Abstract The spatial distribution of polar lakes is governed by the lithology and structure of the geological units, whereas their shape and size is modified by later fluctuations of the polar ice sheet and glacier movements. Polar lakes are the downwards integrator of the sediment action in their catchment areas. Bathymetric profiles of some proglacial, land-locked and epi-shelf lakes spread across the Schirmacher Oasis were obtained during the austral winter period of 2008 using ground-penetrating radar (GPR) to understand their evolution through space and time. Morphometric characteristics of these lakes show different depth ratios and they have different surface areas and volume proportions with depth. This study shows that the largest land-locked lake, L-49, appears to have been formed by the fusion of three sub-basins. Some of the proglacial lakes which are becoming detached from the polar ice sheet, such as lake P-11, are kettle lakes. Water from the polar ice-sheet melt has been accumulating at these lake basins, but no further aggradation processes have affected them. Some of the larger lakes such as lakes L-27, L-49 and P-9 are reworked lakes which evolved during a phase of glacial advancement and were later modified during another advancement phase. The other larger lakes such as lake E-14, E-15 and L-75 are very deep and were produced by the glacial activity associated with some weak structural fabric.

Abstract Rare jökulhlaup events, also known as subglacial lake outburst flood events, have been observed at the Law Dome ice margin and provide an insight into the physical characteristics of subglacial meltwater and drainage. The subglacial topography based on data from the BEDMAP2 and ICECAP projects, together with subsurface transects of the ice margin obtained using ground-penetrating radar, reveal several lakes and lake-like depressions and the drainage pathways of two jökulhlaup events. Oxygen isotope typing of the meltwater during the most recent (2014) jökulhlaup event, combined with ice margin stratigraphy, enable the identification of ice tunnel melt pathways that exploit the 30–90° dipping basal ice layering. The presence of subglacial meltwater beneath Law Dome during the Holocene to Glacial periods is confirmed by the dendritic drainage pattern in the subglacial morphology and extensive layers of basal regelation ice and subglacial carbonate precipitate deposits found within the Løken Moraines sediments. These subglacial carbonates, including ooid layers, formed from the mixing of glacial meltwater and seawater at 72 ka BP. The combined evidence indicates that the ocean discharge of subglacial meltwater may be variable and/or is periodically blocked by basal freezing events near the ice sheet terminus.

Abstract Long-range airborne geophysical measurements were carried out in the ICEGRAV campaigns, covering hitherto unexplored parts of interior East Antarctica and part of the Antarctic Peninsula. The airborne surveys provided a regional coverage of gravity, magnetic and ice-penetrating radar measurements for major Dronning Maud Land ice stream systems, from the grounding lines up to the Recovery Lakes drainage basin, and filled in major data voids in Antarctic data compilations, such as AntGP for gravity data, ADMAP for magnetic data and BEDMAP2 for ice thickness data and the sub-ice topography. We present the first maps of gravity, magnetic and ice thickness data and bedrock topography for the region and show examples of bedrock topography and basal reflectivity patterns. The 2013 Recovery Lakes campaign was carried out with a British Antarctic Survey Twin Otter aircraft operating from the Halley and Belgrano II stations, as well as a remote field camp located at the Recovery subglacial Lake B site. Gravity measurements were the primary driver for the survey, with two airborne gravimeters (Lacoste and Romberg and Chekan-AM) providing measurements at an accuracy level of around 2 mGal r.m.s., supplementing GOCE (Gravity Field and Steady-State Ocean Circulation Explorer) satellite data and confirming an excellent sub-milligal agreement between satellite and airborne data at longer wavelengths.

Abstract We examine three questions concerning the post-glacial geological history of the eastern Gulf of Finland: (1) the amplitude of the Holocene sea-level regressions; (2) the time and mechanism of the development of large sand accretion forms (bars and spits), including dunes; and (3) the sea-level changes and coastal development over the last 4 kyr. Recent on-land geoarchaeological studies, as well as detailed marine geological research of the Gulf of Finland nearshore bottom, have provided new data for developing a hypothesis about the palaeogeographical development of the area. Geoarchaeological studies carried out around Sestroretsky Artificial Lake and within Okhta Cape, as well as analyses of previous studies of the Neolithic–Early Metal settlements, have shed new light on some aspects of coastal system development. Geographical information system (GIS)-based modelling of Holocene shorelines for the different time periods can be useful for future archaeological research. A series of submarine terraces was found at the bottom of the Gulf (sea depths from 10 to 2 m). The analysis of marine geological data (submarine terraces) and distribution of archaeological sites can be explained by a possible rise in relative sea level in the Gulf of Finland at 5 ka BP and a regression around 3 ka BP.

