The Laschamps geomagnetic excursion recorded in continental sediments from southern Germany

SUMMARY

We present palaeo- and rock magnetic results from a well-dated, 21 m-thick, Late Pleistocene continental sedimentary section located in southern Germany. Rock magnetic measurements reveal a complex magnetic mineralogy dominated by low coercivity minerals likely related to single domain biogenic magnetite and biogenic or early diagenetic greigite. In the lower part of the section also detrital haematite is present. The stable remanence shows predominantly normal polarity with two marked deviations at ∼1280-1200 cm and at 886 cm profile depth. Whereas the lower excursion is well established by several samples and documented also by detrital haematite, the upper one is only represented by a single sample and revealed by magnetite and greigite. Using the radiocarbon-based age model for the section, the lower excursion yields an age of 42.8–41.3 ka cal BP and is interpreted to represent the Laschamps geomagnetic excursion. The increased abundance of greigite in the upper part of the section, especially in the sample responsible for the upper anomalous direction, renders the interpretation of an actual excursion problematic due to the reducing environment necessary for the greigite formation.

Palaeomagnetism, Europe, Geomagnetic excursions, Rock and mineral magnetism

1 INTRODUCTION

Geomagnetic excursions are disruptions of the geomagnetic dipole field (Roberts 2008; Laj & Channell 2015), defined as brief departures from normal secular variation, lasting for less than 10⁵ years (Laj & Channell 2015). Apart from their potential for correlating sedimentary and volcanic sequences, it has been proposed that excursions might play an important role in the evolution of the biosphere (Channell & Vigliotti 2019; Cooper et al. 2021), which makes it even more important to study them. It was proposed to define an excursion by palaeomagnetic directions with respective virtual geomagnetic poles (VGPs) with latitudes below 45° (Jacobs 2007). The mechanism that generates these events is still not entirely understood but they might represent incomplete field reversals related to the Earth's inner core dynamics (Gubbins 1999; Labrosse & Macouin 2003; Amit et al. 2010). During the Brunhes normal polarity chron (781 ka–present; Ogg 2012) several excursions occurred, which affected both palaeomagnetic directions and geomagnetic palaeointensity (e.g. Roberts 2008; Singer 2014; Laj & Channell 2015). The two youngest and most accepted excursions are the Laschamps (we use the original name of the village after which the excursion was named; Kornprobst & Lénat 2019) and the Mono Lake excursions, date respectively c. 41 ka and c. 33 ka (Laj & Channell 2015). Recent compilations based on palaeomagnetic data (Korte et al. 2019) and Be isotope-based palaeomagnetic field intensity (Simon et al. 2020) show that the Laschamps excursion is the result of a nearly complete suppression of the dipolar field. In contrast, the Mono Lake excursion seems to be represented by a much less pronounced palaeointensity drop not so frequently observed (Channell et al. 2020). It was first recorded in the sediments of the Wilson Creek Formation at Mono Lake (California, USA; Denham & Cox 1971), where only one excursion is visible which was later correlated with the Laschamps excursion (Kent et al. 2002). A review of the sediment's age suggesting again a younger age of the Wilson Creek excursion (Cassata et al. 2010) was similarly questioned by new data (Vazquez & Lidzbarski, 2012) making it highly likely that only the Laschamps excursion is present at Mono Lake. Later works from both sedimentary and volcanic rocks, however, did suggest the distinct existence of the Mono Lake excursion (Channell 2006; Cassata et al. 2008; Sirocko et al. 2013; Singer 2014; Ingham et al. 2017), characterised by a smaller magnitude compared to the Lachamps excursion, and VGP paths from different sites that appear to be inconsistent (Channell 2006; Liu et al. 2019; Liu et al. 2020). In contrast to the Mono Lake excursion, the Laschamps event seems to be prominent in several lacustrine records of Central and Western Europe, for instance in the Les Echets sediment record near Lyon (de Beaulieu & Reille 1984). Here, originally, anomalies in VGP latitude between 25 and 26 m depth were assigned to the Lake Mungo excursion (Mörner 1999). Referring to a new age model from the ACER project (ACER project members et al. 2017), this excursion between 25 and 26 m depth in the Les Echets profile should be correlative with the Laschamps event, at 41 ka. This re-correlation should also apply to a time-equivalent excursion found by Mörner (1999) in the pollen record from Grande Pile (Woillard 1978; Woillard & Mook 1982). The palaeomagnetic field intensity lows caused by the Laschamps excursion were also registered in the sediment succession of Lac du Bouchet (Creer et al. 1990; Williams et al. 1993).

