In this study, we combine erosion and anthropogenic proxies (Ti, Pb) from calibrated portable XRF with pollen and radiocarbon chronologies in peat from mires of the Kleinwalser Valley (Kleinwalsertal, Vorarlberg, Austria) to reconstruct palaeoenvironmental change and human impact in the northern central Alps. Favoured by a wetter climate, two analysed mires formed 6200 years ago in a densely forested valley. Landscape opening suggests that the first anthropogenic impact emerged around 5700 to 5300 cal BP. Contemporaneously, lead enrichment factors (Pb EFs) indicate metallurgical activities, predating the earliest archaeological evidence in the region. Pollen and erosion proxies show that large-scale deforestation and land use by agro-pastoralists took place from the mid- to late Bronze Age (3500 to 2800 cal BP). This period was directly followed by a prominent peak in Pb EF, pointing to metallurgical activities again. After 200 cal CE, a rising human impact was interrupted by climatic deteriorations in the first half of the 6th century CE, probably linked to the Late Antique Little Ice Age. The use of the characteristic Pb EF pattern of modern pollution as a time marker allows us to draw conclusions about the last centuries. These saw the influence of the Walser people, arriving in the valley after 1300 cal CE. Later, the beginning of tourism is reflected in increased erosion signals after 1950 cal CE. Our study demonstrates that prehistoric humans were intensively shaping the Kleinwalser Valley's landscape, well before the arrival of the Walser people. It also demonstrates the importance of palaeoenvironmental multiproxy studies to fill knowledge gaps where archaeological evidence is lacking.
In dieser Studie kombinieren wir Paläoproxies für Erosion und menschliche Aktivitäten (Ti, Pb) aus Messungen eines mobilen RFA- bzw. XRF-Analysators mit Pollen und Radiokohlenstoffdatierungen in Torfproben aus Mooren des Kleinwalsertals (Vorarlberg, Österreich). Auf diese Weise sollen der Wandel der Paläoumweltbedingungen sowie der Einfluss des Menschen auf die Landschaft der nördlichen Zentralalpen untersucht werden. Begünstigt durch feuchteres Klima, begann vor ca. 6200 Jahren im damals dicht bewaldeten Tal die Torfbildung in zwei der untersuchten Moore. Auflichtungen des Waldes deuten auf den ersten menschlichen Einfluss zwischen cal B.P. 5700 und 5300 hin. Gleichzeitig suggerieren Bleianreicherungen (Pb EF) metallurgische Aktivitäten, die weit vor jeglichen archäologischen Beweisen in der Region liegen. Pollen und Erosionsproxies bezeugen eine weitläufige Entwaldung und Landnutzung durch Landwirtschaft und Beweidung in der mittleren und späten Bronzezeit (3500 bis 2800 cal B.P.). Dieser Phase folgte ein vorläufiger Höchstwert des Pb EF zum Beginn der Eisenzeit, der erneut auf Bergbau oder Metallurgie in der näheren Umgebung hindeutet. Eine im cal A.D. 3. Jahrhundert einsetzende verstärkte Landnutzung wird im cal A.D. 6. Jahrhundert unvermittelt durch klimatisch ungünstigere Bedingungen unterbrochen, welche der sogenannten spätantiken kleinen Eiszeit zugeschrieben werden können. Unter Zuhilfenahme des charakteristischen Verlaufs der modernen Bleiemissionen als Zeitmarker, konnten die Analyseergebnisse der letzten Jahrhunderte eingeordnet werden. Diese waren von der Ansiedlung der Walser im cal A.D. 14. Jahrhundert geprägt. Nach cal A.D. 1950 zeigt sich außerdem der erstarkende Tourismus anhand höherer Erosionsspuren. Unsere Ergebnisse machen deutlich, dass die Landschaft des Kleinwalsertals schon weit vor Ankunft der Walser vom Menschen genutzt und verändert wurde. Unsere Studie betont zudem die Wichtigkeit der Anwendung unterschiedlicher Paläoproxies, insbesondere, wenn archäologische Anhaltspunkte zur Landnutzungsgeschichte fehlen.
