The legacy of millennial-scale land-use practices on landscape composition,diversity and slope erosion in the subalpine areas of Eastern Carpathians,Romania

Abstract

We conducted multi-proxy analyses in two subalpine lakes, Lake Gropile, in the Rodna Mountains and Lake Vinderelu in the Maramureș Mountains, the Eastern Carpathians, Romania, to investigate the effect of different land-use practices on landscape composition, diversity and slope erosion. In the case of Lake Gropile, results evidenced a more extended tree and shrub cover and high fire activity between 6400 and 2800 cal yr BP, accompanied by reduced soil erosion, which appeared more strongly regulated by climatic conditions. Anthropogenic impact became evident 2800 years ago, when landscape openness, pasturing, and disturbance of soil cover increased and intensified over the last four centuries. In the case of Lake Vinderelu, intensified burning was followed by grazing around 1200–1100 cal yr BP and continued throughout the last millennium. Results also highlight the site-specific effects of land-use on vegetation composition and landscape diversity. For Lake Vinderelu, a combined effect of local burning and grazing in removing shrub cover appear to be the main drivers of changes in landscape diversity and structure. At Lake Gropile, fire was more connected to shrub cover changes while grazing to herbaceous cover diversity. Moderate to low grazing appeared to benefit both subalpine ecosystems, creating rich grassland-shrub mosaic communities, while overgrazing reduced landscape diversity and exacerbated erosion. Our findings document the millennial-scale legacy of land-use practices on the subalpine landscapes in this region. We propose that these semi-natural ecosystems hold important ecological and cultural value, and recommend their maintenance through controlled, low intensity pasturing and/or burning.

Keywords

Carpathians charcoal coprophilous fungi erosion Holocene subalpine landscape

Introduction

Mountain landscapes are characterised by topographic variability and steep climatic and environmental gradients (Körner, 2003). They therefore support a wide variety of tree, shrub and grass communities including many pioneer and endemic species. These communities coexist in a spatially constrained land surface, particularly towards higher elevations that is, in the subalpine and alpine belts (Nagy et al., 2003; Noroozi et al., 2018). Their high habitat and species diversity make high elevation vegetation belts important in preserving biodiversity and ecosystem services including soil conservation and water regulation (Noroozi et al., 2018; Stein et al., 2014). Furthermore, changes in such environments can cascade to lower elevations in terms of slope processes, sediment fluxes and river flow regimes (e.g. Gómez-Villar et al., 2014; Knowles et al., 2015; Tasser et al., 2003).

Many subalpine and alpine areas in the Mediterranean and Central Europe reflect a legacy of long-term (centuries to millennia) anthropogenic land-use (e.g. pasturing, forestry, burning) and have achieved cultural significance for local communities (Dietre et al., 2020; Finsinger et al., 2018; Pini et al., 2017; Vandewalle et al., 2014). Therefore, the sustainable management of these landscapes relies on understanding the long-term influence of particular land use practices on both landscape composition and soil development (Arnaud et al., 2016; Giguet-Covex et al., 2014; Jones et al., 2013). Such an approach is essential to understand present-day ecosystem dynamics and to maintain ecosystem functioning and biodiversity.

By contrast, there is a paucity of information on the long-term drivers of subalpine landscape dynamics in the Carpathian Mountains, despite these areas hosting some of the richest mosaic habitats of grassland and shrub communities in Europe (Dítě et al., 2018; Werners et al., 2016). For example, of the 85 plant endemic species in Romania, almost 60% are found in the subalpine-alpine belt (typically as a mix of shrub and grasslands) of the Romanian Carpathians (Coldea et al., 2009). However, these subalpine areas presently face several ecological problems. Firstly, short to medium term studies have already shown for many Carpathian Mountain groups such as the Rodna and Maramures Mts (Romania), Chornohora and Svydovets Mts (Ukraine) and the High Tatras (Slovakia), a climate-induced increase in the distribution of more warm-adapted plant species and a decline in those that are cold-adapted (Kobiv, 2017; Pauli et al., 2012). Secondly, treeline movement upslope and shrub encroachment are visible and expected to accelerate due to land abandonment and changes in seasonal temperature and precipitation patterns (Coldea et al., 2009; Durak et al., 2015; Mihai et al., 2007; Sitko and Troll, 2008). Thirdly, poor land-use practices (e.g. uncontrolled vegetation clearing, unsustainable pasturing) have contributed to the degradation of subalpine landscapes, due to the extension of nitrophilous and ruderal species such as Nardus stricta, and increased soil erosion over the last century (Baur et al., 2007; Kucsicsa, 2013; Maramures Mts Management Plan - PNPNMM, 2016; Novak & Horváthová, 2019; Rodna Mts Management Plan - PNPNMR, 2019; Werners et al., 2016).

A long-term connection between the dynamics of the timberline and treeline and climate, wildfire and anthropogenic land-use has already been demonstrated by palaeoecological studies in the Carpathians (e.g. Carter et al., 2020; Feurdean et al., 2016, 2017; Finsinger et al., 2018; Florescu et al., 2018; Geantă et al., 2014; Haliuc et al., 2016; Kołaczek et al., 2021; Tanţău et al., 2011, 2014a, 2014b; Vincze et al., 2017). However, none of these studies specifically assessed the effects of land-use practices such as pastoralism, woody vegetation clearance and the use of fire on the dynamics and compositional change of subalpine landscapes. Furthermore, the limited integration of palaeoecological data with sedimentology has hampered the exploration of the land-use effects on soil cover and erosion (Dearing et al., 2006; Dearing and Jones, 2003). Accelerated erosion due to clearance or more intensive land use will inhibit soil formation and the subsequent development of vegetation cover, particularly in higher elevation mountain areas where soils are less well developed and easily erodible (Giguet-Covex et al., 2011; Rothacker et al., 2018). This may subsequently activate hydro-geomorphological processes leading to the erosion of topsoil and bedrock. Such processes can transition into a wide array of mass-wasting events, with downstream consequences such as riverbed aggradation, channel instability, the siltation of lakes and reservoirs (Foster et al., 2003; Hall et al., 2014; Perșoiu et al., 2022). Given that the Carpathians are the headwaters of several major rivers, these long-term potential developments as a result of erosion should be investigated with a view to their mitigation.

In this study, we investigated the millennial-scale effect of different land-use practices (grazing, burning, clearance) on landscape composition and diversity, and slope erosion in two subalpine areas in the Eastern Carpathians, northern Romania (Rodna Mts National Park and Maramures Mts Natural Park) over the last 6400 and 1500 years respectively, using a multi-proxy approach. This approach includes a combination of biotic indicators, that is, pollen, plant macro-remains, charcoal, coprophilous fungi records and abiotic indicators such as sediment mineral magnetic properties, geochemistry and grain-size, in addition to the appraisal of published, independent data on past climate conditions and human habitation. We have addressed the following research questions:

(i) Is the present-day subalpine landscape a consequence of particular land use practices e.g. clearing, burning or grazing, or was alternatively shaped by climate change?

(ii) How do changes in long-term erosion patterns reflect past intensifications in land-use?

Extending the spatio-temporal perspective of human influence on the diversity and composition of these landscapes can provide critical insights into environmental baseline conditions, the mechanisms driving landscape dynamics and the long-term sustainability of land-use practices.

Study area and sites

To increase the spatial representation of our study, we cored two sedimentary sequences located in the subalpine vegetation belt of the Rodna National Park and Biosphere Reserve (Lake Gropile – GR; 1920 m asl) and the Maramureș Natural Park (Lake Vinderelu – VD; 1684 m asl) (Figure 1). The main characteristics of the sites and their corresponding catchments (e.g. geology, morphometry, local climate, vegetation) are listed in Table A1. Rodna and Maramureș Mts are characterised by a moderate temperate continental climate with Atlantic and Baltic influences (Geografia României (Geography of Romania), 1987). Mean annual temperatures of 1.4°C and an annual precipitation of 1240 mm were recorded between 1963 and 2010 at the Iezerul Pietrosului meteorological station (1785 m asl; Figure 1). There is greater precipitation during summer, mostly in the 1400–1700 m elevation belt (Dragotă and Kucsicsa, 2011; Mîndrescu, 2001). Soil cover in both subalpine areas is patchy and skeletal, with podzolic soils developed on gentler topography, interrupted by exposed rock on steeper slopes (Geografia României (Geography of Romania), 1987).

Figure 1.

Study area and location of the study sites in Europe (a) and Romania (b). Archaeological evidence of human habitation (settlements, fortifications, ruins of religious sites, tumuli, etc.) is grouped into historical periods. Archaeological data was extracted from the National Archaeological Database (Repertoriul Arheologic National, available at http://ran.cimec.ro). Digital Elevation Model was derived from the Shuttle Radar Topography Mission data set (SRTM; http://www2.jpl.nasa.gov/srtm/).

