Syn‐ and post‐eruptive gully formation near the Laacher See volcano

The Laacher See volcano (LSV) is located at the western margin of the Neuwied Basin, the central part of the Middle Rhine Basin of Germany. Its paroxysmal Plinian eruption c. 13 ka ago (Laacher See event; LSE) deposited a complex tephra sequence in the Neuwied Basin, whilst the distal ashes became one of the most important chronostratigraphic markers in Central Europe. However, some other impacts on landscape formation have thus far been largely neglected, such as buried gully structures in the proximity of the LSV. In this contribution, we map and discuss the spatial extent of these landforms at the site Lungenkärchen c. 4 km south of the LSV based on geophysical prospection as well as contrasting pedo‐sedimentary characteristics of the gully infill (particle‐size distribution, bulk‐sediment density, thin‐section analysis, saturated hydraulic conductivity) and the surrounding soils and tephra layers. These data are combined with a luminescence‐ and carbon‐14 (14C)‐based age model that relates them to the LSE. It is demonstrated how these gullies seem to have been formed and rapidly infilled by rainfall and surface discharge both during and subsequent to the eruptive phase, with modern analog processes documented for the 1980 Mount St Helens eruption (Washington State, USA). Given the density of the gullies at the site and their deviating pedo‐sedimentary properties compared to the surrounding soils, we propose a significant influence on agricultural production in the proximity of the LSV, which remains to be tested in future studies. Finally, in contrast, gullies of similar lateral and vertical dimensions identified in post‐LSE reworked loess and tephra deposits of the Wingertsbergwand (close to the main study site and proximal to the LSV) have shown to be unrelated to the LSE and can either be attributed to periglacial processes at the Younger Dryas‐Preboreal transition or to linear incision during the early Holocene.


| INTRODUCTION
The Laacher See volcano (LSV) at the western margin of the Neuwied Basin is part of the Volcanic Eifel (Rhineland-Palatinate, Germany) ( Figure 1). Its paroxysmal Plinian eruption c. 13 ka ago (Reinig et al., 2020) was nourished by a highly stratified magma chamber and controlled by a variety of volcanic and other environmental factors. These include the aquifer level, crustal lithology, the volcano-tectonic alignment, as well as vent widening, collapse, downward erosion, and migration (Schmincke et al., 1999).
Two thirds of the compositionally zoned phonolitic magma volume of c. 6.3 km 3 were erupted within a period of 10 h (Litt et al., 2003). The ejecta comprised mainly lapilli tephra, but also tremendous amounts of ashes were distributed in three primary directions of propagation across Central Europe (Figure 1b). The Laacher See event (LSE) had a significant impact on topography and regional soils (Schmincke et al., 1999;Nowell et al., 2006;Hahn & Opp, 2011;Schmincke, 2014;Viereck, 2019), and triggered temporary lake reservoir formation in the Neuwied Basin with catastrophic dam outburst towards the end of the event (e.g., Schmincke et al., 1999;Litt et al., 2003).
One peculiar geomorphic feature associated with tephra deposits of the LSE has remained unstudied so far, that is gullies, which are up to a few meters wide and up to c. 2 m deep. Gully formation in central-western Europe has been usually associated with combined anthropogenic land degradation and intense rainfalls over timescales of decades up to centuries, in particular during the 14th and 18th centuries (e.g., Vanwalleghem et al., 2005;Lang & Mauz, 2006;Stolz & Grunert, 2006;Dotterweich, 2008Dotterweich, , 2013. In some very specific cases, large gullies have been associated with single paroxysmal rainfall events, such as the one of 1342 (Bork et al., 2006a;Dotterweich, 2008). The gullies incised into the tephra of the LSE seem to have also formed very rapidly. They occur at prominent proximal tephra sequences, such as the Wingertsbergwand (WBW) immediately south of the LSV (Schmincke, 2014). Yet, very few data are available about their spatial extent, density and occurrence pattern in the wider area of the LSV. The present contribution addresses this knowledge gap and, for the very first time, documents the lateral distribution and stratigraphical context of gullies adjacent to the LSV. Based on a new geophysical, sedimentological and geochronological dataset, we test their genetical relation to the LSE and discuss further implications of eruption-related gully incision and infill for agricultural land use.