Abstract Determining how autogenic and allogenic processes and responses in deltas scale up from meter-scale laboratory experiments to actual field examples remains a challenge. This study was devised to bridge that scale gap using field data from small, hundreds of meter-scale natural deltas. Ground-penetrating radar and core data were collected from four different river-dominated delta morphotypes developing at the margins of freshwater coastal lagoons in southern Brazil. Since the sediment supplying these deltas is sourced from a nearby dune field and is similar between the deltas, it is hypothesized that major morphological differences in the four deltas are primarily the result of differences in sediment discharge rates (sediment-water ratio). As observed in published tank experiments, channel cross-section and distributary channel patterns in the deltas vary as a function of sediment discharge, from shallow sheet-like flow at high discharge to well-established, stable distributary channels (i.e., birdsfoot pattern). A contributing factor may be the development of vegetation on the slower growing deltas influencing sediment cohesion, a key control in laboratory-scale deltas. As in many tank experiments, these lagoon deltas are steep and sandy, with the Froude number modulated to just below Froude critical flow (i.e., they are Froude-scaled). Ground-penetrating radar sections were processed, interpreted, and integrated with cores, allowing the definition of radar units. Analysis of the radar units demonstrates the presence of both allogenic and autogenic signals. Allogenic control is identified in the stacking of clinoforms and is perceptible in both sides of a single delta (delta 4), as well as in two other deltas (deltas 1 and 2). An autogenic signal varies according to delta planform shape and was identified by the stacking of lobe elements, both in dip and strike. Base-level change (lake level) and autogenic avulsion cycles occur on similar timescales, and therefore it is a significant challenge to separate these different processes in the stratigraphy. The potential uses of these types of data include understanding the link between delta dynamics, channel patterns, and stratigraphy to develop improved genetic models of steep sandy deltas common in the stratigraphic record.

It has long been hypothesized that flux of fresh meltwater from glacial Lake Minong in North America's Superior Basin to the North Atlantic Ocean triggered rapid climatic shifts during the early Holocene. The spatial context of recent support for this idea demands a reevaluation of the exit point of meltwater from the Superior Basin. We used ground penetrating radar (GPR), foundation borings from six highway bridges, a GIS model of surface topography, geologic maps, U.S. Department of Agriculture–Natural Resources Conservation Service soils maps, and well logs to investigate the possible linkage of Lake Minong with Lake Chippewa in the Lake Michigan Basin across eastern Upper Michigan. GPR suggests that a connecting channel lies buried beneath the present interlake divide at Danaher. A single optical age hints that the channel aggraded to 225 m as elevated receipt of Lake Agassiz meltwater in the Superior Basin began to wane <10.6 ka. The large supply of sediment required to accommodate aggradation was immediately available at the channel's edge in the littoral shelves of abandoned Lake Algonquin and in distal parts of post-Algonquin fans. As discharge decreased further, the aggraded channel floor was quickly breached and interbasin flow to Lake Chippewa was restored. Basal radiocarbon ages on wood from small lakes along the discharge path and a GIS model of Minong's shoreline are consistent with another transgression of Minong after ca. 9.5 ka. At the peak of the latter transgression, the southeastern rim of the Superior Basin (Nadoway Drift Barrier) failed, ending Lake Minong. Upon Minong's final drop, aggradational sediments were deposited at Danaher, infilling the prior breach.

The Nipissing phase of ancestral Lakes Michigan, Huron, and Superior was the last pre-modern highstand of the upper Great Lakes. Reconstructions of past lake-level change and glacial isostatic adjustment (GIA), as well as activation and abandonment of outlets, is dependent on an understanding of the elevation of the lake at each outlet. More than 100 years of study has established the gross elevation of the Nipissing phase at each outlet, but the mixing of geomorphic and sedimentologic data has produced interpreted outlet elevations varying by at least several meters. Vibracore facies, optically stimulated luminescence and radiocarbon age control, and ground-penetrating radar transects from new and published studies were collected to determine peak Nipissing water-level elevations for the Port Huron (Lake Huron), Chicago (Lake Michigan), and Sault (Lake Superior) outlets. Contemporary elevations are 183.3, 182.1, and 195.7 m (International Great Lakes Datum of 1985 [IGLD85]), respectively. These data and published relative hydrographs were combined to produce one residual hydrograph for the Port Huron outlet that best defines the rise, peak, and rapid fall of the Nipissing phase from 6000–3500 calendar years ago. Establishing accurate elevations at the only present-day unregulated outlet of the Great Lakes and the only ancient outlet that has played a critical role in draining the upper Great Lakes since the middle Holocene is a critical step to better understand GIA and water-level change geologically and historically. The geologic context may provide the insight required for water managers to make informed decisions to best manage the largest freshwater system in the world.

