SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Study site
  5. GPR methods
  6. Results
  7. Discussion and conclusions
  8. Acknowledgements
  9. References

Sea-level rise has been related to global warming. The modern system on the northern coast of Anholt, Denmark, may well be analogous to other beach ridge systems formed in microtidal regimes and our results should have impact on estimation of past sea-level variation. Ground-penetrating radar data collected across the modern (<30 years old) berm, beach ridge and swale deposits resolve downlapping reflections interpreted to mark sea level at the time of deposition. Existing time series of sea-level data constrain actual sea-level variation. Nineteen readings of sea-level markers made along our profile fluctuate within −0.42 and 0.57 m above present mean sea level, consistent with 95% of the sea-level data. These fluctuations reflect tidal effects and meteorological conditions. Main data uncertainties are well-known and the sea-level markers may be identified with a high degree of confidence.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Study site
  5. GPR methods
  6. Results
  7. Discussion and conclusions
  8. Acknowledgements
  9. References

Sea-level rise has been related to global warming and a semi-empirical relationship between global mean temperatures and sea level has been established based on sea-level data from the last 120 years (e.g. Rahmstorf, 2007). From a wider palaeoclimatic perspective, it is desirable to test this relationship against a longer time series of sea-level variation. Estimates of Holocene sea-level variation have been based on a wide range of records (Long, 2000, 2001). Well-developed beach ridge systems form one attractive archive of past sea-level variation (e.g. Bjørnsen et al., 2008) and if beach ridge characteristics can be related precisely to sea level, it should be possible to construct reliable sea-level curves for the past 6000–7000 years.

Beach ridges form in coastal zones (e.g. Tanner, 1995; Otvos, 2000), and their elevation and internal architecture may be used as proxies of relative palaeo-sea-level and -wave climate (Goy et al., 2003). Ground-penetrating radar (GPR) methods have been used for outlining the internal sedimentary architecture of coastal features (including beach ridge systems) with the aim of unravelling the development of such features in response to relative sea-level fluctuations, sediment supply and changes in wind and wave characteristics (van Heteren et al., 1998; Clemmensen et al., 2001; Neal et al., 2003; Fraser et al., 2005; Engels and Roberts, 2005; Rodriguez and Meyer, 2006; Bristow and Pucillo, 2006; Nielsen, 2006; Bjørnsen et al., 2008; Tamura et al., 2008).

Identification of good-quality proxies or markers of palaeo-sea-level based on observed GPR beach ridge system reflectivity architecture may be difficult because ridge architecture first has to be interpreted in relation to wash-over and wave run-up processes and, secondly, in relation to mean sea level. Rodriguez and Meyer (2006) also list factors, which may make it difficult to use beach ridge morphology as a precise sea-level marker. In a recent study from the Kujukuri strand plain at the Pacific coast of Japan, Tamura et al. (2008) identified a downlapping surface in reflection GPR data. They interpreted this downlapping surface to separate foreshore reflections from reflections originating from sedimentary layers deposited in the upper shoreface regime. Tamura et al. (2008) interpreted the downlapping reflections to be markers showing a depth level of ∼1 m below the mean sea level in the environment they were investigating. Other features such as, for example, transitions from dune to beach deposits, have also been suggested as sea-level proxies (van Heteren et al., 2000) and identified in GPR data (Bristow and Pucillo, 2006).

In this study, we interpreted high-resolution GPR reflection data acquired across the youngest part (<30 years old) of a recently formed beach ridge system on the island of Anholt, Denmark. Our GPR line crosses berm ridge, beach ridge and swale deposits. We identified downlap points interpreted to mark the transition from the beach to the upper shoreface regime identified in our GPR data (cf. Tamura et al., 2008). We tested if the downlap points can be used as sea-level markers by comparing the variability of the downlap points to recent sea-level data. We could not trench the deposits to sufficient depth to get direct measurements of the downlap surfaces because of the high stone content of the deposits (cf. Johnston et al., 2007).

