The added value of biomarker analysis to the genesis of plaggic Anthrosols; the identification of stable fillings used for the production of plaggic manure

Plaggic Anthrosols are the result of historical forms of land management in cultural landscapes on chemically poor sandy substrates. Application of plaggic manure was responsible for the development of the plaggic horizons of these agricultural soils. Pollen diagrams reflect aspects of the environmental development but the interpretation of the pollen spectra is complicated due to the mix of the aeolian pollen influx of crop species and species in the surroundings, and of pollen occurring in the used stable fillings. Pollen diagrams and radiocarbon dates of plaggic Anthrosols suggested a development period of more than a millennium. Calluna is present in almost all the pollen spectra, indicating the presence of heath in the landscape during the whole period of soil development. Optically stimulated luminescence dating of the plaggic horizon made clear that the deposition of plaggic covers started in the 16th century and accelerated in the 18th century. The stable fillings, used for the production of plaggic manure and responsible for the rise of the soil surface, cannot be identified with pollen diagrams alone. Biomarker analyses provide more evidence about the sources of stable fillings. The oldest biomarker spectra of the plaggic horizons of three typical plaggic Anthrosols examined in this study were dominated by biomarkers of forest species such as Quercus and Betula while the spectra of middle part of the plaggic horizons were dominated by biomarkers of stem tissue of crop species such as Secale and Avena. Only the youngest spectra of the plaggic horizons were dominated by biomarkers of Calluna. This indicates that the use of heath sods as stable filling was most likely introduced very late in the development of the Anthrosols. Before the 19th century the mineral component in plaggic manure cannot be explained by the use of heath sods. We conclude that other sources of materials, containing mineral grains must have been responsible for the raise of the plaggic horizon.


Introduction
Plaggic Anthrosols occur in cultural landscapes, developed on coversands. These chemical poor Late-glacial aeolian sand deposits dominate the surface geology of an extensive area in northwestern Europe. Plaggic Anthrosols are the characteristic soils that developed on ancient arable fields, fertilized with plaggic stable manure. Plaggic 5 Anthrosols have a complex genesis and are valuable records of environmental and agricultural history (van Mourik et al., 2011).
In previous palaeopedological studies of such soil records in The Netherlands (van Mourik et al, 2011(van Mourik et al, , 2012(van Mourik et al, , 2013, information was unlocked by application of pollen analysis, radiocarbon ( 14 C) and Optically Stimulated Luminescence (OSL) dating. Radiocarbon dates of soil organic carbon, extracted from humic horizons from plaggic Anthrosols, suggested the start of sedentary agriculture between 3000 and 2000 BP but are not indicative for the age of the plaggic sediments due to the complexity of soil organic carbon in plaggic sediments (Mook and Streurman, 1983;van Mourik et al., 1995). It was assumed that farmers used organic sods as stable filling, firstly dug on 15 forest soils and later on heaths for the production of stable manure to fertilize the fields. The mineral fraction of the sods was supposed to be responsible for the development of the plaggic horizon and the raise of the land surface. OSL dating applied on quartz grains extracted from plaggic sediments provides more reliable ages of the plaggic sediments. The OSL dates suggested that the rise of the plaggic horizons started in Introduction Heaths were already present in the Late Paleolithic landscape (Doorenbosch, 2013) and played a ceremonial role in the society of our ancestors. They already had the knowledge to manage the heath as sustainable grazing areas for cattle (Doorenbosch, 2013).
The use of heath for sheep grazing and other purposes as honey production could 5 continue until the middle of the 18th century (Vera, 2011). In the course of the 18th century, the population growth resulted in an increasing food demand. The deep stable economy was introduced and the booming demand for manure resulted in intensivation of manure production. Farmers started with the use of heath sods as (additional) stable filling (Spek, 2004). This caused heath degradation and initiated the second ex- 10 tension of sand drifting. The use of sods finished at the end of the 19th century after the introduction of chemical fertilizers (Spek, 2004). Through the combination of OSL and 14 C dating, historical records and the conventional paleoecological proxy of fossil pollen analysis we have a good impression of the paleoecological environment and the age of such deposits. However, it remains prob-15 lematic to reconstruct the combination of crop residues and various materials used by farmers as stable filling to produce the stable manure, together responsible for the rise of the surface of Anthrosols. This is also hindering a detailed interpretation of the agricultural practices and shifts therein related to the plaggic agriculture system, and specifically the timing of the onset of the intensive heath sod driven deep stable agri-20 culture with which plaggic Anthrosols are most commonly associated. To address this issue, in the present study we expanded our paleoecological toolset with an adapted application of the recently developed biomarker approach (Jansen et al., 2010). This biomarker approach consists of a combination of analytical chemical analysis and modelling with the VERHIB model to unravel concentration patterns of higher chain length the reconstruction of past local vegetation composition of a specific site, but also in studies where the emphasis lies not on the vegetation per se, but rather on reconstructing various sources of soil organic matter input (van Mourik and Jansen, 2013). Goal of the present study was to further explore the applicability of biomarker analysis as part of a multi-proxy reconstruction aimed at unraveling the sources of stable 10 fillings used for the production of plaggic manure in the context of the historic development of the plaggic agriculture ecosystem.

