Henk J.A. Berendsen

Factors Controlling the Holocene Avulsion History of the Rhine-Meuse Delta (The Netherlands)

ABSTRACT The avulsion history of the Holocene Rhine-Meuse delta (The Netherlands) was reconstructed, on a timescale of millenia, using detailed paleogeographic maps based on approximately 200,000 lithological borehole descriptions, over 1150 14C ages, and 36,000 archaeological artifacts. Gradient lines were constructed for all channel belts. These allowed determination of gradients, paleo-flow direction, relative age of channel belts, and time correlation of undated channel-belt fragments. Avulsion sites were inferred from the paleogeographic reconstruction. At least 91 avulsions occurred over the past 10,000 years, of which 82 could be dated with an accuracy of ± 200 14C years. The location and shifting of Holocene avulsion sites in the Rhine-Meuse delta in space and time are related to: (1) Relative sealevel rise. In the Early Holocene, avulsions could not take place, because rivers were still incised. Around 7500 yr BP, avulsions occurred in the western part of the present delta as a result of backfilling of the Late Weichselian valley. Between 7500 and 3700 yr BP, the zone where avulsions occurred shifted inland as a result of relative sealevel rise. (2) Neotectonics. Between 4900 and 1700 yr BP, the location of avulsion sites seems to have been influenced by neotectonic movements of the upthrown Peel Horst. Four out of six avulsion nodes in the Rhine-Meuse delta were located in the Peel Horst fault zones. (3) Increased discharge and/or within-channel sedimentation. From 2800 until about 1500 yr BP, avulsion sites were located all over the delta. During this time, the number of channels was high, and avulsion frequency reached a maximum, at a time when aggradation rate decreased with a reduction in the rate of sealevel rise. After 2000 yr BP meander wavelength of alluvial channels increased considerably. The increased meander wavelength and the high avulsion frequency are attributed to increased bankfull discharge or within-channel sedimentation (leading to channel widening), or both. (4) Human influence. Between 1000 and 650 yr BP, all the rivers were embanked, and avulsions could no longer take place. The few that occurred were induced by humans.

Avulsion Frequency, Avulsion Duration, and Interavulsion Period of Holocene Channel Belts in the Rhine-Meuse Delta, the Netherlands

ABSTRACT Avulsion parameters greatly influence alluvial architecture, because they determine channel density and interconnectedness. Reliable data on avulsion parameters is still scarce. From a detailed reconstruction of the Holocene avulsion history of the whole Rhine-Meuse delta we determined quantitative values for the avulsion parameters: avulsion frequency, avulsion duration, and interavulsion period. In the Rhine-Meuse delta the number of coeval channels is related to the avulsion frequency. Calibrated 14C intensity histograms are used to analyze 14C dates of beginning and ending river activity. The histograms show that instantaneous and gradual avulsions were almost equally important in the Rhine-Meuse delta. A high avulsion frequency occurred from 8000 to 7300 cal yr BP (a total of 17 avulsions, i.e., 2.43 avulsions/100 years). During this period the avulsion frequency was related to the high rate of sea-level rise, which induced fluvial sedimentation in the present near-coastal area. After 7300 cal yr BP avulsion frequency decreased as a result of a continuously decreasing rate of sea-level rise. From 7300 to 3200 cal yr BP avulsion frequency was low; 35 avulsions took place within 4100 years (0.85 avulsions/100 years). Approximately 5000 cal yr BP the coastal barriers became closed and large-scale peat formation occurred. This resulted in fixation of the river channels and low cross-valley gradients, reducing the chances for avulsion. Between 3200 and 1400 cal yr BP avulsion frequency was high again (a total of 34 avulsions, i.e., 1.89 avulsions/100 years) as a result of increased discharge and/or within-channel sedimentation, which enhanced chances for avulsion. The period of activity of channel belts shows no significant trend on the time scale of the Holocene. It is highly variable and averages 1280 ± 820 cal yr. The avulsion duration fluctuates between less than 200 and 1250 cal years and averages 335 cal years. The avulsion duration shows no significant trend through time and remained constant until at least 1900 cal yr BP. Avulsion duration seems not to be influenced by aggradation rate. Average interavulsion period is shorter than the period of activity of channel belts, and is approximately 945 cal years. In the Rhine-Meuse delta, on the time scale of the Holocene, interavulsion period and avulsion duration are on average constant. Therefore, the number of coeval channels is directly related to avulsion frequency. Available evidence from other rivers and deltas around the world, however, suggests that the relationship between these avulsion parameters is not everywhere the same.