Study site

  1. Top of page
  2. Abstract
  3. Introduction
  4. Study site
  5. GPR methods
  6. Results
  7. Discussion and conclusions
  8. Acknowledgements
  9. References

Anholt is located in the sea of Kattegat, which constitutes a microtidal regime with a maximum tidal range of ∼0.4 m (measured by the Danish Maritime Safety Administration). The island is part of the gateway between the North Sea and the Baltic Sea (Fig. 1). Although tidal variation is small, sea level may vary considerably on daily to weekly basis because of meteorological conditions (in particular, wind set-up and barometric pressure) (Sørensen et al., 2007) and during the winter months, short-term variations in sea level in the order of 1 m are not uncommon (data from the Danish Maritime Safety Administration).

image

Figure 1.  Big inset: Location map. DK: Denmark. K: sea of Kattegat. Box marks Anholt. Dot marks sea-level measurement station. Small inset: Anholt; small box indicates study area. Background map shows a zoom of the prograding beach ridge plain of the study areas. The youngest deposits including those investigated here were formed after ad 1979. Ages of coastlines and their positions are based on maps of the National Survey and Cadastre (Kort & Matrikelstyrelsen) and interpretations of Larsen and Kronborg (1994). Topographic data collected by COWI A/S during spring 2007. Inset maps were based on the Generic Mapping Tools (GMT) (Wessel and Smith, 1991).

Download figure to PowerPoint

The strandplain deposits of the northwestern part of the island were formed after ad 1887, and the sediments, which we investigate here, were deposited after ad 1979 (Fig. 1; Larsen and Kronborg, 1994). From the shoreline towards the inland, the modern strandplain can be divided into a beach, a berm ridge, a narrow swale, an inner beach ridge and a wider swale (Fig. 2). No aeolian dune deposits are observed on top of these features. The beach, which is covered by sand and scattered stones, is about 10–15 m wide and composed of a relatively smooth backshore with seaward dips between 3° and 5°, and a foreshore with slightly lower seaward dips; low-relief ridge and runnel structures are developed in the shoreface near present mean sea level. The boundary between the backshore and the foreshore is defined by the mean high-water level. A trench in the lower part of the foreshore shows seaward-dipping beach deposits overlying landward-dipping ridge sediments. The summits of the berm and beach ridges are found at 1.8 and 1.5 m above present mean sea level (m asl) respectively. The two swales are located at ∼0.9 m asl. The berm ridge (Otvos, 2000) is slightly asymmetric with slightly steeper dips towards the sea than towards land (flank dips of 5–10°) and it is covered with stones with diameters of up to ∼10 cm. Locally, however, the landward-facing part of the berm ridge is developed as steeply dipping washover deposits. The outer swale is covered by sand with scattered stones. A trench in the swale indicates that these deposits are composed of alternating sandy units with faint horizontal to low-angle-dipping lamination (units of up to 0.20 m in thickness) and stone-rich horizons. The inner beach ridge is stone-covered like the berm, but may have been somewhat deflated. A trench reveals that the beach ridge is composed of alternating sandy and stone-rich units; the sandy units typically have low-angle, landward-dipping or near-horizontal lamination. The inland-lying swale is rather wide, but similar to the outer swale with respect to dominant grain sizes. When we studied the site in August 2007, the berm ridge showed evidence of recent activation, and we suggest that the ridge was formed and/or reactivated at three events between 1 November 2006 and 19 January 2007 when sea level reached heights between 1.20 and 1.47 m asl according to the Danish Maritime Safety Administration.

image

Figure 2.  Photographs of the study site. (A) Note the berm and the smooth, seaward-dipping beachface composed of lower backshore and foreshore. Photographer is facing west. (B) The modern berm-swale-beach ridge system at the study site, Anholt. Photographer is standing on the central part of the modern berm and is facing southwest. Rucksack for scale.