Materials and methods
The distribution area of plaggic Anthrososl in NW-Europe is indicated in Fig. 1. Pape (1972) published the first map of the distribution of plaggic agriculture in NW-Europe. 15 Bastiaens and van Mourik (1995) found traces of intensivation and extension of this area in Vlaanderen (Belgium) while van Mourik (1999b) also reported plaggic Anthrosols in Schleswig (Germany). Beside this area with "real" plaggic Anthrosols, Spek (2004, p. 724) summarized information about the occurrence of soils with some evidence of application of plaggic manure in the Atlantic coastal zones of Norway, Den-20 mark, France, Galicia, Scotland and Ireland.
We selected three previously investigated plaggic Anthrosols with an undisturbed plaggic Horizon In The Netherlands: Valenakker, Nabbegat and Posteles (Fig. 2). Pollen diagrams, radiocarbon and OSL dates were available. The plaggic horizons of these profiles were resampled for pollen and biomarker analysis as part of the present Introduction   (van Mourik et al., 2012) is situated southwest of the city Weert (middle Limburg) on the sport fields of a former college; during the 20th century the soil has never been ploughed or subjected to land consolidation. This profile has never been affected by roots of Zea mays, introduced in the Netherlands in the middle of the 20th century (van Mourik and Horsten, 1995).

5
Nabbegat (van Mourik et al., 2013) is situated on the Maashorst (eastern North-Brabant). The plaggic deposits were buried by drift sand around 1800 AD. Consequently, the plaggic deposits have perfectly been protected against damage by land consolidation or pollution afterwards (van Mourik et al., 2013). The site is now vegetated by oak and birch trees. Roots of these trees may have caused input of organic matter by decomposed roots in the upper part of the plaggic horizon.
Posteles (van Mourik et al., 2011) is situated in Twente (eastern Overijssel). The landowner informed us that during the last three generations this land was never subjected to deep ploughing or land consolidation but since 1960 Zea mays was regular sowed. In contrast to Valenakker and Nabbegat we can expect biomarkers of this deep 15 rooting cultivated plant.

Pollen analysis
Pollen diagrams of plaggic Anthrosols provide paleoecological information about plant species, present on site and in the region during the formation of the plaggic horizon. Previous research showed that pollen grains, infiltrated in soils and incorporated in 20 plaggic deposits, are well preserved in the anaerobic and acid microenvironment of excremental aggregates (van Mourik, 1999a(van Mourik, , 2001. Samples for pollen extraction were collected in 10 ml tubes in profile pits. For a correct matching of pollen and biomarker spectra of the plaggic deposits, the same samples were treated for both pollen and biomarker extraction and analysis. The pollen 25 extractions were carried out using the tufa extraction method (Moore et al., 1991, p. 50). For the identification of pollen grains, the pollen key of Moore et al. (1991, 83-166) was applied. Pollen scores were based on the total pollen sum of arboreal and non- arboreal plant species. For the estimation of the pollen concentrations of the various soil horizons, the exotic marker grain method was applied (Moore et al., 1991, p. 53).