Korte et al. (2019) developed a suite of spherical harmonics field models that fit published palaeomagnetic data, concluding that the Mono Lake excursion might represents a period when dipole and non-dipole field strengths at the core-mantle boundary were similar, leading to a state of transitional field behaviour with multiple minor excursions between 36 and 30 ka. Apart from the sedimentary sequences in the western USA (Channell et al. 2020), palaeomagnetic sedimentary records of the Mono Lake event are still mainly documented in marine sediment cores of the Black Sea (Nowaczyk et al. 2012; Nowaczyk et al. 2013; Liu et al. 2019) and the Ocean Drilling Program (e.g. ODP Site 919 drilled offshore east Greenland, Channell 2006).

Here, we present a palaeomagnetic investigation on a sediment record from the Nesseltalgraben section, situated in an Alpine ravine about 5 km NE of the town of Berchtesgaden in SE Germany (Fig. 1). The aim of this study is to further characterise the Laschamps palaeomagnetic excursion within a well-dated sedimentary record. The outcrop exposes a continuous 21 m long composite profile predominantly consisting of alternating layers of sand, calcareous silt, organic-rich silt, and compressed peat, deposited in a narrow ravine surrounded by Mesozoic carbonate rocks (Fig. 1C, Mayr et al. 2019).

Figure 1.

Geological map and lithostratigraphy (modified from Mayr et al. 2019). Simplified geological map of the Berchtesgaden area (a). The position of the Nesseltalgraben site is indicated by the black dot. The insert (b) shows the position of the area in south-easternmost Germany (black circle). Lithostratigraphic column of the Nesseltalgraben section with boxes indicating previously sampled sections (c).

These limnic and telmatic sediments were deposited in a sinkhole on top of late Permian–Early Triassic evaporitic bedrock (Mayr et al. 2017). The catchment area of the Nesseltalgraben is confined to the immediate surroundings, which is dominated by calcareous and evaporitic Alpine sediment rocks as well as glacial and glacifluvial deposits of Pleistocene age (Fig. 1). No major volcanic outcrop is reported from this region (Pfiffner 2014).

Previous analyses include high-resolution geochemical analysis, pollen analysis, and stable isotope analysis of bulk organic matter and organic compounds (Mayr et al. 2017; Mayr et al. 2019; Stojakowits et al. 2020). The age model of the composite profile was based on 29 radiocarbon dates of terrestrial plant remains (mosses, monocots, wood) and placed the section between 29 and 59 calibrated kiloyears before present (ka cal BP; Mayr et al. 2019). Thus, it is so far the only sediment profile almost continuously covering Marine Isotope Stage (MIS) 3 in the northern Alps. First investigations on a small part of the section revealed a subsection with a palaeomagnetic anomaly that was correlated with the Laschamps event (Mayr et al. 2017). This work, however, was limited to a small part of the composite profile that was not adequately dated at that time. Here, the results of a second sampling campaign are presented that covers most parts of the profile and provides a much more detailed insight into the palaeomagnetic behaviour during MIS 3 within the frame of a precisely calibrated age model.

