Ground-penetrating radar and its use in sedimentology: principles, problems and progress

Abstract

Ground-penetrating radar (GPR, also referred to as ground-probing radar, surface-penetrating radar, subsurface radar, georadar or impulse radar) is a noninvasive geophysical technique that detects electrical discontinuities in the shallow subsurface (<50 m). It does this by generation, transmission, propagation, reflection and reception of discrete pulses of high-frequency (MHz) electromagnetic energy. During the 1980s radar systems became commercially available, but it was not until the mid-1990s that sedimentary geologists and others began to widely exploit the technique. During the last decade numerous sedimentological studies have used GPR to reconstruct past depositional environments and the nature of sedimentary processes in a variety of environmental settings; to aid hydrogeological investigations, including groundwater reservoir characterisation, and to assist in hydrocarbon reservoir analogue studies. This is because in correctly processed radar profiles, and at the resolution of a survey, primary reflections usually parallel primary depositional structure. Despite the wide use of GPR, a number of fundamental problems remain in its application to sedimentary research. In particular, there are a wide range of approaches to the processing of radar data and interpretation techniques used on the final subsurface images vary widely, with little consensus over a common methodology. This review attempts to illustrate that methods for the collection, processing and interpretation of radar data are intimately linked and that thorough understanding of the nature, limitations and implications of each step is required if realistic sedimentological data are to be generated. In order to extract the maximum amount of meaningful information, the user must understand the scientific principles that underlie the technique, the effects of the data collection regime employed, the implications of the technique's finite resolution and depth of penetration, the nature and causes of reflections unrelated to primary sedimentary structure, and the appropriateness of each processing step with respect to the overall aim of the study. Following suitable processing, a radar stratigraphy approach to reflection profile interpretation should be adopted. New or modified terminologies and techniques to define a radar stratigraphy are also recommended, in order to make the interpretation process more transparent and to avoid confusion with related methodologies such as seismic stratigraphy and sequence stratigraphy. The full potential of GPR in sedimentary research will only be realised if more thorough and systematic approaches to data collection, processing and interpretation are adopted.

Introduction

For humans, the shallow subsurface is perhaps the most important geological layer in the earth. This layer contains many of the earth's natural resources (e.g. building aggregates/stone, placer deposits, drinking water aquifers, soils) and also acts as a sink for human waste (e.g. landfill sites). In addition, through the study of rocks and unconsolidated sediment accumulations at or near the surface we have discovered much about earth history and behaviour of its dynamic landforms. These insights have aided environmental management, such as prediction of natural disasters, helped exploration for more remote natural resources such as oil and gas, and increased understanding of the geological development of other planets in our solar system. Given the large rise in human population predicted for the 21st century, a more detailed understanding of the shallow subsurface will be required if humans are to sustainably manage many of the earth's finite resources.

As Grasmück and Green (1996) note, given the importance of the earth's upper layers to human development, it is surprising that during much of the 20th century techniques for exploring them did not change significantly. Analysis of field exposures linked via data from limited numbers of widely spaced boreholes, shallow excavations and geophysical surveys is still typical. Drilling and trial pits are time-consuming and expensive, often yielding only limited additional information that is difficult to correlate between distant sampling points. In some instances such invasive techniques cannot be implemented due to environmental or conservation considerations.

The most common geophysical techniques employed in shallow subsurface investigations are seismic reflection and seismic refraction, although these have a vertical resolution that does not normally meet the submetre resolution required in many practical situations. Consequently, during the 1970s attention increasingly turned to using other, higher resolution, geophysical techniques. One technique that has proved extremely useful is ground-penetrating radar (GPR or ground-probing radar, surface-penetrating radar, subsurface radar, georadar or impulse radar). GPR detects electrical discontinuities in the shallow subsurface (typically <50 m) by generation, transmission, propagation, reflection and reception of discrete pulses of high-frequency electromagnetic energy in the megahertz (MHz=10⁶ Hz, 1 Hz=1 cycle/s) frequency range. GPR's origins lie in research carried out during the early 20th century by German scientists trying to patent techniques to investigate the nature of various buried features Daniels, 1996, Reynolds, 1997. Pulsed electromagnetic waves were first used in the mid-1920s. Following these initial developments, much early work using radar was in glaciology (Plewes and Hubbard, 2001), with civil engineering, archaeological and geological applications becoming more frequent from the 1970s onwards Daniels, 1996, Conyers and Goodman, 1997, Reynolds, 1997. However, it was not until the 1980s that GPR systems became commercially available and digital data acquisition was feasible (Annan and Davis, 1992). Since the mid-1990s there has been an explosion of interest in GPR, with an ever-increasing number of research articles published on the technique each year (Fig. 1). Many publications relate to geological applications of GPR, a significant subset of which have a strong sedimentological component (Fig. 1). GPR has been used by sedimentologists to reconstruct past depositional environments and the nature of sedimentary processes in a variety of environmental settings, aid hydrogeological investigations (including groundwater reservoir characterisation), and assist hydrocarbon reservoir analogue studies (Table 1). It is these applications in sedimentology that form the basis of this review paper, although much of the GPR research performed is of direct or indirect relevance.