The vegetation in the subalpine belt in the Rodna Mts (1700–2000 m asl) generally, and near Lake Gropile specifically, is characterised by a mixture of grasses, herbs and patches of shrubs (Pinus mugo, Juniperus communis ssp. nana and Ericaceae for example, Rhododendron, Vaccinium spp.) (Coldea, 1990; PNPNMR, 2019). In the Maramures Mts, the subalpine belt is presently located above 1600 m asl and the dominant local vegetation consists of grasses dominated by Nardus stricta, various herbaceous species and scattered Juniperus communis and Pinus mugo (PNPNMM, 2016). The surrounding landscape at VD is presently grazed by sheep or cattle during the summer, while grazing has been strictly forbidden around GR over the last 25 years. A general trend of abandonment of grazing areas is apparent in more remote locations with intensified use in more easily accessible locations (PNPNMM, 2016; PNPNMR, 2019). Locally, the timberline rises to 1500–1600 m asl at GR and to 1400–1500 m asl at VD, while local treelines reach 1700 and 1600 m asl respectively (Figure A1). It should be noted that the position of the timberline and treeline ecotone in this region has been subjected to climatic and anthropogenic changes over the Holocene (Feurdean et al., 2016; Geantă et al., 2014; Tanţău et al., 2014b).

Material and methods

Sediment dating and age-depth models

Sediment coring for both study sites is described in Florescu et al. (2018). The chronology was established based on six AMS radiocarbon ages at GR and four AMS radiocarbon ages at VD, as well as on two sets of 210Pb age estimates (Florescu et al., 2018). Calibrated ages were obtained with the Northern Hemisphere IntCal20 dataset (Reimer et al., 2020) and were used to construct age-depth models for each site using the smooth spline function of the Clam package in R (Blaauw, 2010). Calendar age point estimates for depths were based on weighted average of all age-depth curves and on the error range of the calibrated ages (Table 1).

Table 1.

AMS 14C and 210Pb measurements of the cores from Lake Gropile (GR) and Lake Vinderelu (VD).

Isotope	Laboratory code	Depth range (cm)	Material dated	AMS ¹⁴C yr BP/²¹⁰Pb cal yr BP^a
Gropile
210Pb		0–1	Bulk sediment	−62 ± 1
210Pb		5–6	Bulk sediment	−26 ± 2
¹⁴C	UBA-25545	47–48	Peat	1758 ± 32
¹⁴C	UBA-25546	57–58	Peat	2565 ± 43
¹⁴C	UBA-25547^b	78–79	Peat	2339 ± 27
¹⁴C	UBA-24392	85–86	Peat	4973 ± 37
¹⁴C	UBA-24391	97–98	Peat	5123 ± 32
¹⁴C	UBA-24390	105–106	Peat	5564 ± 33
Lake Vinderelu
210 Pb		0–1	Bulk sediment	−62 ± 1
210 Pb		2.25–2.75	Bulk sediment	−25 ± 5
210 Pb		7.25–8.25	Bulk sediment	51 ± 26
¹⁴C	UBA-19697	29–30	Peat	604 ± 24
¹⁴C	UBA-19698	31–32	Peat	784 ± 28
¹⁴C	UBA-27298	48–49	Bulk sediment	1212 ± 30
¹⁴C	RoAMS 125.45^b	50–51	Organic material	374 ± 40

Present considered AD 1950.

Radiocarbon dates excluded from the model (outliers).

Analysis of pollen, coprophilous fungal spores and plant macro-remains

Pollen analysis was used to reconstruct changes in vegetation composition through time (standard preparation procedure, detailed in Appendix 1). The frequencies of each taxon were calculated as percentages of the total terrestrial pollen sum. To illustrate changes in the main vegetation types, we have grouped the pollen types into the categories used by Feurdean et al. (2013) and Haliuc et al. (2016). This categorisation enables the differentiation of taxa based on the elevation gradient and the degree of human impact. These groups are: (i) upper montane trees, which include tree taxa from the timber- and treeline ecotone (e.g. Betula, Salix, Pinus diploxylon-type, Pinus haploxylon, Picea abies, Larix decidua, Abies alba); (ii) upper montane and subalpine shrubs (Juniperus, Alnus viridis, Ericaceae; the subalpine shrub Pinus mugo is included in the Pinus diploxylon pollen type.); (iii) low to mid montane (Fagus sylvatica) and submontane trees and shrubs (e.g. Quercus, Ulmus, Alnus, Tilia, Corylus, Carpinus betulus, Vitis); (iv) primary anthropogenic indicators (PAI), represented by the pollen of crops and arable weed taxa growing on cultivated land (Cerealia-type, Secale, Zea, Centaurea); (v) secondary anthropogenic indicators (SAI) including plant taxa related to meadows and pastures and more likely to increase with human impact, including livestock grazing (e.g. Chenopodiaceae, Asteraceae, Urtica, Plantago lanceolata, Plantago major/media, Rumex). Detailed pollen diagrams for GR and VD sediment records are shown in Appendix 1 (Figures A3 and A4). Coprophilous fungal spores (e.g. Sporormiella, Sordaria, Podospora) were identified following van Geel et al. (2003) and counted on pollen slides. Calculation of coprophilous spores’ percentages was done by adding their sum to the total terrestrial pollen sum. Plant macro-remains were analysed only at GR, at 2 cm sampling intervals using samples of ca. 15 cm³ sediment volume (Gałka et al., 2020). The samples were washed and wet sieved through a 200 µm mesh size. Fossil carpological remains and vegetative fragments (leaves, rootlets, epidermis) were identified under the stereomicroscope using the available keys (Bojňanský and Fargašová, 2007; Smith and Smith, 2004; Tobolski, 2000; Velichkevich and Zastawniak, 2006) and are presented as concentrations. For a full list of species see Gałka et al. (2020).

Charcoal analysis

Analysis of microscopic and macroscopic charcoal abundance was employed to reconstruct disturbance by fire. Microscopic charcoal has been shown to be indicative of regional fire activity within a radius of 20–100 km of a site, while macroscopic particles (>150 µm) are predominately representative of local fire events occurring within the catchment (Peters and Higuera, 2007; Whitlock and Larsen, 2001). Microscopic charcoal was identified and counted on pollen slides. To quantify macroscopic charcoal, contiguous 2–3 cm³ sub-samples were retrieved at 0.5–1 cm intervals and gently wet-sieved through a 150 μm mesh. The total number of macro-charcoal particles in each sample was estimated under a stereomicroscope (40×) (Feurdean et al., 2013; Florescu et al., 2018). Macro- and micro-charcoal were expressed as accumulation rates (particles cm⁻² year⁻¹) based on macro- and micro-charcoal counts, sample volume, Lycopodium added (micro-charcoal) and sediment deposition time as derived from the age-depth models (detailed description in Appendix 1).

Mineral magnetic properties, geochemistry and grain-size

Magnetic properties (susceptibility, remanence) were employed as indicators of past episodes of soil and bedrock erosion (Thompson and Oldfield, 1986). Magnetic remanence was measured at both sites and magnetic susceptibility only at VD. To derive low-frequency mass specific susceptibility (χ), individual samples were screened with a Bartington Instruments Ltd MS2 metre and MS2B sensor (Dearing, 1999). To determine anhysteretic remanent magnetisation (ARM), we used a Molspin AF Demagnetiser while the saturation isothermal remanent magnetisation (SIRM; magnetic field 1.0 T) was imparted with a Molspin Ltd Pulse Magnetiser. At GR, SIRM backfields (−20, −40, −100 and −300 mT) were additionally induced. For all parameters, the resulting magnetic remanence was measured with a Minispin Fluxgate Magnetometer (Hutchinson et al., 2016). All resulting values were mass normalised. To discriminate between types and sources of magnetic minerals, we calculated the following ratios: SIRM/χ to highlight paramagnetic input, ARM/SIRM as magnetic grain-size indicator, IRM_-100 mT/IRM_-300 mT (L-ratio) as proxy for variation in sources of hard magnetic minerals and S-ratio ((1-IRM-₁₀₀ _mT/SIRM)/2) for the relative contribution of soft, low coercivity minerals that saturate at low magnetic field intensities (Bloemendal et al., 1992; Heslop, 2009; Liu et al., 2007; Thompson and Oldfield, 1986).

Sediment elemental geochemistry and ratios were used to infer detrital input and rock weathering at both sites (Boyle, 2002; Croudace and Rothwell, 2015; Engstrom and Wright, 1984). Elemental concentration was measured using a non-destructive Niton XL3t 900 X-Ray Fluorescence analyser (fpXRF), following Hutchinson et al. (2016). NCS DC73308 was employed as a Certified Reference Material (CRM). Organic matter content was estimated using the loss on ignition (LOI) method (Heiri et al., 2001). Particle size analysis was undertaken with a Horiba Laser Particle Size Analyser (Partica LA-950). We retained median grain size as a proxy for runoff intensity and erosion input (Hutchinson et al., 2016). At VD, particle-size analysis did not include the lowermost 300 years of the profile, for which we used the Zr:Rb ratio as a qualitative indicator of particle-size changes (Biskaborn et al., 2019).