| STUDY AREA
The study area in the East Eifel region is part of the Hercynian Rhenish Massif, characterized by folded Lower Devonian schists and sandstones (Siegenium). It extends into the tectonic, intramontane Neuwied Basin, where the Palaeozoic is overlain by a sequence of Tertiary to Quaternary fluvio-lacustrine sediments (Rath, 2003;Schmincke, 2014). The Rhenish Massif was subjected to tectonic uplift, lithospheric thinning, and peneplanation since the early Palaeogene (Ilies et al., 1979;Semmel, 1996;Nowell et al., 2006), associated with an earlier, Eocene phase of volcanism (Nowell et al., 2006).
Maximum rates of uplift were reached during the Pleistocene with up to 140 m since the formation of the Main Terrace of the Middle Rhine Valley system c. 730 ka ago (Demoulin & Hallot, 2009). This gave rise to lithospheric thinning and the Pleistocene formation of the Laacher See volcanic field with c. 70 vents distributed over an area of 400 km 2 . Around 700-370 ka ago, it erupted phonolithic tuffs and lavas, followed by the formation of cinder cones, basaltic volcanoes and lava flows (c. 200-100 ka ago) (Ilies et al., 1979;Nowell et al., 2006). The short, terminal period was characterized by phonolithic and trachytic lapilli tuffs and ashes, which were distributed over large areas (Schmincke et al., 1999). The last major eruption, the LSE, is dated to around 13 ka ago (Reinig et al., 2020). Massive loads of pumice tuff were emitted during this explosive event and the Laacher See caldera with a diameter of 3 km was created. Its pumice lapilli and ashes were transported over large parts of Central Europe forming a major stratigraphic marker horizon for the terminal Pleistocene (Schmincke et al., 1999;Litt et al., 2003;Nowell et al., 2006;Schmincke, 2014). Recent gas exhalations and deep low-frequency earthquake activity beneath the Laacher See caldera may indicate active magma recharge (Hensch et al., 2019).
The WBW site (Wingertsbergwand), located less than 1.5 km south of the caldera (Figure 1a; Supporting Information Figure S1), F I G U R E 1 (a) Spatial context of the core study site "Lungenkärchen" (white rectangle indicates location of the maps in Figure 2), the Wingertsbergwand and the Laacher See (based on Landsat 8 data, courtesy of US Geological Survey [USGS], 2013). (b) Wider regional context of the study area (dark grey, semi-transparent rectangle represents the map in (a)), including the location of the Laacher See (magenta) and isopachs of depositional thickness of Laacher See tephra in meters after Schmincke et al. (1999) (based on GTOPO30 dataset, courtesy of USGS, 1996. For both maps (a) and (b), the transverse Mercator projection is applied; coordinates represent UTM zone 32N. (c) Location of map (b) [Color figure can be viewed at wileyonlinelibrary.com] exposes more than 30 m of highly stratified LSE deposits, which reflect various stages of the eruption. The LSE deposits are subdivided into the Lower Laacher See tephra (LLST), Middle Laacher See tephra (MLST) and Upper Laacher See tephra (ULST), representing the three main stages of the eruption, which lasted a total of a few months only.
The LLST comprises a poorly sorted ash bed with plant remains and accretionary lapilli resulting from the initial phreatomagmatic explosion. This ash bed is overlain by whitish inflated pumice from the Plinian eruption. The MLST comprises a phreatomagmatic surge deposit combined with small-scale Plinian fallout units (MLST-A), loose ash-flow deposits topped by pumice lapilli (MLST-B), and dense, greyish pumice fallout layers interbedded with minor ash flows (MLST-C). The ULST consists of moderately sorted breccias with crystal-rich lapilli and dark phonolithic clasts, as well as massive ashflow deposits (van den Boogard & Schmincke, 1985;Schmincke et al., 1999;Viereck, 2019). Cross-bedding structures within the proximal ash-flow deposit represent a radially disposed upcurrent migration of antidunes, which reflects high-velocity, cohesive phreatomagmatic flows in the range of > 100 km/h (Schmincke et al., 1973).