Humans have been recurringly present in Alpine environments since the last deglaciation (e.g. Cornelissen and Reitmaier, 2016). These harsh landscapes are heterogeneous and sensitive to climate (Barry, 2002), which requires specific human adaptation (Clegg et al., 1970). Half nomadic lifestyles or transhumance have been strategies to survive, and are still today the basis for seasonal livestock management practice in mountainous regions (e.g. Reitmaier et al., 2018). There is however no consensus on the human colonisation of European mountains during the Holocene. In the Alps, the onset of human impact is still not fully understood because occupation pulses were radiating from different regions and societies at different time periods (e.g. Bätzing, 2015; Carcaillet, 1998; Dietre et al., 2017; Oeggl and Nicolussi, 2009; Valese et al., 2014). It is therefore important to document human occupation and its impact on mountain environments to a certain level of detail, as each region or each valley reveals pieces of information on the complex spatial linkages between humans and environmental and climatic conditions.
Another challenge in reconstructing past human impacts in the Alps is the general scarcity of suitable palaeoenvironmental archives in high mountain areas. Archaeological records and historical sources cannot provide continuous information and may bias interpretations towards separate findings. These gaps are generally closed by environmental archives, such as trees, lakes, glaciers or mires, which potentially provide uninterrupted records of past environmental changes (anthropogenic or natural). However, except for pollen or dendrochronological studies, mountain mires are so far rarely used as environmental archives.
In this paper, we present a multiproxy study of the Kleinwalser Valley (Kleinwalsertal) in the Austrian northern central Alps, using small mountain mires as environmental archives. The valley is historically known for its human occupation, animal husbandry and forestry only since the Late Middle Ages (Fink and von Klenze, 1891; Wagner, 1950). Before that, the human impact on this valley is unclear. While Romans were present in the Alpine foreland (Mackensen, 1995; Weber, 1995), there are no records of their presence in the valley. Palaeovegetation information suggests prehistoric land use (Dieffenbach-Fries, 1981; Grosse-Brauckmann, 2002) but chronologies and data are limited. Other evidence points to early activities at archaeological sites in the valley (Bachnetzer, 2017; Gulisano, 1994, 1995; Leitner, 2003), which are located close to our study sites (Fig. 1). These spots may have acted as strategic points between Alpine foreland and the surrounding mountain ranges, leading prehistoric humans to cross and occupy the Kleinwalser Valley.
By combining geochemical and palynological data together with radiocarbon chronologies, we aim at better understanding the points in time and impacts of human occupation in the Kleinwalser Valley's landscape and beyond. We also aim at detecting early metallurgical activities, possibly where archaeological evidence is lacking. By looking into the past, using multiple proxies on chronologically constrained peat sequences from a key area, we provide new insights into the development and interaction of landscape, climate and humans from mid-Holocene to modern times in the northern central Alps.
The Kleinwalser Valley belongs to the federal state of Vorarlberg in the north-western part of Austria (Fig. 1) and is located at the junction of the geological units of the Northern Calcareous Alps, Penninic flysch and Helvetic (Völk, 2001). The valley floor elevates around 1100 m a.s.l. and is surrounded by mountains ranging from 2000 to 2500 m a.s.l. Geologically speaking, the watershed is composed of calcareous as well as silicate rocks. During the late Pleistocene, the valley was glaciated (Völk, 2001) and several moraines are still present. Iron (Fe), lead (Pb), zinc (Zn) and copper (Cu) ores are present outside the valley. Within 20 km N-NE, iron had been exploited since 1471 CE in Sonthofen (Merbeler, 1995). Pb–Zn deposits are known to the NE at Himmelschrofen (Fig. 1) (von Gümbel, 1861) and in the Ostrach Valley since 1620 cal CE (Oblinger, 1996). To the south, Zn–Pb deposits exist at Zug, at St. Anton at Arlberg, and at St. Christoph at Arlberg and Cu can be found at Bartholomäberg (Weber, 1997) within 35 km off the lower Kleinwalser Valley.