Numerical methods

To statistically determine changes in pollen composition over time, we used stratigraphically-constrained cluster analysis based on Chord distance in PAST (Hammer et al., 2001). Palynological richness, determined from rarefaction analysis, was employed to infer changes in vegetation diversity at the landscape-scale (Birks and Line, 1992; Weng et al., 2006). Rarefaction analysis was applied to the pollen percentages of terrestrial taxa – excluding lowland tree and shrub taxa – and the lowest pollen sums were used to standardise the size of the pollen counts (T₃₁₂ at GR and T₃₉₀ at VD). Rate of change was used to estimate the amount of compositional change in terrestrial pollen per unit time (Grimm and Jacobson, 1992). Chord distance was used to compute dissimilarity between adjacent samples. The datasets were interpolated to equal time intervals (median sampling resolution) before the analysis. Both analyses were performed in Psimpoll (Bennett, 2005). Compositional turnover, estimated by detrended canonical correspondence analysis (DCCA) was employed to statistically measure the amount of change in pollen stratigraphical data (Birks, 2006). DCCA analysis was implemented in CANOCO 4.5 on square root percentages of all terrestrial pollen taxa, excluding lowland arboreal and shrub pollen taxa, with the following settings: detrending by segments, non-linear rescaling and non-downweighting of rare taxa (Ter Braak and Smilauer, 2002).

We used a principal component analysis (PCA) to explore the relationships between biotic proxies (pollen categories and biodiversity indices, charcoal, coprophilous spores) at both sites. PCA was run on the correlation matrix of the data and the significance of the principal components was tested with the broken-stick model (Bennett, 1996). We then performed a redundancy analysis (RDA) to assess the explanatory power of major vegetation types (determined from pollen), grazing (coprophilous fungal spores) and fire (charcoal) over erosion and changes within the depositional environment (inferred from abiotic data) (Legendre and Legendre, 2012; Lepš and Šmilauer, 2003). Both analyses were run in CANOCO 4.5 (Ter Braak and Smilauer, 2002), while the RDA was run on the square-root transformed, centred and standardised species data. Detailed description of statistical choices and data treatment in ordination analysis is provided in Appendix 1.

Results and proxy interpretation

Chronology of sedimentation and lithology

The age-depth models showed that the Lake Gropile (GR) core spans the last 6400 years and the Lake Vinderelu (VD) core covers the last ~1500 years (Figure 2 and Table 1). Deposition time averages to 60 years/cm at GR and 25 years/cm at VD. Both sites show no sign of hiatus, however, there are gradual lithological transitions within the records, ranging from gyttja to peat at GR and from compact gyttja clay to soft gyttja at VD. These lithological transitions are represented graphically in Figures 2 –6.

Figure 2.

Age-depth models for Lake Gropile (GR; left) and Lake Vinderelu (VD; right). Data points and the error range of the calibrated radiocarbon ages used for the construction of the age-depth models are indicated in blue and for 210Pb in green. Data points rejected from the models are drawn in red.

Figure 3.

Lake Gropile (GR) multi-proxy record. Biotic proxies represented by pollen categories, pollen diversity indices (richness, rate of change, turnover), microscopic and macroscopic charcoal influx, dung fungal spores. SAI - Secondary anthropogenic indicators; PAI – Primary anthropogenic indicators. Abiotic proxies are ARM/SIRM, S-ratio, L-ratio, titanium (Ti), K:Rb, Ca:Ti, median particle size (D50). Lithological units as in Figure 2. Archaeological sites comprise only the evidence of human habitation in the study area and their number represents the sum for each historical period. Regional climate periods (blue - cold periods, red - warm periods) are indicated as follow: 1 – Little Ice Age (Christiansen and Ljungqvist, 2012), 2 – Mediaeval Climate Anomaly (Christiansen and Ljungqvist, 2012), 3 – Dak Ages Cold Period (Ljungqvist, 2010), 4 – Roman Warm Period (Ljungqvist, 2010), 5 – Iron Age Cold Epoch (Plunkett and Swindles, 2008), 6 – Bronze Age Optimum (Vinther et al., 2009), 7 – Middle Bronze Age Cold Epoch (Vinther et al., 2009), 8 – Piora Oscillaton (Magny and Haas, 2004), 9 – Holocene Climate Optimum (Davis et al., 2003). The arrow marks the end of the African Humid Period (Collins et al., 2017). The Roman numbers I–IV show the zonation of the record, based on the pollen data.

Figure 4.

Selected pollen taxa for Lake Gropile (GR) (top) and Lake Vinderelu (VD) (bottom).

Figure 5.

Plant macro-remains diagram for Lake Gropile (GR) record. Results are expressed as concentrations and standardised to an equal volume of 1 cm³. Lithological units as in Figure 2.

Figure 6.

Lake Vinderelu (VD) multi-proxy record. Red circles mark burnt wood and blue squares burnt conifer needles in the macro-charcoal record. Lithological units as in Figure 2. Archaeological sites and regional climate periods are represented as in Figure 3. Roman numbers I–III show the zonation of the record, based on pollen data.

Vegetation dynamics, land-use practices and erosion inferred from the multi-proxy records

Stratigraphically constrained cluster analysis indicated the following statistically-determined zones of change for GR: 6400–5200 cal yr BP, 5200–2800, 2800–1000 and 1000 cal yr BP – present; and 1500–1100 cal yr BP, 1100–400, 400 cal yr BP – present for VD respectively. For simplicity, the same time zonation was applied to charcoal and the abiotic proxies (concentration of magnetic minerals, geochemical composition, grain-size).

GR sequence

Zone I (6400–5200 cal yr BP): extended tree cover – high fire occurrence – low grazing – variable erosion

Pollen of upper montane trees accounted for 40–57% in the assemblage and was dominated by Picea abies (30%), Pinus diploxylon (10–20%), Betula (5%), and Abies alba (2%) (Figures 3 and 4; A3). Upper montane and subalpine shrub pollen represented less than 10%, and there was also the maximum concentration of Pinus mugo macro-remains and the only occurrence of Betula nana, Salix sp. and Rubus sp. macro-remains (Figure 5). Secondary anthropogenic indicators (pollen of ruderals and pastures, SAI) were around 7–10% and pollen of the primary anthropogenic indicators (pollen of cultivated plants; PAI) was found with a scattered presence (Figure 3; A3). Palynological richness showed average values (18) compared to the entire record. This time period was characterised by a high rate of change and a major compositional turnover (0.5 SD) towards 5000 cal yr BP (Figure 3). A maximum influx for both micro- and macro-charcoal characterised Zone I, associated with a low proportion (<0.5%) but continuous occurrence of coprophilous fungal spores (Figure 3). Most of this zone overlaps the early in-filling and shallow lake phase of the record (Figure 3), which assumes an easier pathway of catchment sediment fluxes into the depositional environment. Enhanced variability in the ARM/SIRM ratio suggests variable inputs of magnetic mineral grains of single-domain/pseudosingle-domain (SD/PSD) size – typical of soil (increase in ARM/SIRM) and multi-domain (MD) size – typical of parent material/bedrock (decrease in ARM/SIRM) (Figure 3). A variable S-ratio further indicates successive inputs of high-coercivity (hard) antiferromagnetic components such as haematite and goethite, whereby this ratio decreased, and low-coercivity (soft) minerals such as magnetite and maghemite (Figure 3). This variability was further reflected by the L-ratio, showing major changes in the sources of antiferromagnetic minerals. Regarding elemental geochemistry, all the conventional detrital proxies (Ti, Zr, Rb, K) and Fe had similar trends along the entire profile, whereas Ca showed an opposite trend (Figure 3; A2). Elevated detrital input and physical erosion in Zone I are indicated by the overall high Ti concentration and K:Rb ratio. Since K is relatively water soluble, and usually becomes depleted in soils, while Rb is less mobile during mineral weathering, variability in K:Rb ratio is generally interpreted as variations in physical versus chemical rock weathering (e.g. Arnaud et al., 2016; Croudace and Rothwell, 2015). Particle-size was at a maximum at the base of the sedimentary sequence and generally low, that is, <75 µm throughout the zone (Figure 3).

Zone II (5200–2800 cal yr BP): extended tree cover – moderate fire occurrence – low grazing – low erosion

Pollen of upper montane trees varied between 35% and 50%, with an episodic decrease to 25% around 4600 cal yr BP, and was mainly composed of Picea abies (30–35%), Pinus diploxylon (10–20%) and Abies alba (5%). Among the shrubs, Alnus viridis increased to 10%, parallel to a higher abundance of Juniperus sp. (Figure 4; A3). However, Pinus mugo macro-remains declined from ca. 5200 cal yr BP and disappeared completely from the record around 4000 cal yr BP (Figure 5). In comparison to Zone I, a first episodic increase in SAI (10–15%), PAI (<0.5%) and Poaceae (5–7%) occurred. Palynological richness rose gradually until 3500 cal yr BP, then decreased. Rate of change remained remarkably stable throughout zone II, while compositional change occurred around 4800 and 4600 cal yr BP (Figure 3). Zone II displays the maximum concentration of herb macro-remains, which decreased abruptly around 3000 cal yr BP. This was parallelled by a moderate to low micro- and macro-charcoal influx, with a peak around 4600 cal yr BP which coincided with a higher occurrence of coprophilous fungal spores (Figure 3). Regarding erosion proxies, the magnetic mineral types shifted towards the fine-grained and softer assemblages of SD/PSD-type typical of soil (increasing ARM/SIRM and elevated S-ratio). L-ratio gradually reduced, indicating a stabilisation in the erosion sources of antiferromagnetic minerals, except for an episodic increase around 4800 cal yr BP, when particle-size also increased (Figure 3). Detrital fluxes remained elevated, while a shift to lower K:Rb ratios indicates a progression of chemical weathering (i.e. more leaching in soils) (Figure 3).