The main study site of Lungenkärchen in the eastern part of the  (Schmincke & van den Boogard, 1990;Mangartz, 1993 Organization of the United Nations [FAO] and German soil systematics). The narrow floodplain is characterized by gleysols and vegas (Beck, 2003). The main study site covers an area of c. 7 ha at the archaeological site of Lungenkärchen in the eastern Segbach Valley (Figure 1a). At this site, a Roman villa rustica complex with several individual buildings, associated with the local basalt quarries exploited for the Roman millstone industry, was detected and excavated (Grünewald, 2011(Grünewald, , 2012Totschnig & Seren, 2011). The area is characterized by intense agricultural land use and covers a slightly convex, southerly exposed slope inclining towards the east-west running Segbach ( Figure 2a).

| METHODS
The subsurface gully structures investigated here were first detected during geophysical investigations of the Roman villa rustica site (northwest part of the Lungenkärchen site, Figure 2). As any detailed regional accounts on these structures were lacking at that time, the geophysical survey, in particular magnetometer prospection, was extended and supplemented by pedo-sedimentary and geochronological investigations in order to characterize these relict landforms in detail, to assess the spatial extent and to verify their volcanic origin.

| Grain size and lithological analyses
Selected units were sampled in plastic bags for pedo-sedimentary analyse at the Laboratories for Physical Geography at the Universities of Mainz and Cologne, Germany. Samples were air-dried, carefully handpestled and dry-sieved (2 mm). The fraction 0.063-2 mm was wetsieved, whereas pipette analysis (Köhn, 1928) was used to determine the grain-size distribution < 0.063 mm. Samples for grain-size analysis were dried at 105 C before adding hydrogen peroxide (H 2 O 2 ) (30%) to remove organic carbon. After dilution, (NaPO 3 ) 6 and NaCO 3 were added for aggregate dispersion. Statistical parameters were calculated using GRADISTAT (Blott & Pye, 2001). For each sample from LU-A5, the ratio between the coarse siliciclastic (mainly schist, greywacke, quartzite) and tephra components > 2 mm was recorded.

| Analysis of bulk density and saturated water conductivity
To compare the density and saturated water conductivity of the gully infill and the surrounding soils, undisturbed samples were taken from trench LU-A5 using 5 cm × 5 cm aluminum cylinders. A UMS KSAT device was used following the manual of UMS GmbH (2012); saturation of samples took between < 1 (sandy substrates) and 24 h (clay-and silt-rich substrates).

| Thin-section analysis
Two undisturbed samples from trench LU-A1 were taken for thinsection analyses using Kubiëna boxes (5 cm × 7 cm), one inside and one outside the channel structure. Samples were dried at 40 C for several days and impregnated using Oldopal P80-21, then cut and polished (final thickness = 30 μm) following Beckmann (1997). Flatbed scans of the thin sections were obtained at 1200 dpi and investigated at low magnification. Each thin section was scanned under ordinary light and using two polarization foils mounted above and below the thin section at an angle of 90 . Furthermore, the thin sections were inspected under plane polarized light (PPL), crossed polarizers (XPL), and oblique incident light (OIL) at magnifications of 12.5× to 500× using a polarization microscope (Axioskop 40, Zeiss). Micromorphological description followed the terminology of Stoops (2003), which includes a distinction between properties of pore space, groundmass, and pedofeatures. In general, the size, shape, and type of mineral and organic matter components were recorded and the nature of the pore space defined.

| Luminescence dating
At both study sites, samples were taken for optically stimulated luminescence (OSL) dating. While the four samples from trench LU-A1 and from core LU-L6 ( Figure 2) at Lungenkärchen were dated in the luminescence laboratory of the Institute of Geography at Humboldt University of Berlin, Germany, the two samples later taken from the WBW were dated at the Cologne Luminescence Laboratory (CLL) of the Institute of Geography, University of Cologne, as the laboratory in Berlin had ceased operations shortly after processing the first batch.
Therefore, slightly different protocols were applied for both sites.
