Navigation – Plan du site

AccueilNumérosvol. 30/3Assessing the vegetation history ...

Assessing the vegetation history of european chernozems through qualitative near infrared spectroscopy

Reconstitution de l’histoire de la couverture végétale des chernozems d’Europe par la spectroscopie proche infrarouge qualitative
Barbora Strouhalová, Damien Ertlen, Luděk Šefrna, Tibor József Novák, Klára Virágh et Dominique Schwartz
p. 227-241

Résumés

Les chernozems sont réputés comme étant des sols typiques des steppes continentales. Cependant, on en trouve dans de nombreuses régions d’Europe Centrale au climat favorable au développement de la forêt. Par ailleurs, ces mêmes régions sont soumises à d’importantes influences anthropiques depuis le Néolithique. Il n’est donc pas possible de discuter des conditions naturelles de répartition des chernozems sans prendre en compte de manière approfondie des données paléoenvironnementales significatives. La spectrométrie proche infrarouge qualitative (NIRS) s’est révélée être une méthode innovante et robuste pour identifier l’origine végétale de la matière organique des sols (SOM). Dans notre étude, nous avons développé une bibliothèque de référence spécifique pour les chernozems et autres sols développés sur loess, fondée sur deux groupes de référence : les sols sous végétation prairiale et ceux sous végétation arborée. Afin d’identifier leur histoire paléoenvironnementale, nous avons comparé la signature NIRS de 23 chernozems européens à cette base de références. Il en ressort que la SOM de la majorité des chernozems a bien une signature prairiale, mais que certains de ces sols ont évolué sous végétation forestière. Il apparaît également que les chernozems qui sont actuellement recouverts de forêt ont eu dans le passé une histoire prairiale. Il apparaît également que le type d’occupation du sol est un facteur crucial pour la pédogenèse des chernozems depuis le Néolithique. Ainsi, l’évolution de ces sols ne correspond pas à un seul scénario et il n’est pas possible de les considérer comme un témoin indubitable de la présence de steppes pendant l’Holocène.

Haut de page

Texte intégral

We gratefully acknowledge financial support from the GESSOL Project funded by the ADEME/MEDD and the French Ministry of Foreign Affairs. We thank Maria Jelenska (Polish Academy of Sciences) for providing the soil samples from the H1 and the MD sites, Tereza Zádorová (Czech University of Life Sciences in Prague) for providing the soil samples from the BRC site and Klaus Kaiser (Martin Luther University of Halle-Wittenberg, Germany) for providing the soil samples from the ETZ site.

1 - Introduction

1Various environmental characteristics influence the development of soils. The factors interacting in the pedogenesis of soils are climate, parent material, biota, topography, time and man (Jenny, 1994). When the strength of the climate factor overrides the influence of the other factors over a sufficient length of time, the soils are named as zonal soils (Dokuchaev, 1883). Consequently, the zonality associates the soil types to specific environments. The environmental conditions of chernozems are characterised by a continental climate and by a tall-grass steppe, which are typical for vast zonal belts of chernozems in Eurasia and North America (IUSS Working Group WRB, 2015).

2Chernozem is defined as dark brown or black soil because of its richness in organic matter. Organic horizons are at least 40 cm thick. The saturation in bases, in particular Ca2 + and Mg2 + is high, the pH close to 7; the structure of the aggregates is stable and the bioturbation well expressed (Duchaufour, 1977; Altermann et al., 2005, Němeček et al., 2011). Chernozem develops on a carbonate bedrock, usually loess.

3In Eurasia, the distribution of chernozems is connected to the areas of middle latitudes with an annual of precipitation varying from 250 to 600 mm (fig. 1). From west to east, the area they cover increases (Ganssen, 1972). The westernmost tip of the chernozem belt is formed by discontinuous areas of chernozems in Central Europe. Since the Neolithic, the environmental conditions of this zone have been strongly influenced by the human factor. So, we may ask what the proportion of man’s influence and other factors of pedogenesis is. Nowadays most of chernozems are used for agriculture. Only very tiny areas of chernozems are covered with natural or semi-natural grassland. From a botanical point of view, some of them are considered as primary steppes, some as secondary grassland (e.g., Molnár & Biró, 1996; Fekete et al., 2000; Horváth, 2010; Hejcman et al., 2013). They form isolated islands in the landscape and they are usually protected by law. The biotope of the loess steppe (the steppe on chernozem) in Central Europe is considered endangered due to (1) the extension of agricultural land, (2) the invasion of trees and bushes and (3) the decrease in botanical diversity (Virágh et al., 2008; Ložek, 2011).

4Despite the general perception of chernozems as steppe soils, chernozems are present in zones suitable for the natural existence of forests. Nowadays, the woodlands of thermophile oaks or oaks-hornbeams are considered to be characteristic of the climactic vegetation in the zones of the current distribution of chernozems (Neuhäuslová et al., 1997). In literature, we find references about woods on chernozems in Germany (Ehwald et al., 1999; Eckmeier et al., 2007); south Moravia (Chytrý et al., 2010, Vysloužilová et al., 2014), south Slovakia (Hrabovský et al., 2010) or southwestern Ukraine and Moldova (Mucina et al., 2016).

5The persistence of the steppe environment during the entire Holocene in Central Europe was subject to many scientific discussions (see, e.g., Ložek, 1973; Magyari et al., 2010; Hejcman et al., 2013; Pokorný et al., 2015). The origin of steppes in the Eastern European zone was also discussed (Boonman & Mikhalev, 2005). Similar to Central Europe, there are only a few nature reserves in Eastern Europe which protect chernozems under the steppe vegetation (Chendev et al., 2015).

6The palaeoenvironmental data concerning chernozems in the Holocene and in the Pleistocene in Central Europe demonstrates the ambiguity of the typical vegetation cover of chernozems. The palaeomalacological, phytolite and pedoanthracological data shows the environmental conditions of the site on a local scale. The research of chernozems in Central Europe gives various results. Malacology (Ehwald et al., 1999; Ložek, 2004) is indicative of the steppe environment. On the contrary, pedoanthracological research (Pécsi et al., 1984; Kreuz, 2008; Vysloužilová et al., 2014) or phytolite analysis (Barczi et al., 2009) provide evidence for the presence of trees on the areas of chernozems. Palynological studies refer to the presence of trees on chernozems (Havinga, 1972; Svobodová, 1997; Magyari et al., 2010). Nonetheless, it has to be mentioned that palynology is not well adapted to a large spatial scale, which is a weakness, especially in the zones where chernozems and luvisols coexist. Geochemical studies of German chernozems indicated the importance of fire in their pedogenesis (Schmidt et al., 1999). According to this study, black color of chernozem is due to past fires with possible human influence. Their conclusion is supported by analyses of chernozems in Saskatchewan (Ponomarenko & Anderson, 2001).

7In this paper, we explore the role of vegetation for the formation of chernozems. Focusing on the existing studies on the influence of (palaeo)environmental conditions and vegetation in the pedogenesis of chernozems, there is, to our knowledge, no study that is concerned with soil characteristics directly. Soil organic matter (SOM) represents a kind of pedological geoarchive (Ertlen et al., 2010; Ertlen et al., 2015; Vysloužilová et al., 2015). Our approach is based on the method of the near infrared reflectance spectroscopy (NIRS) analysis of the SOM. This study of the soils follows the study by Vysloužilová et al. (2015) of buried chernozems. There is a fundamental difference between these two studies: the evolution of the buried chernozems was stopped at the moment of their burial; on the contrary, the evolution of the (unburied) chernozems, which are active from a biological or chemical point of view, has been continuing with human interference and SOM has been renewing. The near infrared spectra reflect the present stand of the SOM. Nevertheless, SOM can be conserved in chernozems for thousands of years as a consequence of the low rate of SOM renewal (Larionova et al., 2008).