Zone III (2800–1000 cal yr BP): maximum landscape openness – low fire occurrence – maximum grazing – high erosion

There was a marked decrease in upper montane tree pollen – from 40% to 15% (particularly between 1700 and 1100 cal yr BP), mostly related to the decline of Picea abies (10–20%) and Pinus diploxylon (5–10%) (Figures 3 and 4; A3). Subalpine shrubs such as Alnus viridis, Juniperus and Salix showed increased frequencies up to 15–20%. Only macro-remains of herbaceous plants were detected in this zone (Figure 5). Zone III is marked by the expansion of pastoral and ruderal indicators to 15–20%, with higher frequencies of Artemisia, Plantago lanceolata, Rumex acetosa, Urtica, Filipendula, Chenopodiaceae, Fabaceae and the maximum concentration of coprophilous fungal spores (2%) between 2000 and1000 cal yr BP. Marked fluctuations in pollen diversity were recorded in this zone, with high values between 2800 and 1700 cal yr BP and around 1100 cal yr BP, and minimum values around 1500 cal yr BP, the latter corresponding with a major shift in rate of change. However, the compositional turnover depicts gradual changes during zone III. Micro- and macro-charcoal influx was minimal in this zone (Figure 3). Regarding abiotic proxies, the gradual decrease in ARM/SIRM suggests a switch in dominance from SD/PSD to MD size of magnetic grains, and an increased contribution of hard magnetic minerals, as inferred from the S-ratio, suggesting an increase in erosion depth to bedrock/parent material (Figure 3). Although detrital fluxes remained minimum and stable (all detrital proxies), a new increase in the K:Rb ratio indicated that physical erosion prevailed over chemical weathering, suggesting soil cover disturbance. Particle-size generally decreased, except for a short-lived rise around 1500 cal yr BP. These changes overlapped the stable phase of the depositional environment, with accumulation of gyttja peat until 2400 cal yr BP and peat thereafter, whereby the geomorphological connectivity between site and catchment was low (Figure 3).

Zone IV (1000 cal yr BP – present): variable tree cover – high fire occurrence – moderate grazing – moderate erosion

The last millennium of the record was characterised by an initial rise in the abundance of upper montane tree taxa (to 45%), represented by Picea abies (20–30%) and Abies alba (7%) and a subsequent drop to 25% over the last four centuries (Figure 3). The proportion of subalpine shrubs, in particular Juniperus, Salix and Ericaceae, declined to ~10%, but Alnus viridis re-increased markedly over the past four centuries (Figure 4; A3). Herbaceous taxa indicative of human impact showed the highest frequencies in the record (up to 30–35%) over the last 300 years and were represented by cereals (~2%), Poaceae (10–15%), Artemisia (5%), Apiaceae, Asteraceae, Fabaceae, Rosaceae, Chenopodiaceae, Plantago lanceolata, Rumex acetosa, Urtica (each < 3%). Palynological richness varied markedly, with lowest values around 900 and 400 cal yr BP and the maximum of the entire record over the last four centuries. Rate of change showed large fluctuations, however compositional turnover declined gradually. Coprophilous fungal spores were detected in low concentration in most samples, whereas micro- and macro-charcoal influx increased notably over the last four centuries (Figure 3; A3). Abiotic proxies (available only until 400 cal yr BP) showed an increase in the proportion of SD/PSD, hard magnetic grains that originate from upper soil layers, in line with lower K:Rb ratios indicative of a reduction in physical weathering, all supporting reduced slope disturbance (Figure 3).

VD sequence

Zone I (1500–1100 cal yr BP): maximum tree cover – high fire occurrence – low grazing – unstable slopes

Pollen spectra showed the greatest abundance of upper montane arboreal taxa (~20-25%), represented by Picea abies (~15%), Abies alba (2–4%), Pinus diploxylon and P. haploxylon (~2% each) (Figures 4 and 6; A4). Subalpine shrubs, mainly represented by Alnus viridis, accounted for ~10%. of the pollen sum. The micro-charcoal influx was elevated between 1500 and 1200 cal yr BP and decreased thereafter, whereas the macro-charcoal influx rose markedly at 1200 cal yr BP, shortly followed by a maximum concentration in coprophilous fungal spores (Figure 6). Regarding sedimentological parameters, both saturation remanence (SIRM) and SIRM/X were low, and together with elevated magnetic susceptibility (X) indicated an important paramagnetic and/or diamagnetic component to the material deposited in the lake (Thompson and Oldfield, 1986). Low ARM/SIRM further suggested a predominately MD size of the magnetic grains, likely related to bedrock material. Detrital elements (e.g. Ti, Rb, Zr) had the lowest concentration during this zone, while elevated Zr:Rb and K:Rb ratios suggest increased particle-size and the prevalence of poorly weathered material (Figure 6; A2).

Zone II (1100–400 cal yr BP): maximum landscape openness – high fire occurrence – maximum grazing – high erosion

The abundance of all upper montane tree pollen types abruptly decreased to 10% around 1100 cal yr BP and reached a minimum (3%) around 600–700 cal yr BP (Figures 4 and 6; A4). This coincided with a rise in subalpine shrub pollen to >20%, mainly in Alnus viridis (~20%) and in SAI and PAI. Palynological richness showed the lowest values in the profile with a small increase between 600 and 400 cal yr BP. The rate of change displayed stable values whereas compositional turnover only showed a small decline. The micro and macro-charcoal influx episodically increased around 700 cal yr BP, while the concentration of coprophilous fungal spores decreased first around 900 cal yr BP and then around 500 cal yr BP (Figure 6). This zone is characterised by a lithological change from compact gyttja-clay, deposited shortly after the lake formation, to soft gyttja around 1000 cal yr BP, interrupted by a moss-rich layer deposited between ca. 800 and 600 cal yr BP (Figure 6). Starting 900 cal yr BP, there was a decrease in the size of magnetic grains (rise in ARM/SIRM), and an increase in the hard mineral content in sediments, inferred from X, SIRM and SIRM/X (Figure 6; A2), suggesting a change in sediment sources towards soil material (Thompson and Oldfield, 1986). This pattern is replicated by K:Rb and particle-size, which also reduced from 900 cal yr BP, which supports a shift towards increased input of weathered material (Figure 6).

Zone III (400 cal yr BP – present): high landscape openness – moderate fire and grazing – high erosion

Pollen percentages of upper montane trees slightly increased to ~10% (mainly represented by Betula) over the last 200 years. A maximum proportion in SAI, mainly Urtica, Plantago lanceolata, P. major, Filipendula and Rumex, characterised this zone, along with high macro-charcoal influx and low to moderate concentrations of coprophilous fungal spores (Figures 4 and 6; A4). Zone III was characterised by a maximum in palynological richness, high rate of change and moderate shifts in compositional turnover around 400 and 200 cal yr BP. Under a stable lithology, the magnetic signal was dominated by SD/PSD minerals with ferrimagnetic behaviour over the last 100–150 years, as indicated by a maximum in ARM/SIRM, X and SIRM/X, likely as a result of enhancement with magnetic minerals of anthropogenic origin (Hutchinson et al., 2016). Elevated detrital input (all detrital proxies) and particle-size added to a local intensification of physical erosion of soil horizons (K:Rb) (Figure 6).

PCA and RDA outputs

The first PCA axis accounts for 31% of the total variance in the data at GR and 37% at VD and the second axis represents 19% of the variance at GR and 18% at VD respectively (Figure 7a). At GR, compositional turnover, charcoal, Pinus mugo macrofossils and upper montane tree pollen showed strong, positive correlations with PC1, while shrub pollen, coprophilous fungi, PAI, SAI and richness were negatively correlated with PC1. On PC2, palynological richness clustered together with SAI and herbs, while rate of change varied with herb macrofossils and with coprophilous fungi. At VD, herb and shrub pollen, PAI and SAI showed strong positive correlations with PC1, while upper montane tree pollen and compositional turnover were negatively correlated with PC1. On PC2, coprophilous fungal spores clustered with rate of change and macro-charcoal on the positive side, while upper montane shrubs clustered together with palynological richness on the negative side.

Figure 7.

(a) Biplot diagram of the principal component analysis (PCA) for the GR (left) and VD (right) biotic datasets. (b) Triplot diagram of the redundancy analysis (RDA). Abiotic proxies (response variables) are indicated in the smaller blue font and arrows, and land-use and disturbance proxies (explanatory variables) in the larger red font and arrows. Dashed arrows show explanatory variables that had an explanatory power but were not statistically significant at the 95% level and thus not included in the model. Type II scaling was chosen to illustrate correlative relationships between variables.

