Scholarly article on topic 'Luminescence dating of the Rissian type section in southern Germany as a base for correlation'

Luminescence dating of the Rissian type section in southern Germany as a base for correlation Academic research paper on "History and archaeology"

CC BY-NC-ND
0
0
Share paper
Academic journal
Quaternary International
OECD Field of science
Keywords
{"Northern Alpine Foreland" / "Feldspar single grain" / "European Alps" / "Luminescence dating" / Hochterrasse / Rissian}

Abstract of research paper on History and archaeology, author of scientific article — Eike F. Rades, Markus Fiebig, Christopher Lüthgens

Abstract The exact timing of the Rissian has been under discussion since being established by Penck and Brückner (1909) at the beginning of the 20th century. Difficulties in correlating and especially dating the sediments associated with the Rissian have led to different nomenclatures in the different regions of the Northern Alpine Foreland (NAF). Various dating approaches have led so far to often unsatisfying results. In this study we successfully dated the “High Terrace Gravels” of the Rissian type section. Using single grain feldspar luminescence we were able to evade the problems of incomplete bleaching in fluvioglacial sediments. Using the post Infrared IRSL protocol (at 225 °C) we were able to show that the feldspar in the research area only shows low rates of anomalous fading. We were able to show that these low fading rates have an influence on the age determination. This is remarkable because many of the signals were close to saturation. The conclusive age range of 149 ± 15–179 ± 17 ka corresponds to Marine Isotope Stage 6. Defining an age for the Rissian is a first step to consolidate the heterogeneous nomenclature and to reconstruct the chronology of past major Alpine glaciations.

Academic research paper on topic "Luminescence dating of the Rissian type section in southern Germany as a base for correlation"

Quaternary International xxx (2016) 1—13

Contents lists available at ScienceDirect

Quaternary International

journal homepage: www.elsevier.com/locate/quaint

Luminescence dating of the Rissian type section in southern Germany as a base for correlation

Eike F. Rades*, Markus Fiebig, Christopher Lüthgens

Institute for Applied Geology, University of Natural Resources and Life Sciences, Peter Jordan Straße 70, 1190, Vienna, Austria ARTICLE INFO ABSTRACT

Article history: Available online xxx

Keywords:

Northern Alpine Foreland Feldspar single grain European Alps Luminescence dating Hochterrasse Rissian

The exact timing of the Rissian has been under discussion since being established by Penck and Brückner (1909) at the beginning of the 20th century. Difficulties in correlating and especially dating the sediments associated with the Rissian have led to different nomenclatures in the different regions of the Northern Alpine Foreland (NAF). Various dating approaches have led so far to often unsatisfying results. In this study we successfully dated the "High Terrace Gravels" of the Rissian type section. Using single grain feldspar luminescence we were able to evade the problems of incomplete bleaching in fluvioglacial sediments. Using the post Infrared IRSL protocol (at 225 °C) we were able to show that the feldspar in the research area only shows low rates of anomalous fading. We were able to show that these low fading rates have an influence on the age determination. This is remarkable because many of the signals were close to saturation. The conclusive age range of 149 ± 15—179 ± 17 ka corresponds to Marine Isotope Stage 6. Defining an age for the Rissian is a first step to consolidate the heterogeneous nomenclature and to reconstruct the chronology of past major Alpine glaciations.

© 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND

license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

The Northern Alpine Foreland (NAF) has been a key region for Quaternary research since Penck and Brückner (1909) established their concept of the "glacial series" which consists of four alpine glaciations in the area corresponding to a set of four terraces primarily defined as "Older Cover Gravels" (German: "Altere Deckenschotter", associated with the Günzian), "Younger Cover Gravels" (German: "Jüngere Deckenschotter", associated with the Minde-lian), "High Terrace Gravels" (German: "Hochterrassenschotter", associated with the Rissian) and "Lower Terrace Gravels" (German: "Niederterasseschotter", associated with the Würmian), which were each deposited during the respective glacial cycle. The basic concept of the glaciations and the resulting terrace formation has since been extended by three more glacials (Biber, Donau, Haslach) by Eberl (1930), Schafer (1952) and Schreiner and Haag (1982), but not fundamentally challenged. However, the timing of the sedimentation and subsequent erosion is in ongoing debate (Ellwanger et al., 2011). Assignment of the terrace sediments to a specific

* Corresponding author. E-mail address: e.f.rades@gmail.com (E.F. Rades).

glacial was mostly achieved by morphostratigraphic correlation and relative dating which is often not trivial, because especially the older terraces show only poor preservation. Numerical dating of the sedimentation of these landforms would be a tremendous help to clarify the status of key sites in the NAF, providing a chronological framework for the system of glacials and interglacials in this region.

Studies presenting numerical ages from the central NAF are scarce (Doppler et al., 2011; Ellwanger et al., 2011). Different methods were used to assess the age of the sediments including U/ Th-dating, radiocarbon dating, cosmogenic nuclide burial dating and ESR dating (Jerz and Mangelsdorf, 1989; Hauselmann et al., 2007), but mostly luminescence dating (Rogner et al., 1988; Fiebig and Preusser, 2003; Klasen et al., 2006, 2007, in press; Klasen 2008; Rentzel et al., 2009; Fiebig et al., 2014; Lowick et al., 2015; Salcher et al., 2015; Schielein et al., 2015; Bickel et al., 2015a, 2015b). The ages presented in these studies do not all yield consistent results, which may be caused by the limitations of the individual methodological approaches. However, Fiebig and Preusser (2003) dated fluvial terrace sediments from the Ingolstadt area and showed that correlating terrace deposits in the central NAF may not be straight forward. One of the most comprehensive studies yet, is that of Klasen et al. (in press) who tried to characterize the quartz and feldspar signals from various

http://dx.doi.org/10.1016/j.quaint.2016.07.055

1040-6182/© 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

E.F. Rades et al. / Quaternary International xxx (2016) 1—13

sites of the Northern Alpine Foreland and concluded that quartz and feldspar luminescence is highly challenging and the dating results from the area remain questionable. For the main site reinvestigated in this study, Klasen et al. (in press) determined two ages: a quartz OSLage of 72.1 ± 8.2 ka and an IRSLage of 173 ± 15 ka (not corrected for fading). However, the authors point out that the reliability of both ages may be questionable because of methodological reasons, possibly resulting in age underestimation of quartz based ages caused by the occurrence of unstable signal components, and a possible age overestimation of feldspar based ages because of incomplete bleaching.

However, Klasen et al. (in press) conclude that future studies in the research area should rather focus on feldspar as a dosimeter and should consider single grain techniques as a viable tool for dating. This is exactly the issue we want to address in this study, aiming at providing a more reliable numerical dating framework for the central NAF. For this purpose we selected sediments from the gravelpit

exposed in the gravelpit Scholterhaus are the "High Terrace Gravels". These sediments as well as both terminal moraines are associated with the maximum extent of the Rissian ("Doppelwallriss"). Exposed in a nearly 40 m high profile they mostly consist of matrix supported gravels and therefore are not suitable for luminescence dating (Figs. 3 and 4). In an elevation between 561 and 563 asl. a diamicton layer of glacial origin is present (Figs. 3 and 4). Sand lenses can be found at few locations in the profile with thicknesses between 0.1 and 0.6 m. The sand proved to consist of grain size fractions which are suitable for luminescence dating. Samples were taken from four different sand lenses between 550 and 570 m asl. (Fig. 4; Table 1) using steel tubes. Samples SHSn 1 —3 are located below the diamicton, while sample SHSn-4 is located above it. Material directly surrounding the sampled material was taken for gamma spectrometry analyses to determine the radiation by naturally occurring radionuclides. Fig. 4 also shows the relative position of the sample analysed by Klasen (2008).

Table 1

Location of the samples, external and internal dose rate.

Field-ID VLL-ID

Latitude Longitude WGS 84

Elevation of Depth below sample surface

Grain size fraction

Radionuclide concentration

Cosmic dose rate

Total dose rate

K content

(Bq/kg)

(Bq/kg)

(Bq/kg)

(mGy/ka) (Gy/ka)a (W%)

Gravelpit Scholterhaus

SHSn-1 VLL-0173-L 9.7897 48.1108 554 27

SHSn-2 VLL-0174-L 9.7905 48.1122 559 27

SHSn-3 VLL-0175-L 9.7905 48.1122 559 25

SHSn-4 VLL-0176-L 9.7879 48.1105 574 14.5 Gravelpit Gärtner

GÄRn-1 VLL-0172-L 10.3900 48.4418 475 6.5

200—250 12.76 ± 0.34 9.16 ± 0.31 239.5 ± 5.3 13.0 ± 1.3 1.9

200—250 14.11 ± 0.37 10.41 ± 0.35 229.0 ± 5.1 13.0 ± 1.3 1.9

200—250 16.00 ± 0.12 11.84 ± 0.39 227.1 ± 5.0 15.0 ± 1.5 1.9

200—250 17.79 ± 0.39 12.61 ± 0.39 257.1 ± 5.6 34.0 ± 3.4 2.0

13.37 ± 0.42 13.49 ± 0.36 13.56 ± 0.20 13.42 ± 0.37

150—250 27.29 ± 0.65 26.8

: 0.76 349.4 ± 7.6 81.0 ± 8.1 2.7

a For all samples, a water content of 15 ± 10% was assumed following Bickel et al. (2015a,b) were measured. The calculation of the dose rate was carried out using the software ADELE (Kulig, 2005) which does not provide individual errors for the dose rates, but they are included in the error of the final age calculation.