2 METHODS

2.1 Sampling

A preliminary low-resolution set of palaeomagnetic directions from two separated parts of the section has been included in Mayr et al. (2017, Fig. 1c). In order to further investigate the section, samples for palaeomagnetism were obtained with a resolution between 10 and 50 cm in July 2018. A total of 83 standard (2.54 cm of diameter) samples were drilled using a battery-powered drill mounted with a diamond bit, and oriented using a Pomeroy orientation device and a Brunton compass. The local declination was calculated using the IGRF 11 model (Finlay et al. 2010). Based on the initial age model of Mayr et al. (2017), we planned the palaeomagnetic sampling with a special focus on the expected position of the Laschamps excursion. The sampling procedure turned out to be complicated especially for the interval between ∼500 and 800 cm. Many of the collected samples contained too much unconsolidated sand or were too brittle and disintegrated before they could be measured in the laboratory. Accordingly, from the total of 83 samples 37 broke and from nine additional samples the orientation mark could not be maintained in the laboratory.

2.2 Rock magnetic methods

To identify the magnetic carriers in the sampled material, we measured the temperature dependence of the magnetic susceptibility (up to 700 °C) on a total of eight representative specimens of the various lithologies throughout the section using an AGICO MFK-1 Kappabridge with integrated furnace (Figs 2a–c and i). Thermomagnetic curves of four samples were measured using a Variable Field Translation Balance (VFTB) at the Ludwig-Maximilians University of Munich for comparison with the susceptibility versus temperature curves. Additionally, representative samples were used for obtaining stepwise isothermal remanent magnetisation (IRM) acquisition curves at the palaeomagnetic laboratory in Bremen and for magnetic hysteresis, continuous IRM acquisition, backfield and first order reversal curves (FORCs) at the Ludwig–Maximilians University of Munich. Stepwise IRM acquisition curves on 10 representative specimens were initially imparted in 24 steps up to a field of 700 mT and measured automatically with an in-line 2G Enterprises cryogenic magnetometer at the University of Bremen (Figs 3a–f, Mullender et al. 2016). Because full magnetic saturation was not reached in all cases, we imparted 8 more IRM steps up to 2.5 T with an external pulse magnetiser. Hysteresis cycles, backfield, IRM acquisition curves and FORC diagrams were acquired using a Princeton Measurements vibrating sample magnetometer (VSM). In order to explore the possibility of determining the relative palaeointensity (RPI) of the geomagnetic field, a total of 40 unoriented specimen samples were first measured for their mass-normalised magnetic susceptibility using an MFK-1 Kappabridge at the University of Tuebingen. The same samples were measured for their original natural remanent magnetisation (NRM) monitoring the magnetisation decay after 10 alternate field (AF) steps between 3 and 40 mT. We did not exceed this field because of the occasional occurrence of gyromagnetic remanence acquisition during AF demagnetisation (described below), which would have compromised the results. After this, we imparted an anhysteretic remanent magnetisation (ARM) to the samples using an AF of 40 mT and a direct current field of 50 µT. The acquired ARM was subsequently AF demagnetised by using the same steps as above between 3 and 40 mT. The NRM/ARM ratio measured for the steps within the linear portions of the magnetisation decay should be proportional to the field intensity at the time of the sediment deposition (Valet & Meynadier 1998; Valet et al. 2020). The NRM and ARM analysis for palaeointensity were performed at the University of Bremen.

Figure 2.

Rock magnetic results I: Representative susceptibility and magnetisation versus temperature curves. Arrows indicate heating and cooling cycles. Stratigraphic height and sample number are shown. (a–c, i) Susceptibility versus temperature curves obtained using an MFK-1 Kappabridge in air. (d–h) Magnetisation versus temperature curves obtained using a VFTB. (h) is the heating cycle of (g) after using a 5-point running average and an automated calculation of the Curie temperature (indicated with vertical lines), using the method of Moskowitz (1981) and the RockMagAnalyzer (Leonhardt 2006).

Figure 3.

Rock magnetic results II: (a–f) representative results of component analysis of detailed IRM acquisition curves (Maxbauer et al. 2016). (a) Representative results of detailed IRM acquisition curves using a squid-magnetometer, (b–f) representative component analysis. (g–i) FORC diagrams for representative end members. (i) shows a mixture of (g) and (h).