The first two RDA axes at GR explain 31.1% and 4.4% of the total variance respectively (Figure 7b). Forward selection indicated that macro-charcoal had the strongest explanatory power (15% of the cumulated explained variance) on abiotic proxies, followed by upper montane tree pollen and coprophilous fungal spores (13.1 and 10% respectively). All three explanatory variables were statistically significant at the 95% level (p = 0.002, 0.006 and 0.04). The corresponding biplot grouped macro-charcoal with Fe and L-ratio, likely indicating a positive relationship between fire and the input of hard magnetic minerals in the sediments, which may also suggest charcoal transport through runoff and thus taphonomic effects. The positive association of upper montane trees with detrital elements and negative with K:Rb further reflects environmental conditions suitable for tree growth at high elevation, which simultaneously promote chemical weathering and affect erosion and detrital fluxes to the site. Coprophilous fungi clustered together with K:Rb, heavy metals and fine magnetic minerals, suggesting a causal relationship between grazing and the input of poorly weathered and magnetically soft material, but could also be related to increase in grazing pressure as a result of subregional mining.

At VD, the first RDA axis accounted for 37% and the second for 7.3% of the variance (Figure 7b). Upper montane shrub pollen explained 29.7% of the variance in the abiotic proxies and was statistically significant at the 95% level (p = 0.002). Coprophilous fungal spores and micro-charcoal additionally explained 10.4% and 6% of the variance but were not significant at the 95% level. The biplot showed that montane shrubs are positively correlated with heavy metals, Sr, organic matter content and Fe:Mn (proxy for bottom water oxygenation), and negatively related to K:Rb and ARM/SIRM, along the first axis. This implies an association between shrub encroachment and slope stability, with a reduction in physical erosion and detrital fluxes, and an increase in anoxia due to higher organic matter input. Although not statistically significant at 95% level, coprophilous spores seem to explain variation in erosion proxies, while micro-charcoal is associated with the input of paramagnetic material.

Discussion

Long-term changes in land cover and land-use practices in the subalpine landscapes of the Rodna and Maramureș Mts

The two multi-proxy archives demonstrate that prehistoric and historic land use practices of variable intensities strongly altered the structure of the subalpine landscapes in Rodna and Maramures Mts. Archaeological traces of Neolithic and Eneolithic settlements in the surrounding river-valleys (Figure 1) and artefacts at 1400–1600 m asl associated with game hunters and livestock grazing (Bobînă, 2015, 2018; Dragoman et al., 2015) support the occurrence of occasional human activities at high elevation in our study area since at least 6400 cal yr BP. These align with regional palynological evidence of montane forest grazing by livestock since 6000–6500 cal yr BP across the Carpathians (e.g. Fărcaş et al., 2013; Feurdean and Astalos, 2005; Peters et al., 2020; Schumacher et al., 2016). The pollen record from Rodna Mts suggests a greater extent and/or density of montane tree cover between 6400 and 2800 cal yr BP, compared to the rest of the record, represented by Pinus, Picea abies and Abies alba, likely connected to a higher position of the tree- and timberline. The plant macrofossil record does not support the local presence of conifer trees in the GR catchment (at ~1900 m asl), but points to the local occurrence of subalpine shrubs with scattered cold-adapted deciduous trees (Pinus mugo, Salix, Betula alba) prior to 5000 cal yr BP (Figure 4). Disturbance by fire appeared to be important at both local and regional scales during this interval (Figure 3). While warm and dry climatic conditions at this time were conducive to fire (Diaconu et al., 2017), a higher biomass availability associated with a more extensive woody cover, that is, Pinus mugo, also provided sufficient fuel to burn (Figure 7a). Nevertheless, a major compositional change overlapping an episodic decline in the pollen percentages of the upper montane trees, along with the local decline of most subalpine shrubs occurred between 5000 and 4500 cal yr BP (Figures 3 and 5). While the decrease in shrub macro-remains, in particular of Pinus mugo, may be attributed to taphonomic effects, their complete disappearance from 4000 cal yr BP, also visible at other subalpine sites in Rodna Mts (Feurdean et al., 2016), suggests a regional-scale range contraction. Summer temperatures declined at 4500 cal yr BP (Diaconu et al., 2017), which may explain the tree- and timberline depression. On the other hand, the continuous occurrence, albeit in low proportions, of anthropogenic indicators, that is, coprophilous spores, Plantago lanceolata, P. major, Rumex acetosa, Filipendula, Urtica and Fabaceae (Figure A3), suggest the existence of pastures. Therefore, landscape grazing and burning (high charcoal influx), in addition to decreased summer temperatures, may explain the Pinus mugo decline at 5000 cal yr BP and its disappearance after 4000 cal yr BP at GR and other subalpine sites in Rodna Mts (Feurdean et al., 2016). Interestingly, the episodic decline in high-elevation tree cover at 4500–5000 cal yr BP visible in our record is also apparent in other high elevation vegetation records in Rodna Mts around 4500 cal yr BP (Feurdean et al., 2016; Tanţău et al., 2011, 2014b) and in Southern Carpathians (Vincze et al., 2017) around 4200 cal yr BP. Generally, this reduction in the tree cover, and extension of grazed habitats, matches the beginning of the Early Bronze Age seen from the sub-regional archaeological evidence in Rodna Mts, altogether supporting pasturing as a significant driver of vegetation change.

From 2800 cal yr BP, the marked decline in the abundance of high-elevation trees and the extension of shrubs and grasslands in the Rodna Mts, along with a decline in charcoal influx and the increased abundance of coprophilous spores highlight a change in land-use practices towards more intensive grazing and reduced burning (Figure 3). This may suggest that the grazing pressure alone was sufficient to keep the landscape open and prevent treeline advance and shrub encroachment in subalpine grazing areas. In the Maramures Mts the VD record spans only the last 1500 years and shows intensive burning, followed by grazing intensification from 1200 to 1100 cal yr BP. Rate of change and rarefaction analysis demonstrate dynamic changes in vegetation composition and diversity over the past 2000 years at GR and around 1100 cal yr BP and last 400 years at VD, as an effect of land-use intensification. Archaeological evidence shows numerous artefacts dating to the Early Iron Age at both sites, though more common on the southern slopes of the Rodna Mts, and from the High Middle Ages across the Maramureș Mts, and linked to mining and pastoral activities (Bobînă, 2015, 2018; Dragoman et al., 2015; Kacsó et al., 2010). Documentary sources also mention mining and pasturing in the area since the Iron Age (Arhiva Someșană, 1928; Borcoş and Udubaşa, 2012), which may have increased the demand for wood. In fact, the initiation of more sustained forest loss and the extension of grazed habitats around 3200–2800 cal yr BP at high elevation is also recorded on the eastern and northern sides of the Rodna Mts (Geantă et al., 2014; Tanţău et al., 2011, 2014b) and the Maramureș Mts (Fărcaş et al., 2013). Landscape grazing and burning including woody vegetation as evidenced by charred fragments of conifer wood and needles, continued throughout the last millennium in the Maramureș Mts, while in the Rodna Mts it became the most pronounced over the last four centuries (Figures 3 and 6), parallelling human population increases and the intensification of summer transhumance as suggested by archaeological and documentary sources (Coldea and Cristea, 1998; Kubijovič, 1934; PNPNMM, 2016; PNPNMR, 2019). This shift to marked land-use intensification in high-elevation areas in the last millennium, and particularly over the last 400–500 years, is captured by palaeoecological evidence across the Romanian Carpathians (Árvai et al., 2016; Fărcaş et al., 1999; 2013; Feurdean et al., 2009; Magyari et al., 2018; Vincze et al., 2017).

The effects of land-use practices on landscape diversity and change

Our results show that, at the landscape scale, pollen diversity generally increased with vegetation openness represented by herbaceous communities created or maintained by anthropogenic activities. This finding of increased pollen diversity with landscape openness and the level of human impact agrees with previous studies in the Carpathians (Feurdean et al., 2013) and central-eastern Europe (Giesecke et al., 2019). Although our pollen richness excludes taxa typical of lowlands, the picture of more floristically diverse high-elevation landscapes with human impact is likely representative of vegetation assemblages from a wider elevational range, due to the upslope pollen transport effects (Feurdean et al., 2013). A closer inspection of pollen richness and the anthropogenic indicators shows a more nuanced relationship between changes in landscape diversity and land-use intensity. At GR, high levels of pollen richness occurred in a series of episodes (6400–5200, 4600–3800, 2600–2000 and last 400 years) associated with the expansion of meadows and pastures. At VD, a sequence that covers a period with an already intensive human impact (the last 1500 years), small peaks in pollen richness (at 1400, 1200–1100, 900 and 500–150 cal yr BP correspond to extension of meadows, pastures and shrubs (Figures 3, 6 and 7a). At both sites, episodes of elevated diversity are associated with low or declining grazing pressure, indicated by the coprophilous spores, whereas episodes of reduced pollen diversity (1500–1000 cal yr BP at GR, and 1100–900 cal yr BP at VD) correspond to indications of grazing intensification (Figures 3 and 6; A3–A4). These relationships suggest low to moderate grazing practices supported plant diversity of subalpine habitats. In contrast, grazing intensification resulted in a loss of plant diversity and likely contributed to the extension of species-poor Nardetum associations, which are unpalatable to livestock, but now occupy wide areas in the subalpine habitats of both mountain ranges (Novak and Horváthová, 2019; PNPNMM, 2016; PNPNMR, 2019).

