QUANTIFYING ARCHAEOLOGICAL PRESERVATION BIAS USING A UNIVERSAL MODEL OF SOIL EROSION

Erosion and weathering are destructive processes that shape the archaeological record, leaving relatively young cultural materials over-represented at the land surface. Archaeologists often accept this taphonomic or preservation bias as a limitation to archaeological knowledge. Yet, archaeologists base human population estimates on radiocarbon frequency distributions, which depend on site discovery. With population estimates tied to site discovery and numerical dating, it is critical to measure erosion bias and correct human population estimates based on potential sites lost. Unlike the global volcanic model of Surovell et al. (JAS, 36, 1715-1724), or the stratigraphy-based model of Ballenger and Mabry (JAS, 38, 1314-1325), we approach the preservation bias problem at the exposed surface, or the archaeological landscape encountered during systematic surface surveys. In our model, we use the Revised Universal Soil Loss Equation (RUSLE) to estimate average annual soil loss, and we use prehistoric population estimates in North America based on Peros et al. (JAS, 37, 656-664). We present a method for calculating prehistoric human demographic changes where we correct for preservation bias after determining the density of hearth features from landform surfaces of known ages. We tested our model in a 33,000-acre portion of the Caprock Canyonlands in northwest Texas where surface survey spanning five landform surfaces yielded 385 hearth features. Model results show population rapidly expanding during the Middle Archaic and then steadily increasing through the pre-contact period. This result does not compare to the current radiocarbon frequency distribution of hearths from the study area. Thus, we believe that erosion is removing Archaic-aged hearths, but in this first approximation, we cannot rule out the possibility that population density is changing differently than the model expects, or that there is a change in human adaptation or use of the landscape. Nevertheless, our model provides a methodology for quantifying archaeological preservation bias. When we understand the extent to which the archaeological record has been affected by erosion, we can make more substantiated conclusions about the archaeological patterns on the surface that inform us about human behavior.