2.3 Palaeomagnetic methods

In order to dissect the vector component of the NRM, one specimen per sample was treated with stepwise thermal demagnetisation up to a maximum temperature of 680 °C, adopting 50 °C steps up to 250 °C, onward reduced to 30 °C. A total of 10 additional specimens (30 per cent) from the same samples were used for stepwise AF demagnetisation up to a maximum field of 110 mT. The NRM was measured after each demagnetisation step with a 2-G Enterprises longcore cryogenic magnetometer with in-line AF demagnetising device. Thermal demagnetisation was performed using an ASC single-chamber thermal demagnetiser. Demagnetisation data were visually inspected by means of vector endpoint diagrams (Fig. 4, Zijderveld 1967), and palaeomagnetic directions were isolated with the principal component analysis (PCA) of Kirschvink (1980) by interpolating at least four consecutive demagnetisation steps. All demagnetisation experiments were carried out at the palaeomagnetic laboratory at the University of Tuebingen within a magnetically shielded room designed and built by Wolfgang Rösler. Palaeomagnetic data were analysed with the open source web-based tools of ‘paleomagnetism.org’ (version 2.1.0, Koymans et al. 2020). Rock magnetic data were analysed using the AGICO software package Cureval 8.0.2. Previously published radiocarbon ages (Mayr et al. 2019) were recalibrated using the latest calibration curve IntCal20 (Reimer et al. 2020). An updated age model was calculated using the software Bacon 3.6 (Blaauw & Christen 2011). The use of IntCal20 in the revised age model results in up to 7390 yr younger ages at the base of the profile (2090 cm) and on average 370 yr older ages above 1230 cm compared to the age model presented in Mayr et al. (2019). All depths are given as vertical distances from the base of glacifluvial gravels that indicate the advance of glaciers during the initial phase of the last glacial maximum 29.6 kyr cal BP that are overlying the fine-grained sediments investigated here (Mayr et al. 2019). For better comparability, the stratigraphic depths of previous palaeomagnetic results (Mayr et al. 2017) were adapted to the revised stratigraphy in Mayr et al. (2019, see also Table S1).

Figure 4.

Representative demagnetisation results of the Nesseltalgraben samples (black and white circles are projections onto horizontal and vertical plane, respectively). Depths correspond to Mayr et al. (2019). Laschamps and Mono Lake excursions are represented by samples 10A, 9A (1215 cm, 1230 cm) and 45A (886 cm), respectively. (k) Sample where both great-circle trend (orange arrow in the stereographic projection) and best-fitting line (orange dot in stereographic projection and red line in Zijderveld diagram) was used.

3 RESULTS

3.1 Rock-magnetism

3.1.1 Thermomagnetic and susceptibility curves

The behaviour of the susceptibility and magnetisation versus temperature curves can generally be categorised into two groups, group I: 1319–965 cm and group II: 958–477 cm (sample numbers pm1–34 represent group I, sample numbers pm35–76 represent group II).

1. Samples pm1–34 from the lower part of the section exhibit a more or less pronounced hump between ∼450 and 600 °C and a significant increase of intensity after cooling (Figs 2a, b, d and e); The occasional hump between ∼450 and 600 °C is likely related to the formation of magnetite from a Fe-rich precursor such as clay minerals (e.g. Sant et al. 2017) or pyrite (Passier et al. 2001; Wang et al. 2008). Due to this formation process, it is very difficult to recognise any primary Curie temperature.

2. Samples pm35–76 show results suggesting at least two separate Curie temperatures at ∼400 and 580 °C (Figs 2f and h). Most samples also show an increase of intensity after cooling (Figs 2c and f), but occasionally also almost reversible curves are observed, with only one Curie temperature at ∼400 °C (Fig. 2i).

