Remote Sensing in Archaeology: An Explicitly North American Perspective

11. Multiple Methods Surveys: Case Studies

11 Multiple Methods Surveys: Case Studies KennethL.Kvamme,JayK.Johnson,andBryanS.Haley The foregoing chapters have demonstrated that a geophysical survey with a single instrument can provide much insight. Yet, most instruments respond primarily to a single physical property of the earth: magnetometry to soil magnetism, resistivity and electromagnetic induction to soil conductivity, and ground-penetrating radar (GPR) primarily to soil dielectric properties (Weymouth 1986:371). It is easy to conclude that surveys with multiple methods must offer greater insights because different dimensions of the subsurface are being quantified. Buried cultural features not revealed by one method might be made visible by another (Clay 2001; Piro et al. 2000). Additionally , subsurface features indicated by one method may be very different from and add complementary information to patterns revealed by another, as when a hearth and burned elements of a house are shown by magnetometry while soil changes that make up the house floor are portrayed by resistivity, conductivity, or radar methods. Recent advances in computer graphics and geographic information system (GIS) technology allow various data sets to be overlain or combined, making possible simultaneous visualization of features in each. Moreover, new GIS-based “data fusion” methods seek ways to combine the multidimensional imagery into composite data sets that enhance interpretability through the use of principal components analysis, statistical, or context -based methods (Johnson and Haley 2004; Kvamme 2001; Piro et al. 2000). 252 ~ Kenneth L. Kvamme, Jay K. Johnson, and Bryan S. Haley Case Study 1 Prehistoric Earthlodges at Whistling Elk Village, South Dakota Kenneth L. Kvamme Whistling Elk (39HU242), located on the Missouri River in South Dakota, is a fortified earthlodge village assigned to the Initial Coalescent variant of the Plains Village tradition (ca. a.d. 1300). This settlement lies beneath nearly a meter of Missouri River sediments and lacks any surface expression aside from artifacts that occasionally spill out of the embankment and vegetation markings indicating the locus of a fortification ditch in a 1968 aerial photograph. This site was completely surveyed by electrical resistivity using a Geoscan Research RM15 twin-probe array with 1-m probe separation (Figure 11.1a), electromagnetic conductivity using a Geonics Limited EM38 conductivity meter (Figure 11.1b), and magnetic gradiometry using a Geoscan Research FM36 fluxgate gradiometer (Figure 11.1c). Sampling density was 1 × 0.5 m for the first two methods and 1 × 0.25 m for the last. Resistivity probe separation was 1 m (see Kvamme 2003; Toom and Kvamme 2002). Electrical contrasts between the occupation surface and more resistant sediments filling the abandoned ditches and houses produced the best definition of the site’s structure in the resistivity data (Figure 11.1a). An outer fortification ditch containing five uniformly spaced bastions encompassing approximately 1.45 ha is readily seen, and approximately 64 anomalies within the area enclosed by this ditch are interpreted as likely houses, based on the co-occurrence of the various forms of geophysical evidence. Although similar results were obtained by the conductivity survey (Figure 11.1b), cultural features were less well defined owing to peak instrument sensitivity at 0.4 m depth, somewhat above the cultural features (the conductivity data also contained a large plow mark response owing to its shallowdepth focus, signs of which were removed by Fourier methods as described in Kvamme, this volume, Chapter 10). The magnetic gradiometry data appear very noisy at the global scale of Figure 11.1c but nevertheless at a large map scale contain much detail about internal components of the village and individual houses. For example, the presence of large-magnitude positive anomalies suggests that as many as 34 (53 percent) of the houses may have been burned, quite possibly by a prehistoric sack of the village (see Kvamme 2003). The foregoing is best illustrated by focusing on one house, referred to as the “Big House” because of its unusual size (measuring 100 m2 , three to four times the area of other houses). The resistivity data unambiguously define its square form and southeast-facing entryway, characteristic of houses of the Initial Coalescent variant (Figure 11.1d). The magnetometry data, on the other hand, provide information about such internal features as the locus of the centrally placed Multiple Methods Surveys: Case Studies ~ 253 hearth and four roof support posts characteristic of this variant (Figure 11.1e; these magnetic data are from a second high-density survey placed over the house...