Interpreting Ground-penetrating Radar for Archaeology

Lawrence B Conyers

• Verlag: Routledge
• Jahr: 2016
• ISBN:
  9781315426310
• Auflage: 1
• Thema: Social Sciences
• Thema: Archaeology

Ground-penetrating radar (GPR) has become one of the standard tools in the archaeologist's array of methods, but users still struggle to understand what the images tell us. In this book—illustrated with over 200 full-color photographs—Lawrence Conyers shows how results of geophysical surveys can test ideas regarding people, history, and cultures, as well as be used to prospect for buried remains. Using 20 years of data from more than 600 GPR surveys in a wide array of settings, Conyers, one of the first archaeological specialists in GPR, provides the consumer of GPR studies with basic information on how the process works. He show how the plots are generated, what subsurface factors influence specific profiles, how the archaeologist can help the surveyor collect optimal data, and how to translate the results into useable archaeological information.

1

Introduction

Most of the commonly used near-surface geophysical methods, such as earth resistance, magnetics, and electromagnetic induction, produce images generated from instruments that gather averages of measurements related to the various physical and chemical properties of some volume of materials in the ground, with only a moderate amount of depth control. While these data sets can be processed in a way that can estimate depth (or distance), they still represent averages of what are often large volumes of materials. In contrast, GPR generates a data set of reflections from specific materials along the interfaces between units in the ground. Radar travel time is measured precisely, and these measurements can be converted to depth, yielding an often complex three-dimensional data set of reflection amplitudes over a surveyed area. This is the GPR method’s utility, but also a potential pitfall, especially if the interpreter does not understand and cannot correctly read the images produced from those raw GPR data. For this reason, my usual interpretation method includes spending a good deal of time at the outset collecting a data set that is calibrated for specific field conditions, and then even more time interpreting the resulting images that are produced in many different ways prior to reaching any conclusions. I spend what others consider an inordinate amount of time analyzing reflection profiles, long before I process them into amplitude maps. I have found that other GPR users often proceed directly to map making when starting their interpretation, often paying little attention to the actual data themselves; in this procedure, the most readily interpretable images might be the amplitude maps, but there could be more information to aid in overall interpretation within the standard reflection profiles. This book will therefore be heavily weighted to the interpretation of two-dimensional profiles, as a way leading ultimately to the interpretation of many of the other types of images, such as amplitude slice-maps and isosurfaces, which are created from multiple profiles in a grid.

FACTORS AFFECTING GPR INTERPRETATION

In order to interpret both two- and three-dimensional images of GPR data, users must think geologically, bio logically, geophysically, and archaeologically, often at the same time. An understanding of these variables is therefore important (but rarely possible for any individual practitioner):

Soil changes and types (pedology)

Soil chemistry (usually the greatest unknown in most GPR studies)

Stratigraphy of different depositional environments (near-surface geology and geomorphology)

How energy is propagated, reflected, refracted, and attenuated in the ground (physics and chemistry)

The types of cultural features that might be present and their geometry, distribution, and origin (archaeology)

How water is distributed and retained in the ground (hydrology)

The nature and distribution of other materials in the ground, such as tree roots and animal burrows (biology)

An understanding of GPR equipment components, how they are powered, and how electronics affects the type of data collected (electrical engineering)

After a GPR data set is collected and ready to be processed, all the above factors must be considered, even if not completely understood. In addition, one must have an understanding of computer processing methods and map making, as well as of image output, before interpretation can begin. That said, few GPR scientists understand and appreciate all these different variables. I grew up with a generation of scientists who knew computers only as mainframes that took our punch cards to process commands, and we obtained our results after waiting in line with others for thick paper printouts. This laborious one-way communication for data processing was obviously not conducive to active interpretation of complicated data sets, and we are fortunate to have long ago moved beyond such methods. I first used a personal computer (an IBM PC with two “floppy” drives, one for software and one for data storage) when I was 30 years old and remember marveling at how fast results could be obtained and then modified on the screen. Because of my very antiquated computer background, I had a steep learning curve in applying computers for GPR processing when I first began in the 1980s. In contrast, my younger colleagues today seem to have been raised with a computer mouse in their hands, and the computer aspects of GPR processing come quite naturally to them. Dean Goodman wrote the first usable GPR processing and image production program on an IBM XT computer similar to the one I first dealt with in late 1988. He told me that he usually had to let the computer run all night to test some portions of his software code, and then in the morning search through lines of error messages to see where the bugs were to be corrected that day. Fortunately, those days are behind us and we can let Dean and other smart programmers deal with our software problems quickly and efficiently.