14 C and OSL dating
The determination of the age of plaggic deposits is subjected to various complications (Spek, 2004). Pollen stratification is disturbed by bioturbation and ploughing. Besides, 5 the pollen content is a mix of the regular pollen influx and pollen in stable fillings, used for the production of stable manure (van Mourik et al., 2011). The ages of humic horizons of buried Podzols cannot be determined by 14 C dating due to the complex composition of soil organic carbon (van Mourik et al., 1995). During a period of active soil formation, organic carbon can accumulate in the Ah horizon, especially in the humin 10 fraction but also in the humic acid fraction. The age of the humic acids were considered to approach rather well the moment of fossilization of the Ah horizon after burying by driftsand. The difference between humin and humic acids ages was interpreted as a measure for the period of soil activity and humin accumulation. Later, OSL dating confirmed that radiocarbon dates, not only of the humin fraction but also of the humic 15 acids, overestimate the true ages (Bokhorts et al., 2005). Conventional radiocarbon dating of humin and humic acids showed in presented diagrams, extracted from plaggic deposits, was performed in the CIO (Centre for Isotope Research of the University of Groningen).
OSL dates provide reliable information about the moment of fossilization of plaggic 20 material under the rising furrow because the quartz grain were perfectly bleached during active ploughing (Bokhorts et al., 2005). OSL dating of quartz grains, extracted from plaggic deposits, was performed in the NCL (Netherland Centre for Luminesce Dating, Wageningen University).

Biomarker analysis
A detailed description of the biomarker approach using the VERHIB method is presented in our previous publications (Jansen et al., 2010(Jansen et al., , 2013Van Mourik and Jansen 2013). Briefly, the basis of the method lies in the unraveling of the preserved concentration patterns of C 20 -C 36 n-alkanes, which are exclusive to the epicuticular wax layers 5 on leaves and roots of higher plants (Kolattukudy et al., 1976). While such an application in itself is not new,(e.g. Pancost et al., 2002;Hughen et al., 2004) the novelty of our approach lies in the application of the VERHIB model that we specifically developed to unravel the mixed n-alkane signal encountered in soil or sedimentary archives (Jansen et al., 2010). The VERHIB model consists of a linear regression model that describes how a certain input of plant derived compounds such as n-alkanes over time in a certain archive at a certain location, results in accumulation of these compounds. An inversion of the forward model is used to reconstruct the accumulation encountered with depth into its most likely vegetation origin (Jansen et al., 2010).

15
An important matter of debate when using n-alkane patterns to reconstruct past vegetation input is the genotypic plasticity of the n-alkane patterns, in particular in relation to prevailing environmental factors such as climate (e.g. Shepherd and Griffiths, 2006). In a previous study focusing on vegetation of relevance for reconstructions in ecosystems in North-Western Europe where plaggic agriculture occurred, we found 20 that while genotypic plasticity related to climatic factors may influence the signal, such influence does not eradicate the different vegetation origins (Kirkels et al., 2013). To limit external influences as much as possible, relevant vegetation was sampled in close vicinity to the three study sites as much as possible.
The first group of selected plant species concerned the main sources of stable 25 fillings, used for the manure production: fermented litter from deciduous forest soils (Quercus robur, Betula pendula), grass sods from brook valleys (Molinia caerulea) and heath sods (Calluna vulgaris).  (Jansen et al., 2006). The extracts were subsequently fractionated into three fractions containing the n-alkanes, the esters and the combination of alcohols and fatty acids respectively. For this, a silica column consisting of extracted cotton wool and silica gel was used, followed by elution with hexane, hexane/DCM (4 : 1) and DCM/Methanol (9 : 1) respectively. Separation of the n-alkanes took place by on-column injection of followed electron impact ionization (70 eV). The n-alkanes were identified from the total ion current (TIC) signal by their mass spectra (dominant fragment ion represented by m / z = 57) and retention times and quantified using a deuterated internal standard (d 42 -n-C 20 alkane (Jansen et al., 2010) as well as a conventional external n-alkane standard.

