Radar facies of paraglacial barrier systems: coastal New England, USA

Sedimentology (1998) 45, 181±200

Radar facies of paraglacial barrier systems: coastal New England, USA SYTZE VAN HETEREN*, DUNCAN M. FITZGERALD, PAUL A. M CKINLAY and ILYA V. BUYNEVICH Department of Earth Sciences, Boston University, Boston, MA 02215, USA

ABSTRACT Analysis of a large data base of ground-penetrating-radar (GPR) pro®les from both natural and developed paraglacial barriers along the coast of New England has allowed identi®cation of eight re¯ection con®gurations that characterize this type of mid- to high-latitude coastal environment. Bedrock anchor points yield primarily hyperbolic con®gurations, whereas glacial anchor points and sediment-source areas are characterized by chaotic, parallel, and tangential-oblique con®gurations. Beaches and dunes produce predominantly sigmoidal oblique, hummocky, re¯ection-free, and bounding-surface con®gurations. Back-barrier sediments may yield basin-®ll con®gurations, but generally include abundant signal-attenuating units. The GPR data, calibrated with information from cores, were collected across swashaligned and drift-aligned barriers in a variety of wave- and tidal-energy settings. Application of a 120-MHz antenna, as used in this study, enables portrayal of a range of sedimentary units, from individual bedforms (on single records) to entire barrier elements (using large numbers of intersecting GPR sections), at maximum vertical resolutions that vary between 0á2 m and 0á7 m. The most important drawback of GPR in the coastal environment is attenuation of the electromagnetic (EM) signal by layers of salt-marsh peat or by brackish or salty groundwater, primarily along barrier edges. This disadvantage is offset by many bene®ts. Data can be collected at rates of several km per day, making GPR an excellent reconnaissance tool. A core that is used in the calibration of GPR data can be matched with great accuracy to its position on the complementary GPR record, allowing detailed correlation between lithostratigraphy and re¯ection con®guration.

INTRODUCTION Knowledge of the stratigraphy of modern coastal barriers (barrier islands, barrier spits, welded barriers, and baymouth barriers) has steadily increased since the 1950s, partly because barrier models provide important analogues for the interpretation of ancient rocks. As a result, it is now possible to identify progradational (regressive), aggradational (vertically building), and *Present address: Faculty of Earth Sciences, Vrije Universiteit, De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands (E-mail: [email protected]) Ó 1998 International Association of Sedimentologists

retrogradational (transgressive) barrier sequences on the basis of shore-perpendicular stratigraphic cross sections (e.g. Dickinson et al., 1972; Galloway, 1986). Three types of shore-parallel transects have been recognized: one in which tidal-inlet facies are absent, one characterized by evidence of inlet stability, and one that shows the results of lateral inlet migration (e.g. Heron et al., 1984). With few exceptions, these existing barrier models are based on core, trench and surface data from barriers along low-relief, trailing-edge and marginal-sea coasts (de®ned by Inman & Nordstrom (1971) as coasts on the noncollision side of continents and coasts 181

182

S. van Heteren et al.

protected from the ocean by islands) that were not glaciated during the last glacial stage, or that have subsequently received large amounts of ¯uvial and deltaic sediment, erasing the effects of glaciation (Gulf and Atlantic coasts of the United States, North Sea coast of north-western Europe, parts of the southern and western coast of Australia). Barriers along collision and paraglacial (cf. Forbes & Syvitski, 1994) indented coasts differ markedly from most mid-latitude barrier islands. They are commonly small and isolated, and their occurrence is tied to the local or regional availability of sediment. In New England, the most prominent and continuous barrier systems have derived their sediment from the transgressive reworking of relatively large palaeodeltas and from updrift (pro)glacial and riverine sources. Small, localized onshore and offshore sources of glacial origin have been the primary contributors to single barriers of limited size (FitzGerald et al., 1994). Most of the barriers that occur along this part of the Atlantic coast are spits or baymouth barriers, anchored to the mainland on one or both sides, respectively. Barriers along indented coasts, although much studied for their historic and short-term morphologic changes, have generally not been the focus of extensive stratigraphic work, except in South Australia, New South Wales and Queensland (Roy et al., 1994). As these Australian barriers formed in a regime of relative sea-level (RSL) stillstand and fall, they cannot necessarily serve as analogues for barriers that formed in a regime of RSL rise. The relatively few published stratigraphic studies of barrier systems along submergent indented shorelines, which address for the most part the paraglacial New EnglandAcadian coast, are not detailed enough to enable three-dimensional characterization of the barrier stratigraphy (cf. Nichol & Boyd, 1993; FitzGerald et al., 1994; Ollerhead & Davidson-Arnott, 1995). Additional studies are necessary to allow the development of widely applicable stratigraphic models for this type of indented-coast barrier, but detailed core-based stratigraphic study of any barrier is costly and/or labour intensive. A rapid, inexpensive means of collecting stratigraphic data on barriers and in other non-marine environments is ground-penetrating radar (GPR), a high-resolution geophysical method that utilizes electromagnetic (EM) waves to probe the subsurface. GPR data do not eliminate the importance of cores, because individual units outlined by re¯ections on GPR pro®les still need to be

characterized in terms of their lithologic attributes. Nevertheless, large numbers of cores need no longer be a prerequisite for detailed stratigraphic work on coastal barriers when GPR data are available. Analysis of GPR data must be based on an approach that is similar to seismic stratigraphy, which relies not only on thorough knowledge of the behaviour of (acoustic) waves in common earth materials, but also on a well-de®ned set of re¯ection patterns and associated interpretations (Mitchum et al., 1977). Like seismic facies, most GPR re¯ection patterns are not uniquely tied to one particular lithofacies, creating the need for comprehensive sets of radar facies and interpretations for each depositional environment. With this fact in mind, we present a synopsis of a decade of GPR work along the paraglacial coast of New England (Fig. 1) through (1) de®nition and characterization of eight GPR facies, and (2) presentation of examples of barrier units that were analysed in three dimensions. The data discussed here come from a variety of wavedominated and mixed-energy swash-aligned and drift-aligned barriers. Narrow tide-dominated barriers, present in the northern part of the study area, were not studied because saltwater in¯uence, even at shallow depths, limited penetration of the EM waves. METHODS We used a two-component Geophysical Survey Systems Inc. SIR System 3 that includes a monostatic antenna and a control and display unit. The GPR was powered by a 12-V battery and assembled on a small cart to increase mobility. The small size of the cart allowed us to access woody patches and narrow paths. Three transceivers were available, with frequencies of 120, 300 and 500 MHz. We used the 120-MHz transceiver for two reasons: (1) it provided the depth range required (10 m±15 m) for our survey of barrier sediments, and (2) it offered the resolution necessary (0á2 m±0á7 m) to allow identi®cation of the most pertinent facies boundaries, as well as characterization of internal re¯ection con®guration. The GPR data were collected in a continuous recording mode, at two-way travel-time settings that varied between 100 ns and 600 ns and using various gains selected to optimize the output on the display unit. The sections we present here, which are corrected for topography but otherwise unprocessed, differ slightly in appearance from