While a multitude of variables that affect many aspects of GPR collection, processing, and interpretation must be considered, none of us can be expected to be expert in all these subjects. I feel blessed to have had my university training in geology, soils, and geophysics before I started in archaeology, long before I used my first personal computer. My first two college degrees were in geology and geophysics, and only when I was in my thirties did I become exposed to the complexities of the archaeological record within a geological matrix. As a result, when I entered the academic world I often struggled with the details of archaeological interpretation, and my colleagues and students ended up teaching me a great deal about these subjects as I became involved in different projects around the world. I am also hopeless in the subject of electrical engineering and have yet to open up a radar antenna to look at its contents, including the actual antennas and shielding components. My fear is that I would never be able to put it back together again. So, while some GPR experts have in the past tried to engage me in discussions that revolve around capacitors and resisters, I usually try to change the subject quickly so as to not appear as clueless as I really am in this subject, as I know just enough to be dangerous.

The key for any GPR interpretation is therefore to recognize the complexity of the method and at least know something about the variables that might come into play in an area where data have been collected. It is also important to search out other experts as consultants or collaborators when necessary, to help fill in the knowledge voids that will likely exist. I have been very fortunate in my own career to have had two wonderfully talented computer programmers (Jeff Lucius and Dean Goodman) as friends and colleagues. I can let them do what they do best, which is write software code that is useful for the rest of us. I am also constantly requiring the help of engineers at GSSI to keep me out of trouble with my equipment (and fix things when they break), and have relied on hundreds of others for local expertise when gathering data and interpreting results. My warning (and encouragement) to all those who do geophysical archaeology, and particularly GPR, is therefore to know your limits with regard to these variables, constantly ask for help from others who know more than you do, and go out and make mistakes, knowing that this is part of the learning process. On the board in my lab, I keep an archive of “mistakes” that my graduate students make, and we refer to them often; some we know by heart (and we are running out of space on the board!). Then I demand that we all think about other potential mistakes, what might cause other problems, and how to avoid them and learn from them in the future.

In writing this book, I have brought attention to many of my own mistakes, and those of my students, discussing how they were made and what we learned from them. This is not the way most scholarly writing is usually done, as most of us publish only our successes. I can’t remember ever reading an article on “mistakes I have made,” but with a subject as complicated as GPR I think it is necessary. So, in this book, I am “coming clean” by discussing a wealth of blunders and areas of ignorance, and what was learned from them. Perhaps this will keep others out of the same kind of trouble.

My background in geophysics, prior to GPR, was in seismic exploration. When I saw my first GPR reflection profile in 1989, I remember immediately thinking that I could relate easily to the types of displays that were commonly produced—that is, reflection profiles. I was happy to work with images where all the reflections along a line were visible on one piece of paper, ready for interpretation, and there were depth and distance scales to tell me the location of what I was seeing. This feeling is probably quite rare for others, as most off my students and other newcomers to GPR look at profiles and other images and comment that they “look like a bunch of squiggles.” I hope this book will help to make those “squiggles” into something more meaningful and useful.

THE IMPORTANCE OF UNDERSTANDING THE BASIC GPR IMAGES

My bias with GPR interpretation, which comes from my background, is therefore to begin by interpreting the basic raw data from which images are produced. This means studying and understanding in a general way the components of reflection profiles (the traces composed of reflections within a time-window from one spot on the ground) and then the reflection profiles themselves (composites of traces stacked and displayed next to one another along one transect). Only when these components are understood can other displays such as amplitude maps, isosurfaces, and videos be interpreted. This is often not an easy undertaking, but I think it is the most crucial step in GPR interpretation, and one that is most frequently overlooked by many because of its perceived complexity. In this book, I therefore begin with basic interpretations of individual traces and reflection profiles. Only then do I proceed to the production of amplitude maps produced from many profiles. When this is complete, and there is at least a basic understanding of what conditions are producing reflections in the ground, some of the more sophisticated images produced from GPR data can be meaningfully begun.

I have recently witnessed two interesting (and embarrassing) examples of public displays where very experienced GPR scientists presented incorrect results because of their neglect or inability (or both) to interpret the basic data used to generate those images. The pictures they presented were quite “pretty” and contained much that was fascinating, but the results were flawed because the two-dimensional profiles from which their maps and videos were produced had not been interpreted correctly. Fortunately, very few people in the audience understood their data input either, so there was no real public embarrassment. I didn’t point out these errors at the time, because the data from which the final products were produced were presented very quickly and not commented on, and only the results were emphasized. Later I talked to others in the audience and asked them what they thought of the interpretation. Those who understood how GPR data are collected and what produces reflections had the same impression as I did about the erroneous interpretation. For all the others in the audience, the images must have appeared to convey only a very complex data set that was difficult to understand but fun to look at. I discuss some of those interpretation problems in general in the chapters that follow. I am not making this point in order to make fun of others’ mistakes but to point out that even very experienced geophysicists can produce erroneous results when the raw data and basic products used to produce more complex images are not interpreted correctly.

INTERPRETATION THEMES

Throughout the chapters in this book, I discuss themes that relate to my own experiences in GPR interpretation. These topics are by no means comprehensive, deriving only from the surveys that my students and I have collected and the conditions we were confronted with. Some of the interpretive themes presented here are geo...