Fluvial deposits as a record for Late Quaternary neotectonic activity in the Rhine-Meuse delta,The Netherlands

Neotectonic movements have caused differential subsidence in the Lower Rhine Embayment during the Quaternary. The Late Weichselian and Holocene Rhine-Meuse fluvial archive in the central Netherlands was used to quantify neotectonic movements in a setting that was primarily controlled by sea-level rise and climate change. Evidence for neotectonic activity in the central Netherlands is reviewed. Sedimentary evidence shows that fluvial deposits of Late Weichselian and Holocene Rhine and Meuse (Maas) distributaries are vertically displaced along the northern shoulder of the Roer Valley Graben system. Elevation differences in the longitudinal profiles of Late Weichselian terrace deposits were used to quantify tectonic displacements. New results for the southeastern Rhine-Meuse delta (Maaskant area) show that displacements in the top of the Pleniglacial terrace along the Peel Boundary Fault are up to 1.4 m. The maximum displacement between the Peel Horst and the Roer Valley Graben is 2.3 m. This is equivalent to relative tectonic movement rates of 0.09-0.15 mm/yr, averaged over the last 15,000 years.

Alluvial architecture of the human-influenced river Rhine, The Netherlands

Abstract The alluvial architecture of the embanked Rhine distributaries in The Netherlands is largely controlled by human interference with fluvial processes. Embankment, straightening, and stabilization of the rivers have influenced both in-channel and overbank sedimentation. Embanked river floodplain deposits are lithologically and sedimentologically different from natural (not human-influenced) floodplain deposits. After the embankments between 1050 and 1350 AD, the dominant process remained lateral within-channel accretion. However, the morphology of the point bars changed: in the natural situation, the bars had a convex shape (as seen from the channel), whereas in the embanked situation, they are concave (as seen from the channel). The latter is due to the development of bars along the downstream end of a previous formed bar. Deposits associated with lateral within-channel accretion are typically very fine to medium or coarse sand. The infill of residual channels is characterized by the absence of peat. After the normalization of the river channel to a fixed width (1850 AD), concave bank erosion, and development of scroll bars were constricted, and overbank deposition became the dominant embanked floodplain-forming process. Thick layers of sandy and clayey deposits were deposited on the embanked floodplain. Splay-like sand lenses are present in the overbank deposits. In case of dyke failures, a deep scour-hole was formed. Typical deposits associated with dyke-breaches are sandy dyke-breach deposits and scour-hole fills consisting of silty to sandy clay within the embanked floodplain and gyttja within the polder.

Overbank sand deposition in relation to transport volumes during large‐magnitude floods in the Dutch sand‐bed Rhine river system

Using aerial photographs and field measurements, sandy overbank deposits formed by the large-magnitude floods of 1993/94 and 1995 were quantified along two branches of the Dutch Rhine river system: the Waal (1993/94 and 1995) and the IJssel (1995). These deposits were laid down intermittently all along the length of these rivers on the top and landward slope of the natural levees, and covered about 4 per cent of the embanked floodplain on the Waal and about 1 per cent on the IJssel. The overbanks and transport mechanism is basically convective by nature. The spatial variability of overbank sedimentation points to the important role played by helicoidal currents in determining overbank deposition. The presence of embankments and training works appears to influence the sand transport to and morphological development of the floodplains along the Dutch Rhine river system. Overbank deposition volumes about equal present estimates of sand transport during a large-magnitude flood. It appears that studies on sand transport in the Dutch Rhine carried out so far underestimate sand transport during floods.

Calibration of 14C histograms: a comparison of methods

The interpretation of C-14 histograms is complicated by the non-linearity of the C-14 time scale in terms of Calendar years, which may result in clustering of C-14 ages in certain time intervals unrelated to the (geologic or archaeologic) phenomenon of interest. One can calibrate C-14 histograms for such distortions using two basic approaches. The KORHIS method constructs a C-14 histogram before calibration is performed by means of a correction factor. We present the CALHIS method based on the Groningen calibration program for individual C-14 ages. CALHIS first calibrates Single C-14 ages and then sums the resulting calibration distributions, thus yielding a calibrated C-14 histogram. The individual calibration distributions are normalized to a standard Gaussian distribution before superposition, thus allowing direct comparison among various C-14 histograms. Several experiments with test data sets demonstrate that CALHIS produces significantly better results than KORHIS. Although some problems remain (part of the distortions due to C-14 variations cannot be eliminated), we show that CALHIS offers good prospects for using C-14 histograms, particularly with highly precise and accurate C-14 ages.