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 181±200

Radar facies of paraglacial barrier systems

183

Fig. 1. Location of the study areas.

those obtained using a discontinuous recording mode and displayed in a wiggle-trace format. Re¯ection positions or con®gurations, however, are affected neither by display format, nor by the type of recording mode used (van Overmeeren, 1994), as long as continuously recorded sections are collected at constant speeds. Interpretations of GPR-re¯ection con®gurations are based on information from trenches and cores (see Appendix 1). These data enabled correlation of individual re¯ections with lithologic changes, determination of facies depths and thicknesses, and calculation of the velocity of the 120-MHz EM waves in some of the facies. Cores were also used to con®rm and further outline the large-scale continuity of units and boundaries, where suggested by GPR records. APPLICATION OF GPR IN THE STRATIGRAPHIC ANALYSIS OF MODERN AND ANCIENT DEPOSITIONAL SYSTEMS Considering its potential as a high-resolution stratigraphic tool, GPR has been used remarkably little in sedimentary facies analysis. Most of its geological applications have been related to environmental and groundwater, glacier and per-

mafrost, and geotechnical issues (e.g. HaÈnninen & Autio, 1992; Redman et al., 1994). Nevertheless, GPR sections depicting characteristic features of many non-marine and coastal sedimentary environments have now been published, and GPRdata collection speci®cally for facies-analytical purposes has increased rapidly since the early 1990s. Beres and Haeni (1991) and Jol and Smith (1991) de®ned radar facies by analogy to seismic facies and provided examples and interpretations of the most common re¯ection con®gurations. Other studies focused not only on two-dimensional re¯ection characteristics, but also on threedimensional geometry of architectural elements (e.g. Gawthorpe et al., 1993; Stephens, 1994; Olsen & Andreasen, 1995). Commenting on this three-dimensional characterization, Bristow (1995) emphasized that the scale factor must be kept in mind when comparing radar facies and sequences to otherwise very similar seismic facies and sequences. Even the highest-resolution seismic data (1 m under very favourable circumstances; McCann et al., 1988) typically allow identi®cation of only those units that are equal in size to or larger than complete macroforms in major elements of depositional systems (Groups 6±10 in Miall's (1991) classi®cation), formed over periods of hundreds of years or more. GPR data, on the

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 181±200

184

S. van Heteren et al.

other hand, portray a range of small- to large-scale units (Miall's (1991) Groups 4±7), from megaripples that formed in a few days (Beres et al., 1995) to coastal barriers that developed over many thousands of years (van Heteren et al., 1996; van Heteren & van de Plassche, 1997). During the past few years, GPR has been increasingly applied in the stratigraphic analysis of modern non-deltaic coastal systems. A pilot study was conducted by Leatherman (1987). Subsequently, FitzGerald et al. (1987, 1992) and van Heteren et al. (1994) analysed in more detail the stratigraphy of some paraglacial barriers along the Atlantic coast of eastern New England. Elsewhere, Baker (1991), Smith and Jol (1992) and Jol et al. (1994) inferred depositional processes of several Australian and North American coastal spits, Meyers et al. (1996) studied the internal structure of a barrier spit along the coast of Washington, and Dott and Mickelson (1995) used GPR in the analysis of a lacustrine beachridge complex in Michigan. Recently, Jol et al. (1996) presented a selection of GPR sections from the Paci®c, Atlantic and Gulf coasts of North America. THEORETICAL ASPECTS OF GPR Individual GPR records can be interpreted in terms of parameters such as re¯ection con®guration, continuity, amplitude, spacing (frequency), interval velocity and external form, analogous to seismic facies analysis (e.g. Mitchum et al., 1977). The most important factors determining the re¯ection characteristics of a GPR signal in sediments are: (1) pore-water content (a function of the position of the water table and of texturerelated porosity and permeability); (2) pore-water chemistry (type and quantity of dissolved matter); (3) clay content and chemistry; and (4) heavymineral concentration (e.g. Topp et al., 1980). Interpretation of GPR records is based on knowledge of the behaviour of EM waves in different lithologic units (von Hippel, 1954; Daniels et al., 1988; McCann et al., 1988; Davis & Annan, 1989). The two most important attributes of EM waves in rock and sediment, velocity and attenuation, are functions of relative permittivity (or dielectric constant) and conductivity, respectively. Attenuation is important because it affects the penetration depth of the EM waves. It limits the usefulness of GPR in sedimentary facies characterized by salty and brackish groundwater or by

clay-rich strata thicker than 0á5 m to 1á0 m (e.g. Cook, 1975). The maximum penetration depth of returning, measurable EM waves, however, is controlled not only by the electric properties of the stratigraphic units, but also by GPR system and antenna characteristics, including the initial signal amplitude, receiver sensitivity, antenna ef®ciency and gain, and frequency (cf. Davis & Annan, 1989). In any particular area, a higher frequency gives less penetration but better horizontal and vertical resolution than a lower one (cf. GrasmuÈck, 1992; Jol, 1995). Vertical resolution depends not only on the frequency used but also on the velocity of the EM signal in a lithologic unit. Its value may be as small as one eighth to one quarter of the EM-wave length in a lithologic unit (Widess, 1973; Sheriff, 1977), but background noise, wave-shape changes and velocity uncertainties commonly limit resolution to one half of the EM wave length. Ringing may further limit vertical resolution. Horizontal resolution depends on the frequency used, on the depth of a re¯ecting horizon, and on the relative dielectric permittivity (cf. GrasmuÈck, 1992). At depths of 10 m, well within the penetration range of many GPR sections collected with 50-MHz to 300-MHz antennas, typical fresnel zones are at least 2 m (saturated sand) to 3 m (unsaturated sand) in diameter (GrasmuÈck, 1992). If the volume percentage of magnetite and other magnetic minerals is low, the magnetic permeability will not be a key component in the behaviour of the EM waves in a lithologic unit. Through laboratory and ®eld experiments, the ranges of EM properties have been determined for a number of common lithologic units (Table 1). Although measured values provide a good overall indication of the behaviour of EM waves in the subsurface, the variability of sediments and rocks precludes precise characterization of most stratigraphic units. For each study, the value ranges of the most pertinent EM parameters have to be constrained by means of on-site measurements. GPR FACIES: INTERNAL CONFIGURATION, SIGNAL ATTENUATION AND EXTERNAL FORM Ideally, the description of GPR facies or facies sequences should address two components: internal re¯ection con®guration and external form. Both components can be analysed from a twodimensional perspective, based on individual

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 181±200

Radar facies of paraglacial barrier systems

185

Table 1. Electromagnetic properties of units occurring in the study areas (for 80±120-MHz waves). Material Air Freshwater Brackish/saltwater Ice Saturated clay Clay (undifferentiated) Unsaturated clay Saturated silt Silt (undifferentiated) Unsaturated silt Saturated sand Unsaturated sand Saturated sand and gravel Unsaturated sand and gravel Saturated till Till (undifferentiated) Unsaturated till Freshwater peat Bedrock