8In general, SOM is heterochronous and consists of fractions ranging in age from days to millennia (Scharpenseel & Pietig, 1970; Gregorich et al., 1994). The measurable 14C age provides a life expectancy of the SOM in the soil, called mean residence time (MRT), which is a non-arithmetic average of the ages of the SOM fractions (von Lützow et al., 2007; Schwartz, 2012). In hernozems, there is a very strong MRT gradient with relation to the depth: at a depth of 10 cm, the MRT can reach 1000 years; at a depth of 25 cm, it varies between 1000 and 3000 years, at a depth of 60 cm, it varies between 1000 and 4000 years; at a depth of 90 cm, it varies between 2000 and 6000 years (see figure 2, references inside). The average value of the gradient reaches 566 years every 10 cm in depth. Our study takes advantage of the high values of the persistence of the SOM. Thanks to NIRS, it is possible to determinate the vegetational origin of the SOM (Ertlen & Schwartz, 2010). Consequently, according to Ertlen et al. (2015), it is possible to model the changes in vegetation. This helps us to understand the environmental conditions that are relevant for the process of pedogenesis. Consequently, we tested the hypothesis that the presence of chernozems could be an indicator for the distribution of steppes in the past.

Fig. 1: Distribution of chernozems (European Commission & European Soil Bureau Network, 2005) and of studied soils in Europe.

Fig. 1: Distribution of chernozems (European Commission & European Soil Bureau Network, 2005) and of studied soils in Europe.

Chernozems are concentrated along the 50th parallel north in a band of increasing latitudinal extension to the east. The NIRS model of vegetation changes at each site is simplified in 5 cases: (1) stable grassland vegetation, (2) stable forest-steppe vegetation, (3) grassland to woodland transition, (4) woodland to grassland transition, (5) complex evolution between grassland and woodland.

Tab. 1: Characteristics of the studied soils.

Tab. 1: Characteristics of the studied soils.

Fig. 2: Mean residence time (MRT) of chernozem soil organic matter versus depth.

Fig. 2: Mean residence time (MRT) of chernozem soil organic matter versus depth.

2 - Material and methods

2.1 - Regional setting

9Our study is focused on Central Europe, which is a marginal region belonging to the chernozem belt of Eurasia. We preferred to study sites which are not used as arable land, so they may better reflect natural soil development. The analysed soils come mainly from the Czech Republic, Slovakia and Hungary. Then, to a lesser extent, we also studied chernozems located in Poland, Germany, Ukraine, and Russia. In total, we investigated 23 chernozems. Their locations and characteristics are listed in table 1. Each soil has a three-character sign that is used throughout the whole study.

10With the exception of the BRC and the POP, these soils are not cultivated nowadays. Most of the sites are situated under grassland, but four sites are situated under woodland (BAB, SEN, BUL, DUB).

11This study is focused on chernozems and on soils that have similar properties to chernozems defined by the World Reference Base (IUSS Working Group WRB, 2015). We admit that not all of the studied soils would perfectly fit to the WRB 2015 quantitative criteria for chernozems. This deficiency is given by the fact that we consequently looked for localities where chernozems exist under the natural vegetation. Such localities are very rare in Central Europe. The studied soils come from territories where they are classified as chernozems in the national soil classification systems (Földvári & Darab, 1966; Sobocká, 2000; Němeček et al., 2011). These classifications may be slightly different to WRB 2015 (Vysloužilová et al., 2016). The studied soils present the main pedogenic features of chernozems – i.e., the presence of a dark chernic horizon, a crumb structure, a calcareous parent material, high base saturation or strong biological activity. For each of the examined soils, we systematically measured the total organic carbon (TOC) content, the CaCO3 content, the pH and the particle size distribution (tab. 1).

2.2 - Near infrared spectroscopy: a tool to study vegetational origin of the Som

2.2.1 - Principle of the method

12NIRS is an analytical method involving diffuse reflectance measurement in the near infrared region (780-2500 nm). In pedology, it has been used for assessment of various qualitative and quantitative soil parameters (Cécillon et al., 2009). Reflectance signals result from vibrations in C-H, N-H and O-H chemical bonds which are typical for soil organic matter. The near infrared (NIR) spectra are considered as the fingerprint of the SOM (Palmborg & Nordgren, 1996; Coûteaux et al., 2003). Direct qualitative interpretation of soil characteristics is almost impossible, therefore the reading of the spectra requires using of multivariate statistics. Principal component analysis (PCA) was first success-fully tested by Velasquez et al. (2005) to distinguish different land-uses. Qualitative soil parameters were further successfully evaluated by other authors (Cohen et al., 2006; Awiti et al., 2008; Cécillon et al., 2009).

13Using of NIRS for evaluation of the vegetational origin of the SOM was established by Ertlen (2009) and Ertlen et al. (2010), further developed in Ertlen et al. (2015). The approach is based on comparison of near infrared spectrum of a certain sample of an unknown origin with a set of samples from the spectral reference library using multivariate discriminant analysis.

14The spectral reference library was established on the basis of samples from the topsoil horizons that are covered with well-known vegetation (Vysloužilová et al., 2015). Two main vegetation groups are distinguished: woodland and grassland. The two poles, woodland and grassland, were chosen in order to verify the hypothesis of presence of steppe in the Holocene in Central Europe. The library is designed in a way that makes the statistical model focus on the qualitative parameters that we have observed: the vegetational origin of the SOM. That is to say, we selected a large number of sites and samples for the two groups, with a wide range of other soil parameters such as particle size, pH, total organic carbon (TOC) content and, more generally, a wide range of geographical context. So, the origin of the organic matter is the only common denominator within the two groups. Moreover, in order to avoid the specific effects of the physical parameters such as grain size distribution and of the amount of TOC, specific mathematical treatments are applied (§ 2.2.4).

2.2.2 - Development of the two-sided NIRS spectral reference library

15Building the spectral libraries improves the determinative ability of the NIRS (Cécillon et al., 2009). We put a high effort in construction of a robust model of vegetational origin of the soil organic matter adapted on chernozems. It was difficult to find chernozems with stable natural or semi-natural vegetation for the construction of the spectral reference library (Vysloužilová et al., 2015), because chernozem is a soil which is nearly always used as agricultural land. Consequently, we extended the reference soils to all kind of soils developed on loess. 24 sites across Europe (tab. 2) were chosen according to the stability of their vegetation (woodland or grassland). In this manner, at least 95% of the SOM comes from the present ecosystem and the purity of the spectra could be guaranteed (Ertlen et al., 2010). As recommended by Ertlen et al. (2010), we considered the vegetation to be stable if it has not changed for at least the past 150 years. This stability was verified from maps starting from the second military survey maps to the maps of today. The second military survey maps were conducted in the Austrian Empire - in Bohemia (1842-1852), Moravia (1836-1840), and Slovakia and Hungary (1819-1869). The maps, from the different time moments, are available on the map servers
(https://mapire.eu, www.mapy.cz, www.archivportal.hu).
The sites found in France are documented in the Cadastre survey (1760) and in the Archives of National Forest Office (1860 to present) (Ertlen & Schwartz, 2010). The history of the sites in Russia and in Ukraine was acquired from the literature (Jelenska et al., 2008; Khitrov et al., 2013). In some cases, ancient botanical descriptions were available. Moreover, we discussed the past land use and the vegetation changes with the local researchers who have good knowledge of the sites. It should also be noted that in some cases, the number of old documents (maps, then aerial photos) is such that the risk of breaking ecological continuity for 100-150 years is very low. This is particularly the case for forests. Thus, a set of 15 to 30 topsoil samples from each site was collected. The reference library contains 428 samples in total. In the field we used a 4 cm deep steel cylinder for taking the samples according to the method that had been set up by Ertlen et al. (2010).

16For each examined reference site, we systematically measured the total organic carbon (TOC) content, the CaCO3 content, the pH and the particle size distribution in order to check the relation between the soil properties and the vegetation.

Tab. 2: Sites that were used for the construction of the NIRS reference library.

Tab. 2: Sites that were used for the construction of the NIRS reference library.

2.2.3 - NIRS analysis of soils with an unknown history

17According to the MRT measurements of the SOM in the chernozems (see Scharpenseel & Pietig, 1970; Vysloužilová et al., 2014), it is evident that the SOM is, on average, several millennia old in the inferior parts of the humic horizon of the chernozems. Therefore, we suppose to reconstruct the vegetation history of the soils by studying the SOM from the entire profile (Ertlen et al., 2015). In every studied site (tab. 1), we systematically sampled three replicates every 5 cm from the surface to the limit of the C horizon.

