paper on the topic titled The Effects of Rodent Burrowing on Archaeological Deposits and their Implications for California Prehistory. I passed this draft paper around to some of my colleagues at DRI, including Steve Durand and Jonathan Davis, for their comments, but was too busy to take it any farther.

Shortly after I had completed that first draft paper, I learned from Jon Erlandson (then a graduate student at U.C. Santa Barbara), that he was about to publish an article in American Antiquity on the effects of burrowing by rodents on archaeological sites. He told me that the original version of his article had included references to my earlier work at Talepop, but the editors of American Antiquity had made him cut the length of his paper so much that he had to drop the reference to my work. I remember thinking "Oh no! Jon has beaten me to it", but it turned out that his article was much narrower in scope than the one I was working on. This happened again in 1985 when I heard that Barbara Bocek at Stanford University had submitted an article to American Antiquity on burrowing rodents and site formation. But again, her article was focused on a single site on the Stanford campus and lacked a broader perspective. Clearly, this was a topic who's time had come, and I had better get to work if I was going to publish my ideas.

Consequently, in the Spring of 1986 as I was completing my BA in Anthropology at the University of Nevada in Reno, I took an independent study course with Professor Don Hardesty to force me to write a new, more complete version of my rodent disturbance paper. I did a tremendous amount of background research for this version, delving deeply into the literature on burrowing rodents. Not only was there a great deal known about the behavior of various burrowing rodent species, but I discovered that some investigators had actually excavated burrow systems completely and documented them quantitatively. These published observations allowed be to begin specifying quantitatively some aspects of my disturbance model. With this version of the paper, I also began to focus more on pocket gophers as it became clear that they were a more significant force than other burrowing rodents in the disturbance of most sites in California.

Although I completed the paper for the independent study that Spring, I was not satisfied with it. I was able to quantify the rate at which small materials (artifacts and sediment) would be moved up to the surface using data primarily from an article by Milton Miller published in Hilgardia in 1957 (vol. 26, no. 8, pp. 431-452) titled Burrows of the Sacramento Valley Pocket Gopher in Flood-Irrigated Alfalfa Fields. However, figuring out the rates of downward movement and disturbance of stratification still eluded me.

During the Winter of 1986-87, I decided to take a different approach to the problem. I had been learning to program computers, and thought I would try to write a program that would simulate the effects of gopher burrowing. With the data on gopher burrowing I had compiled while writing the last version of the paper, I had enough measures of burrowing rates and volumes of sediment moved to accomplish a simulation. I first had to work out the logic for moving material up and down within a deposit. Starting with the concept that burrowing creates voids that are filled primarily by collapse, I used Miller's data on burrow volume by depth and rates of sediment deposition on the surface to determine the percentage of empty space (open burrow voids) created by active burrowing in a given year by 10 cm depth intervals.

The first step in running the simulation was to remove the appropriate amount of material (sediment and artifacts) smaller than the average burrow diameter from each 10 cm level, and add this material to the top most 10 cm level. This creates voids (sediment volumes less than 100%) in the lower levels and an excess of material in the top most level. In the next simulation step, the voids in each level are filled with small and large sediment and artifacts from the level directly above. By iterating these two steps over a specified number of years, I could track the vertical movement of small and large artifacts, and record the proportion of each level likely to have been disturbed by burrowing. I wrote this simulation program using Borland's turbo Pascal language on a Compaq "lunchbox" style portable computer with a massive 10 mb hard drive.

Once the program was done, I ran the simulation starting with all large and small artifacts deposited on the surface, and thus contained within the 0-10 cm level. By running the simulation over longer time intervals, I was able to graph changes in the vertical distribution of large and small artifacts, and disruption of stratification by mixing through time. Although the large artifacts behaved as expect by accumulating in a narrow zone near the base of the burrow system, I was surprised by the vertical distribution of small artifacts. For some reason, I had assumed that the small artifacts would remain concentrated near the surface. The simulation produced a more even vertical distribution for obvious reasons that had not been clear to me until I ran the simulations.