Burning, in addition to grazing, has also shaped landscape composition and diversity in the study area. A link between enhanced burning and increased landscape diversity was evident during 6400–5200 cal yr BP at GR and over the last four centuries at both sites (Figures 3 and 6). However, the low proportions of coprophilous spores during these intervals suggest that, after periods of pasture abandonment, fire was often used to manage woody encroachment. To sum up, it appears that at VD the combined effect of local burning and grazing in removing shrub cover drove changes in landscape diversity and structure (Figure 7a). At GR, fire was more closely associated with changes in shrub cover, while grazing aligns more with the diversity of the herbaceous cover. Based on the evidence above, we infer that the present subalpine vegetation in both mountain areas appears to have been closely connected to traditional land-use practices for millennia and is thus partly dependent on land management (e.g. Feurdean et al., 2018; Halada et al., 2011).

Linkages between changes in land-use and patterns of catchment erosion

Changes in magnetic mineralogy and sediment geochemical composition facilitate a distinction between inputs of soil, subsoil and parent material, and provide a qualitative measure of the development of erosion at a catchment scale (e.g. Dearing et al., 2001; Giguet-Covex et al., 2011; Hutchinson et al., 2016). Our results illustrate that erosion patterns inferred from magnetic and geochemical assemblages differ with the major changes in past land-cover and land-use practices. The catchments of both sites are developed on metamorphic and sedimentary rocks (Table A1), which are rich in Fe but weakly ferrimagnetic (Oncescu, 1959; PNPNMM, 2016; PNPNMR, 2019). At GR, Fe concentration increases with coarser and harder magnetic minerals, while at VD with coarser and more paramagnetic material (Figure 7b; A2). Coarse-grained magnetic minerals are usually associated with primary sources (parent material, bedrock), while fine-grained assemblages are typical of soil (Thompson and Oldfield, 1986). Consequently, at both sites, the soil-derived material appears to be dominated by secondary fine-grained minerals displaying soft magnetic behaviour which were likely converted from primary assemblages through weathering and paedogenesis (Haliuc et al., 2016; Thompson and Oldfield, 1986), while the parent material contains coarser and harder (e.g. haematite and goethite) magnetic assemblages at GR, coarser and paramagnetic at VD respectively.

The 6400–2800 cal yr BP interval at GR (Zones I and II) is characterised by a more extended tree cover, compared to the rest of the record, and few indications of human land-use. Here, the gradual decrease in the proportion of coarse, hard magnetic minerals and the increase in fine-grained, soft/mixed minerals indicate that erosion of bedrock and parent material gradually decreased (Figure 3). Slope stability and reduced soil disturbance are also supported by the relative stability in detrital input (Figure 3; A2). A shift from physical erosion to more chemical weathering occurred around 5400 cal yr BP and prevailed until 2800 cal yr BP, as further inferred from lower K:Rb ratios (Figure 3). This change indicates soil development inside the catchment and may be connected to the gradually wetter conditions showed by the sub-regional palaeoclimatic records (Diaconu et al., 2017; Feurdean et al., 2008; Magyari et al., 2009). However, grain-size variability illustrates two instances of hydrological disturbance, accompanied by short-lived increases in the depth of local erosion: the first at 5000–4400 cal yr BP, coinciding with the episodic decline in tree cover, and the second at 4000–3500 cal yr BP, overlapping with the disappearance of local populations of Pinus mugo from the record, both attributable to anthropogenic disturbance (Figures 3 and 5).

Zones III and IV of land-cover change at GR, starting at 2800 cal yr BP and continuing to the present, showed increased landscape openness and more intensive grazing, with a maximum between 2000 and 1000 cal yr BP. These changes correspond to a shift from predominantly weathered, soil-derived material (fine-grained, soft magnetic assemblages) to more enhanced sediment input from the poorly weathered parent material (coarse-grained, harder) (Figure 3). It should be noted that these changes also overlapped a low geomorphological connectivity between site and catchment, which diminished the erosion signal. Local climate reconstructions suggest warmer/drier conditions between 3200 and 1000 cal yr BP (Diaconu et al., 2017; Gałka et al., 2020), while an increase in wind-driven dust and silt fluxes in the area around 3000 and between 2000 and 1000 cal yr BP was attributed to local disturbances (Panait et al., 2019). Similar patterns are observed at VD, where erosion and the input of unweathered material increased at 1100 cal yr BP with landscape opening and intensification in land-use practices, and diminished only when grazing and burning reduced, around 600–400 cal yr BP. Erosion proxies at GR do not span the marked land-use intensification of the last 400 years. Despite this limitation, grazing seems to have a similar effect at both sites, linked to increases in the fluxes of unweathered, deeper material (Figure 7b). This demonstrates that overgrazing amplified soil erosion at both sites, by exposing the lower soil horizons and bedrock to eroding agents. Erosion increase with intensified land-use over the last 3000 years was also documented across elevations in the Eastern Carpathians (Florescu et al., 2017; Haliuc et al., 2016; Magyari et al., 2009; Tapody et al., 2021) and may have contributed to the increase in sediment load and aggradation of the rivers draining the opposite flanks of the Eastern Carpathians (Rădoane et al., 2019).

Concluding remarks and strategies for optimising management of subalpine landscapes

This multi-proxy, interdisciplinary study reveals a complex picture of millennial-scale changes in land use and management practices, and their subsequent effects on landscape composition, diversity and slope erosion in the subalpine areas of the Rodna and Maramureș Mts, northern Romania. The legacy of past land-use practices is presently reflected by more open subalpine habitats that have evolved under a combination of pasturing and burning. For the Rodna Mts, results evidenced a more extended tree cover and high fire activity between 6400 and 2800 cal yr BP, when soil erosion was reduced and more regulated by climatic factors. Although the palaeoecological evidence suggests low-intensity pasturing and occasional burning since the late Neolithic, the decline in local Pinus mugo populations at 5000 cal yr BP and their disappearance after 4000 cal yr BP both locally and regionally are partly attributable to anthropogenic disturbance. Anthropogenic impact became clear 2800 years ago, when landscape openness and pasturing increased, along with soil cover disturbance. In the Maramureș Mts, intensive burning followed by grazing occurred around 1200–1100 cal yr BP and continued throughout the last millennium. In contrast, in the Rodna Mts burning and grazing became most pronounced over the last four centuries, parallelling the human population increase and the intensification of summer transhumance. Therefore, we found that the effects of pastoralism and burning on changes in landscape-scale diversity were site-specific. At VD, the combined effect of local burning and grazing in removing shrubs drove changes in landscape diversity and structure. At GR, fire was more connected to shrub cover change, while grazing reflected the diversity of the herbaceous cover. Notably, moderate to low grazing appeared beneficial for the diversity of both subalpine vegetation assemblages through the creation of rich montane grassland-shrub mosaic communities. Conversely, overgrazing has reduced landscape and probably grassland diversity and exacerbated erosion at both sites, by exposing the lower soil horizons and bedrock to eroding agents.

This long-term perspective underscores the need to reconsider the conservation targets of these habitats. We propose that these semi-natural ecosystems hold significant ecological and cultural value. As current land-management strategies for both mountain areas involve legal restrictions on the use of fire to control shrub encroachment and treeline advance, coupled with continuously decreasing and highly localised grazing pressure, there is a risk of losing a habitat produced by grazing and/or burning-dependent grasslands. Therefore, we recommend the management of these ecosystems through controlled, low intensity pasturing or occasional burning. It is also important to acknowledge the necessity for more local scale studies to investigate the interactions between disturbances by grazing and fire, and their potential effects on both subalpine vegetation structure and composition at a landscape scale, and local sediment fluxes.

Footnotes

Appendix

Acknowledgements

Mihaly Braun, Anca Geantă, Daniel Vereș, and Tudor Tămaș, are thanked for their assistance in collecting the sediment cores.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: the Romanian National Authority for Scientific Research and Innovation, CNCS – UEFISCDI (grant numbers PN-III-P1-1.1-TE-2019-1628 and PN-III-P4-IDPCE-2016-0711), the Romania National Council for Higher Education Funding, CNFIS (grant number CNFIS-FDI-2023-F-0579) and the Deutsche Forschungsgemeinschaft (grant number FE_1096/9).

ORCID iDs

Gabriela Florescu

Mariusz Gałka

References

Bennett

Willis

(2001). Pollen. In: Smol

Birks

HJP

Last

(eds) Tracking Environmental Change Using Lake Sediments: Volume 3 Terrestrial, Algal, and Siliceous indicators. Springer, Dordrecht (pp. 5-32).