2.2.4 - Acquisition and treatment of the spectra

18The samples of the reference library and from soils with an unknown history were analysed the same way. They were dried in an oven at 40°C and sieved to pass through a 2 mm mesh. The samples were placed one by one in a rotating cup (diameter 9 cm) and scanned in the wavelength range of 10000 to 4000 cm-1 with a resolution of 2 cm-1. An FT-IR Frontier Spectrometer (PerkinElmer, USA) with a CaF2 beam-splitter, with an integrating sphere and with an InGaAs detector was used. Each spectrum was measured out of an average of 99 scans of different subsamples on a surface of circa 1 cm2. It is possible to measure this way due to the rotating cup. The measured reflectance (R) was transformed into absorbance (A) by using the following equation: A = log10 (1/R) (fig. 3a). The resulting data matrix with 3001 columns was averaged on 8 cm-1 in order to reduce the data matrix to 751 columns. The standardisation and the first derivative are known to reduce the effect of the physical properties and to reduce the effect of the quantity of the organic matter, respectively (McClure, 2001; Shenk et al., 2001; Coûteaux et al., 2003; Ertlen, 2009). The first derivative and the Standard Normal Variate (SNV) were applied as the mathematical pre-treatment (figs. 3b & 3c) to improve the spectral discrimination of the reference library. The aim of this procedure is to distinguish between the two kinds of the origin of the SOM: woodland and grassland.

19The data matrix of the reference library is the object of a multiple discriminant analysis in order to detect the spectral differences between the soils under the woodland and under the grassland (Ertlen et al., 2010). In this way, the wavelengths contribute to the signal discrimination of the two groups (Viscarra Rossel & Webster, 2011). The number of varieties plus the number of groups must be lower than the number of individuals. That is why the data matrix must be reduced. The segment of the spectra between 10000 and 7304 cm-1 was removed because there is not any relevant information contained there. Therefore, the matrix has 427 rows and 415 columns (wavenumber bands). The matrix is divided in two sub-matrixes: there is one for the grassland with 292 rows and one for the woodland with 135 rows. The Mahalanobis distance is used to evaluate the discrimination between the two groups (Ertlen et al., 2010; Vysloužilová et al., 2015). Consequently, the discriminant function calculated from the reference library is applied on the unknown spectra collected from the studied soils.

20The calculated two-sided model of the reference library (see above) was applied on a set of 614 spectra samples from the 23 studied soils with unknown history.

21The comparison of values from the multiple discriminant analysis of the unknown data with the reference library data enables one to interpret the origin of the SOM (grassland or woodland). To study the evolution of the vegetation over time, we examined the score of the multiple discriminant analysis of the NIR spectra at every sampled depth. This approach allows for the identification of relevant tendencies on a millennia scale but it does not give an exact age of the changes in vegetation (Ertlen, 2009; Ertlen et al., 2015).

Fig. 3: Three examples of NIRS spectra.

Fig. 3: Three examples of NIRS spectra.

A/ Spectra without any pre-treatment; B/ Spectra after standardisation; C/ Spectra after standardisation and application of the first derivative.

2.3 - Analytical properties

22Samples of about 50 g were taken at every 5 cm depth from each studied soil. They were dried in the oven at 40°C and sieved to pass through a 2 mm mesh. The particle-size distribution was measured with a laser granulometer (type Beckmann-Coulter LS230). Before the measurement of the particle-size distribution was made, the samples had been treated by H2O2 to destroy the SOM. The samples had been washed by KCl also, distilled water and sodium hexametaphosphate to deflocculate the aggregates without destroying the carbonates. The total organic carbon (TOC) was measured by the wet oxidation method (Walkey & Black, 1934). The content of carbonate was quantified by measuring the volume of the CO2 lost in the reaction with HCl in a closed atmosphere.

3 - Results

3.1 - Properties of studied soils

23In the investigated chernozems, the maximum concentration of the TOC is in the upper horizons, where it varies between 1.7% and 7.5% (tab. 1). This percentage decreases with the depth: the gradient is smaller under the grassland vegetation and greater under the woodland. In deep horizons, the content of the TOC remains high (between 1% and 2%), which is typical for isohumic soils.

24Chernozems are formed on a carbonated substrate; the parent material is mostly loess, except for the POP, where it is calcic marl. The parent material of the soils HUU, HUUD, HUS is a sandy loess sediment. The content of the CaCO3 in the A horizons is low – 0 to 2% (tab. 1). At the base of the profiles, an accumulation of secondary carbonates often occurs, which causes a sudden increase of the carbonate content. Depending on the profile, the percentage of the carbonate content can be up to a few percent. In some cases, the studied soils have an extreme level of carbonates: POP is formed on marl where the content of the CaCO3 reaches 15%. However, the arenic chernozems (HUU, HUUD, HUS) reach a percentage of CaCO3 close to zero.

25The pH of the surface horizons varies between 5 and 8 and it generally increases to the maximum of 8.5 in the C horizons. There is a relation between the pH and the land cover. The soils under woods (BAB, BUL, DUB, SEN, HUGF) are more acidic with a pH that goes down to 5. There are two acidic steppe soils: HUR and HUS, respectively, a luvic chernozem and an arenic chernozem that remains carbonated at the base of the profile only.

26The particle-size distribution is dominated by silts or, in a few cases, by very fine sands. In general, the clay content is low.

3.2 - Reliability of the two-sided NIRS reference library

27The values of the TOC, pH, and CaCO3 contents measured separately from each other are independent on the type of the vegetation cover. Still the coupling of the values of the TOC and pH and TOC and CaCO3 reveals a slight separation between the populations of woodland soils and grassland soils (fig. 4).

28If these differences seem to be normal, they may affect the interpretations of the qualitative spectra. However, it is not possible to find soil under forests and grasslands with entirely comparable analytical characteristics.

29The discriminant analysis of the near infrared spectra distinguishes two groups of soils without overlapping (fig. 5): topsoils under woodland and topsoils under grassland. The average discriminant scores for the grassland and the woodland are -5.508 and 11.954, respectively. The difference between these two values – the Mahalanobis distance – reaches 17.462. This confirms that the model from our current reference library can be reliably applied on the 23 studied soils.

Fig. 4: Relationship between soil properties for grasslands and woodlands according to the reference spectral library.

Fig. 4: Relationship between soil properties for grasslands and woodlands according to the reference spectral library.

Fig. 5: Histogram of the canonical scores for grassland and woodland soils

Fig. 5: Histogram of the canonical scores for grassland and woodland soils

3.3 - Application of the NIRS model to the soils with an unknown origin of vegetation

30There are 614 samples with an unknown vegetation origin: 263 samples are classified as grassland, 62 samples as woodland and 289 samples do not fit in one of these categories only. 230 of the 289 samples are concentrated in-between these two categories. These in-between values can be understood as a mixture of the SOM resulting from the succession of grass and woody ecosystems (see also § 3.4). Furthermore, the score of the other 59 samples are beyond the range defined by the class of grassland or by the class of woodland.

31For 56 samples, the value of the canonical function is below -10, so they are very difficult to interpret. This group includes samples of the base of the HUUD and the HUR profiles. Their low ranking is probably related to the very low C content of these samples (< 0.3% TOC for the concerned depths) collected in the A/C or the Bt horizons. Froehlicher (2016) does not recommend the application of NIRS on samples with a TOC lower than 0.25%, because the contribution of the SOM becomes too low, despite Burns and Ciurczak (2001) estimate for NIRS in general that a concentration of the TOC exceeding 0.1% of the total composition is enough. However, in this discussed group, we also found all the samples from the POP profile that are very rich in SOM. Probably, a significant amount of CaCO3 (15% to 30%) in the POP profile disturbs the signal. POP is also the only chernozem developed on marls. In this group, there are also some isolated samples belonging to the profiles H1 and HUB as well. They do not mislead the interpretations of the rest of the profiles. On the right side of the histogram, only 3 samples show values that are slightly higher than 15. Actually, a detailed analysis of the soils can be undertaken profile by profile. All the profiles with the exception of POP, HUR and HUUD can be interpreted from a palaeoenvironmental point of view.

