Navigation – Plan du site

AccueilNuméros45-11. Case studies and archaeologica...An Experimental Use of Ground-Pen...

1. Case studies and archaeological feedback

An Experimental Use of Ground-Penetrating Radar to Identify Human Footprints

Adam Wiewel, Lawrence B. Conyers, Luca Piroddi et Nikos Papadopoulos
p. 143-146

Résumé

– Features like human footprints can be identified in two-dimensional radar profiles.

– Amplitude variations associated with footprints have been demonstrated with slices.

– Horizon detection methods are another avenue for detecting subsurface footprints.

Haut de page

Texte intégral

Introduction

1Reports of trace fossil footprints of modern humans and our extinct hominin ancestors are increasingly common. The best-known example occurs at Laetoli, where the footprints of several Australopithecus afarensis individuals are telling of foot morphology, bipedal biomechanics, and body size variation within the species (Leakey & Hay, 1979; Masao et al., 2016). Subsequent studies have yielded similar evidence for other hominin species as well as information about group composition, habitat use, and interaction with nonhuman mammalian species (see numerous references in Bennett and Reynolds, 2021).

2Due to the environments in which footprints are commonly found and their susceptibility to erosion following exposure, rapid and accurate documentation is crucial. Laser scanning and digital photogrammetry are routinely used for this purpose (Bennett and Reynolds, 2021: 16-20). Although these tools would not be replaced by ground-based geophysics, methods like ground- penetrating radar (GPR) may identify trackways prior to their excavation, yet such applications are rare. For instance, Urban and colleagues (2019) recently identified human, mammoth, and sloth footprints as part of a survey designed to evaluate GPR’s effectiveness for mapping trackways in New Mexico, USA.

Footprint Model and GPR Survey

3To better understand the potential of GPR for mapping buried human footprints, a 1.8 m × 1 m model meant to mimic the Laetoli footprints was created on a level, vegetation-free, wet basaltic clay soil (Fig. 1). Parallel trackways were made by two adults, with footprint depth varying between 1.5 and 2.8 cm. Some footprints exhibited distinct toe and heel imprints while others were obscured by the lifting and subsequent settling of the undertrack sediment. The track surface was allowed to dry for nearly 24 hours before it was covered with 5.5-6.5 cm of basaltic sand. The sand was then covered by a ¼ inch (8 mm) thick poly-vinyl-chloride plastic mat, which was gridded with transect lines to facilitate the survey.

Figure 1. The footprint model. Transects were surveyed every 10 cm.

Figure 1. The footprint model. Transects were surveyed every 10 cm.

4The model was subsequently surveyed in both the x- and y-directions at 10 cm intervals, a total of 30 transects, with a GSSI SIR-3000 controller with 900 MHz, 2 GHz, and 2.6 GHz antennas. Additional data collection settings for the latter antenna, which resolved the track surface and footprints best, included a 5 ns time window with 512 samples/trace and 300 traces/meter. The following results are limited to this antenna.

Results

5With standard processing steps including time zero correction, background removal, and gaining, the clay track surface is quite apparent in each two-dimensional reflection profile (Fig. 2). The surface is characterized by a high-amplitude planar reflection that slopes downward from the perimeter of the model but otherwise parallels the sand layer that covers the footprints.

Figure 2. Reflection profile showing the clay surface on which the trackways were created and two subtle depressions caused by footprints. Graphs illustrate the depths and reflection amplitudes of hundreds of samples along the horizon.

Figure 2. Reflection profile showing the clay surface on which the trackways were created and two subtle depressions caused by footprints. Graphs illustrate the depths and reflection amplitudes of hundreds of samples along the horizon.

6Footprints are also visible in many reflection profiles as subtle depressions in the clay soil. Some exhibit weak- or moderate-amplitude reflection hyperbolas below the depression, as is typical for “point sources,” while others create higher amplitude reflections that lack hyperbolas. Both reflections relate to the radar energy’s propagation, but the difference is likely due to footprint morphology. Certain features like the heel depression cause the dispersing energy to focus rather than spread.

7Analysis of each profile shows correspondence between them and confirms that the footprints consistently create reflections. Urban and colleagues’ (2019) study detected footprints, which produced higher amplitudes, in three-dimensional slice maps. In our case, footprints are also apparent in slice maps, although they generate reflections that are both higher and lower in amplitude than the surrounding clay soil. This complex patterning may in part result from the way each horizontal amplitude slice truncates the bowl-shaped clay surface, which itself generates high-amplitude reflections, and prints (Fig. 2). As this endeavor was experimental, we created depth slices with varying thicknesses but found that thicknesses in the range of the footprints’ depths (about 1.5-2.8 cm) produced the clearest results. To account for the bowl-shaped clay surface, each reflection profile could be “corrected” so that slices follow that soil layer (rather than horizontal to the ground surface).

8However, we opted for a different (but related) approach in identifying the clay soil horizon on which the tracks were imprinted given their visibility in each reflection profile (Fig. 2).

9Several processing software packages have such capabilities, which allow a user to manually “pick” the horizon or automatically detect a wave’s peak response across each profile. In this way, the undulating track surface can be mapped across the survey area. The results are subsequently gridded to generate a three-dimensional model of the horizon, with most footprints noticeable as subtle depressions (Fig. 3).

Figure 3. Three-dimensional trackway surface generated with horizon detection methods and the same result following subtraction of the trend surface.

Figure 3. Three-dimensional trackway surface generated with horizon detection methods and the same result following subtraction of the trend surface.