25
The concentration patterns of the n-alkanes with carbon numbers 20-36 in vegetation and soil samples were subsequently used as input for the VERHIB model. In the application of VERHIB, assumptions must be made with respect to rooting depth of the relevant vegetation and the ratio of leaf to root biomass input in the soil (Jansen  , 2010). VERHIB considers the species specific n-alkane patterns in plant roots separately from the patterns in plant leaves, and uses this to deal with the input of young root material at depth. With respect to the ratio of input of leaf vs. root biomass as required by the model, no exact information is available for the soil profile under study. Therefore, in line with the exploratory nature of the present study, we applied an 5 assumed leaf/root biomass input ratio of 1.0 and assumed that while input of leaf material always occurred at the top of the soil profile, root input also occurred with depth.
In the present study leaf material is represented not by natural leaf influx from standing vegetation as that was harvested, but by leaf material from stable fillings deposited on the soil. Since our pilot study in polycyclic driftsand deposits showed that VERHIB 10 was unable to filter out root input sufficiently (Van Mourik and Jansen, 2013), when interpreting the occurrence of a certain species with depth in the profiles under study as modelled by VERHIB, the possibility of young root input being responsible for the signal was explicitly taken into account.

Profile Valenakker
Profile Valenakker is a plaggic Anthrosol (Aan), overlying a ploughed umbric Podzol (2ABp, 2Bs). The pollen diagram (Fig. 3) and the absolute dates (Table 1)  Micromorphological observations (Fig. 5a, b) of the plaggic deposits show the complexity of soil organic matter. There are various sources of organic carbon as plants roots, tissue of table fillings and sods. Also the composition of pollen spectra is complex, a mix of the regular pollen influx of plants on the fields and in the surrounding infiltrating into the soil and pollen, and pollen present in various stable fillings. 5 In previous studies the origin of stable fillings, used in plaggic agriculture, was reconstructed on the base of pollen diagrams (Spek, 2004;van Mourik et al., 2012a, b). The pollen spectra of the Aan horizon show very low scores of arboreal trees but reasonable scores of Ericaceae and Poaceae. Ericaceae pollen may indicate the use of heath sods, Poaceae pollen the use of grassland sods, the combination of sods 10 from degrading heath and the rise of the land surface by plaggic manure is caused by the mineral fraction in such sods. However, the rise of the plaggic horizon of ≈ 60 cm cannot be explained by the use of heath sods if it is true that the use of heath sods (with a mineral fraction) was introduced in the course of the 18th century when better construction materials enabled the farmers to build deep stables (Vera, 2011). In fact, 15 the sources of stable fillings cannot be satisfactorily detected with pollen diagrams.
The biomarker spectrum of the base is dominated by Quercus. Despite the low percentages Quercus pollen it is very likely that the farmers used forest litter as stable filling. The middle spectrum is dominated by markers of Avena and Secale. This points to the use of straw from these crop species as stable filling. Pollen of Cerealia is present in 20 the whole diagram. In the upper spectrum biomarkers of Calluna are present together with Avena and Secale. This points to the use of heath sods as additional stable filling during the last phase in the development of the plaggic horizon.