Relative Permittivity e

Conductivity r (S/m)

Velocity m (m/ns)

Attenuation A (dB/m)

1 87á8±0á37 T (°C) 80±81 3±4 15±40

0 10)5±10)3 4±30 10)5 0á02±1

0á30 0á033 0á01±0á033 0á16 0á05±0á07

0 0á002±0á18 320 ± >1000 0á01

2á5±5 22±30

0á002±0á02  0á1

0á09±0á12 0á05±0á07

2á5±5 20±31á6 2á55±7á5 15á5±17á5 3á5±6á5

0á09±0á12 0á05±0á08 0á10±0á20 0á06 0á085±0á13

24±34

0á001±0á1 10)4±0á01 10)7±0á001 7 ´ 10)4±0á009 7 ´ 10)6± 6 ´ 10)5 0á002±0á005

7á4±21á1 57á4±69á4 4±6

0á0025±0á01 0á003±0á01 10)8±0á04

0á28±300

Penetration Dmax (m)

0á3 70 1á5 2±7

1±100 0á03±2á3 0á01±0á14 0á03±0á5 0á01±0á1

0á10±0á12 0á04±0á057 0á049±0á083 0á12±0á13 7 ´ 10)6±24

23±29 23±29

2±4

Source: van Heteren, 1996.

sections, or from a three-dimensional one, which requires a number of intersecting sections. External form can best be characterized through the construction of thickness-isopach or boundarydepth-contour maps that are based on a dense network of GPR track lines. It is preferable that a facies be characterized from a three-dimensional point of view, because form, and in many cases con®guration, are dependent on the orientation of a section relative to the depositional dip of the probed sediments. Records parallel to depositional dip show the widest range of variations, and are therefore the most diagnostic individual transects in terms of depositional processes and environments. We have identi®ed eight different facies types in GPR sections of eastern New England's paraglacial barriers, their sources and their anchor points (Fig. 2). Because GPR data are not always collected along intersecting track lines, the identi®cations are based on two-dimensional characteristics; however, each facies description includes a reference to the third dimension. Seven of the facies types are de®ned exclusively in terms of internal facies con®guration; one also includes the aspect of two-dimensional, external facies form. Abundant sediment cores exist for most GPR facies types (see Appendix 1); they enabled classi®cation of the facies in terms of lithology and depositional environment.

Hyperbolic facies Description A large-scale hyperbolic con®guration was identi®ed on many GPR sections, most of which were collected in areas with nearby bedrock outcrops. Sections with hyperbolic facies are characterized by very prominent, irregular re¯ections that mark the top of a re¯ection-poor unit (Fig. 3). Penetration depths are limited to several metres.

Interpretation Hyperbolic con®gurations result from the response of an extremely irregular boundary to the diverging GPR beam. EM energy returns to the transceiver not only through re¯ection on nearhorizontal surfaces, or when the antenna is above a boundary peak, but also through re¯ection of that part of the peripheral GPR beam that forms a fresnel zone on subhorizontal to sharply inclined boundary surfaces when the antenna has been moved away from such a boundary peak. As long as a ridge, pinnacle or other positive-relief feature forms a detectable fresnel zone within the footprint of the GPR beam, it will contribute to the production of a re¯ection that takes the shape of a hyperbola as the distance between the stationary feature and the moving antenna decreases at ®rst,

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 181±200

186

S. van Heteren et al.

Fig. 2. GPR facies and their interpretation in a paraglacial coastal setting. Scales, which differ for each con®guration, are given on the corresponding GPR pro®les of Figs 3±14.

then increases after the antenna has moved overhead or alongside. Each large-scale hyperbola on the collected GPR pro®les represents part of a buried bedrock ridge or pinnacle. Some of these positive-relief features have functioned as nuclei to barrier formation and as anchor points to barriers during part or all of their existence. Hyperbolic density depends on both the spacing and the irregularity of ridges and pinnacles, which are functions of

the local geology. Aside from the density aspect, however, the re¯ection con®guration is not dependent on track-line orientation.

Chaotic facies Description Two types of chaotic facies occur on GPR pro®les from the study areas. Least common is a

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 181±200

Radar facies of paraglacial barrier systems

187

Fig. 3. Hyperbolic re¯ection con®guration, produced by an irregular bedrock surface underlying water-saturated estuarine sand, on a GPR pro®le from Popham Beach. See Appendix 1 for a description of core PB±15. Bedrock outcrops are located about 10 m south of the pro®le. NGVD-29 is the National Geodetic Vertical Datum of 1929. Here and in all other GPR pro®les, the vertical exaggeration (V.E.) refers to unsaturated facies (lowest number) and saturated facies (highest number).

high-frequency con®guration. It was identi®ed on small parts of only a few sections, commonly in association with adjacent hyperbolic GPR facies. Discontinuous re¯ections dominate. These re¯ections exhibit some small-scale hyperbolic characteristics and some cross-cutting patterns that indicate the merger or divergence of layers. They also may include near-vertical shifts of individual re¯ections (Fig. 4). The second, low-frequency type is much more prevalent. It occupies large sections on records taken along paved roads, constructed on top of texturally variable arti®cial ®ll. Individual re¯ections are mostly discontinuous and vary in dip angle and direction. They constitute a re¯ectionpoor internal con®guration that is independent of track-line orientation (Fig. 5).

Interpretation Chaotic re¯ection con®gurations imply the presence of a lithologic unit that is massive or heterogeneous. The degree of heterogeneity determines re¯ection abundance and variability, whereas textural differences determine re¯ection amplitude. The minor hyperbolas in this facies type result from relatively small point objects that act as near-stationary individual fresnel zones as the GPR footprint sweeps across, in much the same fashion as the bedrock pinnacles of the hyperbolic facies type. Generally, these point objects are large clasts, such as boulders. Alternatively, hyperbolas in low-frequency chaotic facies may re¯ect buried (sewer) pipes (Fig. 5).

Fig. 4. GPR pro®le from Thompson Island. Disrupted and contorted re¯ections dominate. At the location of this pro®le, collapsed and faulted outwash facies are present at the surface and identi®able in trenches. Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 181±200

188

S. van Heteren et al.

Fig. 5. Scattered hyperbolas are the most prominent re¯ections of the low-frequency chaotic facies shown on this GPR pro®le from the Saco Bay area. Town records indicate that at this site, utility cables and pipes occur to a depth of 3 m.

Cores and trenches link the high-frequency chaotic facies to proglacial outwash facies dominated by deformation structures that resulted from the melting of buried stagnant ice (Fig. 4), to saprolites, and to glacial diamict, which are all sources of barrier sediment as well as potential barrier anchor points. Near-vertical re¯ection shifts mark the locations of faults. Other re¯ection patterns indicate the presence of a variety of slumping and other deformation features. The low-frequency chaotic facies occurs exclusively on GPR pro®les from areas where the upper portion of the natural stratigraphy has been disturbed by man. Diffraction patterns caused by overhead power lines and other large objects on or above the surface are common. Varying penetration implies a variety of sediment types.