Scholterhaus, a type section of the Rissian (Penck and Brückner, 1909), which was also one of the sites sampled by Klasen et al. (in press). The sediments of this section can conclusively be tied to the penultimate glaciation, owing to the fact that they are located between two terminal moraines of the Rissian. In this study single grain feldspar luminescence dating is applied for the first time to the "High Terrace Gravels" of the NAF to clarify the timing of the Rissian.

2. Geological setting and sampling

2.1. Lake Constance Area and the gravelpit Scholterhaus

During past glaciations large ice masses built up in the European Alps extending wide into the alpine foreland. Based on glacial landforms and glacial sediments the extent of these glaciations was reconstructed (Fig. 1). One of the biggest foreland glaciers was the Rhineglacier extending north of the alpine front into the NAF (Fig. 1). It had a great impact on the Lake Constance Area. Each glacial advance incised deeper, lowering the hydraulic base level and shaping an amphitheatre like structure in the glaciated area. In the proglacial area the rivers incised to deeper levels with each glaciation which lead to the development of a terrace staircase. These distinct terraces can be correlated over long distances and have widely been used to build relative chronologies in the NAF.

The gravelpit Scholterhaus (Biberach am Riß) is located in the northeastern margin of the area once covered by the Rhineglacier and therefore in direct proximity to the glacier forefront (Fig. 1). The gravelpit is located between two terminal moraines associated with the penultimate glaciation, i.e. the "Rissian" (Fig. 2). The sediments

The sedimentological characteristics of the sand lenses sampled in the Scholterhaus section (cross bedding, coarse grain size: mainly medium to coarse sand) point towards a depositional environment with high sedimentation rates and rapid deposition. Such characteristics may strongly enhance the chance of incomplete resetting of the luminescence signal prior to deposition. For that reason we chose to include an additional sample for methodological comparison from another section. Sample GARn-1 was taken from a sand lens from the gravelpit Gärtner (Fig. 3). The gravelpit Gärtner is located near Burgau in the NAF which is located ~60 km NE of Scholterhaus. The gravelpit is not located in the Riss but in the Mindel valley. The reason why we chose this sample as a comparison is that although the general depositional environment is most probably similar to the Scholterhaus site, the sediment from this sand lens is finer and much better sorted than the sediments of the Scholterhaus samples. The finer grain size composition and the bedding structure suggest that a local aeolian re-deposition may have taken place in the drying braided river system. Aeolian transport, even over short distances, increases the chances of bleaching tremendously because the light can reach individual grains much easier. The drawback of the Gaärtner sample from a stratigraphical point of view is, that the deposit where the sample was taken from is generally associated with the "Lower Cover Gravels" and correlated with the Mindelian glaciation (Habbe and Rägner, 1989). We are aware that sample GARn-1 as a single sample only has restricted possibilities to reliably date the sediments from gravelpit Gaärtner, therefore we stress the point that this sample is intended to function as a control sample from a methodological point of view. A more suitable sample from the Riss valley could unfortunately not be obtained.

E.F. Rades et al. / Quaternary International xxx (2016) 1—13

Mindel-Moraine Doppelwall Wurmian-Moraine tectonic cross section Riss-Moraine Alpine Border Risstal (Fg. 2)

Fig. 1. Map showing the advances of the Rhine glacier in the Lake Constance area. The red triangle shows the location of Biberach where the gravelpit Scholterhaus is located. The inset shows the glacial extent during the last glacier maximum in the area of the European Alps and the location of the larger map. The exact positioning of the two Rissian moraines in respect to the gravelpit Scholterhaus are shown in Fig. 2. Modified after Penck and Bruckner (1909) and Ehlers and Gibbard (2004) (inset). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2. Cross section through the sediments of the Scholterhaus area (changed after Schreiner and Haag, 1982). In green the two terminal moraines of the Rissian glaciation are shown. The terminal moraine of the Mindelian glaciation is marked in yellow. The dashed lines show the inclination of the sedimentary layers to better correlate the different terrace levels of "High Terrace" and "Younger Cover Gravels". An idealised profile of the Scholterhaus profile (Fig. 4) is shown to indicate the position between the Rissian Moraines (green). On the right side of the profile an idealised profile of the gravel pit Gartner is given, from which sample GARn-1 was taken. The sediments of gravel pit Gartner are correlated to the Mindelian in which the third moraine (yellow) was deposited please note that the deposits are located -60 km to the NE of gravelpit Scholterhaus. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 3. (a) Overview picture of gravelpit Scholterhaus. Red arrows indicate the positioning and elevation of the samples in the gravelpit. The location of the diamicton is also identified for better correlation with Fig. 4. (b—d) Close-ups showing the sand lenses from which samples SHSn-1(b), SHSn-2 (b), SHSn-3 (c) and SHSn-4 (d) were taken. (e) Sand lens in the section of gravelpit Gärtner from which samples GARn-1 was taken. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

2.2. The Rissian and its stratigraphical concepts

The concept of a "Rissian" glaciation has first been proposed by Penck and Brückner (1909) as the name for the penultimate glaciation (Fig. 1). The type locality for the Rissian is the gravelpit Scholterhaus located in the city of Biberach (Baden-Württemberg, Germany) at the side of the river Riß, from which the name of the glaciation originates. Today the Rissian is used in different ways in the countries bearing sediments associated with this period of time. In Ellwanger et al. (2011) the official stratigraphic definitions from Baden-Württemberg are summarised. The Rissian is chronologically confined by the Eemian and Holsteinian interglacials and the sediments corresponding to this glacial period are subdivided using marker horizons defining three glacial phases ("Innenwall-Riss", "Aussenwall-Riss" and "Jüngere-Riss"). The official definition of the Bavarian Landesamt, defines the "Riß" as the time period between the "Mindel/Riß Interglacial" and the "Riß/Würm Interglacial", which correspond to the Eemian and the Holsteinian interglacials as used in Baden Württemberg. The Riß is again divided three distinct periods "Older Riß", "Middle Riß" and "Younger Riß" (Doppler et al., 2011). These three Riß periods do not entirely correspond to the Baden-Württembergian definitions with respect to their timing. The authors of the Bavarian study state that "a classification accepted by everyone and applicable to all 'Riß'-de-posits is still lacking" (Doppler et al., 2011). For the adjoining area in

Austria the sediments are assigned to the respective glaciations by correlation of the terrace bodies as in the original study by Penck and Brückner (1909). This proposition is more complicated in the Austrian part of the NAF because of the higher tectonic activity in this area. In Austria the Rissian is also correlated with Marine Isotope Stage (MIS) 6. This view was recently strengthened by two luminescence dating studies by Bickel et al. (2015a, 2015b), but directly correlating the respective ages to the rest of the terraces in the NAF over longer distances without further studies is rather presumptuous. The Austrian part of the NAF is the most eastern one, while the Swiss part is on the western side of the Alps. In Switzerland the usage of the classical Penck and Brückner (1909) nomenclature has been challenged since Schlüchter (1988) proposed at least 15 glaciations during the Quaternary in the Swiss Lowlands. The most current terminology consists of a series of five glaciations (Moählin, Habsburg, Hagenholz, Beringen and Birrfeld), which cannot yet be correlated to other classifications (Graf, 2009; Keller and Krayss, 2010; Preusser et al., 2011; Lowick et al., 2015). Therefore the Rissian is not being used in the latest literature regarding the Swiss part of the NAF.

3. Luminescence dating

For luminescence dating the commonly found minerals quartz and feldspar are used as natural dosimeters in sediments. While

E.F. Rades et al. / Quaternary International xxx (2016) 1—13

Gravelpit Scholterhaus

sampling Klasen et al. (in press)

m asl.