3.4 - Evolution of the landcover evolution

32We have put the canonical scores and the depth of the samples (tab. 1) into relation. As we mentioned in section 3.2, some values rank between the class of woodland and the class of grassland. Two scenarios could explain this mixture: i) the SOM originates from a succession of two different ecosystems like, i.e., the change from woodland to grassland (or vice versa), or ii) there is a single source of the organic matter from a forest-steppe vegetation type. Depending on how the scores change in relation to the depth, we can tell if there is an intermediate vegetation (forest-steppe) of a steady character (the value is expected to be stable throughout the profile), or if the value represents a mixture of the SOM that is successive to a vegetation change from one type to another one (the value is expected to show a trend from one kind of vegetation to the other within the profile as a consequence of the MRT gradient with regards to depth).

33The forest-steppe ecosystem represents in our study an intermediate degree between grassland and woodlands. This ecosystem is regarded as an open biotope with the occasional presence of trees, groups of trees or shrubs. The forest-steppe ecosystem on chernozem is currently difficult to find in Central Europe. The Catalogue of biotopes in the Czech Republic (Chytrý et al., 2010) and the Catalogue of biotopes of Slovakia (Stanová & Valachovič, 2002) do not define the forest-steppe ecosystem on chernozem. By contrast the transition zone between forests and steppe is described e.g.by Fekete et al. (2009).

34Five types of evolution that can be observed (groups of fig. 1 and fig. 6):

35Group 1: the score of nine profiles (CT, ETZ, HUB, HUM, HUS, HUG, KOC, KUR, MIK) corresponds almost entirely to a grassland reference. This means that the organic matter in these soils has a single origin from grassland.

36Group 2: the full scores of three profiles correspond to the area of the intermediate values between the grassland and woodland reference groups. There are no significant changes throughout the thickness of the profile: BRC, BRO, HUGF. This means that the SOM has a single origin from a forest-steppe.

37Group 3: five profiles show a trend rather linked to forest-steppe at depth and tend to change towards forest in the upper parts: BAB, BUL, DUB, KUC, SEN.

38Group 4: one profile (H1) shows an opposite trend: the woodland score at depth changes to a grassland score in the upper horizons.

39Group 5: two profiles show a complex evolution in-between woodland and grassland: HUU, MD.

Fig. 6: Examples of typical NIRS profiles of the chernozems.

Fig. 6: Examples of typical NIRS profiles of the chernozems.

Group 1 (ETZ): stable grassland vegetation; Group 2 (HUGF): stable vegetation of the forest-steppe; Group 3 (BAB): deep grassland vegetation with a tendency towards a forest in the upper part; Group 4 (H1): deep forest vegetation with a grassland trend in the upper part; Group 5 (HUU): complex evolution of the vegetation between forest and grassland clusters. The shaded areas correspond to the two groups defined by the reference library of chernozems. The POP profile cannot be interpreted.

4 - Discussion

4.1 - The reliability of the nirs approach

40First of all, it is necessary to mention that the physicochemical properties of the soils play a role in the absorbance of spectra, especially for carbonate content (Gaffey, 1986; Stenberg et al., 2010). Therefore, we verified if the physicochemical properties of the samples can disturb the interpretation of the near-infrared spectra on vegetational origin of the SOM or not.

41We analyzed the following soil parameters: TOC content, pH, CaCO3 content and grain-size distribution. We have shown that these two populations overlap when comparing these properties one by one between the forest soils and grassland soils. However, by coupling these characteristics two by two, in particular the pH and the TOC content, we observed that the two populations are, in fact, different. These differences are directly related to the vegetation: in forests, the SOM content is higher at the surface, because the fresh organic matter inputs are on the soil surface as a consequence of originating from dead leaves. This results in a higher acidity in the superficial horizons and lower carbonate content. These differences are impossible to avoid, because they are intrinsically linked to the different ways in how these two types of ecosystems function.

42According to Stenberg et al. (2010), CaCO3 mainly influences the near-infrared spectra around 2335 nm and to a lesser extent around 1870, 1990, 2160, 2335 and 2500 nm. Among the bands that most contribute to the discrimination of our classes (fig. 7), the maximum is between 1888 and 1909 nm, which are close, and may be in relation with the band around 1870 nm. The largest carbonate peak around 2335 nm is also taken into account in our model with a modest coefficient of 1.77. We admit that the CaCO3 content plays a role in the qualitative analyses of the spectra, but that this influence is not prevailing. Nevertheless, if some physicochemical parameters strongly correlate to the vegetational origin, they reinforce the SOM fingerprint in the model.

43Overall, the physicochemical variables of the reference spectral library vary in a shorter range than the amplitudes presented by Ertlen (2009) and Froehlicher (2012). This is explained by the deliberate choice to restrict the reference library to soils developed on loess and chernozems. Thus, it must be kept in mind that this rather homogenous library is intended solely to determine the specific vegetational origin of the organic matter of chernozems. Consequently, the proposed model based on a reference library of chernozem and soils on loess is representative and dedicated only for this type of soil.

44The discriminant analysis of the two-sided reference spectral library built on chernozems confirms that there is an obvious distinction of signatures between grassland and forested soils. This demonstrates the reliability of the two-sided library. Indeed, it allows us to perfectly classify 100% of our reference samples.

45The restriction of our reference library to soils developed on loess leads to a lower range of variation in the physicochemical parameters (carbonates, pH and grain-size distribution), which may affect the spectra and reduce their influence. On the other hand, the content of the SOM in the chernozems is relatively high and homogenous and decreases very slightly with the depth. Moreover, the gradient of the MRT is important and constant. Consequently, the chernozem is a type of soil that is very convenient for the presented approach.

46The quality of the distinction between the two classes is given by the importance of the Mahalanobis distance. In our model, this value reaches 17.462, whereas Ertlen et al. (2010) observe a Mahalanobis distance of 12.27 between grassland soils and forest soils under a more heterogenous set of sites. In contrast to Ertlen et al. (2010), we do not observe an increase in Mahalanobis distance after the application of second derivatives.

47Last but not least, the reliability of the discussed model is given by the fact that the upper part of the A horizon is always correctly classified in comparison with the vegetation that was observed on the sites during the fieldwork. The results are also coherent with ancient maps and with other documents that give evidence about the vegetation cover on a centennial scale.

4.2 - Contributions of the NIRS to the palaeo-environmental questions

48The obtained values of the scores with the depth suggest that:

49- the majority of the studied chernozems, which are currently covered by grassland, have a past of grassland exclusively;

50- some chernozems that are currently under grassland were partly under woodland in the past;

51- all the chernozems that are currently under woodland have a grassland past (none of the studied soils has a past of woodland exclusively).

52Our results confirm that chernozems have developed under grassland mostly. At the same time, the results indicate that they can be preserved under woodlands for a certain period of time. The presence of trees on chernozems in the past is supported by other palaeoecological findings. Especially, the pedoanthracological data gives evidence of trees on chernozems in Glacials (Pécsi, 1984; Willis & van Andel, 2004; Vysloužilová et al., 2014) as well as in the Holocene (Beneš, 2008; Kreuz, 2008; Vysloužilová et al., 2014). The presence of trees is also indirectly proved by the palynological studies (Havinga, 1972; Svobodová, 1997; Kuneš, 2015). The result of a phytolith study from a chernozem of the Great Hungarian Plain (Barczi et al., 2009) agrees with our results. It rejects a continuous presence of woods on the chernozems, but it disapproves the existence of pure steppe as well. Likewise, this outcome is supported by several current findings of chernozem under woods (Eckmeier et al., 2007; Vysloužilová, 2014; see § 1), by conclusions of other studies (Rohdenburg & Meyer, 1968; Ehwald et al., 1999; Fischer-Zujkov, 2000) or by the NIRS results of palaeochernozems (Vysloužilová et al., 2015). On the contrary, our results are in conflict with palaeomalacological studies of Holocene buried chernozems (Smolíková & Ložek, 1978) and Pleistocene loess-palaeosol series (Ložek, 2001). The synthesis of multiple malacological analyses of Central European area concludes that a chernozem is an unambiguous indicator of a steppe environment (Ložek, 1973, 2004).