A temperate climate in the Kleinwalser Valley is reflected by a mean annual
temperature of 5.7
The main study site, Hoefle Mire (HFL, GPS: 47
A second study site, the Ladstatt Mire (LAD, GPS: 47
A third peat profile at “Halden-Hochalpe” (HHA, GPS: 47
The Kleinwalser Valley and surrounding areas.
Archaeological sites (red dotted circles): Egg
Field sampling was performed in early August 2016 in the centre of HFL. A
peat sequence of 240 cm was recovered in three parallel, overlapping cores
(B, C, D) using a Russian peat corer with a chamber of 5 cm in diameter.
Additionally, a
After freezing the cores for 2 d at
List of radiocarbon samples including information about origin,
depth, dated material,
A volume of 2 mL was taken from 14 peat samples of cores HFL-B and HFL-C in
the overlapping section and prepared for pollen analysis, following the
method described by Moore et al. (1991). The material was
pretreated in separate steps in 10 % HCl and 10 % KOH to get rid of
carbonates and humic substances. Macro remains were removed with a 200
This study concentrated on calcium (Ca), lead (Pb) and titanium (Ti) for
interpretation. While Ca can yield information on the trophic state of a
mire, Ti can be used as an erosion or human impact proxy
(Hölzer and Hölzer, 1998). Pb often originates from
anthropogenic sources when it exceeds its natural background
(e.g. Weiss et al., 1999). Using
portable X-ray fluorescence spectrometry (pXRF) on peat samples in
palaeoenvironmental research has rarely been done so far and is hence not well
understood, despite some studies that were assessing its general potential
(Kalnicky
and Singhvi, 2001; Mejía-Piña et al., 2016; Shand and Wendler,
2014; Shuttleworth et al., 2014). Therefore, a regression analysis was
conducted to evaluate and calibrate the semi-quantitative pXRF by several
parallel quantitative measurements with inductively coupled plasma mass
spectrometry (ICP-MS). In HFL-B and HFL-C, a total of 187 and 51 samples were
selected for pXRF scanning and ICP-MS analyses, respectively. A total of 62
samples were selected from the LAD core for pXRF scanning. The samples were
transferred into Falcon tubes together with eight glass beads (4 mm) and
ground
Samples dedicated to pXRF were transferred into 12.5 mL polypropylene vials, closed
with Fluxana TF-240-255 film and rubber band. All samples were measured
using a Thermo Fisher Niton XL3t pXRF equipped with an Au anode and a 50 kV
X-ray tube. The predefined “soil mode” was used with 180 s of measurement
time each for the main and low filter, which is 60 s above the minimum
duration recommended by Shuttleworth
et al. (2014). Every sample was measured at least three times and shaken
after each scan to control the reproducibility of measurements (precision).
In addition to the Certified Reference Materials (CRM) used with ICP-MS (see
below), BCR-060 (aquatic plants), IAEA-336 (lichen), IPE-176
(reed/
Repeated pXRF measurements (
Samples dedicated to ICP-MS were digested using
Quality control of Ca, Ti and Pb. Measurements by ICP-MS of procedural blanks and in organic Certified Reference Materials (CRM). Total deviations (Dev.) from certified values as a percentage.
Median ages were extracted from the age–depth model for further
interpretation. The accumulation rates for the model of each core were
calculated as the median of 1000 estimates by rbacon for each depth. In the
lowermost 60 cm of HFL, an accumulation of almost 0.9 mm a
Age–depth models of Hoefle Mire
Pollen diagram of HFL Mire. Fully coloured profiles
represent pollen percentages, shaded areas enhanced by factor 12.