The real excitement came when I was able to compare the simulated vertical distributions of large and small artifacts to actual distributions in archaeological sites of known age. This took a while because I had to locate site reports that provided three essential pieces of data: 1) absolute dating of the occupation, 2) counts of artifacts by excavated level, and 3) data on artifact size so that I could distinguish between large and small artifacts. I found two reports, one by Chester King on the Sweetwater Mesa Site near Malibu and the other by Makoto Kowta on the Sayles Site in Crowder Canyon between the San Gabriel and San Bernardino Mountains, California. Both site reports presented size ranges for different artifact types, which allowed me to group artifacts types into small and large categories, and then add up the number of small and large artifacts in each level. I did this first for the Sweetwater Mesa Site. My excitement grew as I compiled the data on the living room floor at my house in Reno, and the vertical distributions of Sweetwater Mesa artifacts matched almost exactly the simulated distributions produced by burrowing for the same time period as the age of the site. I knew I was on to something when Sayles Site data produced an even better fit with the simulated data.

All that remained was to turn this new approach into a publishable paper. In the Fall of 1987, I submitted an abstract to present a paper on this topic to the Annual Meeting of the Society for American Archaeology. This paper was accepted and I presented it in the Spring of 1988 to very positive reviews. The next Fall I left Reno for Seattle to attend graduate school at the University of Washington. After completing my two years of core courses, I produced a manuscript to submit to the relatively new journal called Geoarchaeology. The paper was accepted without revision, and published in the third issue of 1992 (Volume 7) under the title Effects of Pocket Gopher Burrowing on Archaeological Deposits: A Simulation Approach.

Several years later, another surface collection project rekindled my interest in the problem of burrowing rodent’s impact on the archaeological record. I was working with Ann Ramenofsky and the University of New Mexico on trying to document the occupational history of San Marcos Pueblo, a large site located south of Santa Fe with evidence of a lengthy occupation. Our work focused on mapping the surface architectural expressions and analyses of artifacts visible on the surface of numerous adjacent trash or midden deposits. Previous excavations at the site had documented that the middens could be as much as a meter deep. Consequently, we needed to understand how the surface record in these trash areas was formed to determine how representative the artifacts collected from the surface are of the range of materials deposited during the entire period of use of a trash dump.

While mapping the pueblo and conducting surface artifact collections, we observed abundant evidence of burrowing animals in the form of backdirt piles and open burrow mouths. This led me to consider the effects of burrowing on the composition of artifacts present of the surface. My earlier work on had focused on the how burrowing can cause artifacts to move down from the surface through burial by burrow backdirt and collapse of tunnels. I needed to turn this on its head, and examine how burrowing could bring artifacts to the surface that had been buried by normal trash accumulation.

I used my simulation program, but rather than start with all of the artifacts on the surface, I spread them evenly across the ten 10cm depth intervals. I also modified the program to identify what depth interval materials came from when the simulated burrowing moved them to the surface. This allowed me to track the composition of the surface layer with regard to the proportion of materials from different depths

I ran the simulation several times, and increased the length of time the deposit was exposure to burrowing with each run up to the 700 years since San Marcos Pueblo was abandoned. Results showed that, although artifacts originally deposited on or near the surface remain the most abundant throughout the period of simulated burrowing, contributions from deeper levels rise rapidly making up slightly less than 70% of the surface assemblage after only 300 years. This gave us confidence that surface artifacts at San Marcos would provide a reliable indictor of the duration of use of most if not all of the trash dumps at the site. We used these data to develop an intriguing picture of the Pueblo’s occupational history, which was published in American Antiquity in 2009 under the title “Measuring Time, Population, and Residential Mobility from the Surface at San Marcos Pueblo, North Central New Mexico."

Last Updated: 09 DEC 2011