Parallel facies Description Both wavy- and even-parallel con®gurations were identi®ed on GPR sections from the New England coast. A wavy-parallel con®guration is generally characterized by limited apparent thickness, whereas more common even-parallel con®gurations exhibit a range of penetration depths, from less than 1 m to more than 10 m. Continuous re¯ections dominate the upper parts of these EM units; discontinuous re¯ections ®gure more prominently in the lower parts of sections characterized by wavy-parallel and re¯ection-poor even-parallel EM facies.

Interpretation The limited apparent thickness of wavy-parallel EM facies is a result of a lack of penetration. It indicates the presence of signal-attenuating clay and silt layers. Sparse re¯ections mark texturally coarse layers within a ®ne matrix (Fig. 6). Cores match this re¯ection pattern with glaciomarine mud alternating with thin sand layers. Only random differences exist between GPR sections of different orientation, because this con®guration type is a function of the subsurface palaeotopography. Wavy-parallel re¯ections signify sedimentation at uniform rates over an irregular subsurface. Two types of even-parallel facies occur in the study areas. A re¯ection-poor type is closely related to the wavy-parallel con®guration. Here, widely spaced high-amplitude horizontal re¯ections mark sandy layers in a muddy matrix draped over a ¯at basement surface. Like its wavy counterpart, this type of parallel facies is not controlled by the direction of the survey line. A second even-parallel facies type is very common in sections taken perpendicular to the depositional dip of lithologic units that include a signi®cant directional component. In this case, closely spaced high-amplitude re¯ections and commonly deep penetration signify mud-poor, distinctly layered sand and/or gravel accumulations in beach and deltaic outwash units. This con®guration is primarily the perpendicular equivalent of the different types of oblique radar

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 181±200

Radar facies of paraglacial barrier systems

189

Fig. 6. Wavy-parallel facies type on a GPR pro®le from a Saco Bay barrier, with two clear re¯ections that mimic each other. Farther to the south-south-west, the bottom re¯ection fades but the top re¯ection continues. Fifty metres to the north-north-east of the section shown, this top re¯ection marks the upper boundary of a glaciomarine sand and mud unit, as con®rmed by core OOB81±22-B103 (see Appendix 1).

facies. Alternatively, closely spaced even-parallel re¯ections mark sand and gravel alternations in tidal-inlet facies, regardless of track-line orientation (Fig. 7). However, not all inlet-channel deposits are marked by even-parallel re¯ections, as a migrating channel may produce an oblique signature that includes re¯ections dipping in the direction of migration.

Oblique facies Description Two main oblique re¯ection con®gurations, marked by high-amplitude re¯ections and moderate-to-deep penetration, were identi®ed on GPR sections that are not perpendicular to depositional dip. Abundant toplap and downlap re¯ection

Fig. 7. GPR section from Sandy Neck, showing a number of different re¯ection con®gurations, including evenparallel (lower right) and sigmoidal-oblique (upper left) facies. The strong re¯ections on the GPR record mark closely spaced transitions between ®ne to medium and medium to coarse sand and gravel in tidal-inlet and spit-beach units. The re¯ection con®gurations shown here are continuous in all directions. Core SN94-PA47, »50 m west of the section shown, links the even-parallel con®guration with inlet-channel facies, and the sigmoidal-oblique con®guration with spit-beach sand and gravel (see Appendix 1). Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 181±200

190

S. van Heteren et al.

Fig. 8. GPR pro®le collected parallel to the depositional dip of a deltaic subaqueous outwash sequence adjacent to the Saco Bay barrier system. See Appendix 1 for a description of core SB96-PA5.

terminations are typical in a tangential-oblique re¯ection con®guration (Fig. 8). This EM facies type, which is uncommon, is characterized by moderate thicknesses, in the range of 2 m to 6 m. The re¯ections exhibit varying amplitudes. Onlap, downlap and internal convergence are de®ning features of both sigmoidal-oblique and complex sigmoidal-oblique re¯ection con®gurations (Figs 7 and 9). The presence of toplap in addition to onlap, downlap and internal convergence characterizes the complex sigmoidal-oblique con®guration, which is much less common than the regular sigmoidal type.

Interpretation A tangential-oblique con®guration indicates nearunidirectional deposition (semi-) parallel to the GPR record. High-amplitude re¯ections represent boundaries between alternating layers of silt and sand, whereas moderately deep penetration signi®es the paucity of clay-size particles. Cores link this con®guration to Pleistocene outwash-delta facies, erosion of which provides sediment to barriers. Subaqueous delta progradation can be recognized in the foresets that occur between near-horizontal top- and bottomsets (cf. Oldale et al., 1993). A similar radar facies marks distal portions of washovers (cf. HeÂquette & Ruz, 1991; Murakoshi & Masuda, 1991; van Heteren & van de Plassche, 1997). A sigmoidal signature re¯ects progradation or accretion of a sedimentary unit. This re¯ection con®guration is usually indicative of beach sand

and gravel units with signi®cant vertical grainsize transitions; alternatively, it may signal the presence of storm layers enriched in heavy minerals (cf. Thom & Roy, 1985), which typi®es many eolian and backshore sequences (van Heteren et al., 1994). A complex sigmoidal-oblique re¯ection con®guration indicates periods of alternating or simultaneous aggradation and progradation, typical of some beach facies. Each sigmoidal accretionary wedge represents a distinct period of seaward (spit-) beach accretion, whereas simultaneous aggradation results in a vertical stacking of accretionary units (cf. Hine's (1979) neap-berm development and Carter's (1986) Mode I beach ridges; Fig. 9). High-amplitude re¯ections and deep penetration signify an alternation of gravel and sand with little or no clay and silt.

Hummocky facies Description A wide variety of small-scale re¯ection forms constitutes a hummocky con®guration (Fig. 10). This facies is independent of GPR-pro®le orientation. Individual high-amplitude re¯ections dip in different directions and at different angles. Penetration depth is limited to several metres. The horizontal dimension of individual hummocks ranges from 10 m to 50 m. The larger hummocks are marked by prominent re¯ections that outline steeper internal re¯ections with consistent apparent dip angles and directions.

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 181±200

Radar facies of paraglacial barrier systems

191

Fig. 9. Progradational and aggradational characteristics of a beach sequence are re¯ected in the complex sigmoidaloblique re¯ection con®guration on this GPR pro®le from Sandy Neck, taken parallel to depositional dip. See Appendix 1 for a description of core SN94-PA6.

Interpretation

Bounding-surface facies

This facies shows the effects of multi-directional migration of coexisting bedforms. The hummocks represent preserved in situ, mixed bedforms developed in a longshore accretionary beach setting (cf. Hine, 1979, berm-ridge development or swash-bar welding; Hayes, 1980, welded-ridge facies; Carter, 1986, Mode II beach ridges; Nielsen et al., 1988, trough and bar deposits). The same con®guration may result from ridge and runnel structures (cf. Davis et al., 1972). Different dip directions and angles are a function of true dip of lithologic units and of their orientation relative to the GPR track line. Limited penetration depth is not a function of grain size, but rather the result of attenuation of the EM signal by brackish or salty pore water.