Black: trees; green: forest border (
A detailed interpretation of the pollen record from HFL Mire (Fig. 3) is
provided in the discussion below. We can, however, point out major changes
in the pollen profile. The deepest part was strongly characterised by
The quantitative and validated ICP-MS results of HFL were compared to the
pXRF scans of the same sample, which allowed us to constrain our
calibration. The regression analysis in the HFL peat samples showed high
adjusted
Cross-plots of measured concentrations of Ca, Ti
and Pb with ICP-MS (mg kg
The element-specific overestimation of the pXRF output illustrates that this method cannot be used to quantify element concentrations in peat without a cross-calibration against quantitative methods. Specification and type of sample need to be considered during such a calibration. Shuttleworth et al. (2014) used a Niton XL3t with an Ag anode to evaluate the applicability of pXRF for Pb contents in peat. Their results suggested a linear relationship between pXRF and ICP-OES (optical emission spectrometry), which was confirmed in this study. However, their observed overestimation by pXRF was much lower and was attributed to an incomplete Pb extraction with aqua regia. As a complete digestion was performed in this study, the overestimation by pXRF for Pb cannot have a similar explanation. Additionally, the high content of light organic matter in peat should rather result in an underestimation of heavier elements in XRF analysis (Löwemark et al., 2011). Nevertheless, the regression analysis allowed us to generate quantitative results of Ca, Pb and Ti in both minerotrophic and ombrotrophic peat. The only exception was the measurement of mineral-rich bottom layer samples with a high element load, which plotted outside the linear regression.
The Pb enrichment factor (Pb EF) was calculated following (Weiss et al., 1999) using Ti as a conservative element, which is not affected by dissolution in acidic environments (Nesbitt and Markovics, 1997).
Here we used the background
The temporal variability of Ca, Pb EF and MAR shown in Fig. 5 is based on
the calibrated pXRF concentrations, calibrated ages, densities and
accumulation rates (tables in Supplement). Ca concentrations
in HFL constantly declined towards the top. After a max. of 15 000 mg kg
The concentrations of Ti in HFL decreased from 1040 to around
50 mg kg
Excluding the deepest samples of HFL, Pb was below the detection limit from
230 to 190 cm (6250 to 5600 cal BP) and did not exceed 9 mg kg
In the LAD core, Ti concentrations were above 1000 mg kg
Peat formation in HFL and LAD began around 6200 cal BP. Pollen proportions
suggest a
Only very little archaeological evidence suggests early human occupation in
the area (Bachnetzer, 2017; Leitner, 2003). Similarly, little
palaeoecological evidence exists on early local land use. Regional studies
showed evidence for fire practices (Clark et al., 1989) and
crop cultivation around Lake Constance (Jacomet,
2009; Rösch, 1992) and at Oberstdorf, north of the valley (Fig. 1)
(Dieffenbach-Fries, 1981). In HFL, increasing
Chronological profile of Ca, Pb and Ti
concentrations in HFL, Pb EF in HFL, and mineral accumulation rates
(g m
Almost simultaneously to the observed pollen signals in HFL, the Pb EF rose significantly to 50 at 5450 cal BP and remained around an average value of 22 until 3500 cal BP. While this elevated Pb EF period falls in the minerotrophic part of the core, several authors showed that Pb, an immobile element, could carefully be used as an anthropogenic indicator in minerotrophic peat (e.g. Baron et al., 2005; Shotyk et al., 2001). Archaeological evidence shows that metal tools were circulating in the central Alps (e.g. Artioli et al., 2017). Metallurgy was dominated by Cu at that time, but Pb can be present as an impurity (Höppner et al., 2005; Lutz and Pernicka, 2013). Despite the existence of different ore deposits in the region (see Sect. 2), archaeological evidence suggests that the closest and earliest copper mining emerged in the Lower Inn Valley in the early to mid-Bronze Age around 3600 cal BP (Breitenlechner et al., 2013; O'Brien, 2015; Tomedi et al., 2013). Our HFL record strongly suggests that metallurgical activities around the Kleinwalser Valley could have already taken place around 5500 cal BP, which would be almost 2000 years earlier. Our findings are, however, in agreement with other studies, which indicate that Alpine metallurgy started well before 5000 cal BP (Bartelheim et al., 2002; Frank and Pernicka, 2012; Höppner et al., 2005).