The use of GIS in reconstructing the Holocene palaeogeography of the Rhine-Meuse delta, The Netherlands

In Holocene palaeogeographical studies, GIS nowadays is used in various ways, embracing either GIS modelling possibilities or GIS spatial data storage functionality. This paper is an example of the latter type of GIS use for reconstruction of the Holocene Rhine–Meuse delta (128×75 km; part of the Dutch coastal plain). Depositional traces of river courses (channel belts) are major features to map. We digitized mapped channel belt fragments and stored their reconstructed ages in a linked table. The GIS data structure allows checking for internal consistency of the implemented reconstruction, enabling iterative improvement while digitizing and labelling. Palaeogeographical maps for any given time during the Holocene result from GIS‐database queries. Coupled to the GIS are high‐resolution digital elevation data (sub‐decimetre accurate laser‐altimetry data covering the entire study area) and a database of ca 100 000 borehole descriptions, allowing for further mapping improvements. The GIS structure and digital data together greatly improved mapping and enable larger areas to be mapped in greater detail with greater accuracy and in less time.

New groundwater-level rise data from the Rhine-Meuse delta – implications for the reconstruction of Holocene relative mean sea-level rise and differential land-level movements

We present new local groundwater-level rise data from two Late Glacial aeolian dunes, located near Barendrecht and Oud-Alblas in the western Rhine-Meuse delta. These data are based on AMS radiocarbon dating of terrestrial macrofossils, collected from the base of peat formed on the slopes of these dunes. This method avoids contamination of bulk peat samples by old soil carbon or younger rootlets and rhizomes, as well as the hardwater effect. The new data are used to assess the reliability of previously published groundwater-level index data based on conventional radiocarbon dating of bulk basal peat samples from the slopes of the Late Glacial aeolian dunes at Barendrecht, Hillegersberg, Bolnes and Wijngaarden, all located in the western Rhine-Meuse delta. Comparison of the new and published groundwater-level data shows no significant systematic difference between conventionally dated bulk peat samples and AMS-dated samples of terrestrial macrofossils. The new data from the dune at Barendrecht confirm the reliability of the younger than 6600 cal yr BP age-depth data from the dunes at Hillegersberg and near Bolnes. This result supports the validity of this part of the mean sea-level (MSL) curve for the western Netherlands. Consequently, the position of the groundwater-level curve for Flevoland (central Netherlands) below this MSL curve can most likely be attributed to differential land-level movement. The available data show that the groundwater-gradient effect in the western Rhine-Meuse delta became less than 5 cm/km after 6600 calyr BP. Finally, temporal correlation between temporary increases in local groundwater-level rise with known shifts of river courses in the delta plain suggests, that avulsions can explain sudden local deviations from the trend in groundwater-level rise. A general conclusion of this study is that a complex relationship exists between sea level and local delta-plain water levels.

The TESS-1 Suction Corer: A New Device to Extract Wet, Uncompacted Sediments

Conventional coring can be an expensive and cumbersome technique in uncompacted, waterlogged sediments, especially when large numbers of samples are required rapidly and when vehicular access is restricted. The existing Van der Staay suction corer (Van de Meene et al. 1979) solves economic and other problems (chiefly logistical and environmental issues), but sample extraction has to be done in the field. The original purposes of the Van der Staay suction corer were to enable rapid drilling through sand and allow field sampling. It was not designed to recover intact samples for subsequent laboratory analysis. Cores from the new TESS-1 suction corer, however, are encapsulated in the field. The TESS-1 therefore combines the advantages of the original corer with the possibility of taking samples for laboratory analysis. Samples collected using the TESS-1 suction corer show little deformation, allow coring of soft uncompacted mud and retention of sedimentary structures. Problems due to sediment fluidization and movement inside the core pipe are usually minimal. The TESS-1 suction corer can be used in both wetland areas (beach, tidal flats, and marshes) and in shallow subaqueous environments (lakes, rivers, and nearshore subtidal areas).

Correcting (super 14) C histograms for the non-linearity of the (super 14) C time scale.

Large numbers of 14 C dates of the base and top of Holocene peat layers may be plotted in 14 C histograms in order to establish statistically a chronology of periods of essentially clastic sedimentation and peat formation. Due to the non-linearity of the 14 C time scale in terms of calendar years, clustering of 14 C dates on random peat growth may occur. This seriously hampers the interpretation of histograms. A quantitative method and computer program were developed to correct the histograms for this effect. The correction factor that has to be applied depends on the calibration curve and the interval width of the correction parameter dy. For peat samples, an interval width of 100 14 C yr and a calibration curve based on a 100-yr moving average seems to be a reasonable choice.