Sand Point is a small cuspate foreland located along the southeastern shore of Lake Superior within Pictured Rocks National Lakeshore near Munising, Michigan. Park managers' concerns for the integrity of historic buildings at the northern periphery of the point during the rising lake levels in the mid-1980s greatly elevated the priority of research into the geomorphic history and age of Sand Point. To pursue this priority, we recovered sediment cores from four ponds on Sand Point, assessed subsurface stratigraphy onshore and offshore using geophysical techniques, and interpreted the chronology of events using radiocarbon and luminescence dating. Sand Point formed at the southwest edge of a subaqueous platform whose base is probably constructed of glacial diamicton and outwash. During the post-glacial Nipissing Transgression, the base was mantled with sand derived from erosion of adjacent sandstone cliffs. An aerial photograph time sequence, 1939–present, shows that the periphery of the platform has evolved considerably during historical time, influenced by transport of sediment into adjacent South Bay. Shallow seismic reflections suggest slump blocks along the leading edge of the platform. Light detection and ranging (LiDAR) and shallow seismic reflections to the northwest of the platform reveal large sand waves within a deep (12 m) channel produced by currents flowing episodically to the northeast into Lake Superior. Ground-penetrating radar profiles show transport and deposition of sand across the upper surface of the platform. Basal radiocarbon dates from ponds between subaerial beach ridges range in age from 540 to 910 cal yr B.P., suggesting that Sand Point became emergent during the last ~1000 years, upon the separation of Lake Superior from Lakes Huron and Michigan. However, optically stimulated luminescence (OSL) ages from the beach ridges were two to three times as old as the radiocarbon ages, implying that emergence of Sand Point may have begun earlier, ~2000 years ago. The age discrepancy appears to be the result of incomplete bleaching of the quartz grains and an exceptionally low paleodose rate for the OSL samples. Given the available data, the younger ages from the radiocarbon analyses are preferred, but further work is necessary to test the two age models.

Interrelationships among late Holocene climate, the dynamics of coastal dunes and sedimentation in adjacent small lakes along coasts of the upper Great Lakes have been studied for over a decade. Nonetheless, many questions remain as to relationships between climate variability and dune activity. The study site is Oxbow Lake, near Saugatuck, Michigan, which formed as an artificial cutoff of the Kalamazoo River in 1906. Stratigraphic control of the infilled western end of the lake is from ground penetrating radar, and lake sediment from Livingstone and Glew cores with age control from 210 Pb/ 137 Cs/ 7 Be analysis. The climate data used included Lake Michigan water levels and temperature, precipitation, drought and evaporation data from a weather station 30 km to the south and wind data from buoys on Lake Michigan. Episodic peaks of eolian sand in the lake sediment are interpreted to be sourced from adjacent small parabolic dunes along the shoreline and from a foredune west of the lake. Linear regressions of the climate data and weight percent sand resulted in a variety of correlations, some conflicting, and with uncertain meanings. It was found through visual correlation that peaks in sand correspond with both peaks in water levels of Lake Michigan and the winter Palmer drought severity index. The implications of this research are that dune activity is linked to periods of wet conditions and storminess, contrary to typical eolian environments, but consistent with other studies in temperate coastal dunes along the Great Lakes. Results can be used as a modern analogue for coastal dune activity during times of high lake level.

Abstract The karst of the central Shenandoah Valley has characteristics of both shallow and deep phreatic formation. This field guide focuses on the region around Harrisonburg, Virginia, where a number of these karst features and their associated geologic context can be examined. Ancient, widespread alluvial deposits cover much of the carbonate bedrock on the western side of the valley, where shallow karstification has resulted in classical fluviokarst development. However, in upland exposures of carbonate rock, isolated caves exist atop hills not affected by surface processes other than exposure during denudation. The upland caves contain phreatic deposits of calcite and fine-grained sediments. They lack any evidence of having been invaded by surface streams. Recent geologic mapping and LIDAR (light detection and ranging) elevation data have enabled interpretive association between bedrock structure, igneous intrusions, silicification and brecciation of host carbonate bedrock, and the location of several caves and karst springs. Geochemistry, water quality, and water temperature data support the broad categorization of springs into those affected primarily by shallow near-surface recharge, and those sourced deeper in the karst aquifer. The deep-seated karst formation occurred in the distant past where subvertical fracture and fault zones intersect thrust faults and/or cross-strike faults, enabling upwelling of deep-circulating meteoric groundwater. Most caves formed in such settings have been overprinted by later circulation of shallow groundwater, thus removing evidence of the history of earliest inception; however, several caves do preserve evidence of an earlier formation.