Although the rise in use of GPR in sedimentological studies can be attributed to its wider availability since the 1980s, its use by the research community is also related to the ease and rapidity of data collection, the ability to collect subsurface information away from outcrops or boreholes, and the apparent familiarity of the images, due to GPR's analogy with the established seismic reflection technique. The power of seismic reflection data was demonstrated to geologists when new interpretation techniques associated with seismic stratigraphy (Mitchum et al., 1977) revolutionised regional sedimentological studies in the late 1970s/early 1980s, and subsequently led directly to the concepts associated with the new geological science of sequence stratigraphy (for a review of the history and controversies surrounding seismic and sequence stratigraphy consult Miall and Miall, 2001).

Seismic reflection and GPR data are often analogous in terms of wave propagation kinematics Ursin, 1983, Carcione and Cavallini, 1995 and reflection and refraction responses to subsurface discontinuities McCann et al., 1988, Fisher et al., 1992a. Consequently, the broad assumptions that underpin processing and interpretation of seismic reflection data Sangree and Widmier, 1979, Yilmaz, 1987, Yilmaz, 2001 should also apply to GPR. With respect to interpretation, the basic assumption in both techniques is that, at the resolution of the survey and after appropriate data processing, reflection profiles will contain accurate information regarding the nature of a sediment body's primary depositional structure. In other words, the form and orientation of bedding and sedimentary structures in the plane of the survey will be adequately represented by recorded reflections, and any nongeological reflections can be readily identified and removed by data processing, or by simply discounting them from the interpretation. Although this assumption is a seemingly simple basis for the interpretation of radar reflection profiles, the degree to which it can be assumed to be true is dependent upon a wide range of factors. These include the nature of the sediment body under investigation, the groundwater regime, the type of terrain immediately adjacent to the survey line, the nature and appropriateness of any data processing undertaken, the interpretation techniques employed, and the overall understanding and experience of the researcher(s) with respect to GPR, and hence their appreciation of the other factors.

Geophysicists and sedimentologists want to extract accurate and meaningful sedimentological information from GPR profiles, but so do geomorphologists, soil scientists, hydrogeologists, archaeologists, environmental scientists and others with an interest in the structure of the shallow subsurface. Due to GPR's relatively recent development and acceptance, and the wide range of potential uses, experience of GPR end-users is wide and their subject backgrounds diverse. This point was emphasised at ‘GPR in Sediments’ (Geological Society, London, 20–21 August 2001) the first international conference on the applications of GPR in sedimentological studies (Bristow and Jol, 2003). From the papers presented and comments made by delegates in open discussion sessions, it was clear that appreciation of the GPR technique was highly variable across the user, and potential-user, base. On the basis of this observation and a comprehensive review of the literature, it is clear that overall understanding of the value and limitations of GPR is not as high as that generally displayed by the seismic reflection community for their analogous technique, where robust data collection, processing and interpretation methods have been developed, particularly since the 1970s.

It is not the purpose of this paper to review the use of GPR in sedimentology in terms of the actual sedimentological information extracted and the conclusions drawn. At this stage in the development of the use of GPR in sedimentological studies, this seems premature. Instead, it will attempt to consider in detail the basis for the use of GPR in sedimentology, the problems and pitfalls of data collection and processing, and the development of techniques for the interpretation of radar reflection profiles that maximise the sedimentological information obtained and prevent or minimise incorrect interpretations. Such an approach appears timely; an increasing number of sedimentological studies are attempting to utilise GPR (Fig. 1, Table 1), and yet the basis for the processing and interpretation of their data is often unclear and many misinterpretations or overinterpretations are evident.

Given the background outlined above, the aims of this paper are as follows:

• to introduce those theoretical aspects that are fundamental to understanding the GPR technique and its use in sedimentology, in a manner that is suitable for the wide user-base;

• to outline the fundamental limitations of the GPR technique, in particular by examining how unprocessed or poorly processed radar reflection profiles can often seriously misrepresent the nature of the subsurface sedimentary structure, and by considering the causes and nature of nongeological reflection events that further complicate subsequent data processing and sedimentological interpretation;

• to examine the ways in which appropriate data processing can enhance the interpretability of radar data, by producing reflection profiles that more accurately depict the subsurface sedimentary structure;

• to critically evaluate the assumptions that underlie interpretation of GPR data for sedimentological research purposes;

• to show how systematic interpretation of appropriately processed GPR profiles, through the application of a strictly defined radar stratigraphy approach, can maximise the sedimentological information extracted and minimise interpretation pitfalls.

To earth scientists familiar with the seismic reflection technique, the need for an understanding of the basic principles underlying data acquisition, processing techniques that convert data into the most meaningful representation of the subsurface possible, and robust interpretation techniques that maximise geological information return, might seem obvious. However, it is clear that, as yet, the GPR community engaged in sedimentological research does not fully share this common vision. This can only inhibit future research and wider appreciation of the technique within the earth-science community. This paper is, therefore, offered as a contribution to the on-going debate regarding the future direction of GPR research in sedimentology.