The start of this period saw the evolution from minerotrophic to
ombrotrophic conditions in HFL, characterised by the occurrence of
The pressure of agro-pastoralism progressively disappeared from 2800 to
2500 cal BP as shown by the disappearance of
While anthropogenic impact in the Kleinwalser Valley declined from 2800 to 2000 cal BP, Pb EF suggests that metallurgical activities continued in the area. A Pb EF of up to 175 hints at intensive metallurgical activities not far from HFL between 2800 and 2600 cal BP. The exact locations of these activities remain unclear, but may be local to regional, as metallurgy became widespread in the eastern Alps around 3500 cal BP (Höppner et al., 2005; Lutz and Pernicka, 2013). The signal is in line with the technological introduction of the first Pb alloys (Tomedi et al., 2013). Afterwards (2600–2300 cal BP), short episodes of moderate Pb EF indicate ongoing metallurgy in the region. Part of it could have been connected to Celtic cultures (e.g. La Tène), as metal artefacts and metallurgy evidence were found in the northern central Alps (Bächtiger, 1982; Mansel, 1989).
During early Roman times (ca. 2000 cal BP in this region) HFL and LAD show a low MAR, indicating little erosion and therefore decreased land use, as Walde and Oeggl (2003, 2004) already suggested. In contrast, Friedmann and Stojakowits (2017) reported higher land use in the Alpine foreland. Later on, the study of Büntgen et al. (2011) suggested warm summers with moderately low precipitation from 200 to 300 cal CE. Simultaneously, rising mineral input in HFL indicates increased land use in the Kleinwalser Valley. The local and regional human activity and connectivity of the valley can only be inferred by connecting historical sources (Dertsch, 1974; Fink and von Klenze, 1891; von Raiser, 1830; Weber, 1995), archaeological finds (Gulisano, 1995) and a Roman trade route completion (Via Decia) through Sonthofen (20 km N) around 250 CE (Heuberger, 1955). We therefore suggest that people may have used the valley's slopes, which led to the observed MAR rise in HFL and LAD around the 3rd century cal CE. We also observe an elevated Pb EF around 100 cal CE, but in contrast to Mackensen (1995), we cannot firmly connect that to local Roman mining around Sonthofen. As strong Pb emissions of Roman origin are recorded across Europe (e.g. De Vleeschouwer et al., 2010b), we tend to attribute the enrichment in HFL to diffuse distal sources.
After 400 cal CE, increased
This period saw the expansion of the Frankish empire to higher elevations,
but little is known about it in this part of the northern central Alps, as
neither Roman nor Middle Age historical sources exist (Babucke,
1995). We nevertheless observe a growing human impact in HFL, reflected in
increasing cultural pollen proportions (
In contrast to HFL, the MAR in LAD seems to have taken another development between the Middle Ages and industrialisation. Even if the accumulation rates in this part need to be considered with caution, the MAR rose quite strongly after around 1400 cal CE. Particularly this western side of the valley started to be used for cattle grazing after 1450 CE (Amann, 2013b). Consequently, the vulnerable slopes suffered from erosion, induced by timber cutting and cattle trampling.
Pb EF has also increased since around 1400 cal CE. Part of this increase may be attributed to the Sonthofen mining district, where exploitation was first documented for 1471 cal CE (Merbeler, 1995). However, the emergence of widespread European mining activities in the Middle Ages (e.g. Forel et al., 2010; Le Roux et al., 2005) could have contributed to the signal in HFL. The interpretation of the late Middle Ages in the LAD record is however limited by the age constraint. A more detailed discussion would therefore be speculative. We can however observe that the trends described above for Pb EF, the MAR, and pollen in HFL continued towards the 19th century. We can moreover suggest several interpretations from our geochemical data for the late 19th and 20th centuries. As a result of heavy industry and the introduction of leaded gasoline in Europe, the Pb EF in HFL strongly increased continuously from 1850 cal CE to a maximum of 250 in the 1980s and a sharp decline thereafter, perfectly fitting to the maximum use of leaded fuel and its subsequent ban (Pacyna and Pacyna, 2001). The chronology of the last centuries is however not well constrained. This could have been caused partly by drainage of Hoefle Mire, which temporarily reduced peat accumulation and enhanced decomposition at intermediate depth. However, our Pb EF profile is strikingly similar to the well-dated Pb EF profile of several Swiss mires (Shotyk et al., 1998; Weiss et al., 1999) and, hence, provides another chronological reference point for the topmost part of the Hoefle core. We can therefore say that, during the first half of the 20th century, the MAR remained low in both mires. The tourism intensification in the valley (Fritz, 1981) and the construction of related infrastructure increased the MAR in both mires again after 1950 cal CE.