Download figure to PowerPoint

GPR methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Study site
  5. GPR methods
  6. Results
  7. Discussion and conclusions
  8. Acknowledgements
  9. References

GPR data were acquired with shielded 250 MHz Sensors & Software antennae. The antennae were mounted on a skid plate in contact with the ground surface during recording. Spatial and temporal sampling intervals were 0.05 m and 0.4 ns respectively. At each recording location, 16 recordings were stacked into a single trace. Initial processing involved dewowing of the GPR signals using a software package provided by Sensors & Software. Subsequently, the GPR data set was processed using the ProMAX software of Landmark. Gain corrections were made using an automatic gain control algorithm (maximum scaling factor of 200). A bandpass filter was applied to reduce noise above 350 MHz. The GPR wave velocity of the dominantly sandy and gravely sediments is mainly determined by the water saturation of the sediment (Reynolds, 1997). Analysis of the curvature of reflection hyperbolas observed in the acquired reflection section constrained root-mean-square (rms) velocities to vary between 0.065 m ns−1 in the swales to 0.10 m ns−1 in the uppermost parts of the ridges; the rms velocities gradually dropped to 0.08 m ns−1 below the ridges. Migration of the reflection section was performed using a modified version of the Stolt migration algorithm (Stolt, 1978), which allowed for a varying velocity field. Steeply dipping ‘noise’ (presumably energy that did not migrate properly) was dampened using an F-K filter (cf. Yilmaz, 1987). Average velocities calculated from the laterally and horizontally varying rms-velocity structure were used for depth-conversion of the individual traces. Topography of swale bases and ridge summits was measured with an accuracy of ∼0.02 m using a Trimble GPS system. Spectral analysis showed that the peak frequency of the reflection signals was in the 250–300 MHz range. The vertical resolution is about 1/4 of the dominant wavelength (Yilmaz, 1987; Jol, 1995). In our case, this resolution is ∼0.1 m.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Study site
  5. GPR methods
  6. Results
  7. Discussion and conclusions
  8. Acknowledgements
  9. References

Our GPR line is 159 m long. Here, we have chosen to interpret the 129 m of the profile closest to the sea along which we have good constraints on topography (Fig. 3). The upper ∼1 m of the subsurface below the swales crossed by the GPR line predominantly shows seaward dipping (∼2–7°) reflections. Similar dipping events can be observed in approximately the same depth range below the berm and beach ridges, although the reflection patterns are more heterogeneous below the ridges. Relatively strong, continuous reflections, which are horizontal or dip moderately towards the sea, are observed at about −1 to 0 m asl below the swales. Below these reflections, the observed reflection patterns become slightly more chaotic with sea- and landward-dipping, mounded and occasional concave-down reflections. A transition to more chaotic reflection patterns appears to occur at about the same depth level below the beach ridge and the berm ridge. However, below the ridges, this transition is less evident than below the swales, because of the more heterogeneous reflectivity patterns in the upper ∼2 m of the subsurface below the ridges. Relatively strong seaward-dipping (∼10°) reflections are observed from ∼0 to 1 m asl below the seaward-dipping flank of the berm ridge. Landward-dipping reflections are evident in the upper ∼1 m of the subsurface below the southern flank of the berm ridge. Below the seaward-dipping flank of the beach ridge, the uppermost ∼0.5 m of the subsurface is characterised by weakly seaward-dipping to almost horizontal reflections truncated by the ground surface.

image

Figure 3.  Ground-penetrating radar section collected across modern features on the northwestern tip of Anholt (cf. Fig. 1). For display purposes, the profile has been split into two parts: the southern part (A), and the northern part (B). In A and B, the data are shown without (top) and with (bottom) interpretation of features (thick black lines). Dotted black line indicated by ‘m’ may represent a multiple reflection (cf. Bristow and Pucillo, 2006). Vertical exaggeration is 1 : 2. Thick red line indicates present mean sea level as defined by Danish Vertical Reference (DVR90). Dotted red lines indicate levels corresponding to 0.5 m above and below present mean sea level. Circles indicate interpreted points of downlap discussed in text. Horizontal axes represent distance to the shoreline at the time of recording.