For Lungenkärchen, the grain fraction between 63 and 250 μm was prepared applying standard laboratory procedures of the quartz coarse-grain method: wet sieving, removal of carbonates and organic material with hydrochloric acid (HCl) (10%) and H 2 O 2 (10%), mineral separation with heteropolytungstate (densities of 2.7, 2.62 and 2.58 g/cm 3 ), and etching of the quartz fraction with hydrogen fluoride (HF) (40%) for 60 min. For quartz-OSL measurements, 1 mm-and 2 mm-aliquots were prepared. When signal efficiency was sufficient, F I G U R E 2 Overview of the core study area "Lungenkärchen" (see white rectangle in Figure 1a for location). It shows the location of trenches (yellow bars) and sediment cores (yellow dots) as well as Roman archaeological structures (part of a villa rustica) deduced from the geophysical prospection (red lines) and archaeological excavations (purple signature) (for details on the archaeological structures see Grünewald, 2012). (a) Aerial image from 2009 provided by LVermGeo Rheinland-Pfalz showing some of the linear gullies through differential soil moisture. (b) Magnetometer prospection of the study area indicating the spatial distribution of the gullies oriented towards the Segbach running in a west-east direction. Coordinates represent UTM zone 32N [Color figure can be viewed at wileyonlinelibrary.com] 1 mm-aliquots were preferred to reduce the number of grains per aliquot in order to allow a better identification of the best-bleached grain population.
Equivalent doses (D e ) were determined using a Risø TL-DA-15C/D reader equipped with a 90 Sr/ 90 Y beta source delivering 0.097 Gy/s at the sample position. Sets of 18 to 45 aliquots per sample were stimulated with blue light emitting diodes (LEDs) (λ = 470 ± 30 nm) at 125 C for 40 s ( Figure S2). The resulting OSL signals were detected in the near-ultraviolet (UV) range through a 7.5 mm Hoya U340 filter (transmission between 290 and 370 nm).
For D e determination, the initial integral of 0.48 s was used subtracting the last 4 s of the OSL decay curve. A standard single-aliquot regenerative (SAR) dose protocol following Murray and Wintle (2000) was used for OSL measurements, preheating at 220 or 240 C for 10 s and a test-dose cut-heat at 160 C (Supporting Information Table S1). The water content of the samples (relating to the dry weight) was determined after drying at 105 C for 24 h and an error of 5% was considered. High-resolution gamma spectrometry was applied to estimate the sediment dose rates arising from the decay of primordial radionuclides (Tables S2 and S3). The cosmic-dose rate was estimated considering the geographic position (50.5 N, 7 E), the altitude (285 m) and half of the sampling depth. Luminescence ages were calculated using Adele software v2017 (Degering & Degering, 2020, https:// www.add-ideas.com). In core LU-L6, the dose rate could not be determined directly from the layer of the OSL sample, but from the adjacent layers. To receive the dose rates for luminescence age determination of the samples of core LU-L6, layer models were calculated in Adele. All luminescence data of LU-A1 and LU-L6 are summarized in Tables 1 and S4. At the WBW, one sample was collected from the gully infill (WBW-B), a second sample from the reworked loess and tephra deposits into which the gully had incised (WBW-A). Luminescence signals of the quartz population were tested first, but they turned out to be relatively dim and decay-curve fitting with the R package 'Luminescence' version 0.9.0.88 (Kreutzer, 2019) indicated quartz signals not dominated by the fast component for more than 50% of the measured aliquots ( Figure S4a). Furthermore, analysis of the signal curves in the Analyst software (Duller, 2015) showed that depletion ratios were inadequate (< 0.8) for more than 50% of the measured aliquots as well ( Figure S4b), indicating problems with feldspar contamination.