3.1.2 IRM acquisition and backfield curves

The IRM curves were ‘unmixed’ for their coercivity component using the Skewed Generalised Gaussian (SGG) approach (Egli 2003). This analysis was performed with the open source software MAX UnMix (Maxbauer et al. 2016). All samples are dominated by a low magnetic coercivity phase that reaches half of the saturation between approximately 50 and 70 mT (Fig. 3a). Only in the lower part of the section (group I, Fig. 3b) this phase coexists with a high magnetic coercivity phase that does not reach complete saturation at the highest field. Also in samples of group I there might be another low coercivity phase present (Figs 3b and c). IRM acquisition curves obtained using a VSM do show similar results (Fig. S1). Backfield curves indicate Hcr values between ∼40 and ∼90 mT, with most values around 60–70 mT (Fig. S1).

3.1.3 Hysteresis curves and FORC diagrams

The hysteresis loops are characterised by a strong paramagnetic component which, after correction, yield homogeneous, single-domain magnetite-like hysteresis loops with Hc values between ∼15 and ∼30 mT sometimes reaching ∼60 mT (Fig. S2). Anomalous hysteresis loops were only rarely observed (e.g. PM-76 in figure S2), and might be related to a weak magnetisation signal in these samples. Using the bulk hysteresis parameters to determine the domain state of the magnetic particles in sediments has been recently questioned (Roberts et al. 2018). We thus explored the coercivity behaviour within the main hysteresis branches by means of FORC diagrams (Figs 3g–i). They can again be subdivide into the two main groups:

1. Slightly v-shaped maxima close to the Bi axis (Fig. 3g).

2. A fairly narrow central ridge with peaks between 30 and 50 mT (Fig. 3h).

Occasionally, also mixtures between the two end-members are observed (Fig. 3i). The FORC diagrams of the two groups can mainly be related to the presence of stable single domain (SD) and super paramagnetic (SP) magnetite (Roberts et al. 2000), as well as interacting SD greigite (Roberts et al. 2011). The peak of the central ridge at lower values (close to 30 mT) suggests biogenic single domain magnetite (Chang et al. 2014), but a confined spread of around ±5 mT would be expected (Roberts et al. 2012). The spread along the y-axis together with the elevated values of the central ridge peaks and the slight shift of the central ridge below the B_i = 0 axis might be related to the presence of interacting SD greigite (Roberts et al. 2011). It has been shown that the SP magnetite FORC diagram obscures the single domain central ridge, which, with more detailed FORC measurements, could be separated from the SP peak (Roberts et al. 2000).

3.1.4 Rock magnetic summary

Based on all performed rock magnetic measurements, a clear picture emerges regarding the magnetic mineral assemblage in the Nesseltalgraben sediments: (i) low coercivity phases occur throughout the section, which can be identified as SD biogenic magnetite and biogenic or early diagenetic greigite represented mainly by a central ridge in most FORC diagrams between ∼30 and 50 mT (Fig. 3). This phase could be biogenic in origin, as magnetotactic bacteria are also commonly observed in perialpine freshwater environment (Pan et al. 2005a). Its signal, however, is masked in most thermal experiments by other magnetic phases forming during heating (Fig. 2) and in some of the FORC diagrams by SP minerals (Fig. 3g); (ii) samples in the lower part of the studied section (group I) contain a more complicated magnetic mineralogy with the contribution of minor amounts of haematite, revealed by the high-coercivity component of the IRM acquisition curves (Fig. 3b) and SP magnetite shown by FORC diagrams (Fig. 3g); (iii) samples from the upper part of the section (group II) show a more distinct presence of greigite indicated also by thermomagnetic curves (Fig. 2f + h).

3.2 Palaeomagnetism

As expected from the rock-magnetic results, the NRM signal and its behaviour during demagnetisation is not homogenous through the section (Fig. 4). During AF demagnetisation, samples from group II show gyro-remanent behaviour in agreement with the presence of greigite (Fig. 4h, Snowball 1997; Stephenson & Snowball 2001; Fu et al. 2008). Thermal demagnetisation of samples from group II, however, do show steady and high-quality trends towards the origin of the projection plane between ∼200 and ∼540 °C (Figs 4a–d). Using the intensity decay plot of sample 35 (Fig. 4c), two separate unblocking temperatures can be inferred at 360 and at 480 °C with roughly the same directional information. The lower part of the section (group I) is characterised by generally lower quality of demagnetisation diagrams. Demagnetisation in this part of the section is stable until 680°C, approximately the Néel temperature of haematite (O'Reilly 1984). This is in agreement with the presence of haematite inferred from the IRM acquisition curves.