Description Away from the sea or the back-barrier area, this facies is easily penetrated by the EM signal, to depths of more than 15 m. Individual re¯ections may dip in either direction on any section, showing occasional bounding surfaces (cf. Schenk et al., 1993; Fig. 11). The orientation of the GPR track line does not affect this con®guration.

Interpretation Relatively low-amplitude re¯ections as compared to those of the oblique and parallel con®gurations on the same sections imply relatively small re¯ection coef®cients, characteristic of boundaries between unsaturated or texturally similar sediments. Surface and core data link the

Fig. 10. GPR pro®le from the portion of Sandy Neck that is adjacent to the tidal inlet, showing a hummocky re¯ection con®guration. See Appendix 1 for a description of core SN94-PA63. Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 181±200

192

S. van Heteren et al.

Fig. 11. GPR pro®le of a thick eolian unit at Sandy Neck, consisting mostly of well-sorted ®ne-medium sand. Highamplitude re¯ections mark the water table and a palaeosol (labelled) that is continuous with a bounding surface. See Appendix 1 for a description of core SN94-PA2, which is nearby and places the base of the eolian unit at about 4 m below NGVD±29. Abundant bounding surfaces and several palaeosols characterize eroding dune slopes surrounding the GPR-pro®le site.

boundary-surface con®guration to sandy eolian units that allow excellent penetration. Occasional medium-amplitude re¯ections signify coarse lag deposits (coarse sand and granules) and palaeosols, whereas the absence of hyperbolas re¯ects a corresponding lack of large clasts. Re¯ection terminations marking changes in dip direction illustrate cross-cutting relationships.

Re¯ection-free facies Interpretation A (virtually) re¯ection-free con®guration may signify: (1) massive homogeneous lithologic units; (2) the presence of highly conductive dissolved minerals in groundwater; and (3) the presence of a layer of clay- and gas-bearing saltmarsh peat that attenuates all of the EM signal, preventing it from penetrating deeper units that would otherwise be characterized by distinct re¯ection patterns. Attenuation of the EM signal is exhibited in some very prominent ways in units characterized by high conductivity losses. On many barrieredge sections, a sharp, inclined transition from a re¯ection-rich to a re¯ection-free GPR facies marks an abrupt attenuation increase, ascribed to the boundary between the freshwater aquifer and a surrounding zone of brackish or salty groundwater (Fig. 12). In transects near salt marshes, wedges of peat as thin as 0á1 m mask underlying strata through attenuation of most or all of the signal. Sharp, vertical transitions between re¯ection-rich and re¯ection-free con®gurations were tied to pinch-out locations of saltmarsh peat (Fig. 13). The extent of salt-marsh peat

underneath (or intercalated with) a barrier lithosome is easily mapped using this feature.

Basin-®ll facies Description Prominent non-parallel bounding re¯ections are the basis for identi®cation of a complex basin-®ll re¯ection con®guration. A subhorizontal top re¯ection overlies an irregular bottom re¯ection, resulting in a pattern that is not tied to track-line orientation. Penetration is limited to 2 m or less. In Fig. 14, the complexity of the basin ®ll is accentuated by the presence of a small divergent ®ll pattern that forms an integral part of the con®guration.

Interpretation A basin-®ll con®guration suggests a low-energy depositional environment with non-uniform sedimentation. The limited penetration depth that characterizes this con®guration implies the presence of clay-rich sediments or other attenuating units. The irregular bottom re¯ection of this facies type marks a palaeotopographic surface that outlines the top of Pleistocene sediments. Cores that relate to the GPR pro®le of Fig. 14 show that the subhorizontal top re¯ection marks the upper boundary of ¯at-topped back-barrier facies that overlie the irregular palaeotopography and are covered by washover sand and gravel (Appendix 1). The small ®ll pattern on Fig. 14 re¯ects the deposits of a small tidal creek. High-resolution seismic sections of ponded glaciomarine and glaciolacustrine sequences from

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 181±200

Radar facies of paraglacial barrier systems

193

Fig. 12. GPR pro®le from a barrier along Saco Bay, showing the outline of a freshwater aquifer of limited size in barrier-spit sediments (cf. Bokuniewicz & Pavlik, 1990).

Fig. 13. GPR pro®le from Sandy Neck, illustrating the effect of EM-signal attenuation by a wedge of salt-marsh peat that extends underneath part of the present-day barrier. See Appendix 1 for a description of core SN94-PA3.

some of the study areas' offshore regions exhibit ®ll patterns also, but these are divergent, not complex (Belknap et al., 1989). Divergent ®lls have thus far not been observed on any onshore GPR records, likely because they commonly occur within and underneath thick draped sequences of mud (parallel radar facies), which limit penetration of the GPR signal.

External form External facies form, as mentioned brie¯y in de®ning the basin-®ll facies type, is primarily

important in the three-dimensional description of stratigraphic units. A large number of intersecting GPR sections that provide good coverage of a stratigraphic feature, when integrated with core and morphologic data, enables identi®cation of facies form as well as dimensions. Determination of shape, extent and thickness of stratigraphic units is essential to an understanding of the processes responsible for their deposition. Equally important in this respect is the delineation of individual facies boundaries. In one of our studies, shape and dimensions of the barrier lithosome along Saco Bay in Maine

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 181±200

194

S. van Heteren et al.

Fig. 14. GPR section from the Saco Bay barrier system, showing a complex basin-®ll re¯ection pattern. See Appendix 1 for a description of core SB93-VC32.

(Fig. 1) were de®ned by means of more than 100 shore-normal and shore-parallel GPR transects, calibrated by about 40 cores (van Heteren et al., 1996). Bedrock ridges and pinnacles that are presently buried but served as barrier anchor points in the past were easily identi®ed on GPR sections (Fig. 15). A high degree of small-scale barrier segmentation, which is not immediately evident at the surface, clearly distinguishes this system from typically inlet-bounded mid-latitude barriers formed in wave-dominated or mixedenergy regimes. Analysis of the assemblage of barrier-spit facies at Sandy Neck in Massachusetts (Fig. 1) was based on a similar combination of GPR and core data. Here, about 100 shore-normal and shore-parallel GPR transects and 77 cores were used in conjunction with morphologic and seismic evidence to

map the form of the tidal-inlet ®ll (Fig. 16). The same data base also allowed delineation of the eolian facies, which extends from the highly irregular present-day surface of the spit to a smooth lower boundary. The latter slopes upward in the direction of the modern tidal inlet, re¯ecting a rise in RSL as Sandy Neck accreted through time (van Heteren & van de Plassche, 1997). IMPORTANCE OF GPR APPLICATION IN COASTAL STRATIGRAPHIC STUDIES The application of geophysical methods in stratigraphic studies of modern sedimentary environments has been directed primarily to marine and lacustrine environments where coring is dif®cult and expensive. High-resolution seismic data have