Abstract A 3D ground penetrating radar (GPR) survey, using three different frequency antennae, was undertaken to image buried steel culverts at the University of Houston’s La Marque Geophysical Observatory 30 miles south of Houston, Texas. The four culverts, under study, support a road crossing one of the area’s bayous. A 32 m by 4.5 m survey grid was designed on the road above the culverts and data were collected with 100 MHz, 250 MHz, and 1 GHz antennae. We used an orthogonal acquisition geometry for the three surveys. Inline sampling was from 1.0 cm to 10 cm (from 1 GHz to 100 MHz antenna) with inline and crossline spacings ranging from 0.2 m to 0.5 m. We used an initial velocity of 0.1 m/ns (from previous CMP work at the site) for the display purposes. The main objective of the study was to analyze the effect of different frequency antennae on the resultant GPR images. We are also interested in the accuracy and resolution of the various images, in addition to developing an optimal processing flow. The data were initially processed with standard steps that included gain enhancement, dewow and temporal-filtering, background suppression, and 2D migration. Various radar velocities were tried in the 2D migration and ultimately 0.12 m/ns was used. The data are complicated by multipathing from the surface and between culverts (from modeling). Some of this is ameliorated via deconvolution. The top of each of the four culverts was evident in the GPR images acquired with the 250 MHz and 100 MHz antennas. For 1 GHz, the top of the culvert was not clear due to the signal’s attenuation. The 250 MHz shielded antenna provides a vertical resolution of about 0.1 m and is the choice to image the culverts. The 100 MHz antenna provided an increment in depth of penetration, but at the expense of a substantially diminished resolution (0.25 m).

Abstract This field guide focuses on glacial history, dynamics and processes, and postglacial landscape adjustments in the southern Fraser Plateau region. Located between the Coast and Columbia Mountains in south-central British Columbia, Canada, the southern Fraser Plateau was near the geographic center of the last (marine oxygen isotope stage [MIS] 2) Cordilleran Ice Sheet (CIS). The transition from cold to warm-based ice during MIS 2 is recorded in till sedimentology and structural geology. The perceived absence of large deglacial recessional moraines has been used as evidence that ice regionally stagnated because of a rapid rise in equilibrium line altitude. However, glacioisostatic rebound orientations, ice-marginal channel and grounding-line and push moraine distributions, and reconstructions of late-glacial ice-marginal lake evolution suggest a systematic northwestward pattern of active ice-margin retreat toward the Coast Mountains, accompanied by regional thinning. Eskers and erosional corridors record drainage of supraglacial lakes or ice-marginal water sources in or over thin ice. Many ice-dammed lakes drained catastrophically. Following lake drainage, streams incised valley fills, leaving behind terraces capped by paraglacial fans and eolian sediment. In sum, we examine (1) valley-fill sediments that record Quaternary history dating back to the early or mid-Pleistocene; (2) till, moraines, erosional corridors, and eskers that provide evidence for MIS 2 CIS dynamics and hydrology; (3) late-glacial ice-marginal lake sediments and landforms that allow reconstruction of lake evolution and drainage, and changing ice-margin positions; and (4) the character and ages of river terraces, paraglacial fans, and eolian sediments that record the timing and nature of postglacial landscape adjustments.