53If the presented results confirm the presence of woods on the chernozem, the question about the possible duration of the periods with forests comes into consideration. From the time point of view, NIRS is not able to detect a short episode of woodland, because of the very slow turnover of the SOM. In the superficial horizon, woodland can be recorded for some centuries (DUB, BUL). Except in this case, if there is a wood signal over several tens of centimetres, it is because the episode was relatively long. Chernozems are, therefore, able to evolve and persist under woodland for several millennia (KUC, BAB, SEN, HUGF, MD, HUU).

54Palaeoenvironment studies from regions of our interest can give us an idea about the appearance of these woodlands. Pokorný et al. (2015), in their malacological-palynological study from Central Bohemia (lower Ohře region), claim that semi-open forest-steppe landscape was replaced by a Neolithic cultural landscape which could be a suitable environment for the formation and persistence of chernozems. The results from Central Bohemia (BRO, KOC, KUC) support their conclusion, each site shows another result. The pre-Neolithic landscape is described as a forest-steppe with dominating Pinus and Betula with the local occurrence of broadleaf deciduous trees such as Quercus, Acer, Tilia and Ulmus with stably present herbaceous taxa (Pokorný et al., 2015).

55The region of the north-western fringe of the Great Pannonian Plain was recently studied by Kuneš et al. (2015) and Jamrichová et al. (2017). A recent multiproxy palaeoenvironmental study of the Lake Vracov in the southeast Czech Republic (Kuneš et al., 2015) suggests that the landscape was continuously partly opened during the entire Holocene. According to their outcome, favourable climatic conditions (arid summers, cold winters) lead to the evolution of grasslands and open woodlands and to the formation of chernozems. Jamrichová et al. (2017) proved that the pre-Neolithic vegetation was composed by open forest-steppe with dominance of pine. Since the Neolithic, different types of woodland management have created a more open forest structure and expanse of light demanding species, an open environment was favorable to the formation and persistence of the chernozem. The ratios of woodlands and grassland have been fluctuating over the time, which influence the NIRS spectral image of the chernozems from this region (BRC, BAB, BUL, DUB, SEN).

56The origin and continuity of grasslands during the Holocene was questioned in the Great Hungarian Plain as well. The observed Hungarian NIRS profiles (HUB, HUM, HUS, HUG, HUU) confirm that there has been a stable presence of grasslands in the Great Hungarian Plain (Molnár & Biró, 1996; Magyari et al., 2010) for centuries, but do not exclude a singular presence of trees (Barczi et al., 2009). Among the study sites, a stable NIRS signal of forest-steppe was found in the HUGF profile only.

57The samples from the zonal Eastern European chernozems (MIK, KUR, CT) confirm the grassland character of the sites which are considered to be virgin steppe on the chernozem (Jelenska et al., 2007; Khitrov et al., 2013).

4.3 - Pedological implications for the Holocene

58The late Glacial and Early Holocene are designated as the most probable time for chernozem pedogenesis (Ehwald et al., 1999; Ložek, 2004; Lorz & Saile, 2011). Various authors questioned the conditions surrounding their preservation up to today in central Europe (Ehwald et al., 1999; Fischer-Zujkov, 2000; Ložek, 2004; Eckmeier et al., 2007). In most cases, the palynological findings indicate the forestation of Central Europe in the first half of the Holocene. Consequently, that raises the question if the chernozems start forming only after Neolithic clearings (the beginning of the Neolithic ca. 7500 BP) and under continuous cultivation that would simulate dry conditions. The MRT data is not in contradiction with this hypothesis. Consequently, it has to be considered if it fits into the rank of the anthric chernozems that are defined in the classifications as a type of chernozem resulting from long-term cultivation (IUSS Working Group WRB, 2006; IUSS Working Group WRB, 2015; Vysloužilová et al., 2016).

59In a previous study by Vysloužilová et al. (2015), NIRS was used to study chernozems buried in the Holocene. Similar to the results of the soils from this study, certain palaeosoils have shown a steppe character and some others have been forested. Nonetheless, these contradictory outcomes confirm that a mosaic of vegetation of the forest-steppe character used to occupy the chernozems in Central Europe.

60Nowadays, the differences in vegetation are perceived on a very local spatial scale as we can especially observe at the site of the Gödöllő Hills - Szarkaberki Valley in Central Hungary. The score of the HUGF profile is relatively close to the reference class of woodland. The present vegetation covering the soil is a forest-steppe. The stability of the scores with regards to depth suggests that all organic material gives evidence of this kind of ecosystem. However, the HUG profile, which is located very close to HUGF, is marked by a grassland fingerprint. In this case, we should also consider the stability of these ecosystems over lengthy millennia terms.

61In order to bring valuable palaeoenvironmental reconstructions of specific sites, it is crucial to use tools which are adapted to a local scale. Among the approaches adapted to the local scale, a NIRS of the soils can be considered as the best performing method. In the case of a chernozem, NIRS showed to be a powerful tool.

5 - Conclusion

62On the basis of the study of European chernozems by means of NIRS, the following conclusions can be drawn:

63- the NIRS two-sided reference library built under perfectly defined conditions is able to distinguish woodlands and grasslands according to the characteristics of the spectral image of the soil sample. Therefore, it can be reliably applied to study the environmental history of chernozem sites with regard to environmental context of the reference library. When it was possible to compare, the NIRS results correspond with other information about the palaeoenvironment;

64- a chernozem is not exclusive to grassland soil. Even if the grassland is necessary for the development of the chernozem, the chernozem is able to support woodland episodes. These episodes can last up to a few millennia;

65- the variety of our results (groups 1 to 5) and contradictory findings from other palaeoenvironmental methods indicate that stable forest-steppe vegetation was the typical vegetation on the chernozems in the Holocene;

66- it is rather misleading to consider chernozems as an indicator of a Holocene-long grassland vegetation;

67- consideration of the spatial scale (local or regional) in the palaeoenvironmental studies can explain various interpretations of the history of chernozems.

Haut de page

Bibliographie

ALTERMANN M., RINKLEBE J., MERBACH I., KÖRSCHENS M., LANGER U., HOFMANN B., 2005 - Chernozem–Soil of the Year 2005. Journal of Plant Nutrition and Soil Science, 168, 725-740.

AWITI A.O., WALSH M.G., SHEPHERD K.D. & KINYAMARIO J. 2008 - Soil condition classification using infrared spectroscopy: A proposition for assessment of soil condition along a tropical forest-cropland chronosequence. Geoderma, 143, 73-84.

BARCZI A., GOLYEVA A.A. & PETO Á., 2009 - Palaeoenvironmental reconstruction of Hungarian kurgans on the basis of the examination of palaeosoils and phytolith analysis. Quaternary International, 193, 49-60.

BENEŠ J., 2008 - Environmentální archeologie a kultura s lineární keramikou v Čechách. In E. Černá & J. Kuljavceva Hlavová (eds.), Archeologické výzkumy v severozápadních Čechách v letech 2003-2007. Ústav Archeologické Památkové Péče Severozápadních Čech, Most, 33-51.

BOONMAN J.G. & MIKHALEV S.S., 2005 - The Russian Steppe. In J.M. Suttie, S.G. Reynolds & C. Batello (eds.), Grasslands of the World. Food and Agricultural Organization of the United Nations, Rome, 514 p.

BURNS D.A. & CIURCZAK E.W. (ed.), 2001 - Handbook of Near-Infrared Analysis, Second Edition, Revised and Expanded. Practical spectroscopy series, 27. Marcel Dekker, New York, 814 p.

CÉCILLON L., BARTHÈS B.G., GOMEZ C., ERTLEN D., GENOT V., HEDDE M., STEVENS A. & BRUN J.J., 2009 - Assessment and monitoring of soil quality using near-infrared reflectance spectroscopy (NIRS). European Journal of Soil Science, 60, 770-784.

CHENDEV Y., SAUER T., RAMIREZ G. & BURRAS C., 2015 - History of East European Chernozem Soil Degradation; Protection and Restoration by Tree Windbreaks in the Russian Steppe. Sustainability, 7, 705-724.