Fig. 15. Schematic representation of the compartmentalized, bedrockanchored barrier lithosome along the Saco Bay shoreline. The main variables governing the thickness of this lithosome are dune height and bedrock and glacial antecedent topography. Appendix 1 includes descriptions of cores SB95-PA2 and SB85±144-B1. GPR lines showing the variability of dune height and/or antecedent topography are Figs 6 and 14, and those published in van Heteren et al. (1996). Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 181±200

Radar facies of paraglacial barrier systems

195

Fig. 16. External form of the tidalinlet ®ll at Sandy Neck, showing the effects of inlet deepening and narrowing. Appendix 1 includes descriptions of cores SN94-PA47 and SN94-PA67. The GPR line of Fig. 7 shows a 1á5-m-thick inlet-channel unit. Appendix 2 outlines the GPR signature of a 6-m-deep migrating inlet channel. The seismic record of Appendix 3 contains evidence of the tidal-inlet ®ll offshore Sandy Neck.

contributed signi®cantly to most shallow marine and lacustrine facies models. Integration of seismic pro®les with other geophysical data, cores and grab samples has resulted in very useful three-dimensional facies models that commonly include a process-response component, linking sedimentation and erosion to resulting stratigraphy. The resulting multi-component data base is compatible with three-dimensional data from studies of ancient environments. In modern coastal and continental environments, high-resolution geophysical methods have not been used extensively. With few exceptions (e.g. Galloway, 1986; Tyler & Ambrose, 1986), current facies models of these depositional environments are still based largely on (near) pointsource data from cores and trenches. Generally, existing continuous data either do not have the resolution necessary to characterize the intricacies of these laterally variable environments (seismic records), or they do not have the lateral extent to allow delineation of architectural elements (sand and gravel pits). GPR data not only allow high-resolution, largescale stratigraphic analysis of sedimentary environments and facies, permitting development of threedimensional models of entire depositional systems, but they also provide a very useful link to seismic models of offshore marine and lacustrine systems. For example, GPR data from the Saco Bay coastal zone show good correspondence with high-resolution seismic data from Saco Bay itself (cf. Kelley et al., 1989). The glacial units that underlie the Holocene sediments in both areas produce comparable (but not identical) re¯ection con®gurations in the two data sets. Similar correspondence may exist between the re¯ection con®gurations of barrier

facies in these data sets. Recognition of barrier-core facies on seismic records has relied mostly on partly preserved morphologic features (Hovland & Dukefoss, 1981; Rokoengen et al., 1982; Oldale et al., 1993), but de®nition of the re¯ection con®guration of barrier facies on GPR sections may enable delineation of less conspicuous barrier elements on seismic records. For all geophysical methods, data interpretation requires a thorough knowledge of the geology of the area surveyed, core control being particularly important. Whereas core control is scattered in most marine and lacustrine studies, and cores are not accurately located relative to corresponding seismic information, on land each core can be matched accurately to its position on the complementary GPR record. Commonly, individual re¯ections on GPR sections can be tied with relative ease to corresponding facies changes observed in cores, because the re¯ections are interference composites of relatively small vertical ranges compared to those of seismic data (cf. Sheriff, 1977). In conclusion, integration of GPR data in coastal stratigraphic studies has proven to be advantageous in several aspects. Firstly, GPR is important in reconnaissance of large study areas because it allows rapid data collection. Based on initial GPR information, core sites can be planned where they will be most useful and informative. Secondly, GPR data are instrumental in correlating between cores, and in the delineation of facies. Three-dimensional coastal stratigraphic models that include a GPR component are very useful in the analysis of the ancient record, and are comparable to similar highresolution models for shallow-marine and deltaic environments. Thirdly, the directional aspect of

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 181±200

196

S. van Heteren et al.

GPR data can be used to analyse the dynamics of coastal processes and resulting barrier response, as preserved in the sedimentary record, through palaeocurrent analysis (Lawton et al., 1994; Beres et al., 1995). For example, it is frequently possible to distinguish beach sediments formed by progradation, aggradation and longshore accretion. Finally, the detailed and reliable correlations of GPR re¯ections and re¯ection con®gurations with lithologic boundaries and facies types in cores, trenches and quarries may help in the interpretation of high-resolution seismic records that lack adequate and position-sensitive core control, particularly where recognition of preserved coastal and continental facies is concerned. ACKNOWLEDGMENTS This work was funded by the M.I.T./Maine Sea Grant College Program (National Oceanic and Atmospheric Administration, U.S. Department of Commerce), the Netherlands Organization for Scienti®c Research (NWO), and the Massachusetts Coastal Zone Management Program. Martin Konert analysed the textural characteristics of the Sandy Neck cores. We bene®ted from discussions with Daniel F. Belknap, who also operated the seismic pro®ler, Joseph T. Kelley, and Orson van de Plassche. We thank journal reviewers Richard A. Davis, Harry M. Jol and editor A. Guy Plint, who suggested many signi®cant improvements to the original manuscript. REFERENCES Baker, P.L. (1991) Response of ground-penetrating radar to bounding surfaces and lithofacies variation in sand barrier sequences. Explor. Geophys., 22, 19±22. Belknap, D.F., Shipp, R.C., Kelley, J.T. and Schnitker, D. (1989) Depositional sequence modeling of the late Quaternary geologic history, west-central Maine coast. In: Studies in Maine Geology 5: Quaternary Geology (Ed. by R. D. Tucker and R. G. Marvinney), pp. 29±46. Maine Geological Survey, Department of Conservation, Augusta. Beres, M., Jr. and Haeni, F.P. (1991) Application of ground-penetrating-radar methods in hydrogeologic studies. Ground Water, 29, 375±386. Beres, M., Green, A., Huggenberger, P. and Horstmeyer, H. (1995) Mapping the architecture of glacio¯uvial sediments with three-dimensional georadar. Geology, 23, 1087±1090. Bokuniewicz, H. and Pavlik, B. (1990) Groundwater seepage along a barrier island. Biogeochemistry, 10, 257±276.