Bennett

(1996) Determination of the number of zones in a biostratigraphical sequence. New Phytologist 132(1):155–170.

Feurdean

Liakka

Vannière

Marinova

Hutchinson

Mosburgger

Hickler

(2013) 12,000-Years of fire regime drivers in the lowlands of Transylvania (Central-Eastern Europe): a data-model approach. Quaternary Science Reviews (81): 48–61.

Florescu

Vannière

Feurdean

(2018). Exploring the influence of local controls on fire activity using multiple charcoal records from northern Romanian Carpathians. Quaternary International (488): 41–57.

Grimm

(2011) Tilia software, version 2.1.1. Springfield, IL: Illinois State Museum

Legendre

(2012) Numerical ecology. Elsevier.

Lepš

Šmilauer

(2003) Multivariate analysis of ecological data using CANOCO. Cambridge university press.

Moore

Webb

Collison

(1991). Pollen analysis. Blackwell scientific publications.

Reille

(1995) Pollen et spores d’Europe et d’Afrique du Nord. Atlas photographique.

10.

Stockmarr

(1971) Tabletes with spores used in absolute pollen analysis. Pollen spores 13 (pp. 615-621).

11.

Ter Braak

Smilauer

(2002). CANOCO reference manual and CanoDraw for Windows user’s guide: software for canonical community ordination (version 4.5). www.canoco.com.

12.

Zuur

Ieno

Elphick

(2010) A protocol for data exploration to avoid common statistical problems. Methods in ecology and evolution 1(1): 3–14.

13.

Arhiva Someșană (1928) Revistă Istorică Culturală no. 9. Năsăud. Available at: https://documente.bcucluj.ro/web/bibdigit/periodice/arhivasomesana/pdf/BCUCLUJ_FP_279098_1928_005_009.pdf (accessed 20 March 2022).

14.

Arnaud

Poulenard

Giguet-Covex

, et al. (2016) Erosion under climate and human pressures: An alpine lake sediment perspective. Quaternary Science Reviews 152: 1–18.

15.

Árvai

Popa

Mîndrescu

, et al. (2016) Dendrochronology and radiocarbon dating of subfossil conifer logs from a peat bog, Maramureş mts, Romania. Quaternary International 415: 6–14.

16.

Baur

Cremene

Groza

, et al. (2007) Intensified grazing affects endemic plant and gastropod diversity in alpine grasslands of the Southern Carpathian Mountains (Romania). Biologia 62(4): 438–445.

17.

Bennett

(1996) Determination of the number of zones in a biostratigraphical sequence. New Phytologist 132(1): 155–170.

18.

Bennett

(2005) Documentation for Psimpoll 4.25 and Pscomb 1.03: C Programs for Plotting Pollen Diagrams and Analysing Pollen Data. Uppsala, Sweden: Department of Earth Sciences, University of Uppsala.

19.

Birks

HJB

(2006) Estimating the amount of compositional change in late-Quaternary pollen-stratigraphical data. Vegetation History and Archaeobotany 16(2–3): 197–202.

20.

Birks

HJB

Line

(1992) The use of rarefaction analysis for estimating palynological richness from Quaternary pollen-analytical data. Holocene 2(1): 1–10.

21.

Biskaborn

Nazarova

Pestryakova

, et al. (2019) Spatial distribution of environmental indicators in surface sediments of Lake Bolshoe Toko, Yakutia, Russia. Biogeosciences 16(20): 4023–4049.

22.

Blaauw

(2010) Methods and code for ‘classical’ age-modelling of radiocarbon sequences. Quaternary Geochronology 5: 512–518.

23.

Bloemendal

King

Hall

, et al. (1992) Rock magnetism of Late Neogene and pleistocene deep-sea sediments: Relationship to sediment source, diagenetic processes, and sediment lithology. Journal of Geophysical Research 97: 4361–4375.

24.

Bobînă

(2015) Mountain Archaeology in Romania: The status of research. Terra Sebus 7: 149–164.

25.

Bobînă

(2018) Arheologie Montană. Studiu de caz: Munții Maramureșului. Teza de Doctorat. [Mountain Archaeology. Case study: Maramures Mountains. PhD Thesis (in Romanian)]. Universitatea „1 Decembrie 1918” din Alba Iulia. Facultatea de Istorie și Filologie, Școala Doctorală de Istorie.

26.

Bojňanský

Fargašová

(2007) Atlas of Seeds and Fruits of Central and East-European Flora: The Carpathian Mountains Region. Dordrecht, The Netherlands: Springer Science & Business Media.

27.

Borcoş

Udubaşa

(2012) Chronology and characterisation of mining development in Romania. Romanian Journal of Earth Sciences 86(1): 17–26.

28.

Boyle

(2002) Inorganic geochemical methods in palaeolimnology. In: Last

Smol

(eds) Tracking Environmental Change Using Lake Sediments Volume 2 Physical and Geochemical Methods. Dordrecht: Springer, pp.83–141.

29.

Carter

Bobek

Moravcová

, et al. (2020) The role of climate-fuel feedbacks on holocene biomass burning in upper-montane Carpathian forests. Global and Planetary Change 193(7): 103264.

30.

Christiansen

Ljungqvist

(2012) The extra-tropical Northern Hemisphere temperature in the last two millennia: Reconstructions of low-frequency variability. Climate of the Past 8(2): 765–786.

31.

Coldea

(1990) Munţii Rodnei. Studiu Geobotanic [Rodnei Mountains. Geobotanical Study]. Bucureşti: Editura Academiei Române.

32.

Coldea

Cristea

(1998) Floristic and community diversity of sub-alpine and alpine grasslands and grazed dwarf-shrub heaths in the Romanian Carpathians. Pirineos (151–152): 73–82.

33.

Coldea

Stoica

Puşcaş

, et al. (2009) Alpine–subalpine species richness of the Romanian Carpathians and the current conservation status of rare species. Biodiversity and Conservation 18(6): 1441–1458.

34.

Collins

Prange

Caley

, et al. (2017) Rapid termination of the African humid period triggered by northern high-latitude cooling. Nature Communications 8(1): 1372–1411.

35.

Croudace

Rothwell

(2015) Micro-XRF Studies of Sediment Cores: Applications of a Non-destructive Tool for the Environmental Sciences, vol. 17. Dordrecht: Springer, p.668.

36.

Davis

BAS

Brewer

Stevenson

, et al. (2003) The temperature of Europe during the holocene reconstructed from pollen data. Quaternary Science Reviews 22(15–17): 1701–1716.

37.

Dearing

(ed.) (1999) Environmental Magnetic Susceptibility. Using the Bartington MS2 System, 2nd edn. Kenilworth: Chi Publ.

38.

Dearing

Battarbee

Dikau

, et al. (2006) Human–environment interactions: learning from the past. Regional Environmental Change 6(1–2): 1–16.

39.

Dearing

Jones

(2003) Coupling temporal and spatial dimensions of global sediment flux through lake and marine sediment records. Global and Planetary Change 39(1–2): 147–168.

40.

Dearing

Doody

, et al. (2001) Preliminary reconstruction of sediment-source linkages for the past 6000 yr at the Petit lac d’Annecy, France, based on mineral magnetic data. Journal of Paleolimnology 25: 245–258.

41.

Diaconu

Tóth

Lamentowicz

, et al. (2017) How warm? How wet? hydroclimate reconstruction of the past 7500 years in northern Carpathians, Romania. Palaeogeography Palaeoclimatology Palaeoecology 482: 1–12.

42.

Dietre

Reitmaier

Walser

, et al. (2020) Steady transformation of primeval forest into subalpine pasture during the late Neolithic to Early Bronze Age (2300−1700 BC) in the Silvretta alps, Switzerland. Holocene 30(3): 355–368.

43.

Dítě

Hájek

Svitková

, et al. (2018) Glacial-relict symptoms in the western Carpathian flora. Folia Geobotanica 53(3): 277–300.

44.

Dragoman

Pop

Bobînă

, et al. (2015) O arheologie a munților din Maramureș, România: cercetările din 2014 și 2015 [An Archaeology of the Maramureș Mountains, Romania: The Surveys from 2014 and 2015 (in Romanian, with English abstract)]. in Marmatia - publicaţie periodică a Muzeului Judeţean de Istorie şi Arheologie Maramureş (12), Baia Mare. ISSN- 1582-9979.

45.

Dragotă

Kucsicsa

(2011) Global climate change-related particularities in the Rodnei Mountains National Park. Carpathian Journal of Earth and Environmental Sciences 6(1): 43–50.

46.

Durak

Żywiec

Kapusta

et al. (2015) Impact of land use and climate changes on expansion of woody species on subalpine meadows in the Eastern Carpathians. Forest Ecology and Management 339: 127–135.

47.

Engstrom

Wright

Jr (1984) Chemical stratigraphy of lake sediments as a record of environmental change. In: Haworth

Lund

JWG

(eds) Lake Sediments and Environmental History: Studies in Palaeolimnology and Palaeoecology in Honour of Winifred Tutin. University of Minnesota Press, Minneapolis. pp.11–67.

