Geophysical investigations at an ancient Sabaean temple in the Republic of Yemen
Archaeo-Physics LLC was contracted by the American Foundation for the Study of Man to conduct a geophysical investigation at Mahram Bilqis (the Awam Temple), a Sabaean archaeological site near Marib, Yemen. Geophysical fieldwork was conducted from March 18 through March 29, 2006. Geophysical survey was performed by David Maki and Geoffrey Jones of Archaeo-Physics assisted by local Yemeni workers, and by General Organization for Antiquities, Museums and Manuscripts (GOAMM) and AFSM personnel. The AFSM co-field directors were Drs. Juris Zarins and Abdu Ghaleb.The investigation consisted of high-resolution ground penetrating radar (GPR) and magnetic field gradient surveys. The most interesting results were obtained from survey areas to the west and south of the temple. The data analysis tool below allows the viewer to explore the geophysical survey results from these survey areas.
The arrow keys in the lower right portion of the graphic allow the viewer to examine GPR images from different depths below surface. Consecutive clicks create a manual animation of the 3-D radar data. In the upper right portion of the graphic are three buttons. These buttons allow the viewer to display the magnetic field gradient imagery, either as a backdrop to the GPR imagery or by bringing the magnetic imagery into the foreground. A third button allows the viewer to examine the final interpretations of the geophysical data. These interpretations are in the form of a composite graphic constructed after analysis of both the GPR and magnetic data.
Active Content: Internet Explorer users: ActiveX controls must be activated by clicking inside the graphic. Firefox users: popup windows must be enabled
Geophysical Survey Results
Each of the survey methods used at Mahram Bilquis proved to be successful in resolving apparent archaeological patterning in spite of adverse environmental conditions. There can be greater confidence in interpretations where correlations exist between GPR and the magnetic survey results. Because each geophysical method responds to contrasts in different material properties, many suspected features detected by one instrument were not detected by the other. Similarly, one instrument may be unaffected by environmental conditions detrimental to the other.
The GPR signal response from most suspected archaeological features tended to be very subtle, suggesting the buried materials possess a limited contrast with the surrounding soils. Unfired mud brick and limestone blocks are possible sources of this signal. Basalt blocks appear to comprise a primary building material in limited areas. These areas can be identified by the intensity of the signal response in the magnetic data. Sources of uncertainty in the GPR survey data included metal debris on the ground surface, rock and soil redeposited from previous excavations, uneven terrain created by drifting aeolian sediments, and machinery, cell phones, and weapons, all of which were present in the vicinity of the GPR antenna during data collection.
Three primary sources of signal are visible in the GPR data. These are:
(1) Signal created by linear and rectilinear archaeological features. These anomalies are generally aligned either parallel or perpendicular to the alignment of the western gate of the temple, which is very close to true north. This signal is often very subtle when compared with the signal created by (2) and (3) below.
(2) Signal created by the interface between the soil paleosurface(s) and the aeolian overburden. These reflections tend to create large and amorphous anomalies that are best observed by viewing the GPR data as an animated sequence. An interesting trend observed in this signal is the migration of the paleosurface with depth towards an area immediately outside the western gate of the temple.
(3) Signal noise and clutter created by a variety of sources. This signal can generally be recognized in the GPR imagery as strong linear reflections aligned along the direction of data collection (i.e. grid north-south). Signal noise and clutter becomes more apparent in depth slices from below 50 centimeters below surface.
Magnetic Survey Results
In the magnetic data plots, anomalies in the colored range of the scale (both positive and negative) are likely to be caused by ferrous metal or igneous rock at or near the surface. Weaker or more diffuse patterning may not be as readily apparent, but may be discerned in the values closer to the mean, in the white/gray/black range of the scale. Visually detecting archaeological patterning is often a matter of “looking past” the extreme values that are often extraneous to the archaeological record.
Three areas of particular interest are expressed as strong magnetic anomalies with correlations in the GPR data, these are thought very likely to be large architectural features incorporating basalt:
A. A concentration of strong bipolar anomalies (grid) northwest of the western gate of the oval enclosure. Although somewhat amorphous, there are distinct rectilinear elements.
B. A very extensive area of linear patterning that is obviously associated with the cemetery (grid) south of the oval enclosure. This patterning has the same orientation as tombs exposed in the excavation blocks.
C. A large “L-shaped” anomaly located (grid) northwest of the western gate. The very diffuse and monopolar appearance of the magnetic anomaly suggests that the source is relatively deep. Despite its apparent depth, it is expressed with moderate strength, suggesting that the source may be massive basalt architecture. Elements of this feature appear clearly in the GPR data. It appears in the GPR data to be sloping downwards (becoming more deeply buried) as it approaches the western gate of the temple.
Conversely, a lack of magnetic response suggests a different physical composition:
D. Very strong rectilinear patterning in the GPR data suggests massive stone architecture, but without apparent correlation in the magnetic data. This suggests limestone as the principal construction material.
Much of the patterning thought to be of interest is quite subtle. In the magnetic imagery these weaker induced magnetic fields are expressed in the gray shaded range of the color scale very close to the mean. Two examples are given to illustrate different possible sources:
E. Throughout a large but poorly defined area (grid) west of the oval enclosure, rectilinear trends suggest architectural features or roadways. Generally these are weakly expressed in both datasets, suggesting more ephemeral architecture. Mud brick, limestone, or combinations thereof, are suggested as the principal building materials, although some elements appear to contain basalt as well.
It may be noted that the rectilinear patterning discussed or marked on the interpretive map falls with few exceptions within two orientations. One orientation is shared with the axis of the western gate and some of the rectilinear portions of the excavated temple, and tends to occur adjacent to these elements. Another orientation is shared with – and occurs proximal to – excavated tombs within the cemetery. The latter orientation is very close to true cardinal directions. Shared orientation may suggest temporal, functional, or cultural associations between features. Interestingly, a third orientation is suggested by:
F. Two relatively deep GPR anomalies having an orientation distinct from apparent features lying above them.
The 2006 geophysical investigation at Mahram Bilqis yielded high-resolution imagery depicting what appears to be a complex of structures and roadways to the (grid) west and south of the oval enclosure. These buried resources are generally aligned along a common orientation that is very similar to the orientation of the western gate of the temple.
The success of the 2006 geophysical investigation may be attributed to the application of a methodology that has evolved in North America over that past two decades. Due to the subtle low contrast nature of the North American archaeological record geophysical practitioners often find it necessary to employ precise spatial control combined with very high data sample densities - much higher sample densities than are typically employed in an old world context. Spatial control during the Mahram Bilqis investigation was facilitated by an extremely precise site grid put in place by the AFSM team. The high data sample densities collected during this investigation were made possible largely due to the enthusiastic help we received from our local Yemeni colleagues, who quickly learned the intricacies of precision geophysical data collection and assisted us for many long hours - under sometimes trying circumstances - with smiles on their faces.
Panoramic view of a portion of the Mahram Bilqis temple complex.
Additional photos of: the site; the geophysical survey and excavations in progress; and some of our colleagues who assisted us with data collection can be viewed by clicking here.