Discussion

10An important consideration in all GPR surveys is the penetration depth and transmitted wavelength of radar energy produced by different antennas. The 2.6 GHz antenna used here was limited to a depth of about 16.5 cm (with an estimated relative dielectric permittivity [RDP] of 8.5). The antenna is suitable because the footprints were only 5.5-6.5 cm below the “ground surface,” but real-world scenarios are likely more variable. The lower frequency antennas used in this experiment – 2 GHz and 900 MHz – penetrate deeper below the surface, although the resolution of features is diminished because of their longer wavelengths (Conyers, 2013: 62-73). The 2.6 GHz antenna has a wavelength about 4 cm in these soils. To be resolved by an antenna, objects must have a size of at least 40 percent of its wavelength (Conyers, 2013: 62), or 1.6 cm.

11The modeled footprints’ depths range from about 1.5 to 2.8 cm, so they are generally resolvable with the 2.6 GHz antenna but not the others. Urban and colleagues’ (2019) results with a lower frequency antenna (250 MHz), however, illustrate the complex way different factors like overburden thickness, soil RDP, and feature size affect survey outcomes. Transect spacing is also crucial to a survey’s outcome and is often determined in part on the type(s) of suspected features. In this survey, 10-cm transect spacing in both directions meant that most prints were traversed by the antenna at least twice. Presumably, a survey with increased transect density would better resolve the footprints but at the cost of additional survey time with our approach. This issue can potentially be avoided with a multi-channel system, perhaps one with step-frequency capabilities, but further experimentation is needed. Of course, certain environments would impede the use of such instruments, but human footprints have been identified in a wide range of environments (Bennett & Reynolds, 2021: 16-17).

12Archaeological features like human footprints are unquestionably significant, however, and are susceptible to degradation after exposure or excavation. Footprint excavation and other methods of documentation are equally, if not more, labor intensive (Urban et al., 2019). Thus, GPR is an ideal approach for identifying, visualizing, and preserving trackways in a non-destructive manner, but survey strategies and analytical approaches will vary. Horizon detection is a promising method, but standard two-dimensional profile analysis and three-dimensional amplitude slices may work equally well.

Haut de page

Bibliographie

Bennett, M.R., Reynolds, S.C., 2021. Inferences from footprints: Archaeological best practice. In A. Pastoors, T. Lenssen-Erz (ed.), Reading prehistoric human tracks, Springer, Cham, Switzerland, 15-39.

Conyers, L.B., 2013. Ground-penetrating radar for archaeology, 3rd ed., AltaMira Press, Lanham, Maryland.

Leakey, M.D., Hay, R.L., 1979. Pliocene footprints in the Laetolil Beds at Laetoli, northern Tanzania. Nature, 278: 317-323.

Masao, F.T., Ichumbaki, E.B., Cherin, M., Barili, A., Boschian, G., Iurino, D.A., Menconero, S., Moggi-Cecchi, J., Manzi, G., 2016. New footprints from Laetoli (Tanzania) provide evidence for marked body size variation in early hominins. eLife, 5: e19568.

Urban, T.M., Bennett, M.R., Bustos, D., Manning, S.W., Reynolds, S.C., Belvedere, M., Odess, D., Santucci, V.L., 2019. 3-D radar imaging unlocks the untapped behavioral and biomechanical archive of Pleistocene ghost tracks. Scientific Reports, 9: 16470.

Haut de page

Table des illustrations

Titre Figure 1. The footprint model. Transects were surveyed every 10 cm.
URL http://journals.openedition.org/archeosciences/docannexe/image/9144/img-1.jpg
Fichier image/jpeg, 2,6M
Titre Figure 2. Reflection profile showing the clay surface on which the trackways were created and two subtle depressions caused by footprints. Graphs illustrate the depths and reflection amplitudes of hundreds of samples along the horizon.
URL http://journals.openedition.org/archeosciences/docannexe/image/9144/img-2.jpg
Fichier image/jpeg, 2,0M
Titre Figure 3. Three-dimensional trackway surface generated with horizon detection methods and the same result following subtraction of the trend surface.
URL http://journals.openedition.org/archeosciences/docannexe/image/9144/img-3.jpg
Fichier image/jpeg, 1,5M
Haut de page

Pour citer cet article

Référence papier

Adam Wiewel, Lawrence B. Conyers, Luca Piroddi et Nikos Papadopoulos, « An Experimental Use of Ground-Penetrating Radar to Identify Human Footprints »ArcheoSciences, 45-1 | 2021, 143-146.

Référence électronique

Adam Wiewel, Lawrence B. Conyers, Luca Piroddi et Nikos Papadopoulos, « An Experimental Use of Ground-Penetrating Radar to Identify Human Footprints »ArcheoSciences [En ligne], 45-1 | 2021, mis en ligne le 16 août 2021, consulté le 30 janvier 2023. URL : http://journals.openedition.org/archeosciences/9144 ; DOI : https://doi.org/10.4000/archeosciences.9144

Haut de page

Auteurs

Adam Wiewel

National Park Service, Midwest Archeological Center, Lincoln, Nebraska, USA

Lawrence B. Conyers

Department of Anthropology, University of Denver, Denver, Colorado, USA

Articles du même auteur

Luca Piroddi

Department of Civil, Environmental and Architecture Engineering (DICAAR), University of Cagliari, Sardinia, Italy

Nikos Papadopoulos

Laboratory of Geophysical-Satellite Remote Sensing, IMS-FORTH, Rethymno, Greece

Articles du même auteur

Haut de page

Droits d’auteur

CC-BY-NC-ND-4.0

Creative Commons - Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International - CC BY-NC-ND 4.0

https://creativecommons.org/licenses/by-nc-nd/4.0/

Haut de page
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search