From each palaeomagnetic direction we calculated the associated VGP, and the stratigraphic variation of its latitude with respect to the mean palaeomagnetic pole is used to define the magnetic polarity stratigraphy, whereby VGP approaching +90° (–90°) represent normal (reverse) geomagnetic polarity fields. As expected, the VGP latitude is predominantly positive, indicating deposition under normal polarity fields (Figs 5 and 6). The latitude of the VGPs from the lower part of the section is in very good agreement with the preliminary results published by Mayr et al. (2017), supporting the reliability (and repeatability) of the measurements (Fig. 6). In the whole section there are two clear departures from the otherwise normal polarity field: the first at ∼1280–1200 cm and the second at 886 cm, hereafter referred to as lower and upper excursion, respectively (Fig. 6). Whereas the upper excursion is represented by only a single sample (#45, see Table S2 for depths and age model of respective samples), the lower excursion is at least represented by one reversed polarity sample straddled by six shallow VGP latitude points (Fig. 4,6,7). In addition, four samples from this interval show a distinct great circle trend towards a reversed polarity (Figs 4 and 5). These data were combined with the reversed polarity direction (#10) to determine a best-fitting line lying on each circle and a reversed-mode mean direction (McFadden & McElhinny 1988). Since we are not studying an interval of persisting reversed polarity, the lack of reliable reversed directions makes the result of this analysis poorly reliable, so these directions are only shown for comparison. In contrast, we used a best-fitting line for the same samples without including the origin or anchoring the last heating steps (Figs   4i and k). To obtain a normal polarity mean direction, we combined all results based on thermal and AF demagnetisation experiments with an unambiguous normal polarity (Fig. 5). This leads to a mean direction of D = 348.8°, I = 62.8°, k = 10.5 and α₉₅ = 7.2°. The inclination is well within error limits of the expected value assuming a geocentric axial dipole (I_exp = 65.5°).

Figure 5.

Equal area projections of sample mean directions obtained using principle component analysis of Kirschvink (1980). Left-hand panel shows all results based on high quality thermal demagnetisation experiments including best-fiting line solutions of the excursion samples with respective sample numbers. Sample #45 represents the Mono Lake excursion. Middle panel shows excursional samples using remagnetisation great circles (McFadden & McElhinny 1988) together with the best-fitting line of sample #10 and the resulting mean direction with a 95 per cent confidence circle (see text for further explanation). Only the respective part of the great circle is shown where the sample directions actually plot (compare Fig. 4). Right-hand panel shows only normal polarity samples (both using thermal and AF demagnetisation techniques) without excursional results. Mean direction is shown in the inset.

Figure 6.

Plots of palaeomagnetic and relative palaeointensity results versus stratigraphic depths [cm]. From left- to right-hand side: declination, inclination, virtual geomagnetic pole (VGP) latitude (red and blue dots are directions respectively from thermally and AF demagnetised samples, squares from great-circle analysis, and yellow from Mayr et al., 2019), susceptibility (sus), natural remanent magnetisation (NRM) intensity of the unoriented specimens for rock-magnetic analysis, ratios of NRM versus susceptibility, NRM versus bulk anhysteretic remanence magnetisation (ARM¹) and NRM versus selected ARM steps (ARM²; average values are shown by the red dots) after alternate field demagnetisation (see text for further explanation). Grey bars indicate position of the two excursions.

Figure 7.