Profile Nabbegat
Profile Nabbegat is a haplic Arenosol (with Mormoder humus form), overlying a plaggic Introduction The post sedimentary pollen spectra of the 3ABp reflect the start of agriculture (increase of Cerealia) on a former heath (decrease of Ericaceae) in a surrounding with coppice hedges (Quercus, Corylus). Based on radiocarbon dates, the agricultural activities started before ≈ 1000 BC, the OSL dates point to deposition of plaggic material after ≈ 1500 AD. 5 The radiocarbon ages indicate that the farmers used organic matter with very few mineral "contamination" for a long time. The OSL ages indicate that the rise of the plaggic horizon started ≈ 1500 AD due to mineral grains as part of the manure. The plaggic horizon developed between 1500 and 1800 AD. Around 1800 AD, short after the introduction of the deep stable economy (Vera, 2011), the plaggic Anthrosol was overblown by driftsand. Apparently, the use of heath sods resulted in heath degradation, sand drifting and acceleration of the rise of the plaggic horizon (van Mourik et al., 2012a). The sand drifting stabilized under planted Quercus trees; the roots of these trees reached the buried Anthrosol and may have contributed the scores of biomarkers in the upperpart of the buried plaggic horizon. The composition of the pollen spectra of 15 the plaggic horizon is rather uniform, dominated by Ericaceae and Cerealia. Figure 7 shows the results of biomarker analysis. Biomarkers of Quercus were present in all the spectra, dominant in the lower spectra, regular in the other spectra. This points to the use of forest litter as stable filling during the development of the lower part of the plaggic horizon. The main crop species during this time was Spergula. the 2Ap horizon show increasing percentages of Cerealia.
The radiocarbon age of the base of the plaggic deposits (95 cm) is ≈ 850 AD, The OSL age ≈ 1500 AD. The OSL age of the 2Ap (105 cm) is 2035 ± 450 BC, ≈ 3500 year older than sample 95. In this part of the profile we see the effect of bioturbation on the age of the coversand. Grains from the base of the Aan were transported to the 2Ap and

Discussion
Pollen diagrams of plaggic Anthrosols provide valuable paleoecological information to reconstruct the soil dynamics during the plaggic agriculture. However, interpretation of pollen diagrams is complicated. Pollen grains, extracted from plaggic deposits, may originate from two sources (van Mourik et al., 2011). The first source concerns the 5 regional pollen influx from flowering species and local flowering crop species. Pollen grains precipitate on the soil surface and may infiltrate into the Anthrosols by ploughing and bioturbation. This pollen influx will be mixed with the pollen content of materials, used as stable filling to produce manure. Pollen will be preserved in plaggic deposits in the anaerobic and acid micro environment of humic aggregates, produced by worms and micro arthropods (van Mourik, 1999b(van Mourik, , 2001. In general it is not possible to make a clear separation between pollen grains originating from the regular pollen influx or from materials as sods. Therefore, the identification of the various sources of stall fillings cannot be based on pollen analysis alone. Additional information, acquired by biomarker analysis proved very useful 15 for this purpose. In the pollen diagrams Fagopyrum is found in almost all the spectra of the plaggic deposits and in Valenakker and Nabbegat even in the top spectra of the buried ploughed Podzol, probably as result of pollen infiltration. Fagopyrum as crop species on sandy soils was introduced after 1350 AD (Leenders, 1996). Based on this palynological time 20 marker, plaggic deposition started around 1350 AD.
The radiocarbon ages of plaggic deposits are much older. This is caused by (1) older organic carbon, present in the applied stable fillings for the manure production and (2) accumulation of organic carbon during active soil formation. Consequently the radiocarbon dates overestimate the ages of the plaggic sediments, but approach the 25 age of the introduction of agricultural soil management (van Mourik et al., 1995(van Mourik et al., , 2011(van Mourik et al., , 2012a the Celtic fields are an example of a prehistorical agricultural system based on manure management (Spek, 2004). The mineral component of stable manure, applied on the fields, was responsible for the thickening of the plaggic horizon. Ploughing of the furrow will bleach the OSL signal of the mineral grains until the moment that the grains are no longer part of the active 5 soil furrow. For that reason, OSL dating of the plaggic horizon provide reliable ages of the plaggic deposits (Bockhorst et al., 2005). The OSL dates of the profiles Valenakker, Nabbegat and Posteles indicate a start of the thickening ≈ 1550 AD.
It was not possible to determine the sources of stable fillings palynologically. Possible stable fillings were forest litter, sods from moist grass lands and heats sods. But in almost all spectra of the pollen diagrams Ericaceae, Poaceae and arboreal pollen occur. Biomarkers extracted from plaggic deposits, originate from two sources. The first source concerns biomarkers from decomposed roots of crop species, the second source of organic material as straw and sods, used as stable filling for manure production.