• D ' ' O . . o

•0 • ° \ O

SHS m S

Gm m S

mS-lenses

m asl.

new sampling 2014 F S G

o • • o" 0

. A- ' • •—

- O " -

o ' "o •• 1 D ' ' V Oo°o0o°o00o0oooo • • • • • •

o 00 0 0° /

■ 1 • / »

O • O o o •

SHSn 4

SHSn 1-3

Fig. 4. Profiles of the gravelpit Scholterhaus. The left profile was recorded during a first sampling by Klasen (2008). The sampling location by Klasen (2008) is indicated. The second profile shows the situation during the sampling in 2014. The prominent band of diamicton can also be seen in the photograph of the section (Fig. 3).

isolated from light the natural radiation from the surrounding sediments and cosmic radiation build up a signal which can be extracted from the mineral grains stimulating them with electromagnetic radiation of specific wave length (e.g. blue light or infrared light, respectively). This is done for subsamples called aliquots. Using silicon oil to fixate the grains on the sample discs, grain assemblages of various sizes can be measured. Single aliquot

measurements can consist of some hundred grains or less, depending on the radius chosen for the area sprayed with silicon oil. The smallest aliquot size used is single grains. Known laboratory doses are then administered to the aliquots and the corresponding luminescence signals are measured afterwards. The measured samples will lie on a line which can mostly be fitted to a linear or an exponential model. The fitted model can subsequently

E.F. Rades et al. / Quaternary International xxx (2016) 1—13

be used to convert the natural signal into an equivalent dose (De) (Fig. 5). The natural occurring radiation measured from a second sample consisting of the surrounding sediments, together with the cosmic and possible internal radiation, are finally used to calculate the dosage per time to which the sample was exposed. Dividing the De by this environmental dose rate will result in the time since deposition. A crucial perquisite for all luminescence dating approaches is that the signal was thoroughly reset by daylight during the transport prior to deposition, to ensure that only signal build up during the last depositional cycle is measured. The problem that occurs if this perquisite is not met and the signal is only partially bleached is called incomplete bleaching. Incomplete

Dose [s]

• REG points

O REG point repeated O REG point 0

• Natural signal

Single aliquot growthcurve SHSn-4 De= 4216.35 ± 121.87 | fit: EXP

• REG points

O REG point repeated O REG point 0

• Natural signal

Single grain growthcurve SHSn-4 De= 2990.67 ± 141.6 | fit: EXP

6000 Dose [s]

Fig. 5. Examples for two growth curves of sample SHSn-4. The normalised signal is plotted on the y-axis. The x-axis represents the given dose in seconds. The upper growth curve is typical for a single aliquot measurement of the samples in this study. The natural signal plots in the upper part of the growth curve and is in saturation. The lower growth curve is typical for the single grain measurements were most natural doses had lower intensities of the signal compared to the single aliquot measurements (Fig. 7).

bleaching is known to be a problem especially for glacial and river sediments because during transport the different grains can shield each other from the sunlight, leading to different times of exposure for each grain. This problem is more frequently associated with feldspar luminescence because the feldspar signal takes much longer to be reset than the quartz signal (Godfrey-Smith et al., 1988; Murray et al., 2012), especially when using a post infrared elevated temperature protocol (Kars et al., 2014). To counter the effects of incomplete bleaching different statistical age models can be applied (Galbraith et al., 1999; Galbraith and Roberts, 2012). These models rely on the statistical overdispersion parameter determined for the population of equivalent doses derived from the measurements of the single aliquots for each sample. Due to the incomplete bleaching a high scatter of equivalent doses is to be expected for the same sample. Different statistical age models are available for various types of possible depositional scenarios (c.f. Section 3.6).

3.1. Sample preparation

All samples were processed in the Vienna Laboratory for Luminescence Dating (VLL). The light exposed ends of the samples from the steel tubes were discarded. The rest of the samples was dried and then sieved. For further treatment we selected the fraction between 200 and 250 mm for all samples except sample GARn-1 which showed much finer grains than the other samples. For this sample we widened the used fraction to 150—250 mm. The samples where then treated with 10% hydrochloric acid and 15% hydrogen peroxide to remove all carbonate and organic residues. To disperse possible clay aggregates and remove them, the samples were also treated with Sodium oxalate. The potassium rich feldspar fraction (p < 2.58 g/cm3) and the quartz rich fraction (2.62 g/cm3 < p < 2.70 g/cm3) were separated using LST fastfloat. Before measurement the quartz rich fraction was etched for 40 min using 40% Hydrofluoric acid to remove the outermost layer affected by alpha radiation and to remove possible remaining feldspar contamination. After etching, the quartz fraction was rinsed with Hydrochloric acid to remove fluorides from the previous preparation step. The samples were dry sieved after etching for a second time to keep the same fraction, which is essential for the correct dose rate calculation. The feldspar rich fraction was not etched. For single aliquot measurements a monolayer with a diameter of 1 mm (~10 grains) was applied on discs using silicon oil to fix the grains.

All measurements in this study have been conducted on a Ris0 TL/OSL DA-20 reader with an attached 90Sr/90Y beta source (B0tter-Jensen et al., 2000, 2003; Thomsen et al., 2006). The uniformity of the beta source intensity was tested using batch 101 of the Ris0 calibration quartz, confirming that the source is oriented horizontally within error and creates a homogenous radiation field. IR stimulation of the feldspar single aliquots was carried out using IR-diodes (870 nm), single grain measurements were carried out using an IR-laser diode emitting at 830 nm. The signals were detected using a photomultiplier tube and a LOT/Oriel 410/30 interference filter.

3.2. Quality tests

Previous work (Klasen, 2008; Bickel et al., 2015a, 2015b; Klasen et al., 2007, 2006, in press) has shown that dating sediments of the "High Terrace Gravels" is not trivial. Problems with missing fast component and thermal instability in quartz, as well as failed preheat plateau tests and incomplete bleaching in feldspar were ubiquitous. Therefore we employed a series of quality tests to ensure the quality of the used signal. The recycling ratio limit for the single aliquot measurements was at <10% (quartz/feldspar) and

for single grain measurements at <20% (feldspar). The recycling ratio is a comparison of the signal of two measurements of the same given dose at the beginning and at the end of the measurement. The threshold for the recuperation ratio was set to <5% for single aliquot measurements and <10% for all single grain measurements. The recuperation compares the measurement of the

sunlight for ~2.5 weeks on the windowsill. The residual dose was lowest for pIRIR150 (~10 Gy) and doubled for the pIRIR225 (~20 Gy) (Table 2) and highest for pIRIR290 (~40 Gy), whereas most of the pIRIR290 measurements failed due to bad recycling. Mean recycling and recuperation ratios for the final single grain measurements are shown in Table 2.

Table 2

Values obtained for the different quality test measurements.

Field-ID

Residual

Recycling ratio (SG)

Recuperation

ratio (SG) [%]

Fading

rates IR50 [%]

Fading

rates pIRIR225 [%]

recovery (SG)

Overdispersion from

dose recovery (SG) [%]

Overdispersion from

CAM (SG) [%]

Gravelpit Scholterhaus

SHSn-1 — SHSn-2 — SHSn-3 20.59 ± 0.75 SHSn-4 — Gravelpit Gartner

GARn-1

19.2 ± 1.6

0.97 ± 0.12 1.02 ± 0.15 0.97 ± 0.10 0.98 ± 0.11

0.99 ± 0.11

6.71 ± 3.15 5.34 ± 2.54 4.21 ± 3.43 6.05 ± 3.78

4.88 ± 3.30

0.9 0.2 1.9 1.6

0.2 0.2 0.2 0.2

0.1 0.1 1.2 0.5

0.2 0.4 0.2 0.2

2.0 ± 0.3

1.2 ± 0.2

0.93 ± 0.24

0.75 ± 0.20 1.04 ± 0.37

0.92 ± 0.20

pIRIR225 26.83 ± 2.98

19.1 ± 1.74 30.54 ± 6.3

23.28 ± 2.15

pIRIR225

natural signal to the measurement of a zero dose measurement. A dose recovery preheat plateau test for quartz was conducted on sample BITn-3 using temperatures of 180, 200, 220, 240 and 260 °C (given dose 220 Gy). The measuring temperature was 125 °C. The best recycling ratios and dose recovery could be achieved using a preheat of 240 °C. To find the best measurement temperature for the post infrared infrared (pIRIR) measurements we measured single aliquots from samples SHSn-3 and SHSn-4 at elevated temperatures of 150, 225 and 290 °C (Thomsen et al., 2008; Buylaert et al., 2009; Reimann and Tsukamoto, 2012). The tests revealed that the samples show the best recuperation and recycling at an elevated temperature of 225 °C while the signals at the other two temperatures mostly failed the quality criteria. A dose recovery test using single aliquot discs yielded good results with a mean recovery ratio of 1.03 ± 0.12 for the pIRIR225 protocol and 0.99 ± 0.06 for the corresponding infrared stimulated luminescence (IRSL) signals at 50 °C taken from the same measurement (Fig. 6). For single grain measurements dose recovery was tested measuring one SG-disc each. Measurements were conducted with a given dose of ~420 Gy (Fig. 6). The test revealed a possible slight dose underestimation for the single grain pIRIR225 (0.91 ± 0.12) and for the single grain IR5o signal (0.88 ± 0.09). Residual dose measurements were carried out using three aliquots for each of the five samples (SHSn-1-4, GARn-1). The prepared aliquots were exposed to natural

Single - »Single aliquot grain - +10 %

- - ^T0"%

—i-1-1-1-1—

SHSn-1 SHSn-2 SHSn-3 SHSn-4 GÄRn-1

Fig. 6. Dose recovery test showing the normalised equivalent doses for single aliquot and single grain measurements. For Sample SHSn-2 no single grain dose recovery was conducted because of the low number of grains yielding a signal.