We present two peat records of the central Alps covering the last 6200 years. By combining geochemical, palynological and chronological tools, we are able to understand the occupation and high human impact on the landscape in a valley of the northern central Alps and beyond. Calibrating a portable geochemical tool with ICP-MS also allows us to quantify geochemical elements in peat at a resolution that is rarely obtained and demonstrates its potential in (palaeo)environmental studies.
A cooler and wetter climate around 6200 cal BP promoted mire formation in the Kleinwalser Valley. Pollen spectra and erosion suggest human presence in the valley as early as around 5700 cal BP, lasting over several centuries, which is in line with studies on regional occupation. Increased Pb EF values around the same time suggest metallurgical activities in the area, which predates regional archaeological evidence by almost 2000 years. Large-scale deforestation and land use (agro-pastoralism) took place between 3500 and 2800 cal BP causing a drastic landscape opening and high initial erosion rates at both low and high elevations. At the end of this period, a prominent Pb EF points to metallurgical activities. However, landscape stabilised and forests recovered thereafter until well into the Roman period. A second period of increased erosion and land use started after 230 cal CE to a climax at around 700 cal CE. This increased land use period was interrupted by climatic deteriorations around 500 to 600 cal CE. Pb EF increased during the Roman period. It is however challenging to identify an origin as Roman Pb pollution was widespread. The Middle Ages saw progressive land management and the arrival of the Walser people. While historical sources point to a strong Walser influence on the area, our data only allow us to suggest that deforestation, agriculture and pastoralism continued during that period. The extent of those activities remains unclear however. The Pb EF increased during Late Middle Ages. It then rose faster during the industrial revolution and peaked before leaded gasoline was banned, allowing us to use it as a chronological marker to constrain the last century, which saw increasing tourism and its consequences taking place in the Kleinwalser Valley.
Although several archaeological findings and sites gave evidence for three prehistoric hotspots of human activity before, our pollen and geochemical data allow us to detail the evolution of the local and regional landscape and its use in the northern central Alps. The combination of historical sources with erosion indicators and pollen points to local land use and human presence in the valley thousands of years before the Walser arrival. Because the landscape had already undergone strong anthropogenic changes before, the recorded impact of the Walser was ultimately less prominent than could be expected.
The geochemical data underlying this study are given either in the tables of the article or in the Supplement published with this article. The pollen data can be requested by contacting the first author.
The supplement related to this article is available online at:
As corresponding author, CvS was responsible for planning and carrying out field work, geochemical analyses, data processing, and designing and writing of the paper. The pollen analyses were conducted by AL, whereas FDV contributed to sample preparation, geochemical analyses and designing of the paper. Both IU and JS helped in planning and conducting field sampling and contributed to the interpretation process.
The authors declare that they have no conflict of interest.
This article is part of the special issue “Connecting disciplines – Quaternary archives and geomorphological processes in a changing environment”. It is a result of the First Central European Conference on Geomorphology and Quaternary Sciences, Gießen, Germany, 23–27 September 2018.
We would like to thank Karl Keßler (landscape conservation Kleinwalser
Valley), Alexander Suhm (Kiel University), Chuxian Li, Marie-Jo Tavella and
Gaël Le Roux (all EcoLab) for their help and advice during field and
laboratory phases. Clemens von Scheffer benefited from several travel grants from
PPP (no. 57316724)/PHC-PROCOPE, no. 37646SG from DAAD/Campus-France (funded by
the BMBF – German Federal Ministry of Education and Research),
“