CHYTRÝ M., KUČERA T. & KOČÍ M., 2010 - Katalog biotopů České republiky. Agentura ochrany přírody a krajiny ČR, Praha, 445 p.

COHEN M.J., DABRAL S., GRAHAM W.D., PRENGER J.P. & DEBUSK W.F., 2006 - Evaluating ecological condition using soil biogeochemical parameter s and near infrared reflectance spectra. Environmental Monitoring and Assessment, 116, 427-457.

COÛTEAUX M.-M., BERG B. & ROVIRA P., 2003 - Near infrared reflectance spectroscopy for determination of organic matter fractions including microbial biomass in coniferous forest soils. Soil Biology and Biochemistry, 35, 1587-1600.

DOKUCHAEV V.V., 1883 - Russkij Cernozem. Printing house of Dekleron and Evdokimov, Sankt Petersburg, 376 p.

DUCHAUFOUR P., 1977 - Pédologie. 1. Pédogénèse et classification. Masson, Paris & New York, 477 p.

ECKMEIER E., GERLACH R., GEHRT E. & SCHMIDT M.W.I., 2007 - Pedogenesis of Chernozems in Central Europe – A review. Geoderma, 139, 288-299.

EHWALD E., JÄGER K.-D. & LANGE E., 1999 - Das Problem Wald - Offenland im zirkumherzynen Trockengebiet vor der neolithischen Besiedlung sowie die Entstehung der zirkumherzynen Schwarzerden. In R. Rolle & F.M. Andraschko (eds.), Frühe Nutzung pflanzlicher Ressourcen. Internationales Symposium Duderstedt, 12. - 15.5.1994. Hamburger Werkstattreihe zur Archäologie, 4. Lit, Hamburg, 12-34.

ERTLEN D., 2009. - Archivage pédologique et dynamiques environnementales: mise au point d’une méthode de reconnaissance des paléovégétations, fondée sur l’analyse spectroscopique dans le proche infrarouge (SPIR) des matières organiques de sols et paléosols. Thèse de Doctorat, Université de Strasbourg, Strasbourg, 382 p.

ERTLEN D. & SCHWARTZ D., 2010 - Utilisation d’archives géohistorique pour la mise au point d’une nouvelle méthode paléoécologique. Géocarrefour, 81, 55-66.

ERTLEN D., SCHWARTZ D., TRAUTMANN M., WEBSTER R. & BRUNET D., 2010 - Discriminating between organic matter in soil from grass and forest by near-infrared spectroscopy. European Journal of Soil Science, 61, 207-216.

ERTLEN D., SCHWARTZ D., BRUNET D., TRENDEL J.-M., ADAM P. & SCHAEFFER P., 2015 - Qualitative near infrared spectroscopy, a new tool to recognize past vegetation signature in soil organic matter. Soil Biology and Biochemistry, 82, 127-134.

EUROPEAN COMMISSION & EUROPEAN SOIL BUREAU NETWORK, 2005 - Soil Atlas of Europe. Office for Official Publications of the European Communities, Luxembourg, 128 p.

FEKETE G., VIRÁGH K., ASZALÓS R. & PRÉCSÉNYI I., 2009 - Static and dynamic approaches to landscape heterogenity in the Hungarian forest-steppe zone. Journal of Vegetation Science, 11, 375-382.

FISCHER-ZUJKOV U., 2000 - Die Schwarzerden Nordostdeutschlands – ihre Stellung und Entwicklung im holozänen Landschaftswandel. Doktorarbeit Dissertation, Humboldt University, Berlin, 173 p.

FÖLDVÁRI GY. & DARAB K., 1966 - Magyarország genetikus talajtípusainak, altípusainak és változatainak szisztematikus jegyzéke [Systematic register of genetic soil types, subtypes and variants of Hungary]. In I. Szabolcs (ed.), A genetikus üzemi talajtérképezés módszerkönyve [Methodological manual of genetic farm soil mapping]. Országos Mezőgazdasági Minőségvizsgáló Intézet, Budapest, 165-254. [in Hungarian]

FROEHLICHER L., 2012 - Paléovégétations et impacts anthropiques sur le terroir du village disparu de m’Elbersforst. Mémoire de Master 1, Université de Strasbourg, Strasbourg, 192 p.

FROEHLICHER L., 2016 - Les haies, une alternative à l’openfield dans les zones loessiques d’Alsace. Perspectives historiques, systèmes agraires du futur, effets sur le colluvionnement et le stockage du carbone. Thèse de doctorat. Université de Strasbourg, Strasbourg, 466 p.

GAFFEY S.J., 1986 - Spectral reflectance of carbonate minerals in the visible and near infrared (0.35–2.55 μm): calcite, aragonite and dolomite. American Mineralogist, 71, 151-162.

GANSSEN R., 1972 - Bodengeographie mit besonderer Berücksichtigung der Böden Mitteleuropas. K.F. Koehler, Stuttgart, 325 p.

GREGORICH E.G., MONREAL C.M., CARTER M.R., ANGERS D.A. & ELLERT B.H., 1994 - Towards a minimum data set to assess soil organic matter quality in agricultural soils. Canadian Journal of Soil Science, 74, 367-385.

HAVINGA A.J., 1972 - A palynological investigation in the pannonian climate region of lower Austria. Review of Palaeobotany and Palynology, 14, 319-352.

HEJCMAN M., HEJCMANOVÁ P., PAVLŮ V. & BENEŠ, J., 2013 - Origin and history of grasslands in Central Europe - a review. Grass and Forage Science, 68, 345-363.

HORVÁTH A., 2010 - Validation of description of the Xeromesophilous loess grassland association, Euphorbio pannonicae-Brachypodietum pinnati. Acta Botanica Hungarica, 52, 103-122.

HRABOVSKÝ A., BALKOVIČ J. & KOLLAR J., 2010 - Fytocenologicka charakteristika lesnej vegetacie NPR Dubnik. In Študentská vedecká konferencia : Prírodovedeckej fakulty UK v Bratislave 28.4.2010. Zborník Recenzovaných Príspevkov. Univerzita Komenského v Bratislave, Bratislava, 1380-1385.

IUSS WORKING GROUP WRB, 2006 - World reference base for soil resources 2006. World Soil Resources Reports, 103. FAO, Rome, 128 p.

IUSS WORKING GROUP WRB, 2015 - World Reference Base for Soil Resources 2014. International soil classification system for naming soils and creating legends for soil maps. Update 2015. World Soil Resources Reports, 106. FAO, Rome, 192 p.

JAMRICHOVÁ E., HÉDL R., KOLÁŘ J., TÓTH P., BOBEK P., HAJNALOVÁ M., PROCHÁZKA J., KADLEC J. & SZABÓ P., 2017 - Human impact on open temperate woodlands during the middle Holocene in Central Europe. Review of Palaeobotany and Palynology, 245, 55-68.

JELENSKA M., HASSO-AGOPSOWICZ A., KADZIALKO-HOFMOKL M., SUKHORADA A., BONDAR K. & MATVIISHINA Z., 2008 - Magnetic iron oxides occurring in the Chernozem soil from Ukraine and Poland as indicators of pedogenic processes. Studia Geophysica Geodaetica, 52, 255-270.

JENNY H., 1994 - Factors of soil formation: A System of Quantitative Pedology. Dover Publications, New York, 281 p.

KHITROV N.B., GERASIMOVA M.I., BRONNIKOVA M. & ZAZOVSKAYA E.P., 2013 - Chernozem of Kursk Biosphere Station of IG RAS. In Guidebook for Field Excursions: XIIth International Symposium and Field Seminar on Paleopedology. Moscow, 42-50.

KREUZ A., 2008 - Closed forest or open woodland as natural vegetation in the surroundings of Linearbandkeramik settlements? Vegetation History and Archaeobotany, 17, 51-64.

KUNEŠ P., SVOBODOVÁ-SVITAVSKÁ H., KOLÁŘ J., HAJNALOVÁ M., ABRAHAM V., MACEK M., TKÁČ P. & SZABÓ P., 2015 - The origin of grasslands in the temperate forest zone of east-central Europe: long-term legacy of climate and human impact. Quaternary Science Reviews, 116, 15-27.