3.3. Dose rate determination

Environmental dose rate was calculated from samples of the sediment surrounding each individual sample. All samples were stored for at least one month to ensure equilibrium conditions for the 226Ra-222Rn decay (Murray et al., 1987). The sediments were filled in 500 ml Marinelli beakers and the 40K, 238U and 232Th content was measured using high resolution gamma spectrometry with a Canberra germanium coaxial detector (40% n-type). Dose rate conversion and age calculation were executed using the software ADELE (Kulig, 2005) which uses the dose rate conversion factors of Adamiec and Aitken (1998) and calculates cosmic dose following Prescott and Hutton (1994). While the quartz was etched with 40% HF to remove the outer alpha-irradiated layer, the alpha irradiation for feldspar minerals was calculated using an alpha efficiency value of 0.07 ± 0.02 following Klasen (2008).

To estimate the environmental dose rate for the single grain measurements of feldspar the internal 40K value of the single grains is crucial. Previous studies have shown high differences in the K content of individual feldspar grains ranging between 6 and 14% (Barre and Lamothe, 2010; Huot and Lamothe, 2012; Neudorf et al., 2012; Smedley et al., 2012; Trauerstein et al., 2014). 40K is the dominating source of internal radiation of Feldspar. Although, in most cases the effect of internal radiation is small compared to the external radiation variations of the K content can lead to over and underestimation in the calculated ages. Therefore we conducted measurements on single grains for some samples using the microprobe. The measurements were conducted on two different types of microprobe samples. We embedded feldspar grains of the sample fraction used for the measurements into resin (Korapox 439) to measure the variability of the K content of the single grains. Furthermore we fixed actually measured grains from a single grain disc of sample SHSn-1 using the same resin. We fixated the grains in the same order as on the SG-disc to assign the correct signal to each grain. For this purpose we put a strong adhesive tape onto the disc and treated the samples by alternatively heating it up to 80 °C and putting it on a shaking table while cooling down. This way most grains could be recovered and the spatial reference could be kept. The results of the measurement showed that most of the grains consist of two intergrown, but clearly separated feldspar species. Both species can optically be separated on microprobe pictures by their differences in colour, the most abundant feldspar species being nearly white and the second species light grey

E.F. Rades et al. / Quaternary International xxx (2016) 1—13

(Fig. S1). Several of the analysed grains are white throughout. This most abundant feldspar is a potassium endmember with a very consistent K content of 13.42 ± 0.39% (Table 1) and Na or Ca close to zero. The other fraction shows a Na content of 8.49 ± 0.84% and a Ca and K content close to zero making it a sodium feldspar end-member. As the signal of potassium feldspar is normally assumed to be dominant we expect the measured luminescence signal to be emanating from this part of the grains. Comparing the measured luminescence intensities of sample SHSn-1 gave no evidence that changes in the K-feldspar to Na-feldspar had any influence on the emitted signal strength, confirming that the signal of the K-feldspar dominates. Furthermore, the De single grain distributions showed no evidence for two separate populations (Fig. 7). This should be the case if both feldspar species emitted a similarly intense signal, because the grains where close to or in saturation and the different feldspar species would have different 2D0 values.

3.4. Fading tests and implications

A fading test after Lamothe et al. (2012) was conducted for all four samples using 8 aliquots per sample. The Samples were measured using different intervals between irradiation and measurement with a final measurement of at least 7 weeks after the first measurement. The natural signals of our measurements are always plotting beyond the linear part of the growth curve and therefore the g-values obtained cannot reliably be used to correct the ages, but they show that our pIRIR225 measurements (g-values between 0.1 ± 0.2% and 1.2 ± 0.2%) display no significant fading (Table 2). SHSn-1 and SHSn-2 also show low values of 0.9 ± 0.2% and 0.2 ± 0.2%, respectively, in the IR5o measurements from the same sequence. However, SHSn-3, SHSn-4 and GARn-1 show g-values between 1.6 and 2.0 and therefore should show a significant loss of signal with time. On single grain level no fading measurement was conducted.

Table 3

Protocol used for single grain measurement IR50 (step 3 and 7) and pIRIR225 (step 4 and 8).

Step Treatment Observed

1a Give dose

2 Preheat 250 °C for 60s

3 Stimulation with IR for 2 s at 50 °C LX (IRSL50)

4 Stimulate with IR for 2 s at 225 °C LX (pIRIR225)

5 Give test dose ~12/36Gy

6 Preheat 250 °C for 60s

7 Stimulation with IR for 2 s at 50 °C TX (IRSL50)

8 Stimulate with IR for 2 s at 225 °C TX (pIRIR225)

a 0 Gy, -60 Gy, -240 Gy, 600 Gy, 1200 Gy.

For all samples we evaluated the approach of using only the 30% brightest grains proposed by Reimann et al. (2012). They proposed that the brightest grains should be the ones with the most stable signal and therefore showing the least fading. This test did not shift the overall distribution of the De-values. Using only 30% of the measured grains would hence only lower the number of Des which can be used for the statistical age models, lowering their significance.

3.5. Luminescence measurements

As fluvioglacial sediments are quite often incompletely bleached (Luthgens et al., 2011, 2010; Bickel et al., 2015b), we first tested quartz as a dosimeter. After a first test revealed some good signals for 2 mm aliquots for sample GARn-1 we measured 48 aliquots of GARn-1 and SHSn-3. However, this measurement only gave few usable signals (five and one signals respectively). Other signals detected showed feldspar contamination or were too dim to successfully build growth curves. Therefore quartz measurements were disregarded. For feldspar we used a modified pIRIR protocol at 225 °C (Table 3) after Buylaert et al. (2009) which showed the best values for the quality criteria (Table 2). For this protocol we tested two different test doses, because the first measurements suggested that the test dose of ~12 Gy was possibly not high enough and lead to a higher number of recycling ratios above the threshold. Then a higher dose of ~36 Gy did not noticeably alter the number of growth curves that could be established nor did it show any effect on the equivalent dose calculation. However, we considered the higher test dose to be more reliable and therefore it was used thereafter. The single aliquot measurements gave a lot of natural signals plotting above 2Do (Murray and Wintle, 2000) and therefore in the upper area of the growth curve which is unreliable for equivalent dose calculation and indicating a saturated feldspar luminescence signal (Fig. 5). Analogue to quartz grains (Duller, 2012) saturation of feldspar grains is probably not reached at the same dosage for each grain, which will lead to a high variability of Des (Table 4).

We decided using single grain measurements to inspect the dose distribution in more detail. The quantity of luminescent grains on each disc varies a lot, e.g. while on the first disc for sample SHSn-1 17 grains out of 100 gave a measurable signal on another disc only 5 growth curves could be established for 100 grains measured.

Also on the single grain level we detected saturated signals for both IR5o and pIRIR225. To avoid a shift in the distribution towards the smaller Des we included as many values as possible, even if they plotted above 2D0. However, for signals plotting on the upper linear

Table 4

Doserates from single aliquot and single grain measurements calculated using different statistical models and final calculated ages.