Abstract We investigated the viability of ground-penetrating radar (GPR) as a method to estimate the quantity of sediment stored behind the Merrimack Village Dam on the Souhegan River in southern New Hampshire. If the predam riverbed can be imaged, the thickness and volume of the reservoir deposit can be calculated without sampling. Such estimates are necessary to plan sediment management after dam deconstruction. In May 2008, we surveyed six cross sections with a Mala Geosciences ProEx 100 MHz GPR. In a related study, topographic surveys were conducted in 2008–2009 to monitor the sediment flux associated with the removal of the Merrimack Village Dam in August 2008. Within a month of the removal, these surveys mapped the predam riverbed in the uppermost cross sections in the former impoundment. We compared these surveys to our interpreted GPR images for one cross section to determine a calibrated velocity for the impounded sand of 0.043 ± 0.020 m/ns. We also estimated the radar velocity of the deposit by analyzing hyperbolic reflections in the GPR images, and found a similar result (0.039 m/ns). Using the calibrated velocity, we estimated a total volume of sediment stored behind the Merrimack Village Dam of 66,900 ± 9900 m 3 , which compares well to a previous estimate (62,000 m3) based on a depth-to-refusal survey. Our findings indicate that GPR is a useful technique for quantifying impounded sediment prior to dam removal in reservoirs containing 1–10 m of sand overlying a coarser predam riverbed, but it may be less effective in settings with finer and/or thicker impounded sediment.

This study investigates environmental change as a context for the location of occupation in campsites along the lower Nottawasaga River in southern Ontario, Canada. Evidence of Archaic and Woodland period campsites presents a difficult scenario for archaeological interpretation. Did Archaic cultures prefer to occupy higher ground, such as on the Edenvale moraine (206 m above sea level [a.s.l.]), or lower areas within river valleys, such as on the floodplain (181 m a.s.l.)? The aim of this study is to disentangle the evidence from a geoarchaeological perspective, considering cultural sites cross-temporally in their physical setting of a river floodplain in the Great Lakes region. We used geoarchaeological methods and methodologies, including soil-sediment analysis, sedimentological profiling of cutbanks and cores, topographic profiling in the field, river morphometric and hydrometric measurements using various techniques, ground-penetrating radar (GPR), and conceptual modeling. Results indicate that the Edenvale moraine, which topographically confines the channel and floodplain in the study area, may have led to catastrophic stripping of the floodplain during drainage of glacial lakes, causing the Nottawasaga River to downcut into the Edenvale moraine. This occurred following a period of floodplain stability, when natural and cultural deposits would have been preserved and buried in confined sections of the floodplain that are currently being vertically accreted. Any model that considers accretion style, such as the preservation-exposure model presented here, should be interpreted within the context of the environmental history of the area, since accretion style varies temporally and spatially at the river-reach scale.

The Bolonia Bay, close to the Strait of Gibraltar, hosts the Roman ruins of Baelo Claudia. In the first and third century A.D., this ancient town was affected by two earthquakes. Several earthquake-related damages can be found inside the ruins, and the adjacent mountain ranges show features of Quaternary activity. Extensive paleoseismological and archaeoseismological investigations have been conducted at the archaeological site and in its environs. The first 14 C dating results from damaged infrastructure are presented in this paper, together with the preliminary results of fault-trenching on one of the closest suspect seismogenic faults near the archaeological site. The observations have been quantified using the two logic trees for paleoseismology and archaeoseismology. Our results show that a mere paleoseismological classification of the geological features leads to a paleoseismic quality factor (PQF) of 0.03, which is low compared to other studies. Taking into account the additional information from archaeoseismological work (archaeoseismological quality factor [AQF] is 0.5), it becomes clear that the Baelo Claudia study site provides an opportunity for detailed earthquake investigations. Therefore, it has a high potential for reliable seismic hazard analyses. A complementary application of both logic trees is recommended in future studies if sufficient data are available.

A ground-penetrating radar (GPR) survey was carried out in order to characterize reflection patterns at Paluküla Hill, at the NE portion of the rim of the 4-km-diameter, Upper Ordovician Kärdla impact crater. The results allow us to distinguish between the Quaternary overburden and layered postimpact marine sedimentary rocks that cover the crystalline rim. The bedrock is indicated by reflections parallel to bedding that are tilted inward and outward with respect to the crater along the inner and outer rim slope, respectively. It is obvious that the uppermost part of the Paluküla crystalline rim was eroded and leveled in the Late Ordovician prior to having been covered by subhorizontally layered sediments. The unexpected position and low height of the crystalline rim at the northernmost edge of Paluküla Hill indicate that the rim of the Kärdla structure has collapsed to different degrees. This is consistent with the different heights of the crystalline rim. Our results favor the interpretation that the observed geology is not due to erosion by sea resurging into the crater but is rather a result of collapse of the crater rim.