Therefore, ages were based on infrared stimulated luminescence (IRSL) dating of potassium feldspar (KF). KF grains in the 100-200 μm fraction were extracted by dry sieving, chemical treatment with HCl (10%) and H 2 O 2 (10%) to remove carbonates and organic matter, and density separation with sodium-polytungstate (KF < 2.58 g/cm 3 ). The feldspar extracts were measured on 2 mm-aliquots. The D e measurements were performed on a Risø TL/OSL reader equipped with a 90 Sr/ 90 Y beta source delivering 0.15 Gy/s at the sample position, infrared (IR) LEDs (λ = 870 ± 40 nm) and 5 mm blue D410/30x LOT interference filter (λ = 410 ± 30 nm). Measurements followed a feldspar SAR dose protocol (Wallinga et al., 2000) with preheat at 270 C for 60 s and IR signal stimulation at 50 C for 300 s (see Table S5). All dose-response curves were forced through the origin. Protocol performance for the WBW samples was tested using a preheat-plateau experiment with preheat temperatures between 210 and 290 C (doses are independent from thermal pre-treatment for temperatures of 230 to 290 C, Figure S5 (Huntley & Baril, 1997) and a-values of 0.07 ± 0.02 (Kreutzer et al., 2014). The samples showed no indication of significant disequilibria in the Th and U decay chains (Table S6). Ages were corrected for anomalous fading following the approach of Huntley and Lamothe (2001) and experimental g 2day -values (using delay times of up to 6 h and the approach according to Auclair et al. [2003]) of 4.0 ± 0.5 and 3.7 ± 0.8%/decade for WBW-A and WBW-B, respectively ( g-values are the arithmetic mean of five aliquots) ( Figure S7). The luminescence data for both samples are summarized in Tables 2 and S7.

| Radiocarbon dating
Four charcoal samples were dated by carbon-14 ( 14 C)-accelerator mass spectrometry (AMS) at the Institute of Geology and Mineralogy, University of Cologne. The data were calibrated using Calib 7.1 (Stuiver & Reimer, 1993) and the IntCal13 database (Reimer et al., 2013). All data are reported as cal BP using the 2σ confidence interval (Table 3).

| Stratigraphic evidence
Trench LU-A1 Trench LU-A1 (Figure 4) cuts the V-shaped gully structure of the most prominent linear anomaly (Figure 2b), which has a width of c. 2.7 m and a depth of c. 1.5 m. The boundary of the gully (Layers 11-13) is clearly visible and it is incised into stratified tephra-dominated deposits of varying thickness and particle-size distribution (Layers 5-10). Layers 5-8 are relatively fine-grained with clay-and silt-sized particles accounting for roughly 50%. Only Layer 5 shows a significant lapilli component. Layers 9 and 10 are dominated by lapilli and sandsized particles. The thin section from Layer 8 in Figure 4 shows a sequence of five finely stratified and densely packed layers with low to medium porosity (Figure 5a, c). Pores mainly classify as simple   Figure 7) and as stray finds (Grünewald, 2011(Grünewald, , 2012, point to an anthropogenic infill; possibly a La Tène-age pit of a wooden post, disintegrated into charred oak wood. Layer 6 represents the recent agricultural soil overlying the incised channel structure.

Further trenches
Trench LU-A8 represents the proximal part of a terminal fan connected to the gully of LU-A1 (Figures 2 and 7). Both OSL and 14 C data clearly show a hiatus at c. 1.65 m below surface between the terminal F I G U R E 5 Flatbed scans of thin sections LU-A1-M2 (a) and LU-A1-M1 (b) from trench LU-A1 (Figure 4). Panels (c) and (d) Tables 1 and 3), which also contain Prehistoric artifacts, some more precisely dated to the La Tène age (Grünewald, 2012), as well as many charcoal fragments. The dark and humic Layer 4, that contains both 14 C data reported earlier, may represent a fossil A-horizon, that is a temporarily stable surface. At the base of the third meter of core LU-L6 (right next to LU-A8), a bright band of well-stratified, pure volcanic ash was found (Figures 7 and   S9). In contrast to the OSL dating of trench LU-A1, the luminescence properties were less suitable here, while water contents were very high in the two lower samples, which were taken at and below the groundwater table. Many aliquots showed a slow decay of the OSL signal (although no feldspar contamination) or increased recuperation.