Age model of sedimentation for the Nesseltalgraben section (a) based on Mayr et al., (2019) shown with the virtual geomagnetic pole (VGP) latitude variations (b). VGP latitude results are shown for the high-quality thermally (thermal-HQ) demagnetised samples from both campaigns (dark blue and green), AF demagnetised samples of this study (light blue) and mean directions of four remagnetisation great circles combined (see figure 5). Dotted lines indicate the palaeo-equator and the 45° cut-off latitude. Brown bars show supposed intervals straddle by the geomagnetic field excursions. Orange boxes show literature age ranges of the Laschamps and Mono Lake excursions (Guillou et al. 2004; Roberts 2008; Lascu et al. 2016; Liu et al. 2020), whereas light orange box of the Mono Lake excursion represents the interval of excursional behaviour (Korte et al. 2019).

4 DISCUSSION

4.1 Rock magnetism

Diverse magnetic minerals content is expected by the nature of the depositional environment, with complex sources of sediments. However, the performed analyses provide insights regarding the origin of the different minerals. The central ridge of the single domain magnetite revealed in several FORC diagrams can be interpreted for the magnetite as being biogenic in origin related to magnetotactic bacteria (Chang et al. 2014). Freshwater environments at the foot of the Bavarian Alps have been widely studied for their magnetotactic bacteria content (Pan et al. 2005a, b). We might speculate that greigite in the upper part of the section (group II) is generated in a similar way by sulfate-reducing bacteria (Lefèvre et al. 2011) related to a reduced amount of available oxygen. Generally, however, greigite is often regarded as an early diagenetic mineral (Roberts et al. 2011). The occurrence of super paramagnetic magnetite can be related to soil formation processes (Van Oorschot et al. 2002). Finally, because of the high temperatures needed to demagnetise samples of the lower part of the section together with the greyish-black color (missing of a reddish pigment), the observed haematite is interpreted to be of detrital or aeolian origin. A detrital origin from Alpine sediments such as the surrounding Haselgebirge seems most likely (Leitner et al. 2017).

4.2 Palaeomagnetic directions

There is an apparent difference in demagnetisation behaviour between the lower (group I) and upper (group II) part of the section. The upper part shows straightforward results (Fig.4a-d), where the potential greigite and magnetite components contain comparable directions. Not a single sample exists where both phases show significantly different directions, whereas in some samples slight differences can be suggested (Fig. 4a). The difference might be related to slightly different formation times of the biogenic or early diagenetic minerals. The presence of biogenic or early diagenetic greigite complicates the identification of the upper excursion, because reducing conditions necessary for the formation of greigite renders the preservation of biogenic or detrital magnetite unlikely (Larrasoaña et al. 2014). In the lower part of the section we interpret the main remanence carriers as both magnetite and haematite (Figs 4e–k). This part shows more scattered demagnetisation behaviour and is more difficult to interpret. MAD values in this part are elevated but distinctive trends are still recognisable (Figs 4i and j).

4.3 Correlation

Using an updated age model, which is based on 29 recently acquired radiocarbon ages (Mayr et al. 2019) including one additional date (Stojakowits et al. 2020), we compare the obtained VGP latitude record with the expected ages of the Laschamps and Mono Lake excursions (Fig. 7). The lower VGP latitude anomaly coincides well with the expected age of the Laschamps excursion (41 ka expected versus a median modelled age of 42.8–41.3 ka cal BP observed, Fig. 7). The upper VGP latitude anomaly is represented by only one sample and has a median age of 35.7 ka cal BP. This is slightly older than the commonly reported ages for the Mono Lake excursion of 34–32 ka (Singer 2014).

Nevertheless, as pointed out by Korte et al. (2019), the Mono Lake excursion may rather represents a 6-kyr-long interval of persisting VGP anomalies starting about 36 ka and, therefore, our sample could indicate the inception of the Mono Lake excursion. It also has to be noted that during the proposed main Mono Lake excursion interval (e.g. 33 ka, Laj & Channell 2015) not many directional data are available from the Nesseltalgraben section and more investigation is needed to confirm the exact appearance of the Mono Lake event at this section (Fig. 7). Unfortunately, many samples within the 600–800 cm stratigraphic position broke during the sampling campaign. Future sampling might be successful in adding more data in this interval and therefore this section might yield further information regarding the Mono Lake excursion and its period of recurring excursions (Korte et al. 2019).