Georgia Southern University maintains a traditional geology curriculum for both bachelor of science (B.S.) and bachelor of arts (B.A.) degree candidates. Field experiences figure prominently in our curricula, and students have been taught to use traditional means of gathering and recording field data (e.g., Brunton compasses and notebooks with sketches). We have recently introduced high-resolution geophysical investigations that are focused particularly on ground-penetrating radar. A nearby field location, known as Middleground, offers an excellent road cut with sufficient exposure, lithological heterogeneity, and relief to conduct both geological and geophysical investigations. We have shown students how one technique contrasts with the other, and how they can be used to support each other. Student reactions to the Middleground ground-penetrating radar exercise have been positive and enthusiastic, and have led us to formulate new and diverse applications of ground-penetrating radar to assist students in developing their three-dimensional visualization skills and a greater understanding of geophysical techniques in field investigations.

Field, map, and aerial photoreconnaissance in the Lake Alvord basin has focused on identifying late Pleistocene depositional shoreline features (e.g., tombolos, spits, barriers). Features in different areas of the basin are well defined, and their spatial extents are easily mapped; however, absolute—or even relative—ages of shoreline features are not clear. Ground penetrating radar (GPR) was used to distinguish between intermediate and highstand stage shorelines during what is thought to have been the latest Pleistocene, threshold-controlled lake cycle. Radar transects of 280 and 600 m imaged a spit and a baymouth barrier at sites in the northeastern quadrant of the basin where transects were aligned normal to the strike of each depositional geomorphic feature. Signal penetration with 100 MHz antennas was shallow (∼4 m), but resolution was sufficient to locate and identify gross morphostratigraphic features. Flooding surfaces are shown to correspond to intermediate stage lake surface elevations, and the absence of a flooding surface at the elevation of the highest shoreline indicates this to be the maximum lake surface elevation during this cycle. Elevations of intermediate lake stage elevations and highstand stage elevations were consistent at the two sites, with the highstand elevations corresponding closely to the basin threshold at Big Sand Gap. These data provide a first-order approximation of lake stage sequence and the degree of postdepositional neotectonic activity and illustrate the utility of GPR when used in context with field measurements in distinguishing transgressive and highstand features.

The internal architecture of sandbars in modern braided streams has not been adequately documented, especially in medium-scale braided rivers. Identification of the architecture and development of an understanding of the formative processes for these macroforms is important for (1) understanding sedimentation in braided streams, (2) understanding reservoir and aquifer compartmentalization in ancient deposits, and (3) predicting the controls on deposition in similar settings. A 225 MHz GPR survey was conducted within a braided reach of the Wisconsin River near Spring Green, Wisconsin, USA, to characterize the subsurface architecture of a midchannel bar. A 20 × 20 m survey grid consisting of sixteen GPR transects oriented approximately in flow-parallel and flow-transverse directions was established on the bar. Three-dimensional analysis of the GPR profiles resulted in the interpretation of five major radar facies that represent depositional mechanisms that controlled bar growth and modification. Vertical accretion (aggradation) was the primary depositional mechanism for bar growth and was augmented by much smaller amounts of downstream accretion, lateral accretion, and upstream accretion. A channel fill pattern was also recognized and correlated between multiple profiles, and it provided evidence for two preexisting, independent macroforms that converged to form the studied bar. The work provides insight into bar morphology within sandy braided reaches that closely resembles that of similar GPR studies performed in both smaller and larger rivers and supports a scale-independent model for some aspects of bar growth and modification in sandy, braided rivers.

A ground penetrating radar (GPR) survey was used to image the fluvial architecture within a buried paleovalley at Maple Creek, Guyana. The survey was part of a larger study of the stratigraphy and organization of fluvial elements within the valley fill. The survey consisted of 44 km of 50 and 100 MHz GPR profiles collected in a grid pattern. The 100 MHz antennae were used where depth to bedrock was less than 20 m, and the 50 MHz antennae were used where depth to bedrock exceeded 20 m. The survey grid consisted of 28 east-west–trending transects and 6 north-south–trending transects. East-west transects in the southern part of the study area were spaced 100 m apart. Those in the northern part of the study area were spaced 400 m apart, and north-south cross-transects were spaced at 500 m intervals. The survey imaged two strong reflectors interpreted to represent major bounding surfaces. The lower surface was confirmed to be the bedrock-sediment interface defining the valley boundary. The second major surface is the boundary between the fluvial valley fill and overlying bleached sand correlative with the White Sand Formation. In addition to the major surfaces, several minor surfaces were also imaged, resulting in the identification of 21 radar elements. Sixteen of the elements were interpreted to represent fluvial architectural elements. Four distinct morphological zones were recognized and were differentiated by variation in the geometry of the bedrock-sediment interface and by distinctive assemblages of architectural elements.