Field-ID Equivalent Equivalent dose Equivalent dose Equivalent dose Equivalent dose Agec CAM (SG)

dose (SA) CAM (SG) MAMa (SG) CAM (SG)b MAMa (SG)b [ka]

[Gy] [Gy] [Gy] [Gy] [Gy]

Gravelpit Scholterhaus IR50 PIRIR225 IR50 PIRIR225 IR50 PIRIR225 IR50 PIRIR225 PIRIR225

SHSn-1 702 - 207 386 ± 13 333 20 267 21 196 z 23 325 ± 24 364 26 214 32 238 37 179 ± 17

SHSn-2d 557 - 125 347 ± 21 303 41 285 24 225 60 — — — — 161 ± 25

SHSn-3 716 244 417 ± 21 310 18 321 44 203 28 374 ± 27 360 28 316 53 277 44 161 ± 15

SHSn-4 542 126 366 ± 15 302 21 262 27 167 z 23 330 ± 26 271 26 241 42 165 - 29 149 ± 15

Gravelpit Gärtner

GÄRn-1 667 t 154 438 ± 24 389 - 23 343 49 385 34 325 ± 24 412 49 214 94 298 75 146 ± 15

a Minimum age models were calculated assuming an overdispersion of 35%.

b These Des only take into account signals from grains which are present in the IR50 and pIRIR225 signals.

c The calculation of the dose rate was carried out using the software ADELE (Kulig, 2005) which does not provide individual errors for the dose rates, but they are included in the error of the final age calculation. d For Sample SHSn-2 only 12 grains gave a signal in the pIRIR225 therefore the De-values are not considered highly reliable.

E.F. Rades et al. / Quaternary International xxx (2016) 1—13

n = 92 | skewness = 0.71

n = 12 | skewness = 0.31

"I—I—I—r

0 400 800 De (Gy)

n = 70 | skewness = 1.02

"I—I—I—T

0 400 800 De (Gy)

n = 44 | skewness = 1.38

"I—I—I—T

0 400 800 De (Gy)

n = 79 | skewness = 1.44

o y 0

CD o 2

n o —

o +J —

<0 D 1

:3 o —

o ^ 0.

CM O —

— 0

— o o —

"1—T

400 800 De (Gy)

"I-1-1-r

0 400 800 De (Gy)

Fig. 7. pIRIR225 De-distributions of all samples. The Scholterhaus samples (SHSn-1-4) show a shift of the Gaussian distribution to the older ages, indicating incomplete bleaching. Sample GARn-1 shows a nearly even distribution only disturbed by some grains with higher Des.

part of the growth curve no accurate dose could be calculated. These values could not be included dose calculations. However, the percentage of those grains is between 15% for SHSn-1,10% for SHSn-4, and only 1% for SHSn-3. Sample SHSn-2 is at 30% of saturated

grains but sample has only a very low number of successfully established growth curves. These values are in the same limits as the IRSL single aliquot samples with a similar problem by Lowick et al. (2015). They state that the ages from these samples can only

be considered as minimum ages, but they have not compared IR50 and pIRIR225 data.

In a next step we compared the single grain De-values calculated from the IR50 signal and from the pIRIR225 signal (Table 4) and recognised an unexpected phenomenon: The Des derived from the IR50 for samples SHSn-3, SHSn-4 and GAoRn-1 are higher than those from pIRIR225 signals from the same samples (Table 2). Taking the results from the fading tests into account, we normally we would have expected the IR50 values to be lower than the pIRIR225 values due to the significantly higher anomalous fading of the IR50 signal. This effect cannot be observed for samples SHSn-1 and SHSn-2, which show only minimal fading in the IR50 signal. What could be the reason for that effect? We assume that for those samples showing low IR50 fading rates, more aliquots are in saturation, so that no equivalent dose can be calculated. As fading is also low for the pIRIR225 signal for these samples, the resulting De values for both IR50 and pIRIR225 are similar. For the samples with higher fading rates of the IR50 signal, less aliquots plot on the upper linear part of the growth curve, simply because they lose more signal owing to fading. However, the fading rates for the pIRIR225 signal of these samples is again very low, so that more aliquots plot in an area where no De can be calculated. This results in a truncation of the upper part only for the pIRIR225 equivalent doses. Because truncation is not evident for the IR50 signal because of stronger fading, the IR50 equivalent does are higher than those resulting from the truncated pIRIR225 De distribution. To validate this assumption we discarded all single grains for which we could only derive either the IR50 or the pIRIR225 signal. When comparing the new average equivalent doses (Table 4), the unexpected effect described above is not significant anymore. As originally expected, we now observe IR50 Des to be lower or in agreement within error as the corresponding values of the pIRIR225. This corresponds to the expected effect by fading of the IR50 signals.

3.6. Statistical equivalent dose modelling

De-values derived from the feldspar single aliquot measurements using IR50 and pIRIR225 almost entirely plot above the 2D0 value, following Murray and Wintle (2000) this indicates that the samples are in field saturation. The actual values of single aliquots show a high variation within the distribution. This variation in the single aliquots leads to high uncertainties for the overall Des ranging between 23% and 31% (Table 4). Comparing the single aliquot and single grain results, we only identified a very limited number of single grains resulting in similarly high equivalent doses as calculated for the single aliquots. However, it is very much likely that these very few individual grains dominate the averaged signal of the single aliquots. We regard these single grains carrying high equivalent doses as incompletely bleached grains. However, because the signals from these grains dominate the averaged signals from the single aliquots, the latter measurements are not suitable for age determination of the last burial event. Therefore we only used the results from single grain measurements for age determination. To calculate the average Des for the single grain measurements we compared different statistical approaches. First, we used the central age model (CAM) (Galbraith et al., 1999; Galbraith and Roberts, 2012) which is very robust against outliers. It assumes a Gaussian distribution for the acquired Des and calculates the most likely central value. The second model we applied is the three parameter minimum age model (MAM) which is commonly used for samples which are suspected to have been insufficiently bleached prior to deposition (Galbraith et al., 1999). This model tries to fit a Gaussian distribution to the lowest part of De-distribution assuming that higher values represent incompletely bleached signals. The calculation of the MAM highly

depends on the estimated overdispersion of the dose distribution. Therefore, we calculated the overdispersion values of the single grain dose recovery tests. These overdispersion values can be assumed to represent the minimally expected overdispersion for these samples (Fuchs and Lang, 2001), because we used a known dose (Table 2). The calculated overdispersion from the dose recovery is relatively high. However, Duller (2012) showed that the overdispersion for quartz grains increases with increasing age of the sample. This is probably the same for feldspar grains, which can explain the high overdispersion. Noticeably, the overdispersion calculated applying the CAM to the whole dataset is even higher (44-57%) for all Scholterhaus samples (Table 2), while sample GARn-1 shows a lower overdispersion (35%) than the Scholterhaus samples. This strengthens our initial assumption that the Schol-terhaus samples suffer from incomplete bleaching, while sample GARn-1 was reset to a higher level. We calculated all minimum ages using estimate threshold for the overdispersion of 35%. This value is higher than the overdispersion from the dose recovery and lower than the overdispersion from the CAM, at least for the Scholterhaus samples. For sample GAoRn-1 the overdispersion of 35% is equal to the central age overdispersion, therefore the De of the central age and the minimum age are basically the same reducing the MAM to a CAM for this sample.

As shown in Section 3.3 the pIRIR225 signals are most likely already truncated for the higher Des because of saturation effects. In that case, special care needs to be taken when statistical age models are applied. When using the MAM for an already truncated dataset, the model may likely result in underestimated equivalent doses. The naturally truncated distribution on the other hand represents a similar scenario as the distribution used for a Minimum age model, when assuming that those De values that were excluded from the distribution because of saturation effects are actually only saturated, because they were not sufficiently bleached in the first place. Looking at the shape of the dose distributions, all samples show a clear peak in the kernel density estimate on the lower end of the distribution, with higher equivalent doses forming a tail, resulting in right skewed dose distributions typically expected for incompletely bleached samples. Therefore, we propose that the equivalent doses and the resulting ages calculated using the CAM for the single grain De distributions in fact represent the well bleached fraction of grains and may in fact be regarded as a "natural minimum age model". Consequently we only calculated ages for the pIRIR225 using the CAM, which we think represents the best approximation for the most representative ages.

4. Discussion

4.1. Luminescence measurements

In Section 3.6 we explain why we think that the pIRIR225 CAM Des presumably give the best approximation to the depositional ages for the Scholterhaus samples. Strictly speaking as they have been calculated from truncated distributions, these ages have to be considered as minimum ages. However, the process, which is accountable for the overall distortion of the natural distribution, is incomplete bleaching and this is a completely random process and should result in inter-sample scatter in ages. The ages obtained from the pIRIR225 CAM Des of the Scholterhaus section range between 179 ± 17 ka and 149 ± 15 ka, which is the same taking a lo error into account. This agreement in age is a good argument for the validity of our age modelling approach. This could also be an effect of the common saturation level of the feldspar from the Scholter-haus site, but the Des calculated using the CAM are much lower than the calculated saturation levels of the samples. The coincidence of four samples of the same section to give the same age

based on a random process is expected to be quite low. Also the obtained ages all fit perfectly well into the MIS 6 which is assumed to be the timing for the Rissian glaciation as shown by previous studies from the Eastern NAF (cf. Section 2.2) and therefore the timing of sedimentation for the "High Terrace Gravels" (van Husen, 2004; van Husen and Reitner, 2011; Schielein et al., 2015; Bickel et al., 2015a, 2015b). Taking all these arguments into account, we conclude that the ages do not only represent minimum ages for deposition of the Scholterhaus sediments but represent the actual depositional age, although we are aware of the possible implications of the truncation effects discussed above. Our samples provide an age range for the deposition of the Scholterhaus samples of 149 ± 15—179 ± 17 ka. There is a trend in the data with the youngest age in the upper parts of the profile, above the diamicton and the oldest age in the lowermost sample. However, within 1s errors these ages are statistically the same. Therefore the age range is the highest resolution we were able to achieve and no definite statements can be given for e.g. sedimentation rates.

