Moundbuilders of the Amazon: Geophysical Archaeology on Marajo Island, Brazil. Anna Curtenius Roosevelt. Academic Press, San Diego, 1991. xxvii + 495 pp. $89.95 (cloth). ISBN 0-12-595348-8.

Archaeological Prospecting and Remote Sensing. Irwin Scollar, Allain Tabbagh, Albert Hesse, and Irmela Herzog. Topics in Remote Sensing No. 2, Cambridge University Press, Cambridge, 1990. xv + 674 pp., figures, tables, notes, index. $135.00 (cloth). ISBN 052132050X.

Reviewed by James I. Ebert, Ebert & Associates, Albuquerque, NM 87107 USA

Although "archaeological prospection" has long had a respectable following — probably the majority of those who have constituted it are members of this society — a number of recent signs point to a resurgence in interest in applying remote sensing, particularly geophysical methods such as magnetometry, ground penetrating radar and soil resistivity surveying, to the search for and non-destructive characterization of archaeological remains. The 1st and 2nd International Conferences on Archaeological Prospection were well attended in 1995 and 1997, and another is planned for this year. Extensive Web sites dealing with the subject have appeared, and several institutions offer graduate degrees in archaeological prospection and related areas. John Wiley recently announced the publication of a journal, Archaeological Prospection, to cover all manner of archaeological discovery methods including geology, remote sensing, and geophysical and geochemical methods in "urban, rural, and marine environments."

The major incentives for this expansion in interest in non-contact detection of subtle surface or subsurface archaeological remains are probably the same as those for much of what else is happening in archaeology today: The phenomenal increase in the availability of digital methods, and the ever-increasing necessity for increasing the cost-effectiveness of conservation and research in the field. Nonetheless, it is interesting that what are arguably two of the most useful and comprehensive contributions to the literature of archaeological prospection, which are the subjects of this review, were published in 1990. Anna Roosevelt’s Moundbuilders of the Amazon is a conscientiously detailed account of a pioneering and exhaustive program of using geophysical methods to complement survey and excavation; its primary lessons are that such a program (then or now) must be meticulously planned and does not, by any means, produce "data for free." Irwin Scollar and his colleagues’ Archaeological Prospecting and Remote Sensing is a work of such staggering comprehensiveness in its subjects that it will probably serve the profession as our basic manual in these areas for decades to come.

Moundbuilders of the Amazon presents a meticulous, step-by-step description of research integrating geophysical remote sensing with intensive surface survey and excavation at Teso dos Bichos, a large population center of the Marajoara Chiefdom, which from A.D. 400-1300 occupied alluvial floodplains of the Lower Amazon noted for its monumental earthen mounds and elaborate ceramics, particularly funerary urns. It could today, as easily as when it was written, be used almost as a set of "how-to" instructions by archaeologists working virtually anywhere who would like to experiment with such methods in the context of their own site-based or regional fieldwork.

In fact, Brazilian archaeologists were among some of the first to experiment with geophysical survey methods because of their availability: geological exploration in Brazil made heavy use of them. Some early archaeological applications there were encouraging, and beginning in 1983 Roosevelt began a collaborative project with Brazilian archaeologists, under both NSF and Brazilian support, directed toward comparing various geophysical survey methods. These included total station topographic mapping, magnetic survey, conductivity survey, resistively survey, and ground penetrating radar.

The overall impression gained by any reader, especially one interested in applying practical lessons about geophysical prospecting in their own work, is that using these techniques is not to be lightly undertaken. Geophysical surveying is not something done as a "shortcut," nor does it in any way obviate systematic excavation. Geophysical data serve, rather, to bolster and amplify what one learns from later, hands-on digging. One must from the beginning be overwhelmingly systematic and meticulous, particularly in recording the spatial locations of data collection. That Roosevelt and her collaborators did their geophysical surveying just before computer data recording and data analysis became automatic procedures emphasizes these facts, but they are no less true today.

A base map with resolution congruent with that of the geophysical data to be collected is a fundamental prerequisite. The area of Teso dos Bichos Roosevelt and her colleagues studied measures on the order of 140m x 160m, and geophysical measurements were taken at resolutions as high as 1m x 1m, so a high-precision topographic base map had to be compiled first. This was done with EDM (electronic distance measurement) technology, familiar to many archaeologists today but something quite new in the early 1980’s. An EDM theodolite was used to map the site, with its x-y-z readings being manually entered in real time into a laptop computer (a Herculean effort, as I can attest, having done much the same thing over large areas at about this same time). The researchers also set out stakes at 5m x 5m intervals to serve as datums for the geophysical surveys. This was much more difficult than they had anticipated, as setting out always is, for one must use the survey instrument as feedback for plus or minus movement readings. This is still a problem today, and it would behoove us to find ways to automatically track survey instruments such as magnetometers using radio direction finding or some similar technology.

The natural soils of the site are highly conductive because of alkalinity and active-exchange clays, but when sherds, ash and bone are added become less so. When heated, as around hearths, they become highly magnetic. Clay sediments from nearby stream courses were used by the inhabitants of the site to construct their mounds, and ceramic clays were used to construct stoves. These practices created pockets of highly differentiated sediments in the site which were easily detected geophysically.