Download figure to PowerPoint

The predominantly seaward-dipping reflections observed in the upper subsurface below both the berm ridge and the beach ridge are interpreted as representing beach deposits formed by wave run-up (Otvos, 2000). Landward-dipping reflections below the southern flanks of the ridges are interpreted as wash-over deposits (Neal et al., 2003). We interpret the sediments giving rise to the diverse reflection patterns below the continuous reflections found at a depth of about −1 m asl as belonging to the upper shoreface regime (cf. Tamura et al., 2008). Circles in Fig. 3 show where seaward-dipping beach reflections downlap onto upper shoreface deposits and are interpreted as marking a level near sea level at the time of deposition (cf. Tamura et al., 2008). The heights of the 19 downlap points identified in Fig. 3 are given in Table 1. The average height of the 19 readings is 0.063 m asl. Minimum and maximum values read for these points are −0.42 and 0.57 m asl respectively. The average value and standard deviation of the 591682 sea-level data collected by the Danish Maritime Safety Administration between 1991 and 2007 (Fig. 4) are 0.02 and 0.25 m respectively. The average measured sea-level value plus/minus two standard deviations constitutes the interval from −0.48 to 0.52 m. This interval is nearly identical to the interval defined by the minimum and maximum picked downlap values, and 95% of the sea-level measurements fall within this range. Main sources of error on the readings of the downlap points are: (1) the vertical GPR resolution of 0.1 m; (2) errors in topography measurements and interpolation between measured points of topography (estimated to be 0.02–0.1 m); (3) possible undesired frequency filtering effects causing pre-cursers to the GPR signals (probably less than 0.1 m); (4) incorrect average velocity for depth-conversion (assumed to give rise to a depth uncertainty of <10% of the depth reading). Thus, the total uncertainty is assumed to be 0.25–0.27 m.

Table 1.   Height of downlap points (Fig. 3). Average value is 0.063 m asl, standard deviation is 0.23 m.
Downlap point (from left to right in Fig. 3) Height (m asl)
1−0.17
2−0.07
3−0.09
40.15
5−0.34
60.04
70.18
80.12
90.17
100.05
110.19
120.17
130.25
140.31
150.57
16−0.04
170.23
18−0.42
19−0.12
image

Figure 4.  Histogram showing the distribution of 591682 sea-level data values measured by the Danish Maritime Safety Administration from 1991 to 2007. The station is located on the mainland of Denmark ∼50 km west of Anholt and is believed to be representative of the sea-level conditions of Anholt. Vertical line indicates average value extracted from GPR data. Dotted vertical lines indicate minimum and maximum values extracted from the GPR data (Table 1).

Download figure to PowerPoint

The Fennoscandian region has been subject to isostatic rebound (Lambeck et al., 1998). Anholt is situated between the current 0 and 1 mm yr−1 isolines for vertical movement, whereas the station of sea-level measurements is located close to the 0 mm yr−1 isoline (Kakkuri and Chen, 1992). We assume that the total rebound of the modern coastal features we have studied does not exceed 0.03 m. Weak earthquakes are recorded in the Kattegat area (Gregersen et al., 1996). Jensen et al. (2002) found evidence for movement in the uppermost parts of the subsurface in the sea of Kattegat. Thus, we cannot rule out that Anholt has undergone small-scale vertical tectonic movements during the last 30 years, although we know of no evidence for such movements. Recent sea-level rise (Rahmstorf, 2007) would affect the downlap point positions. However, sea-level analysis made by the Danish Meterological Institute does not indicate an accelerating sea-level rise during the last part of the 20th century (Erik Bødtker, personal communication).

Discussion and conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Study site
  5. GPR methods
  6. Results
  7. Discussion and conclusions
  8. Acknowledgements
  9. References