The age obtained for sample GARn-1 (146 ± 15 ka) is also located in this age range. The characteristics of the equivalent dose distribution for this sample suggest sufficient bleaching of this sample before deposition. However, as this sample is a single sample, fewer arguments can be established to rule out effects of signal saturation which would result in age underestimation when applying the CAM. The section where this sample was taken has previously been correlated with the "Younger Cover Gravels", which are associated with the Mindelian. The finding that the sample is not older than the Scholterhaus samples, but even the youngest of this study, is intriguing. As stated before, this sample seems to be better bleached as the other samples. Therefore, a truncation of the highest Des would lead to an age underestimation. However, the distribution of GARn-1 is more centred than the Scholterhaus distributions (Fig. 7). This is again an indication for better bleaching which verifies our choice to use the Gaärtner sample as a methodological control for glaciofluvial sediments of old age. To confirm the age of the Gaärtner section and its correlation with the "Younger Cover Gravels" more samples should be taken for further studies and studied using single grain feldspar luminescence dating.

4.2. Comparison with other dated sites

As mentioned before, not many studies have been conducted in the NAF dating sediments associated with glacial stages. Häauselmann et al. (2007) tested the application of cosmogenic nuclide burial dating to the "Younger Cover Gravels" and "Older Cover Gravels" near Memmingen in the Iller Valley which was also a major study locality of Penck and Brückner (1909). A direct comparison is difficult because only our control sample was sampled in sediments associated with the "Younger Cover Gravels". Our sample GARn-1 (146 ± 15 ka) is much younger than the age of 0.68 (+0.23—0.24) ka of Häuselmann et al. (2007). The high errors of the cosmogenic nuclide dating and the erosion rates questioned by the authors themselves (Haäuselmann et al., 2007) illustrate some of the problems dating the glacial sediments in the NAF. The possible age and composition of the sediments also hinder the appliance of other dating methods frequently used (e.g. radiocarbon-, U/TH dating, and so forth). However, luminescence dating seems to fit all the requirements for these sediments, and in fact some attempts have been made to do so. Roägner et al. (1988) made a first attempt using thermoluminescence dating on a loess like sediment associated to the "Younger Cover Gravels", obtaining minimum ages between 207 and 280 ka. Klasen et al. (in press) dated the "Younger Cover Gravels" from Offingen and Gaärtner to similar ages of 259 ± 24 ka and 122 ± 13—277 ± 24 ka, respectively. Still, these are fading uncorrected IRLS ages. The ages of both

studies have to be considered with caution because the thermoluminescence study of Rogner et al. (1988) is no longer state of the art anymore and Klasen et al. (in press) were very sceptical regarding the obtained ages. However, our sample from the same gravelpit Gärtner is in a matching age range and both studies of the same sediments show much younger ages as the study by Häuselmann et al. (2007).

Studies specifically dating the "High Terrace Gravels" are rare. In a luminescence study using a multiple aliquot IRSL protocol in the surroundings of Ingolstadt (Germany) Fiebig and Preusser (2003) dated sediments including those associated with the "High Terrace". Their study illustrates the importance of dating the sediments in the NAF using novel methodological approaches, because ages obtained for these sediments vary between ~66 and ~81 ka, which makes an association of these sediments to the penultimate glaciation very difficult. The ages of the study are not fading corrected and the age discrepancy to the ages obtained from the gravelpit Scholterhaus in this study could result from anomalous fading. However, if the feldspar from the Ingolstadt area shows signals comparable to the ones from our study, the more important factor should be incomplete bleaching, which leads to age over-estimation. This may suggest that the "High Terrace Gravels" dated by Fiebig and Preusser (2003) may not be correlated to the penultimate glaciation. This implicates that timing of the sedimentation of the "High Terrace Gravels" at different locations can be asynchronous.

Rentzel et al. (2009) reported feldspar luminescence ages between 190 and 245 ka for "High Terrace Gravels", associating them to MIS 7 or even MIS 9. The samples of Rentzel et al. (2009) were taken in Sierenz (France) in vicinity of the city of Basel (~180 km SWW of the Scholterhaus section). As mentioned above the "High Terrace Gravels" do not have to be deposited at the same time. However, in contrast to the current study the ages by Rentzel et al. (2009) are based on the classic single aliquot regenerative dose (SAR) IRSL approach while we used single grains and the pIRlR protocol. However, in our study we also conducted single aliquot measurements as a first approach. The De values of these measurements (Table 2) are in the same range as the Des from Rentzel et al. (2009) (please note that our De values show a much higher error because no CAM was applied to the data). Our single grain measurements show that the grains with the highest natural signal dominate the single aliquot measurements, even though we only used 1 mm aliquots. This means that the single aliquot measurements are probably overestimating the depositional age of the sample due to incomplete bleaching. Therefore, we suggest that most probably the ages of Rentzel et al. (2009) for the "High Terrace Gravels" are overestimating the depositional age due too poor bleaching of the samples.

In the eastern part of the NAF (>260 km east of Scholterhaus) Bickel et al. (2015a, 2015b) used quartz and lRSL50 to analyse samples of several profiles associated to the "High Terrace". The timing for the deposition of the "High Terrace Gravels" and therefore the penultimate glaciation is in error the same in both studies (e.g. 116 ± 17-142 ± 15 ka and 122 ± 18-210 ± 24 ka, respectively). The ages in average are bit lower, but in the same age range as our samples from the gravelpit Scholterhaus. Being able to use quartz OSL and the lRSL50 signals the effect of insufficient bleaching in feldspar luminescence was shown. Furthermore, the study shows the variance of the impact of this effect on individual samples. ln contrast to the samples of this study the quartz fraction mostly yielded good signals and at least some of the feldspar samples were completely bleached. The differences in the signal intensities between the Austrian NAF and Scholterhaus can most probably be attributed to different source areas. However, the accordance of the ages obtained for the "High Terrace" sediments within error over

such a long distance indicates that the timing of the glaciation and the sedimentation of the terrace sediments of the European Alps was synchronous and did not significantly vary spatially.

In another luminescence study Schielein et al. (2015) took samples in two sections in the Lech Valley (southern Germany) north of the city of Augsburg (~100 km NE of the Scholterhaus section). This area was also part of the original study by Penck and Brückner (1909) but the situation is more complex because of repeated aggradation and incision events leading to interlacing and superposition within the terrace sediment bodies. Due to their findings Schielein et al. (2015) divided the "High Terrace Gravels" in this area into two different phases of sedimentation. The upper part of the terrace was dated to ages ranging between 160 ± 15—178 ± 20 ka (Schielein et al., 2015), which coincides with our findings in Scholterhaus. The second, lower part of the terrace seemingly has no correspondence in the Scholterhaus section. In the first and the second section of Schielein et al. (2015) this older part is only dated by one SAR IRSL age each ranging between 219 ± 31 and 291 ± 51 ka. All the ages presented by Schielein et al. (2015) are fading uncorrected SAR IRSL ages. In the study, the authors state that fading is not relevant because the samples are not in the linear part of the growth curve and the quartz and feldspar data match closely. The current study shows that fading can be present and affect the outcome of the measurement, even though the natural signal plots this high on the growth curve. Furthermore, quartz measurements of the lower part are missing. However, these measurements are the authors main argument to discard possible difficulties caused by anomalous fading (Schielein et al., 2015). In the light of our study we would suggest to conduct further measurements of the lower terrace part using single grain feldspar luminescence dating in order to verify the division of the "High Terrace" in the lower Lech valley.