Magnetic survey was the most valuable technique in discriminating the size, number and composition of households in the site. Readings were taken at 1m x 5m and 2m x 2m intervals with proton precession magnetometers; to control for diurnal changes in the earth’s total magnetic field measurements were also taken from a permanent datum station every 15 minutes. Some higher resolution surveys were also done. Burned clay stoves and hearths were pinpointed, as was a burned house. Expectably, anomalies were more easily discriminated in the higher resolution surveys. A manually calculated and drafted map of magnetometer reading locations and magnetic contours on page 206 of this book stands as testimony to the tremendous field and laboratory effort expended in this process, and under magnification appears as meticulous and precise as anything we can make with a computer today.

Conductivity survey was used to derive more general maps of the major stratigraphic units in the site, and to confirm specific features, sensing soil texture, moisture retention, chemistry, and compaction. Native soils are more conductive while cultural inclusions reduce conductivity; an electronic induction meter was used to measure conductivity differences at 1m intervals along transects, and suggested earth constructions in some places, reinforcing the magnetometer surveys.

Electrical resistivity survey was additionally employed along seven sections of the site and resulted in bounding the total disturbed area of the site, as well as confirming large earth constructions in several parts of Teso dos Bichos and some areas outside the site.

Ground penetrating radar survey also was conducted over most of the site, in continuous transects 2m apart to make profile maps of the site’s stratigraphy, and to locate large subsurface anomalies or objects. Radar anomalies occurred only in small patches within the site, probably because of soil moisture and the high conductivity of the clay alluvium. The radar did, however, detect intrusions into the uppermost meter-plus of the soils.

Testing by excavation of the results of the various geophysical surveys is detailed in Chapter 5 of the volume. While excavation is to be regarded as the essence of archaeology, of course, this is perhaps the least interesting part of Moundbuilders of the Amazon. As any archaeologist who has tried to compare such disparate sorts of data as (for instance) those derived from aerial photointerpretation and ground survey of the same site knows, there is often very little comparability, and very little that one can say about such a comparison. The data are different, and they are complementary, but one is not in any sense a "reflection" of the other. It seemed to me, in fact, that Roosevelt may have only been reaching this conclusion, the most important lesson it has to teach the aspiring archaeological "prospector," as she finished her volume.

Archaeological Prospecting and Remote Sensing is written by an archaeologist, two geophysicists, and a mathematician, the authors point out in their preface. Although the archaeologist, Irwin Scollar, clearly provides the organization of this volume, no reader will be able to ignore the influence of the other "hard"-er scientists, and in fact some may be able to wade through parts of this volume only with difficulty. This is not to say that pages of equations weren’t necessary to reach the authors’ goals of presenting a systematic outline of the principles of geophysical methods employed in archaeological prospection, which they have without a doubt fulfilled.

Irwin Scollar has been and continues to be a trailblazer in "high-tech" applications in archaeology beginning some 40 years ago when he studied with the father of British aerial archaeology, O. G. S. Crawford. Scollar later took a research position at the Rheinisches Landesmuseum in Bonn, where, funded by a number of German agencies and foundations, he began using techniques then available only to a very few — including digital scanners, image processing, computer photogrammetry, and geophysical prospecting. He currently continues to occupy the cutting edge of digital archaeology as developer of the Bonn Archaeological Software Package (WinBASP), which incorporates many of the statistical and photogrammetric methods he began working on in Bonn.

Chapter 1 defines archaeological prospecting as using a wide range of passive (aerial and other photography, magnetic, thermal, and gravity prospecting) and active (electrical, electromagnetic, radar, seismic, and induced polarization) methods to detect and map sites and features. Such methods are necessary because excavation is destructive, expensive, and increasingly a "tool of last resort." The data they produce lends itself to visualization once it is transformed and displayed using computer methods, through which it can be organized spatially and temporally, and analyzed statistically to understand associations among types, periods, time, and space.

The archaeological phenomena actually discovered and measured by such techniques are structures, which are created either by bringing in non-native materials, or disturbing native materials (soils and sediments); other aspects of settlements such as fires, animals, agriculture and waste disposal also alter sediments in detectable ways. Properties of soils and sediments and additions, subtraction and alterations to them which produce measurable anomalies are explored in depth in Chapter 2.

The next section of Archaeological Prospecting and Remote Sensing, Chapters 3-5, constitute one of the most exhaustive discussions of aerial photography, image enhancement, and photogrammetry available to archaeologists today. The use of aerial photography for archaeological purposes has been pursued systematically in England and Europe since the 1920’s. The basic method of such "aerial archaeology," as its English-speaking practitioners call it, is to fly around visual indications of archaeological sites and structures until one gets an aerial photo that makes those indications "obvious" to the archaeologist — large and contrasty. Almost all such aerial archaeological air photos are oblique, taking advantage of subtle variations in sun angle, vegetative patterning, moisture markings, and frost and snow marks as viewed from different perspectives. This is in striking contrast to archaeological remote sensing in the United States, where aerial photographs have also been used by archaeologists since the 1920’s, but not very systematically, where archaeologists tend to use vertical axis photographs which are not taken for optimum clarity but rather for engineering mapping purposes. Cameras, films, lenses, filters, exposures, and other factors in taking the clearest possible aerial photos for archaeological purposes are discussed at very useful length.