Identification of sedimentary features that may constitute sea-level proxies has been made in a number of recent studies of GPR data (Bristow and Pucillo, 2006; Rodriguez and Meyer, 2006; Tamura et al., 2008). Favourable conditions (a dry topsoil, no or limited presence of clay and mud, no aeolian sand cover and limited saltwater instrusion) allow us to image the subsurface to a depth of 3–5 m below the surface to a distance of only ∼5 m from the coastline. We are able to identify downlap points interpreted to represent the boundary between beach and upper shoreface deposits (cf. Tamura et al., 2008), and our high-resolution data combined with recent sea-level data allow us to test how these downlap points relate to present sea level. We find that the 19 downlap points we have identified have an average value of 0.063 m asl and a standard deviation of 0.23 m and are consistent with sea-level data acquired at a nearby station. Moreover, the range defined by the minimum and maximum heights of the downlap points represents 95% of the sea-level measurements. Thus, our study indicates that the observed downlap points represent actual sea level at the time of deposition and we expect that these points most likely constitute robust sea-level markers in microtidal regimes similar to our study site. We stress the importance of measuring as many downlap points as possible before a mean sea level is calculated because sea level varies due to meteorological conditions and tidal effects, and because not all parts of the system may be preserved (Johnston et al., 2007).

The high degree of consistency between GPR data and sea-level measurements must be partly coincidental, because the estimated total uncertainty of our investigations exceeds the standard deviations of the GPR downlap readings and the sea-level measurements. However, error sources 2, 3 and 4 give rise to strongly correlated errors, and the uncertainties linked to our experiment cannot explain the short-scale fluctuations observed for the downlap readings.

Tamura et al. (2008) found the downlap level at an average depth of ∼1 m below mean sea level in the Kujukuri strand plain, but they did not resolve variations in the depth to this surface. In their case, the average depth level of the downlap points corresponds to the low-tide level during spring tide. Thus, it may be the case that, for different regions, the boundary between beach (foreshore) and upper shoreface deposits represents a somewhat different depth level with respect to sea level, depending on local beach morphology, the tidal effects as well as wave and wind climate of the region under investigation. We expect that the microtidal regime of the sea of Kattegat may be representative of a large part of the Baltic Sea area and therefore we assume that the sea-level marker tested here can be used for studies of recent and fossil coastal systems as wells as for the construction of relative sea-level curves, at least in this area.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Study site
  5. GPR methods
  6. Results
  7. Discussion and conclusions
  8. Acknowledgements
  9. References