4.4 Relative palaeointensity

Applying a simple normalisation of the NRM using bulk susceptibility or ARM (before AF treatment) shows significant variations throughout the record (Fig. 6), which, however, are very unlikely related to the intensity of the geomagnetic field. The rock-magnetic analyses suggest that the observed variations depend on the variations of the magnetic mineral content other than the field intensity. We selected a number of specimens that were characterised by a minimum of 5 demagnetisation points linearly decaying to zero (i.e. MAD < 5°), and we averaged the NRM/ARM ratio of the selected steps. This approach was recently successfully used by Valet et al. (2020) to determine the RPI of Pliocene–Pleistocene Pacific deep-sea records. Applying this filter significantly decreases the number of reliable points (Fig. 6), which hampers the possibility to monitor the RPI variations through the record. This data highlight that reliable RPI analyses should be performed in absence of lithological/magnetic mineralogy variations and, even within this condition, the samples on which the RPI estimate is based should be properly demagnetised to monitor the NRM behaviour, without applying bulk susceptibility or ARM normalisation. This bulk approach results in unreliable estimates.

5 CONCLUSIONS

We present palaeomagnetic data for the Late Pleistocene Nesseltalgraben continental section located in the Northern Calcareous Alps (southern Germany). The age of the section is well constrained by radiocarbon age estimates between >50 and 30 ka. Two intervals are identified with anomalous palaeomagnetic directional behaviour, at 1280–1200 cm and 886 cm stratigraphic heights. Whereas the lower one (1280–1200 cm) is characterised by at least six different samples, the upper one at 886 cm is only represented by a single sample. Using the most recent recalibrated age model for this section, the excursions have ages of 42.8–41.3 and 35.7 ka cal BP, respectively. Accordingly, we correlate the lower one with the Laschamps geomagnetic excursion. The upper sample, even if it might be related to the Mono Lake excursion, is problematic due to more prominent presence of greigite, which might argue against the preservation of primary or near primary magnetite. The Nesseltalgraben section is a well-calibrated, easily accessible record that puts further constraints on the age and duration of the Laschamps geomagnetic excursion.

SUPPORTING INFORMATION

Figure S1:Isothermal remanent magnetisation (IRM) and backfield curves obtained with the vibrating sampling magnetometer (Princeton Measurements).

Figure S2: Representative hysteresis loops of the Nesseltalgraben section. Original results are shown in purple with left axis labels. Hysteresis loops after correction for paramagnetic contributions are shown in green with right axis labels.

Table S1: Radiocarbon data from the composite profile (Mayr et al. 2019) with one additional date from (Stojakowits et al. 2020). One outlier is marked with italics. The results of the ¹⁴C-measurements given in the table were recalibrated using the IntCal20 calibration curve (Reimer et al. 2020) and the CALIB 8.2 software.

Table S2: Palaeomagnetic results including age model.

Please note: Oxford University Press is not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the paper.

ACKNOWLEDGEMENTS

We cordially thank Maxwell Brown, one anonymous reviewer and the Editor, Prof Andrew Biggin, for the constructive comments that significantly improved the quality of the manuscript. We also thank assistant editor Lousie Alexander for handling of the manuscript. Furthermore, we want to thank B. Lempe, V. Diersche and J. März for their support during many field campaigns and for valuable discussions. This study was funded by the DFG (Deutsche Forschungsgemeinschaft) in the framework of the project ALPWÜRM (MA 4235/10–1). ED is supported by the DFG (Projektnummer 408178503).

DATA AVAILABILITY

Most of the data underlying this paper are available in the paper and in its online supplementary material. All additional data like the relative palaeointensity values will be shared on reasonable request to the corresponding author.