Archaeological image processing, the subject of Chapter 4, is the quantitative used of imaged data, although it is emphasized that digital methods can never replace the judgement of the archaeological mind — it is the archaeologist’s knowledge and judgement that are all important, regardless of new technical methods. How many times has any reader encountered such an argument in any more recent, "upbeat" discussion of promising new methods? Not often enough. This explicit statement goes a long way toward the amelioration of the reader’s potential annoyance at some of the pages on image processing which follow, with perhaps somewhat overly-technical discussions of point spread functions, filtering, histogram modification, correction of motion blur, noise reduction, and edge extraction.

A photograph is the accommodation of a scene in the 3-dimensional world to a 2-dimensional image, and correcting for or perhaps more realistically using some of the resultant geometric distortions through rotation, scaling, projective transforms, and rubber sheeting are found in Chapter 5 where, finally, a number of real archaeological examples are employed to demonstrate how oblique aerial photos of crop marks and other patterning can be made into maps. More equations are also included, which no contemporary archaeologist will ever use in doing photogrammetry because software does it all for us today — in fact, Scollar’s WinBASP software does. Most of the examples are even more dated, however, involving aerial photos and photogrammetry done as long ago as the 1970’s — but the principles are still the same, and the discussions of photogrammetric computation of digital terrain models, contouring, digital photomosaics, scanning and film writing are as valid today as they ever were.

The second half of the volume is directed toward all of those other archaeological prospecting methods: resistivity prospecting (Chapter 6), magnetic prospecting and its scientific basis in soil properties (Chapters 7 and 8), electromagnetic prospecting (including ground penetrating radar)(Chapter 9), and thermal prospecting (Chapter 10). Although like most somewhat technically-oriented archaeologists, I like to feel that I understand such methods intuitively, I quickly began skipping more and more pages of differential equations in these sections, which are clearly the work of the two geophysicists and the mathematician. There is probably much more than you will ever need, or ever wanted, to know in these chapters, although there is also much practical detail here which would benefit anyone actually doing, say, resistivity prospecting — how to orient quadripoles to reduce the effects of apparent anisotropy in a resistivity map, equations to calculate the actual speed of making resistivity measurements, illustrations of various types of resistivity electrodes (including a tractor-towable disc electrode array), and a discussion of parasitic electrical phenomena.

In Chapters 7 and 8, building upon the baseline measurement of the magnetic properties of soils ameliorates the difficulty of determining the shape and nature of buried objects through magnetic survey, which can also be aided by the theoretical calculation of expected anomalies. Such calculations are complex, requiring more double and triple integration. A somewhat less mathematical discussion of the principles of the operation of different sorts of magnetometers and the data products produced by each should, however, be of wider interest, as will sections on practical considerations of magnetometry such as differential magnetometry, avoidance of magnetic contamination, position control of measurements, radio positioning, and data reduction and display. A section on the treatment of data as images in grayscale or color, aided by image processing, allowing the "photointerpretation" of magnetometer data is worthy of close reading. Possibly the most interesting section of Chapter 8, however, is a concluding historical note on the development of magnetic prospecting in archaeology, some of which was laid as early as 1896, culminating in the first archaeological magnetometry experiments in Cambridge University in the mid 1950’s, which involved some of the very first archaeological use of computers.

The last two Chapters of Archaeological Prospecting and Remote Sensing cover electromagnetic remote sensing and "thermal prospecting," and serve to reinforce, as does the conclusion of Roosevelt’s Moundbuilders of the Amazon, the fact that geophysical prospection does not produce the same sort of data as does archaeological surface survey or excavation. Electromagnetic prospecting, in contrast to electrical and magnetic prospecting, measures man-made electromagnetic fields at low frequencies (metal detectors) or high frequencies (ground penetrating radar). To me, the output from ground penetrating radar has always looked almost exactly like what one gets from a Bass Lo-k-tor, which in fact is what it almost exactly is, and anyone who has used such an aid in fishing knows that a Bass Lo-k-tor doesn’t tell you what you are looking at, or even just where it is. A great deal of expertise and judgement (to echo Scollar’s previous point) is needed on the part of the fisherperson or archaeologist to translate between geophysical data, or in fact that derived from aerial photographs, and what one sees or expects to see on or in the ground. This is also true for data derived through thermal imaging, covered in the volume’s last chapter.

Quite surprisingly, there is no concluding chapter or statement in this otherwise seemingly exhaustive volume. This may be because, after examining the complexities of the techniques of archaeological prospecting and the interpretation of the data they produce, the authors might be forced to conclude something slightly different than they optimistically set forth in the book’s introduction: that rather than an alternative to destructive and expensive excavation, the use of geophysical methods in archaeology results in unique, while complementary, information on what is contained on and in the ground.