This study was financed by the Danish Natural Science Research Council and the Carlsberg Foundation. Promax and SeisVision software were available at the Department of Geography and Geology through a Landmark University grant. We thank M. Coleman, J. W. Johnston, I. P. Martini, V. Pascucci and T. Tamura for constructive comments.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Study site
  5. GPR methods
  6. Results
  7. Discussion and conclusions
  8. Acknowledgements
  9. References
  • Bjørnsen, M., Clemmensen, L.B., Murray, A. and Pedersen, K., 2008. New evidence of the Littorina transgressions in the Kattegat: optically stimulated luminescence dating of a beach ridge system on Anholt, Denmark. Boreas, 37, 157168.
  • Bristow, C.S. and Pucillo, K., 2006. Quatifying rates of coastal progadation from sediment volume using GPR and OSL: the Holocene fill of the Guichen Bay, south-east South Austalia. Sedimentology, 53, 769788.
  • Clemmensen, L.B., Andreasen, F., Heinemeier, J. and Murray, A., 2001. A Holocene coastal aeolian system, Vejers, Denmark: landscape evolution and sequence stratigraphy. Terra Nova, 13, 129134.
  • Engels, S. and Roberts, M.C., 2005. The architecture of prograding sandy-gravel beach ridges formed during the last Holocene highstand: Southwestern British Columbia, Canada. J. Sed. Res., 75, 10521064.
  • Fraser, C., Hill, P.R. and Allard, M., 2005. Morphology and facies architecture of a falling sea level strandplain, Umiujaq, Hudson Bay, Canada. Sedimentology, 52, 141160.
  • Goy, J.L., Zazo, C. and Dabrio, C.J., 2003. A beach-ridge progradation complex reflecting periodical sea-level and climate variability during the Holocene (Gulf of Almería, western Mediterranean). Geomorphology, 50, 251268.
  • Gregersen, S., Leth, J., Lind, G. and Lykke-Andersen, H., 1996. Earthquake activity and its relationship with geologically recent motion in Denmark. Tectonophysics, 257, 265273.
  • Van Heteren, S., Fitzgerald, D.M., McKinlay, P.A. and Buynevich, I.V., 1998. Radar facies of paraglacial barrier systems: coastal New England, USA. Sedimentology, 45, 181200.
  • Van Heteren, S., Huntley, D.J., Van De Plassche, O. and Lubberts, R.K., 2000. Optical dating of dune sand for the study of sea-level change. Geology, 28, 411414.
  • Jensen, J.B., Petersen, K.S., Konradi, P., Kuijpers, A., Bennike, O., Lemke, W. and Endler, R., 2002. Neotectonic, sea-level changes and biological evolution in the Fennoscandian Border Zone of the southern Kattegat. Boreas, 31, 133150.
  • Johnston, J.W., Thompson, T.A. and Baedke, S.J., 2007. Systematic pattern of beach-ridge development and preservation: conceptual model and evidence from ground penetrating radar. In: Stratigraphic Analysis Using GPR (G.S.Baker and H.M.Jol eds). Geol. Soc. Am. Spec. Paper, 432, 4758.
  • Jol, H.M., 1995. Ground penetrating antennae frequencies and transmitter powers compared for penetration depth, resolution and reflection continuity. Geophys. Prospect., 43, 693709.
  • Kakkuri, J. and Chen, R., 1992. On horizontal crustal strain in Finland. Bull. Géodésique, 66, 1220.
  • Lambeck, K., Smither, C. and Johnston, P., 1998. Sea-level change, glacial rebound and mantle viscosity for northern Europe. Geophys. J. Int., 134, 102144.
  • Larsen, G. and Kronborg, C., 1994. Geologisk Set – Det mellemste Jylland. Geografforlaget, Brenderup, Denmark, 273 pp.
  • Long, A., 2000. Late Holocene sea-level change and climate. Prog. Phys. Geogr., 24, 415423.
  • Long, A., 2001. Mid-Holocene sea-level change and coastal evolution. Prog. Phys. Geogr., 25, 399408.
  • Neal, A., Richards, J. and Pye, K., 2003. Sedimentology of coarse-clastic beach-ridge deposits, Essex, southeast England. Sed. Geol., 162, 167198.
  • Nielsen, C., 2006. Udviklingen af en lavenergi strandvoldsslette undersøgt med georadar og OSL-dateringer, Mellem til Sen Holocæn, Femø, Danmark. Geol. Tidsskr., 2, 510.
  • Otvos, E.G., 2000. Beach ridges – definitions and significance. Geomorphology, 32, 83108.
  • Rahmstorf, S., 2007. A semi-empirical approach to projecting future sea-level rise. Science, 315, 368370.
  • Reynolds, J.M., 1997. An Introduction to Applied and Environmental Geophysics. John Wiley & Sons Ltd., Chichester, England, 796 pp.
  • Rodriguez, A.B. and Meyer, C.T., 2006. Sea-level variation during the Holocene deduced from the morphologic and stratigraphic evolution of Morgan Peninsula, Alabama, U.S.A. J. Sed. Res., 76, 257269.
  • Sørensen, P., Sørensen, C., Ingvardsen, S.M., Andersen, I. and Kloster, B.B., 2007. Højvandsstatistikker 2007 – Extreme sea level statistics for Denmark, 2007. Dan. Coast. Authority (Accessed 14 September 2009: http://www.kyst.dk).
  • Stolt, R.H., 1978. Migration by Fourier transform. Geophysics, 43, 2348.
  • Tamura, T., Murakami, F., Nanayama, F., Watanabe, K. and Saito, Y., 2008. Ground-penetrating radar profiles of Holocene raised–beach deposits in the Kujukuri strand plain, Pacific coast of eastern Japan. Mar. Geol., 248, 1127.
  • Tanner, W.F., 1995. Origin of beach ridges and swales. Mar. Geol., 129, 149161.
  • Wessel, P. and Smith, W.H.F., 1991. Free software helps map and display data. EOS Trans. Amer. Geophys. U., 72, 445446.
  • Yilmaz, Ö., 1987. Seismic Data Processing. Society of Exploration Geophysicists, Tulsa, OK, USA, 526 pp.