Accordingly, many aliquots had to be excluded from further analysis (Table 1) F I G U R E 6 Trench LU-A5 representing the upper to middle part of the easternmost prominent channel running in south-eastern direction. A sketch of the trench is shown, with labeled samples analyzed for the coarse fraction > 2 mm, saturated water conductivity, sediment bulk density (* = no measurement), and the amount of non-volcanic particles among coarse components. For location of LU-A5 see Figure 2 [Color figure can be viewed at wileyonlinelibrary.com] F I G U R E 7 Trench LU-A8 and parallel core LU-L6 (see also Supporting Information Figure S9) representing the proximal part of the terminal fan of one of the most prominent gullies running towards the Segbach Valley. For location see Figure 2 [Color figure can be viewed at wileyonlinelibrary.com] manganese (Mn) precipitation and strongly contrasts with the underlying light clays and silts ( Figure S11).  Figure 8a). While the gully deposits were IRSL-dated to 9.4 ± 4.1 ka, the incised pre-incision strata were IRSL-dated to 9.7 ± 2.5 ka (Table 2). Similar to Lungenkärchen, asymmetric dose distributions, indicating incomplete bleaching, led to the application of the MAM bs (bootstrapped MAM).

| DISCUSSION
Investigations from this study reveal the characteristics and spatial distribution of the gullies, as well as their genetic relationship to the LSE. The interpretation of the short, linear gullies identified in the magnetograms and GPR depth slices (Figures 2, 3 and S8) leading from the margins of the small plateau of the Roman villa rustica into the Segbach Valley was ground-truthed by the trenches cross-cutting the gullies. Although the pre-incision strata and the gully infill share common sediment sources (i.e., LST, pre-LSE soil, clastic components of local underlying schists, greywacke and quartzite), both units are significantly different in terms of colour, bulk sediment density, water conductivity, micro-lamination, and ratio of volcanoclastic to siliciclastic components. Even though graded bedding is only weakly expressed in the thin section of the channel infill, when combined with the visible layering and surface exposure (i.e., textural crust), it is clearly indicative of transport by water; in contrast to solifluctional dynamics outside the gullies. Each graded sequence potentially relates to a specific discharge event (Vanwalleghem et al., 2005;Bork et al., 2006b). The bow-like shape of the accumulations of opaque grains may point to later movement by gravity microslides inside the gullies. Less distinct macro-scale bedding of the channel infill when compared to the outer tephra has also been observed in the few documented gullies of the more proximal tephra exposures in the Mendig Graben, that unequivocally formed during the LSE (Schmincke, 2014).
The Plinian eruption of the LSVwith several short phreatomagmatic interruptions in betweenhas induced the most significant changes in the wider study area during recent geological times. This is exemplified by the thick, slightly reworked tephra layers on top of the late Pleistocene solifluction layers, and the Miocene clay-silt deposits in the trench profiles. The eruption released an extremely hot gas column and large amounts of volcanic aerosols into the atmosphere, with immediate effects on solar radiation flows and cloud condensation in the lower atmosphere, inevitably resulting in thunderstorms with heavy downpours (Schmincke et al., 1999;Litt et al., 2003;Schmincke, 2014). Subsequently, even if only moderately strong rainfall was assumed, substantial surface discharge would have formed on top of the fresh, marginally permeable fine-grained ash deposits (cf. Collins et al., 1983;Major & Yamakoshi, 2005;Manville et al., 2009), leading to rapid formation of these gullies. By considering the 1σ error ranges, the OSL data obtained from the gully infill Mendig (Schmincke et al., 1999) revealed an infiltration capacity of 2 to 5 mm/h only (Collins et al., 1983). Erosion of loose tephra on these hillslopes occurred in the form of linear erosion, in some places even cutting down into the colluvium. A clear linear relationship F I G U R E 8 (a) Channel structure in the uppermost part of the western Wingertsbergwand (reworked LST), sampled for OSL dating (see details in Table 2). It is incised into horizontally bedded silt and tephra deposits that stratigraphically overlie the ULST (not captured by the profile photograph). (b) Similar channel structures can be observed in other parts of the Wingertsbergwand, cut into the stratigraphically older ULST). In contrast to profile WBW-A1 in (a), these date syn-and immediately post-LSE (see also Schmincke, 2014) [Color figure can be viewed at wileyonlinelibrary.com] between hillslope gradient and erosion rate was documented (Collins et al., 1983). This relationship may be extrapolated to the study site of  (Chinen & Rivière, 1989. Given the well-preserved nature of the edges of the gullies within the trench profiles at Lungenkärchen, it seems that they were filled very quickly, probably at a timescale similar to the Mount St Helens analog ( Figure S12). In one of the very few literature accounts on gullies incised into the ULST deposits, Schmincke (2014, in Abb. 127) presents a V-shaped gully deeper and wider than those identified at Lungenkärchen in combination with syn-eruptive normal faulting adjacent to the WBW, that is closer to the LSV. At that site, a 25 cm-thick layer of primary fallout ashes within the gully infill is indicative of gully formation after violent rainfall which occurred before the end of the eruption (Schmincke, 2014); such syn-eruptive gully formation has also been inferred on Vanuatu (Németh & Cronin, 2007). The lowermost laminated pure ash deposit in the third meter of core LU-L6 (Figures 7 and S9) could be interpreted as primary fallout ash (sensu Schmincke, 2014) and, thus, may also indicate a syn-eruptive start of gully formation at the Lungenkärchen site.