LABAZ B., MUSZTYFAGA E., WAROSZEWSKI J., BOGACZ A., JEZIERSKI P. & KABALA C., 2018 - Landscape-related transformation and differentiation of Chernozems – Catenary approach in the Silesian Lowland, SW Poland. Catena, 161, 63-76.

LARIONOVA A.A., ZOLOTAREVA B.N., YEVDOKIMOV I., KUZYAKOV Y. & BUEGGER F., 2008 - The rates of organic matter renewal in gray forest soils and chernozems. Eurasian Soil Science, 41, 1378-1386.

LORZ C. & SAILE T., 2011 - Anthropogenic pedogenesis of Chernozems in Germany? – A critical review. Quaternary International, 243, 273-279.

LOŽEK V., 1973 - Příroda ve čtvrtohorách. Academia, Praha, 372 p.

LOŽEK V., 2001 - Molluscan fauna from the loess series of Bohemia and Moravia. Quaternary International, 76-77, 141-156.

LOŽEK V., 2004 - Středoevropské bezlesí v čase a prostoru III. Historie lesa a bezlesí v kvartéru. Ochrana Přírody, 59, 71-78.

LOŽEK V., 2011 - Po stopách pravěkých dějů. O silách, které vytvářely naši krajinu. Dokořán, Praha, 184 p.

MAGYARI E.K., CHAPMAN J.C., PASSMORE D.G., ALLEN J.R.M., HUNTLEY J.P. & HUNTLEY B., 2010 -
Holocene persistence of wooded steppe in the Great Hungarian Plain. Journal of Biogeography, 37, 915-935.

MCCLURE W.F., 2001 - Analysis using Fourier transforms. In D.A. Burns & E.W. Ciurczak (eds.), Handbook of Near-Infrared Analysis, Second Edition, Revised and Expanded. Practical spectroscopy series, 27. Marcel Dekker, New York, 261-305.

MOLNÁR Z. & BIRÓ M., 1996 - Vegetation history of the Kardoskut area (S.E. Hungary) I.: Regional versus local history, ancient versus recent habitats. Tiscia, 30, 15-25.

MUCINA L., BÜLTMANN H., DIERßEN K., THEURILLAT J.-P., RAUS T., ČARNI A., ŠUMBEROVÁ K., WILLNER W., DENGLER J., GARCÍA R.G., CHYTRÝ M., HÁJEK M., DI PIETRO R., IAKUSHENKO D., PALLAS J., DANIËLS F.J.A., BERGMEIER E., SANTOS GUERRA A., ERMAKOV N., VALACHOVIČ M., SCHAMINÉE J.H.J., LYSENKO T., DIDUKH Y.P., PIGNATTI S., RODWELL J.S., CAPELO J., WEBER H.E., SOLOMESHCH A., DIMOPOULOS P., AGUIAR C., HENNEKENS S.M. & TICHÝ L., 2016 - Vegetation of Europe: hierarchical floristic classification system of vascular plant, bryophyte, lichen, and algal communities. Applied Vegetation Science, 19, 3-264.

NĚMĚČEK J., MÜHLHANSELOVÁ M., MACKŮ J., VOKOUN J., VAVŘÍČEK D. & NOVÁK P., 2011 - Taxonomický klasifikační systém půd. 2. Upravené vydání. Česká zemědělská univerzita, Praha, 94 p.

NEUHÄUSLOVÁ Z. (ed.), 1997 - Mapa potenciální přirozené vegetace České republiky: Map of potential natural vegetation of the Czech Republic. Botanický ústav Akademie věd České republiky, Praha.

PALMBORG C. & NORDGREN A., 1996 - Portioning the variation of microbial measurements in forest soils into heavy metal and substrate quality dependent parts by use of near infrared spectroscopy and multivariate statistics. Soil Biology and Biochemistry, 26, 711-720.

PÉCSI M. (ed.), 1984 - Lithology and Stratigraphy of Loess and Paleosols: proceedings of the symposium. Geographical Research Institute, Hungarian Academy of Sciences, 229-248.

POKORNÝ P., CHYTRÝ M., JUŘIČKOVÁ L., SÁDLO J., NOVÁK J. & LOŽEK V., 2015 - Mid-Holocene bottleneck for central European dry grasslands: Did steppe survive the forest optimum in northern Bohemia, Czech Republic? The Holocene, 25, 716-726.

PONOMARENKO E.V. & ANDERSON D.W., 2001 - Importance of charred organic matter in Black Chenozem soils of Saskatchewan. Canadian Journal of Soil Science, 81, 285-297.

ROHDENBURG H. & MEYER B., 1968 - Zur Datierung und Bodengeschichte mitteleuropäischer Oberflächenböden (Schwarzerde, Parabraunerde, Kalksteinbraunlehm): Spätglazial oder Holozän? Göttinger Bodenkundliche Berichte, 6, 127-212.

SCHARPENSEEL H.W., PIETIG F. & TAMERS M.A., 1968 - Bonn Radiocarbon Measurements I. Radiocarbon, 10, 8-28.

SCHARPENSEEL H.W., PIETIG F. & TAMERS M.A., 1969 - University of Bonn Natural Radiocarbon Measurements II. Radiocarbon, 11, 3-14.

SCHARPENSEEL H.W. & PIETIG F., 1970 - University of Bonn natural radiocarbon measurements III. Radiocarbon, 12, 19-39.

SCHARPENSEEL H.W. & PIETIG F., 1971 - University of Bonn natural radiocarbon measurements IV. Radiocarbon, 13, 189-212.

SCHARPENSEEL H.W. & PIETIG F., 1973 - University of Bonn natural radiocarbon measurements V. Radiocarbon, 15, 13-41.

SCHARPENSEEL H.W., 1974 - Zur Berechtigung der Interpretation natürlicher Radiokohlenstoffmessungen an organischem Kohlenstoff CaCO3-reicher Böden. Mitteilungen der Deutschen Bodenkundlichen Gesellschaft, 18, 198-201.

SCHMIDT M.W.I., SKJEMSTAD J.O., GEHRT E. & KÖGEL-KNABNER I., 1999 - Charred organic carbon in German chernozemic soils. European Journal of Soil Science, 50 (2), 351-365.

SCHWARTZ D., 2012 - Les temps du sol : Interprétations temporelles de l’archivage pédologique dans les approches paléoenvironnementalistes et géoarchéologiques. Étude et Gestion des Sols, 19, 51-65.

SHENK J.S., WORKMAN J.J. & WESTERHAUS M.O., 2001 - Application of NIR spectroscopy to agricultural products. In D.A. Burns & E.W. Ciurczak (eds.), Handbook of Near-Infrared Analysis, Second Edition, Revised and Expanded. Practical spectroscopy series, 27. Marcel Dekker, New York, 419-474.

SMOLÍKOVÁ L. & LOŽEK V., 1978 - Die nacheiszeitlichen Bodenabfolgen von Poplze und Štětí als Beleg der Boden- und Landschaftentwicklung im Böhmischen Tschernosemgebiet. In H. Nagl (ed.), Beiträge zur Quartär- und Landschaftsforschung Festschrift zum 60. Geburtstag von Julius Fink. Verlag F. Hirt, Wien, 531-549.

SOBOCKÁ J., 2000 - Morfogenetický klasifikačný systém pôd Slovenska: bazálna referečná taxonómia. Výskumný ústav pôdoznalectva a ochrany pôdy v Bratislave, Societas pedologica slovaca, Bratislava, 76 p.

STANOVÁ V. & VALACHOVIČ M. (eds.), 2002 - Katalóg Biotopov Slovenska. DAPHNE - Inštitút aplikovanej ekológie, Bratislava, 225 p.

STENBERG B., VISCARRA ROSSEL R.A., MOUAZEN A.M. & WETTERLIND J., 2010 - Visible and Near Infrared Spectroscopy in Soil Science. In American Society of Agronomy (ed.), Advances in Agronomy. Elsevier, Amsterdam, 163-215.