15
In the three diagrams we find Quercus as dominant marker in the lowest part of the Aan-horizon, indicating the use of forest litter. In Nabbegat, Quercus markers can also originate from roots of the planted Quercus forest after the stabilization of the sand drifting. This is not the case on Valenakker and Posteles. The middle part of the Aanhorizon is dominated by markers of Avena and Secale, indicating the use of straw as 20 stable filling.
Only in the top of the Aan-horizon markers of Calluna are present, indicating the use of heath sods as stable filling. Based on the results of the biomarker analysis we can conclude that heaths sods were used as stable filling only in the 18th and 19th century. This fits with the observations about the use of heaths in historical archives 25 Vera (2011).
So the question rises about heath management before the introduction of the deep stable economy. Some researchers point to careful heath management before the 19th century. In interviews with farmers, born before 1950, Burny (1999) collected essential SOILD doi: 10.5194/soil-2015-81 The added value of biomarker analysis to the genesis of Plaggic Anthrosols information about historical heaths management in the Belgian Kempen. Before the 19th century, heath sods were never dug on the dry Calluna heath, only on the moist Erica heath. These organic sods were not used as stable filling but as fuel for the furnace. Burning of Calluna heaths was the most important management action to rejuvenate the heath. Juvenile heath is food for cows. Sods digging was a bad action 5 due to the resistance and incoherence of these dry sods and also the long recovery period. Mowing of older Calluna shrubs took place. Twigs were used for roofs, burning and also as stable filling (Burny, 1999). Because of the very low nutrient contribution to the manure of mowed Calluna, the farmers preferred the use of twigs of broom (Genista). 10 An additional factor may be the contribution of biomarkers extracted from sheep droppings. According to Simpon et al. (1999) biomarkers survive the congestion process and stay in the manure. But what do sheep consume? Grazing sheep are very selective in collecting food (Oom et al., 2008;Smits and Noordijk, 2013). They prefer grasses (Molinia, Festuca and Corynephorous). Only late in the season they eat shoots 15 of Calluna, at that time nourishing with high concentrations Ca, Mg and but no P.
If it is true that Calluna heath sods were dug only in the 18th and 19th century, how can we explain the mineral component in the plaggic manure, responsible of the rise of the land surface before that time?
According to Smits and Noordijk (2013) there are several sources of minerals. Firstly, 20 an amount of mineral grains will be incorporated in the manure during emptying out the manure of the stable. Secondly, farmers had the knowledge that the addition of sand could improve the fertility of the soil. Not the leached and acid sand from heath sods but not leached sand, dug on sheep walks and in blown out depressions in nearby drift sand landscapes. Introduction

Conclusions
The vertical zoning of biomarkers and pollen in plaggic horizons are different. Palynologically, the plaggic horizon is homogenous, the biomarker diagrams show clear differentiation.
We can identify various stable fillings used, based on the vertical distribution of 5 biomarkers.
The biomarker spectra of the base layer of the plaggic horizon are dominated by biomarkers of deciduous trees litter (dominated by Quercus), indicating the use of organic matter from the forest floor.
The biomarker spectra of the middle part of the plaggic deposits are dominated by 10 crop species (Avena, Secale), indicating the use of straw from these species as stable filling during a relatively long time.
Only the top spectra of the plaggic horizons are dominated by Calluna, indicating that heath sods were used as stable filling only during the last phase in the development of the plaggic horizon.

15
Profile Posteles shows the impact of the contribution of biomarkers of roots of Zea mays, introduced around 1950 AD, suppressing the other species.
The negligible percentages of Calluna in biomarker spectra of plaggic deposits with exception of the top, suggest an overestimating of the use of heath sods in the traditional interpretation of the genesis of plaggic horizons, the dominance of crop species 20 in biomarker spectra of plaggic deposits suggests underestimating of the use of straw as source material for the production of organic stable manure to fertilize ancient arable fields.