Bristow, C. (1995) Facies analysis in the Lower Greensand using ground-penetrating radar. J. geol. Soc., London, 152, 591±598. Carter, R.W.G. (1986) The morphodynamics of beachridge formation, Magilligan, Northern Ireland. Mar. Geol., 73, 191±214. Cook, J.C. (1975) Radar transparencies of mine and tunnel rocks. Geophysics, 40, 865±885. Daniels, D.J., Gunton, D.J. and Scott, H.F. (1988) Introduction to subsurface radar. IEE (Institution of Electrical Engineers) Proc., 135 (F), 278±320. Davis, J.L. and Annan, A.P. (1989) Ground-penetrating radar for high-resolution mapping of soil and rock stratigraphy. Geophys. Prosp., 37, 531±551. Davis, R.A., Jr., Fox, W.T., Hayes, M.O. and Boothroyd, J.C. (1972) Comparison of ridge and runnel systems in tidal and non-tidal environments. J. sedim. Petrol., 42, 413±421. Dickinson, K.A., Berryhill, H.L. and Holmes, C.W. (1972) Criteria for recognizing ancient barrier coastlines. In: Recognition of Ancient Sedimentary Environments (Ed. by J. K. Rigby and W. K. Hamblin), Spec. Pub. Soc. econ. Paleont. Miner., 16, 192±214. Dott, E.R. and Mickelson, D. (1995) Lake-Michigan water levels and the development of Holocene beachridge complexes at Two Rivers, Wisconsin: stratigraphic, geomorphic, and radiocarbon evidence. Bull. geol. Soc. Am., 107, 286±296. FitzGerald, D.M., Baldwin, C.T., Ibrahim, N.A. and Sands, D.R. (1987) Development of the northwestern Buzzards Bay shoreline, Massachusetts. In: Glaciated Coasts (Ed. by D. M. FitzGerald and P. S. Rosen), pp. 327±357. Academic Press, San Diego. FitzGerald, D.M., Baldwin, C.T., Ibrahim, N.A. and Humphries, S.M. (1992) Sedimentologic and morphologic evolution of a beach-ridge barrier along an indented coast: Buzzards Bay, Massachusetts. In: Quaternary coasts of the United States: marine and lacustrine systems (Ed. by C. H. Fletcher and J. F. Wehmiller), Spec. Pub. Soc. econ. Paleont. Miner., 48, 64±75. FitzGerald, D.M., Rosen, P.S. and van Heteren, S. (1994) New England Barriers. In: Geology of Holocene Barrier Island Systems (Ed. by R. A. Davis, Jr.), pp. 305±394. Springer-Verlag, Berlin. Forbes, D.L. and Syvitski, J.P.M. (1994) Paraglacial coasts. In: Coastal Evolution ± Late Quaternary shoreline morphodynamics (Ed. by R. W. G. Carter and C. D. Woodroffe), pp. 373±424. University Press, Cambridge. Galloway, W.E. (1986) Reservoir facies architecture of microtidal barrier systems. Bull. Am. Ass. petrol. Geol., 70, 787±808. Gawthorpe, R.L., Collier, R.E.Ll., Alexander, J., Bridge, J.S. and Leeder, M.R. (1993) Ground penetrating radar: application to sandbody geometry and heterogeneity studies. In: Characterization of Fluvial and Aeolian Reservoirs (Ed. by C. P. North and D. J. Prosser). Spec. Publ. geol. Soc. Lond., 73, 421±432. GrasmuÈck, M.P. (1992) Beispiele zur Anwendung von Georadar in der QuartaÈrgeologie. Eclogae Geol. Helv., 85, 471±490.

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 181±200

Radar facies of paraglacial barrier systems HaÈnninen, P. and Autio, S (eds.) (1992) Proc., Fourth Internat. Conf. Ground Penetrating Radar. Spec. Pap. Geol. Surv. Finland, 16. Kuopio, Finland. Hayes, M.O. (1980) General morphology and sediment patterns in tidal inlets. Sedim. Geol., 26, 139±156. HeÂquette, A. and Ruz, M.-H. (1991) Spit and barrier island migration in the southeastern Canadian Beaufort Sea. J. Coast. Res., 7, 677±698. Heron, S.D., Jr., Moslow, T.F., Berelson, W.M., Herbert, J.R., Steele III, G.A. and Susman, K.R. (1984) Holocene sedimentation of a wave-dominated barrier-island shoreline: Cape Lookout, North Carolina. In: Hydrodynamics and Sedimentation in WaveDominated Coastal Environments (Ed. by B. Greenwood and R. A. Davis, Jr.), Mar. Geol., 60, 413±434. Hine, A.C. (1979) Mechanisms of berm development and resulting beach growth along a barrier spit complex. Sedimentology, 26, 333±351. Hovland, M. and Dukefoss, K.M. (1981) A submerged beach between Norway and Eko®sk in the North Sea. Mar. Geol., 43, M19-M28. Inman, D.L. and Nordstrom, C.E. (1971) On the tectonic and morphologic classi®cation of coasts. J. Geol., 79, 1±21. Jol, H.M. (1995) Ground penetrating radar antennae frequencies and transmitter powers compared for penetration depth, resolution and re¯ection continuity. Geophys. Prosp., 43, 693±709. Jol, H.M. and Smith, D.G. (1991) Ground penetrating radar of northern lacustrine deltas. Can. J. Earth Sci., 28, 1939±1947. Jol, H.M., Smith, D.G. and Meyers, R. (1994) Ground penetrating radar of lakeshore spits in northwestern Saskatchewan, Canada: variable internal structure. In: GPR `94 -Proc., Fifth International Conference on Ground Penetrating Radar (Ed. by J. D. Redman, A. P. Annan, J. P. Greenhouse, T. Klym and J. R. Rossiter), pp. 817±830. Waterloo Centre for Groundwater Research, Kitchener. Jol, H.M., Smith, D.G. and Meyers, R. (1996) Digital ground penetrating radar (GPR): A new geophysical tool for coastal barrier research (examples from the Atlantic, Gulf, and Paci®c coasts, U. S. A.). J. Coast. Res., 12, 960±968. Kelley, J.T., Shipp, R.C. and Belknap, D.F. (1989) Geomorphology and Late Quaternary evolution of the Saco Bay region. In: Studies in Maine Geology 5: Quaternary Geology (Ed. by R. D. Tucker and R. G. Marvinney), pp. 47±65. Maine Geological Survey, Department of Conservation, Augusta. Lawton, D.C., Jol, H.M. and Meyers, R.A. (1994) 3-D ground penetrating radar surveys: a new approach to imaging and visualising the near surface. In: GPR `94 ± Abstracts for Poster Papers, Fifth International Conference on Ground Penetrating Radar (Ed. by J. D. Redman, A. P. Annan, J. P. Greenhouse, T. Klym and J. R. Rossiter), p. 16. Waterloo Centre for Groundwater Research, Kitchener. Leatherman, S.P. (1987) Coastal geomorphological applications of ground-penetrating radar. J. Coast. Res., 3, 397±399.