Klasen et al. (in press) dated one sample in the gravelpit Scholterhaus yielding an IRSL age of 173 ± 15 ka and an quartz OSL age of 72.1 ± 8.2. The feldspar age is within error the range of our findings, the quartz age is much younger. The dating of feldspar without trying pIRlR measurements when fading is detected is not state of the art anymore (the study is based on the results by Klasen (2008). In this preceding study Klasen (2008) also provided a fading corrected age for the sample which dated much higher to 300 ± 30 ka. The quartz age is discarded by the authors themselves due to the bad signal quality of the quartz. Compared to our plRlR225 ages there seems to be an overcorrection of the IRSL ages. This provides evidence that the correction after Huntley and Lamothe (2001) is not applicable for samples with natural signals on the non linear part of the growth curve. On the other hand our findings show that anomalous fading does change the distribution of the Des. The effects of anomalous fading and incomplete bleaching work antagonistically. Perhaps the resulting age is a combination of both effects, coincidentally leading to an age in the correct range. This antagonistic relationship was also pointed out by Lowick et al. (2015) in their from northern Switzerland, which is also part of the NAF. They found that this problem hindered the correct calculation of the burial dose using feldspar luminescence for samples >30 ka. Our current study has shown that single grain feldspar pIRlR dating can be a promising approach to solve this problem.

5. Conclusions

Using single grains and the plRlR225 protocol major difficulties in dating sediments of the NAF could be addressed and valuable new insights could be achieved. The "High Terrace Gravels" of the Rissian type section Scholterhaus could successfully be dated and a conclusive age range of 149 ± 15—179 ± 17 ka was established. The

dating of the Rissian type section allows to correlate the penultimate major glaciation with the MIS 6. The results of this study agree with multiple luminescence ages of the "High Terrace Gravels" in the NAF (Schielein et al., 2015; Bickel et al., 2015a, 2015b) and give possible explanations why other luminescence studies may have overestimated the age of the "High Terrace Gravels" (Rentzel et al., 2009; Schielein et al., 2015).

Our single grain dating clearly shows that the consequences of incomplete bleaching can easily conceal the real De in single aliquot measurements of potassium Feldspar. By identifying a truncation of the De-distribution due to the saturation of the luminescence signal and the resulting "natural minimum age" reliable dating results could be achieved. However, this model has some assumptions, which must be checked for every individual case. The biggest assumption being that the truncated distribution is not shifted towards higher or lower values. The effects of signal saturation discussed above may leave a last bit of doubt whether the ages my still have to be regarded as minimum ages.

Further studies are needed to date other type sections in the NAF to clearly define the timing of the different glacial advances in this area. We suggest that single grain feldspar luminescence dating combined with the pIRlR protocol currently provides the most promising tool to achieve this goal.

Acknowledgments

We thank Lukas Bickel and Johanna Lomax for the pre-work of this study and are grateful to Helene Pfalz-Schwingenschlögl for assistance with the drawings. We thank Federica Zaccarini for the microprobe measurements. This project was funded by the Austrian Science Fund (FWF): P23138.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.quaint.2016.07.055.

References

Adamiec, G., Aitken, M.J., 1998. Dose-rate conversion factor: update. Ancient TL 16, 37-50.

Barré, M., Lamothe, M., 2010. Luminescence dating of archaeosediments: A comparison of K-feldspar and plagioclase IRSL ages. Quaternary Geochronology 5, 324-328. http://dx.doi.org/10.1016/j.quageo.2009.03.004. Bickel, L., Lüthgens, C., Lomax, J., Fiebig, M., 2015a. Luminescence dating of gla-ciofluvial deposits linked to the penultimate glaciation in the eastern Alps. Quaternary International 357, 110-124. http://dx.doi.org/10.1016/ j.quaint.2014.10.013.

Bickel, L., Lüthgens, C., Lomax, J., Fiebig, M., 2015b. The timing of the penultimate glaciation in the northern Alpine Foreland: new insights from luminescence dating. Proceedings of the Geologists' Association 126, 536-550. http:// dx.doi.org/10.1016/j.pgeola.2015.08.002. Botter-Jensen, L., Andersen, C.E., Duller, G.A.T., Murray, A.S., 2003. Developments in radiation, stimulation and observation facilities in luminescence measurements. Radiation Measurements 37, 535-541. http://dx.doi.org/10.1016/S1350-4487(03)00020-9.

Botter-Jensen, L., Bulur, E., Duller, G.a.T., Murray, A.S., 2000. Advances in luminescence instrument systems. Radiation Measurements 32, 523-528. http:// dx.doi.org/10.1016/S1350-4487(00)00039-1. Buylaert, J.-P., Murray, A.S., Thomsen, K.J., Jain, M., 2009. Testing the potential of an elevated temperature IRSL signal from K-feldspar. Radiation Measurements 44, 560-565. http://dx.doi.org/10.1016/j.radmeas.2009.02.007. Doppler, G., Kroemer, E., Rögner, K., Wallner, J., Jerz, H., Grottenthaler, W., 2011. Quaternary stratigraphy of southern Bavaria. Quaternary Science Journal 60, 329-365. http://dx.doi.org/10.3285/eg.60.2-3.08. Duller, G.A.T., 2012. Improving the accuracy and precision of equivalent doses determined using the optically stimulated luminescence signal from single grains of quartz. Radiation Measurements 47, 770-777. http://dx.doi.org/ 10.1016/j.radmeas.2012.01.006. Eberl, B., 1930. Die Eiszeitenfolge im nördlichen Alpenvorlande.—Ihr Ablauf, ihre Chronologie auf Grund der Aufnahmen des Lech-und Illergletschers (Augsburg). Ehlers, J., Gibbard, P.L., 2004. Quaternary Glaciations Extent and Chronology - Part I: Europe, Developments in Quaternary Sciences.

Ellwanger, D., Wielandt-Schuster, U., Franz, M., Simon, T., 2011. The Quaternary of the southwest German Alpine Foreland (Bodensee-Oberschwaben, Baden-Württenberg, southwest Germany). E&G Quaternary Science Journal 60, 306—328. http://dx.doi.org/10.3285/eg.60.2-3.OT.

Fiebig, M., Preusser, F., 2003. Das Alter fluvialer Ablagerungen aus der Region Ingolstadt (Bayern) und ihre Bedeutung für die Eiszeitenchronologie des Alpenvorlandes. Zeitschrift für Geomorphologie, NF 449—467.

Fiebig, M., Herbst, P., Drescher-Schneider, R., Lüthgens, C., Lomax, J., Doppler, G., 2014. Some remarks about a new Last Glacial record from the western Salzach foreland glacier basin (southern Germany). Quaternary International 328—329, 107—119. http://dx.doi.org/10.1016/j.quaint.2013.12.048.

Fuchs, M., Lang, A., 2001. OSL dating of coarse-grain fluvial quartz using singlealiquot protocols on sediments from NE Peloponnese, Greece. Quaternary Science Reviews 20, 783—787. http://dx.doi.org/10.1016/S0277-3791(00)00040-8.

Galbraith, R.F., Rober, R.G., Laslett, G.M., Yoshida, H., Olle, J.M., 1999. Optical dating of single and multiple grains of quartz from Jinmium rock shelter, northern Australia: part I, experimental design and statistical models. Archaeometry 41, 339—364. http://dx.doi.org/10.1111/j.1475-4754.1999.tb00987.x.

Galbraith, R.F., Roberts, R.G., 2012. Statistical aspects of equivalent dose and error calculation and display in OSL dating: An overview and some recommendations. Quaternary Geochronology 11, 1—27. http://dx.doi.org/10.1016/ j.quageo.2012.04.020.

Godfrey-Smith, D.I., Huntley, D.J., Chen, W.H.H., 1988. Optical dating studies of quartz and feldspar sediment extracts. Quaternary Science Reviews 7, 373—380.

Graf, H.R., 2009. Stratigraphie und Morphogenese von frühpleistozanen Ablagerungen zwischen Bodensee und Klettgau. Eiszeitalter und Gegenwart 58,12—53.

Habbe, K.A., Rägner, K., 1989. The Pleistocene Iller glaciers and their outwash fields. Catena Supplement 15, 311—328.

Häuselmann, P., Fiebig, M., Kubik, P.W., Adrian, H., 2007. A first attempt to date the original "Deckenschotter" of Penck and Brückner with cosmogenic nuclides. Quaternary International 164—165, 33—42. http://dx.doi.org/10.1016/ j.quaint.2006.12.013.

Huntley, D.J., Lamothe, M., 2001. Ubiquity of anomalous fading in K-feldspars and the measurement and correction for it in optical dating. Canadian Journal of Earth Sciences 38,1093—1106. http://dx.doi.org/10.1139/e01-013.

Huot, S., Lamothe, M., 2012. The implication of sodium-rich plagioclase minerals contaminating K-rich feldspars aliquots in luminescence dating. Quaternary Geochronology 10, 334—339. http://dx.doi.org/10.1016Zj.quageo.2012.03.003.

Jerz, H., Mangelsdorf, J., 1989. Die Interglaziale Kalksinterbildung bei Hurlach näordlich Landsberg am Lech. Eiszeitalter und Gegenwart 39, 29—32.