SVOBODOVÁ H., 1997 - Die Entwicklung der Vegetation in Südmähren (Tschechien) während des Spätglazials und Holozäns - eine palynologische Studie. Verhandlungen der Zoologisch-Botanischen Gesellschaft in Österreich, 134, 317-356.

VELASQUEZ E., LAVELLE P., BARRIOS E., JOFFRE R. & REVERSAT F. 2005 - Evaluating soil quality in tropical agroecosystems of Colombia using NIRS. Soil Biology and Biochemistry, 37, 889-898.

VIRÁGH K., HORVÁTH A., BARTHA S. & SOMODI I., 2008 -
A multiscale methodological approach for monitoring the effectiveness of grassland management. Community Ecology, 9, 237-246.

VISCARRA ROSSEL R.A. & WEBSTER R., 2011 - Discrimination of Australian soil horizons and classes from their visible-near infrared spectra. European Journal of Soil Science, 62, 637-647.

VON LÜTZOW M., KÖGEL-KNABNER I., EKSCHMITT K., FLESSA H., GUGGENBERGER G., MATZNER E. & MARSCHNER B., 2007 - SOM fractionation methods: Relevance to functional pools and to stabilization mechanisms. Soil Biology and Biochemistry, 39, 2183-2207.

VYSLOUŽILOVÁ B., 2014 - Steppes ou forêts ? Les conditions environnementales de la formation et de l’évolution des chernozems en Europe Centrale. Thèse de Doctorat, Université de Strasbourg, Strasbourg & Univerzita Karlova v Praze, Praha, 398 p.

VYSLOUŽILOVÁ B., DANKOVÁ L., ERTLEN D., NOVÁK J., SCHWARTZ D., ŠEFRNA L., DELHON C. & BERGER J.-F., 2014 - Vegetation history of Chernozems in the Czech Republic. Vegetation History and Archaeobotany, 23, 97-108.

VYSLOUŽILOVÁ B., ERTLEN D., ŠEFRNA L., NOVÁK T., VIRÁGH K., RUÉ M., CAMPANER A., DRESLEROVÁ D. & SCHWARTZ, D., 2015 - Investigation of vegetation history of buried chernozem soils using near-infrared spectroscopy (NIRS). Quaternary International, 365, 203-211.

VYSLOUŽILOVÁ B., ERTLEN D., SCHWARTZ D. & ŠEFRNA L., 2016 - Chernozem. From concept to classification - a review. Acta Geographica Carolinae, 51, 85-95.

WALKEY A. & BLACK I.A., 1934 - An examination of the Degtjareff method for determining organic carbon in soils: Effect of variations in digestion conditions and of inorganic soil constituents. Soil Science, 63, 251-263.

WILLIS K. & VAN ANDEL T., 2004 - Trees or no trees? The environments of central and eastern Europe during the Last Glaciation. Quaternary Science Reviews, 23, 2369-2387.

Haut de page

Table des illustrations

Titre Fig. 1: Distribution of chernozems (European Commission & European Soil Bureau Network, 2005) and of studied soils in Europe.
Légende Chernozems are concentrated along the 50th parallel north in a band of increasing latitudinal extension to the east. The NIRS model of vegetation changes at each site is simplified in 5 cases: (1) stable grassland vegetation, (2) stable forest-steppe vegetation, (3) grassland to woodland transition, (4) woodland to grassland transition, (5) complex evolution between grassland and woodland.
URL http://journals.openedition.org/quaternaire/docannexe/image/12101/img-1.jpg
Fichier image/jpeg, 64k
Titre Tab. 1: Characteristics of the studied soils.
URL http://journals.openedition.org/quaternaire/docannexe/image/12101/img-2.jpg
Fichier image/jpeg, 112k
Titre Fig. 2: Mean residence time (MRT) of chernozem soil organic matter versus depth.
URL http://journals.openedition.org/quaternaire/docannexe/image/12101/img-3.jpg
Fichier image/jpeg, 48k
Titre Tab. 2: Sites that were used for the construction of the NIRS reference library.
URL http://journals.openedition.org/quaternaire/docannexe/image/12101/img-4.jpg
Fichier image/jpeg, 112k
Titre Fig. 3: Three examples of NIRS spectra.
Légende A/ Spectra without any pre-treatment; B/ Spectra after standardisation; C/ Spectra after standardisation and application of the first derivative.
URL http://journals.openedition.org/quaternaire/docannexe/image/12101/img-5.jpg
Fichier image/jpeg, 32k
Titre Fig. 4: Relationship between soil properties for grasslands and woodlands according to the reference spectral library.
URL http://journals.openedition.org/quaternaire/docannexe/image/12101/img-6.jpg
Fichier image/jpeg, 20k
Titre Fig. 5: Histogram of the canonical scores for grassland and woodland soils
URL http://journals.openedition.org/quaternaire/docannexe/image/12101/img-7.jpg
Fichier image/jpeg, 8,0k
Titre Fig. 6: Examples of typical NIRS profiles of the chernozems.
Légende Group 1 (ETZ): stable grassland vegetation; Group 2 (HUGF): stable vegetation of the forest-steppe; Group 3 (BAB): deep grassland vegetation with a tendency towards a forest in the upper part; Group 4 (H1): deep forest vegetation with a grassland trend in the upper part; Group 5 (HUU): complex evolution of the vegetation between forest and grassland clusters. The shaded areas correspond to the two groups defined by the reference library of chernozems. The POP profile cannot be interpreted.
URL http://journals.openedition.org/quaternaire/docannexe/image/12101/img-8.jpg
Fichier image/jpeg, 24k
Haut de page

Pour citer cet article

Référence papier

Barbora Strouhalová, Damien Ertlen, Luděk Šefrna, Tibor József Novák, Klára Virágh et Dominique Schwartz, « Assessing the vegetation history of european chernozems through qualitative near infrared spectroscopy »Quaternaire, vol. 30/3 | 2019, 227-241.

Référence électronique

Barbora Strouhalová, Damien Ertlen, Luděk Šefrna, Tibor József Novák, Klára Virágh et Dominique Schwartz, « Assessing the vegetation history of european chernozems through qualitative near infrared spectroscopy »Quaternaire [En ligne], vol. 30/3 | 2019, mis en ligne le 01 janvier 2021, consulté le 22 mai 2025. URL : http://journals.openedition.org/quaternaire/12101 ; DOI : https://doi.org/10.4000/quaternaire.12101

Haut de page

Auteurs

Barbora Strouhalová

Institute of Archaeology of the Czech Academy of Sciences, Prague, v.v.i., Letenská 4, CZ-118 01 PRAGUE. Email: strouhalova@arup.cas.cz; University of Strasbourg, Faculté de Géographie et d’Aménagement, Laboratoire Image, Ville Environment, 3 rue de l´Argonne, FR-67000 STRASBOURG.

Damien Ertlen

University of Strasbourg, Faculté de Géographie et d’Aménagement, Laboratoire Image, Ville Environment, 3 rue de l´Argonne, FR-67000 STRASBOURG. Email: damien.ertlen@live‑cnrs.unistra.fr

Articles du même auteur

Luděk Šefrna

Charles University in Prague, Faculty of Science, Department of Physical Geography and Geoecology, Albertov 6, CZ-128 43 PRAGUE. Email: ludek.sefrna@natur.cuni.cz

Tibor József Novák

University of Debrecen, Department for Landscape Protection and Environmental Geography, Egyetem tér 1, P. O. Box 400, HU-4002 DEBRECEN. Email: novak.tibor@science.unideb.hu

Klára Virágh

5 Institute of Ecology and Botany, Centre for Ecological Research, Institute of Ecology and Botany, Hungarian Academy of Sciences, Alkotmány 2-4, HU-2163, VÁCRÁTÓT. Email: viragh1951@gmail.com

Dominique Schwartz

 University of Strasbourg, Faculté de Géographie et d’Aménagement, Laboratoire Image, Ville Environment, 3 rue de l´Argonne, FR‑67000 STRASBOURG. Email: dominique.schwartz@live‑cnrs.unistra.fr

Articles du même auteur

Haut de page

Droits d’auteur

Le texte et les autres éléments (illustrations, fichiers annexes importés), sont « Tous droits réservés », sauf mention contraire.

Haut de page
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search