197

McCann, D.M., Jackson, P.D. and Fenning, P.J. (1988) Comparison of the seismic and ground probing radar methods in geological surveying. IEE (Institution of Electrical Engineers) Proc., 135 (F), 380±390. Meyers, R.A., Smith, D.G., Jol, H.M. and Peterson, C.D. (1996) Evidence for eight great earthquake-subsidence events detected with ground-penetrating radar, Willapa barrier, Washington. Geology, 24, 99± 102. Miall, A.D. (1991) Hierarchies of architectural units in terrigenous clastic rocks, and their relationship to sedimentation rate. In: The Three-Dimensional Facies Architecture of Terrigenous Clastic Sediments and Its Implications for Hydrocarbon Discovery and Recovery (Ed. by A. D. Miall and N. Tyler), Concepts Sedim. Paleont., 3, 6±12. Mitchum, R.M., Jr., Vail, P.R. and Sangree, J.B. (1977) Seismic stratigraphy and global changes of sea level, Part 6: stratigraphic interpretation of seismic re¯ection patterns in depositional sequences. In: Seismic Stratigraphy ± applications to hydrocarbon exploration (Ed. by C. E. Payton), Mem. Am. Ass. petrol. Geol., 26, 117±133. Murakoshi, N. and Masuda, F. (1991) A depositional model for a ¯ood-tidal delta and washover sands in the late Pleistocene Paleo-Tokyo Bay, Japan. In: Clastic Tidal Sedimentology (Ed. by D. G. Smith, G. E. Reinson, B. A. Zaitlin and R. A. Rahmani), Mem. Can. Soc. petrol. Geol., 16, 219±226. Nichol, S.L. and Boyd, R. (1993) Morphostratigraphy and facies architecture of sandy barriers along the Eastern Shore of Nova Scotia. Mar. Geol., 114, 59±80. Nielsen, L.H., Johannessen, P.N. and Surlyk, F. (1988) A Late Pleistocene coarse-grained spit-platform sequence in northern Jylland, Denmark. Sedimentology, 35, 915±937. Oldale, R.N., Colman, S.M. and Jones, G.A. (1993) Radiocarbon ages from two submerged strandline features in the western Gulf of Maine and a sea-level curve for the northeastern Massachusetts coastal region. Quat. Res., 40, 38±45. Ollerhead, J. and Davidson-Arnott, R.G.D. (1995) The evolution of Buctouche Spit, New Brunswick, Canada. In: Coastal Evolution in the Quaternary: IGCP Project 274 (Ed. by O. van de Plassche, M. J. Chrzastowski, J. D. Orford, A. C. Hinton and A. J. Long), Mar. Geol., 124, 215±236. Olsen, H. and Andreasen, F. (1995) Sedimentology and ground-penetrating radar characteristics of a Pleistocene sandur deposit. Sedim. Geol., 99, 1±15. Redman, J.D., Annan, A.P., Greenhouse, J.P., Klym, T. and Rossiter, J.R (eds.) (1994) GPR `94 ± Proceedings of the Fifth International Conference on Ground Penetrating Radar. Waterloo Centre for Groundwater Research, Kitchener. Rokoengen, K., Lùfaldli, M., Rise, L., Lùken, T. and Carlsen, R. (1982) Description and dating of a submerged beach in the northern North Sea. Mar. Geol., 50, M21-M28. Roy, P.S., Cowell, P.J., Ferland, M.A. and Thom, B.G. (1994) Wave-dominated coasts. In: Coastal Evolution

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 181±200

198

S. van Heteren et al.

± Late Quaternary shoreline morphodynamics (Ed. by R. W. G. Carter and C. D. Woodroffe), pp. 121±186. University Press, Cambridge. Schenk, C.J., Gautier, D.L., Olhoeft, G.R. and Lucius, J.E. (1993) Internal structure of an aeolian dune using ground-penetrating radar. In: Aeolian sediments, ancient and modern (Ed. by K. Pye and N. Lancaster), Spec. Publ. Int. Ass. Sediment., 16, 61±69. Sheriff, R.E. (1977) Limitations on resolution of seismic re¯ections and geologic detail derivable from them. In: Seismic Stratigraphy ± applications to hydrocarbon exploration (Ed. by C. E. Payton), Mem. Am. Ass. petrol. Geol., 26, 3±14. Smith, D.G. and Jol, H.M. (1992) GPR results used to infer depositional processes of coastal spits in large lakes. In: Proc., Fourth Internat. Conf. Ground Penetrating Radar (Ed. by P. HaÈnninen and S. Autio), Spec. Pap. Geol. Surv. Finland, 16, 169±177. Stephens, M. (1994) Architectural element analysis within the Kayenta Formation (Lower Jurassic) using ground-probing radar and sedimentological pro®ling, southwestern Colorado. Sedim. Geol., 90, 179±211. Thom, B.G. and Roy, P.S. (1985) Relative sea levels and coastal sedimentation in southeast Australia in the Holocene. J. sedim. Petrol., 55, 257±264. Topp, G.C., Davis, J.L. and Annan, A.P. (1980) Electromagnetic determination of soil water content: measurements in coaxial transmission lines. Water Resources Res., 16, 574±582. Tyler, N. and Ambrose, W.A. (1986) Facies architecture and production characteristics of strand-plain reservoirs in North Markham ± North Bay City ®eld, Frio Formation, Texas. Bull. Am. Ass. petrol. Geol., 70, 809±829. van Heteren, S. (1996) Preserved records of coastalmorphologic and sea-level changes in the stratigra-

phy of paraglacial barriers. PhD Dissertation, Boston University. van Heteren, S. and van de Plassche, O. (1997) In¯uence of relative sea-level change and tidal-inlet development on barrier-spit stratigraphy, Sandy Neck, Massachusetts. J. sedim. Res., 67, 350±363. van Heteren, S., FitzGerald, D.M. and McKinlay, P.A. (1994) Application of ground-penetrating radar in coastal stratigraphic studies. In: GPR `94 ± Proc., Fifth International Conference on Ground Penetrating Radar (Ed. by J. D. Redman, A. P. Annan, J. P. Greenhouse, T. Klym and J. R. Rossiter), pp. 869±881. Waterloo Centre for Groundwater Research, Kitchener. van Heteren, S., FitzGerald, D.M., Barber, D.C., Kelley, J.T. and Belknap, D.F. (1996) Volumetric analysis of a New England barrier system using ground-penetrating-radar and coring techniques. J. Geol., 104, 471± 483. van Overmeeren, R.A. (1994) High speed georadar data acquisition for groundwater exploration in The Netherlands. In: GPR '94 ± Proc., Fifth International Conference on Ground Penetrating Radar (Ed. by J. D. Redman, A. P. Annan, J. P. Greenhouse, T. Klym and J. R. Rossiter), pp. 1057±1073. Waterloo Centre for Groundwater Research, Kitchener. von Hippel, A.R (ed.) (1954) Dielectric Materials and Applications. John Wiley and Sons, New York. Widess, M.B. (1973) How thin is a thin bed? Geophysics, 38, 1176±1180.

Manuscript received 7 February 1996; revision accepted 27 May 1997.

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 181±200

Radar facies of paraglacial barrier systems

Appendix 1. Logs of cores mentioned in text and ®gure captions.

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 181±200

199

200

S. van Heteren et al.

Appendix 2. GPR section from Sandy Neck, showing the signature produced by the eastward-migrating inlet channel (arrows). Core SN94-PA66 indicates that channel facies are present between 2á5 m and 8á5 m below NGVD±29.

Appendix 3. Raytheon 3á5-kHz seismic pro®le from the Sandy Neck inlet channel and ebb-tidal-delta edge, showing offshore inlet-channel facies that extend to a depth of 6 m below NGVD±29. The base of the inlet-channel unit is continuous with the base of the present-day inlet channel.

Ó 1998 International Association of Sedimentologists, Sedimentology, 45, 181±200