Along the western part of the Lungenkärchen site, the spacing of the gullies is very narrow, in places even < 10 m (Figures 2 and 3), which is less than reported for the larger, more proximal gullies around the WBW; their spacing is 25-50 m (Schmincke, 2014). This positive relationship between channel size and spacing, however, is similar to other fluvial channel networks (according to the catchmentscale concept of stream orders, e.g., Strahler, 1957), and can be From another perspective, the fluvial infill of the gullies results in a higher bulk density and lower saturated water conductivity, which adds to heterogeneous pedo-sedimentary characteristics at the agricultural field scale (see e.g., differential moisture patterns visible in satellite imagery of Figure 2a). These pedo-sedimentary differences may lead to heterogeneous plant growth, crop yields and demands in fertilization, irrigation and pesticide application (Patzold et al., 2008), particularly where the overlying soil horizons are relatively thin. The impact of infilled gullies on agricultural production may be very significant at the field-scale, for example with a reduction of crop yields of up to 100% in a modern, artificially filled gully in the black soil region of northeast China (Liu et al., 2013). However, it remains to be tested if such findings can be transferred to the relatively dense pattern of fossil, much older volcanic gullies adjacent to the LSV and elsewhere.
In contrast to infilled gully structures on agricultural land that are prone to the adverse effects of re-incision and that may show histories of multiple incision-and-infill cycles on decadal to centennial timescales (e.g., Vanwalleghem et al., 2005), those related to the LSE are protected by several decimeters of overlying brown earths and regosols.
At the WBW, the IRSL date of post-LSE stratified silt and tephra deposits corroborates the interpretation of reworking mainly during the Younger Dryas and/or the early Preboreal. The poor bleaching and dim luminescence properties result in relatively large error ranges of 26% and 44%, respectively. Nevertheless, the stratigraphically younger infill of the gully (9.4 ± 4.1 ka) and the stratified silt and tephra deposits (9.7 ± 2.5 ka) reveal very similar ages. Therefore, it seems likely that incision and infill of this unit occurred soon after deposition of the reworked LST and loess during the terminal Younger Dryas or the early Holocene. The error ranges of the OSL data give way to two possible interpretations: i. The proximally reworked loess with tephra components dated to 9.7 ± 2.5 ka was deposited during and subsequent to the late Younger Dryas, a time with reported aeolian deposition (Janotta et al., 1997;Dotterweich et al., 2013) and enhanced fluvial activity (e.g., Andres et al., 2001). Gully incision occurred under periglacial climate conditions with sparse vegetation cover and substantial generation of runoff (Lang & Mauz, 2006;Kaiser et al., 2007). These factors may have prevailed until the beginning of the Holocene as suggested by Dotterweich et al. (2013).
ii. Gully incision was caused by cutting back from the outlet during the Preboreal (cf. Larsen et al., 2013), that is linear incision during the transition towards warmer temperatures, denser vegetation cover, and more stable hillslopes. This scenario would be supported by the youngest part of the OSL age range of 9.4 ± 4.1 ka from the gully infill cut into the reworked loess and tephra of the WBW.