Keller, O., Krayss, E., 2010. Mittel- und späatpleistozaäne Stratigraphie und Morpho-genese in Schlüsselregionen der Nordschweiz, vol. 59. http://dx.doi.org/ 10.3285/eg.59.1-2.08.

Kars, R.H., Reimann, T., Ankjaergaard, C., Wallinga, J., 2014. Bleaching of the post-IR IRSL signal: new insights for feldspar luminescence dating. Boreas 43, 780—791. http://dx.doi.org/10.1111/bor.12082.

Klasen, N., 2008. Lumineszenzdatierung glazifluvialer Sedimente im näordlichen Alpenvorland (Luminescence dating of glaciofluvial sediments in the Northern Alpine Foreland). Universitaät zu Koäln.

Klasen, N., Fiebig, M., Preusser, F., in press. Applying luminescence methodology to key sites of Alpine glaciations in southern Germany. Quaternary International. http://dx.doi.org/10.1016/j.quaint.2015.11.023.

Klasen, N., Fiebig, M., Preusser, F., Radtke, U., 2006. Luminescence properties of glaciofluvial sediments from the Bavarian Alpine Foreland. Radiation Measurements 41, 866—870. http://dx.doi.org/10.1016/j.radmeas.2006.04.014.

Klasen, N., Fiebig, M., Preusser, F., Reitner, J.M., Radtke, U., 2007. Luminescence dating of proglacial sediments from the eastern Alps. Quaternary International 164—165, 21—32. http://dx.doi.org/10.1016/j.quaint.2006.12.003.

Kulig, G., 2005. Erstellung einer Auswertesoftware zur Altersbes- timmung mittels Lumineszenzverfahren unter spezieller Berücksichtigung des Einflusses radioaktiver Ungleichgewichte in der 238 U-Zerfallsreihe. Technical University Bergakademie Freiberg.

Lamothe, M., Barre, M., Huot, S., Ouimet, S., 2012. Natural luminescence and anomalous fading in K-feldspar. Radiation Measurements 47, 682—687. http:// dx.doi.org/10.1016/j.radmeas.2012.04.018.

Lowick, S.E., Buechi, M.W., Gaar, D., Graf, H.R., Preusser, F., 2015. Luminescence dating of Middle Pleistocene proglacial deposits from northern Switzerland: methodological aspects and stratigraphical conclusions. Boreas 44, 459—482. http://dx.doi.org/10.1111/bor.12114.

Lüthgens, C., Boäse, M., Krbetschek, M., 2010. On the age of the young morainic morphology in the area ascribed to the maximum extent of the Weichselian glaciation in north-eastern Germany. Quaternary International 222, 72—79. http://dx.doi.org/10.1016/j.quaint.2009.06.028.

Lüthgens, C., Boäse, M., Preusser, F., 2011. Age of the Pomeranian ice-marginal position in northeastern Germany determined by Optically Stimulated

Luminescence (OSL) dating of glaciofluvial sediments. Boreas 40, 598—615. http://dx.doi.Org/10.1111/j.1502-3885.2011.00211.x.

Murray, A.S., Thomsen, K.J., Masuda, N., Buylaert, J.-P., Jain, M., 2012. Identifying well-bleached quartz using the different bleaching rates of quartz and feldspar luminescence signals. Radiation Measurements 47, 688—695. http://dx.doi.org/ 10.1016/j.radmeas.2012.05.006.

Murray, A.S., Marten, R., Johnston, A., Martin, P., 1987. Analysis for naturally occurring radionuclides at environmental concentrations by gamma spec-trometry. Journal of Radioanalytical and Nuclear Chemistry Articles 115, 263—288. http://dx.doi.org/10.1007/BF02037443.

Murray, A.S., Wintle, A.G., 2000. Luminescence dating of quartz using an improved single-aliquot regenerative-dose protocol. Radiation Measurements 32, 57—73. http://dx.doi.org/10.1016/S1350-4487(99)00253-X.

Neudorf, C.M., Roberts, R.G., Jacobs, Z., 2012. Sources of overdispersion in a K-rich feldspar sample from north-central India: insights from De, K content and IRSL age distributions for individual grains. Radiation Measurements 47, 696—702. http://dx.doi.org/10.1016/j.radmeas.2012.04.005.

Penck, A., Brückner, E., 1909. Die Alpen im Eiszeitalter. Tauchnitz, Leipzig (Tauchnitz).

Prescott, J.R., Hutton, J.T., 1994. Cosmic ray contributions to dose rates for luminescence and ESR dating: large depths and long-term time variations. Radiation Measurements 23, 497—500. http://dx.doi.org/10.1016/1350-4487(94)90086-8.

Preusser, F., Graf, H.R., Keller, O., Krayss, E., Schlüchter, C., 2011. Quaternary glaciation history of northern Switzerland. Quaternary Science Journal 60,282—305. http://dx.doi.org/10.3285/eg.60.2-3.06.

Reimann, T., Thomsen, K.J., Jain, M., Murray, A.S., Frechen, M., 2012. Single-grain dating of young sediments using the pIRlR signal from feldspar. Quaternary Geochronology 11, 28—41. http://dx.doi.org/10.1016Zj.quageo.2012.04.016.

Reimann, T., Tsukamoto, S., 2012. Dating the recent past (500 years) by post-lR lRSL feldspar — examples from the North Sea and Baltic Sea coast. Quaternary Geochronology 10,180—187. http://dx.doi.org/10.1016/j.quageo.2012.04.011.

Rentzel, P., Preusser, F., Pümpin, C., Wolf, J.-J., 2009. Loess and palaeosols on the high terrace at Sierentz (France), and implications for the chronology of terrace formation in the upper Rhine Graben. Swiss Journal of Geosciences 102, 387—401. http://dx.doi.org/10.1007/s00015-009-1338-9.

Rogner, K., Löscher, M., Zöller, L., 1988. Stratigraphie, Palaogeographie und erste Thermolumineszenzdatierungen in der westlichen Iller-Lech-Platte (Nördliches Alpenvorland). Zeitschrift Für Geomorphologie Supplement 51—73.

Salcher, B.C., Starnberger, R., Götz, J., 2015. The last and penultimate glaciation in the North Alpine Foreland: new stratigraphical and chronological data from the Salzach glacier. Quaternary International 388, 218—231. http://dx.doi.org/ 10.1016/j.quaint.2015.09.076.

Schaöfer, l., 1952. Die donaueiszeitlichen Ablagerungen an Lech und Wertach. Geologica Bavarica 19,13—64.

Schielein, P., Schellmann, G., Lomax, J., Preusser, F., Fiebig, M., 2015. Chro-nostratigraphy of the Hochterrassen in the lower Lech valley (northern Alpine Foreland). Quaternary Science Journal 64, 15—28. http://dx.doi.org/10.3285/ eg.64.1.02.

Schlüchter, C., 1988. A non-classical summary of the Quaternary stratigraphy in the northern Alpine Foreland of Switzerland. Bulletin de la Societé neuchateloise de Géographie 32,143—157.

Schreiner, A., Haag, T., 1982. Zur Gliederung der Rißeiszeit im östlichen Rheingletschergebiet ( Baden-Württemberg ). Eiszeitalter und Gegenwart 32, 137—161.

Smedley, R.K., Duller, G.A.T., Pearce, N.J.G., Roberts, H.M., 2012. Determining the K-content of single-grains of feldspar for luminescence dating. Radiation Measurements 47, 790—796. http://dx.doi.org/10.1016/j.radmeas.2012.01.014.

Thomsen, K.J., Botter-Jensen, L., Denby, P.M., Moska, P., Murray, A.S., 2006. Developments in luminescence measurement techniques. Radiation Measurements 41, 768—773. http://dx.doi.org/10.1016/j.radmeas.2006.06.010.

Thomsen, K.J., Murray, A.S., Jain, M., Botter-Jensen, L., 2008. Laboratory fading rates of various luminescence signals from feldspar-rich sediment extracts. Radiation Measurements 43, 1474—1486. http://dx.doi.org/10.1016/

j.radmeas.2008.06.002.

Trauerstein, M., Lowick, S.E., Preusser, F., Schlunegger, F., 2014. Small aliquot and single grain lRSL and post-lR lRSL dating of fluvial and alluvial sediments from the Pativilca valley, Peru. Quaternary Geochronology 22, 163—174. http:// dx.doi.org/10.1016/j.quageo.2013.12.004.

van Husen, D., 2004. Quaternary glaciations in Austria. ln: Developments in Quaternary Sciences. Elsevier. Elsevier, pp. 1—13. http://dx.doi.org/10.1016/S1571-0866(04)80051-4.

van Husen, D., Reitner, J.M., 2011. An outline of the Quaternary stratigraphy of Austria. Quaternary Science Journal 60, 366—387. http://dx.doi.org/10.3285/ eg.60.2-3.09.