48.

Farcas

de Beaulieu

Reille

, et al. (1999) First 14C datings of late glacial and holocene pollen sequences from Romanian Carpathes. Comptes Rendus de l’Académie des Sciences-Series III-Sciences de la Vie 322(9): 799–807.

49.

Fărcaş

Tanţău

Mîndrescu

, et al. (2013) Holocene vegetation history in the Maramureş mountains (Northern Romanian Carpathians). Quaternary International 293: 92–104.

50.

Feurdean

Astalos

(2005) The impact of human activities in the Gutaiului Mountains, Romania. Studia UBB Geologia 50(1–2): 63–72.

51.

Feurdean

Florescu

Vannière

, et al. (2017) Fire has been an important driver of forest dynamics in the Carpathian Mountains during the holocene. Forest Ecology and Management 389: 15–26.

52.

Feurdean

Gałka

Tanţău

, et al. (2016) Tree and timberline shifts in the northern Romanian Carpathians during the holocene and the responses to environmental changes. Quaternary Science Reviews 134: 100–113.

53.

Feurdean

Klotz

Mosbrugger

, et al. (2008) Pollen-based quantitative reconstructions of holocene climate variability in NW Romania. Palaeogeography Palaeoclimatology Palaeoecology 260(3–4): 494–504.

54.

Feurdean

Liakka

Vannière

, et al. (2013) 12,000-Years of fire regime drivers in the lowlands of Transylvania (Central-Eastern Europe): A data-model approach. Quaternary Science Reviews 81: 48–61.

55.

Feurdean

Willis

Astaloş

(2009) Legacy of the past land-use changes and management on the ‘natural’ upland forest composition in the Apuseni Natural Park, Romania. Holocene 19: 967–981.

56.

Feurdean

Ruprecht

Molnár

, et al. (2018) Biodiversity-rich European grasslands: Ancient, forgotten ecosystems. Biological Conservation 228: 224–232.

57.

Finsinger

Fevre

Orbán

, et al. (2018) Holocene fire-regime changes near the treeline in the Retezat mts. (Southern Carpathians, Romania). Quaternary International 477: 94–105.

58.

Florescu

Hutchinson

Kern

, et al. (2017) Last 1000 years of environmental history in Southern Bucovina, Romania: A high resolution multi-proxy lacustrine archive. Palaeogeography Palaeoclimatology Palaeoecology 473: 26–40.

59.

Florescu

Vannière

Feurdean

(2018) Exploring the influence of local controls on fire activity using multiple charcoal records from northern Romanian Carpathians. Quaternary International 488: 41–57.

60.

Foster

Dearing

Jones

, et al. (2003) Meteorological and land use controls on past and present hydro-geomorphic processes in the pre-alpine environment: An integrated lake–catchment study at the Petit lac d’Annecy, France. Hydrological Processes 17(16): 3287–3305.

61.

Gałka

Tantau

Carter

, et al. (2020) The holocene dynamics of moss communities in subalpine wetland ecosystems in the eastern Carpathian Mountains, Central Europe. Bryologist 123(1): 84–97.

62.

Geantă

Gałka

Tanţău

, et al. (2014) High mountain region of the northern Romanian Carpathians responded sensitively to holocene climate and land use changes: A multi-proxy analysis. Holocene 24: 944–956.

63.

Geografia României (Geography of Romania) (1987) vol. iii, Carpaţii româneşti şi Depresiunea Transilvaniei, Ed. Academiei Republicii Socialiste România.

64.

Giesecke

Wolters

van Leeuwen

JFN

, et al. (2019) Postglacial change of the floristic diversity gradient in Europe. Nature Communications 10(1): 5422–5427.

65.

Giguet-Covex

Arnaud

Poulenard

, et al. (2011) Changes in erosion patterns during the holocene in a currently treeless subalpine catchment inferred from lake sediment geochemistry (Lake anterne, 2063 m a.s.l., NW French alps): The role of climate and human activities. Holocene 21(4): 651–665.

66.

Giguet-Covex

Pansu

Arnaud

, et al. (2014) Long livestock farming history and human landscape shaping revealed by lake sediment DNA. Nature Communications 5(1): 3211–3217.

67.

Gómez-Villar

Sanjuan

Ruiz

JMG

, et al. (2014) Sediment organization and adjustment in a torrential reach of the Upper Ijuez River, Central Spanish Pyrenees. Cuadernos de Investigación Geográfica: Geographical Research Letters 40: 191–214.

68.

Grimm

Jacobson

(1992) Fossil-pollen evidence for abrupt climate changes during the past 18 000 years in eastern North America. Climate Dynamics 6(3–4): 179–184.

69.

Halada

Evans

Romão

, et al. (2011) Which habitats of European importance depend on agricultural practices? Biodiversity and Conservation 20(11): 2365–2378.

70.

Haliuc

Hutchinson

Florescu

, et al. (2016) The role of fire in landscape dynamics: An example of two sediment records from the Rodna Mountains, northern Romanian Carpathians. CATENA 137: 432–440.

71.

Hall

Arheimer

Borga

, et al. (2014) Understanding flood regime changes in Europe: A state-of-the-art assessment. Hydrology and Earth System Sciences 18(7): 2735–2772.

72.

Hammer Harper

Ryan

(2001) PAST: Paleontological statistics software package for education and data analysis. Palaeontologia Electronica 4(1): 9.

73.

Heiri

Lotter

Lemcke

(2001) Loss on ignition as a method for estimating organic and carbonate content in sediments: Reproducibility and comparability of results. Journal of Paleolimnology 25(1): 101–110.

74.

Heslop

(2009) On the statistical analysis of the rock magnetic S-ratio. Geophysical Journal International 178(1): 159–161.

75.

Hutchinson

Akinyemi

Mîndrescu

, et al. (2016) Recent sediment accumulation rates in contrasting lakes in the Carpathians (Romania): Impacts of shifts in socio-economic regime. Regional Environmental Change 16(2): 501–513.

76.

Jones

Reinhardt

Dearing

, et al. (2013) Detecting climatic signals in an anthropogenically disturbed catchment: The late-Holocene record from the Petit lac d’Annecy, French alps. Holocene 23(9): 1329–1339.

77.

Kacsó

Iştvan

Minghiraş

(2010) Cercetări de arheologie montană la Băile Borşa – Montanarchäologische Forschungen in Băile Borşa. Memoria Antiquitatis 25–26: 505–536.

78.

Knowles

Harpold

Cowie

, et al. (2015) The relative contributions of alpine and subalpine ecosystems to the water balance of a mountainous, headwater catchment. Hydrological Processes 29(22): 4794–4808.

79.

Kobiv

(2017) Response of rare alpine plant species to climate change in the Ukrainian Carpathians. Folia Geobotanica 52(2): 217–226.

80.

Körner

(2003) The Alpine Life Zone. In Alpine Plant Life. Berlin, Heidelberg: Springer, pp.9–20.

81.

Kołaczek

Buczek

Margielewski

, et al. (2021) Development and degradation of a submontane forest in the Beskid Wyspowy Mountains (Polish Western Carpathians) during the holocene. Holocene 31: 1716–1732.

82.

Kubijovič

(1934) Păstoritul în Maramureş [Pastoralism in Maramureș. In Romanian]. Buletinul Societăţii Regale Română de Geografie.

83.

Kucsicsa

(2013) Parcul Naţional Munţii Rodnei: Relaţii Om-Mediu. București: Editura Universitară.

84.

Legendre

(2012) Numerical Ecology. Developments in Environmental Modelling. Amsterdam: Elsevier Science.

85.

Lepš

Šmilauer

(2003) Multivariate Analysis of Ecological Data Using CANOCO. Cambridge: Cambridge University Press.

86.

Liu

Roberts

Torrent

, et al. (2007) What do the HIRM and S-ratio really measure in environmental magnetism? Geochemistry Geophysics Geosystems 8(9).

87.

Ljungqvist

(2010) A new reconstruction of temperature variability in the extra-tropical Northern Hemisphere during the last two millennia. Geografiska Annaler Series A Physical Geography 92(3): 339–351.

88.

Magny

Haas

(2004) A major widespread climatic change around 5300 cal. yr BP at the time of the Alpine Iceman. Journal of Quaternary Science 19(5): 423–430.

89.

Magyari

Buczkó

Jakab

, et al. (2009) Palaeolimnology of the last crater lake in the eastern Carpathian Mountains: A multiproxy study of holocene hydrological changes. Hydrobiologia 631: 29–63.

90.

Magyari

Vincze

Orbán

, et al. (2018) Timing of major forest compositional changes and tree expansions in the Retezat mts during the last 16,000 years. Quaternary International 477: 40–58.

91.

Mihai

Savulescu

Sandric

(2007) Change detection analysis (1986–2002) of vegetation cover in Romania. Mountain Research and Development 27(3): 250–258.

92.

Mîndrescu

(2001) Lacul Vinderel. Caracterizare hidro-geomorfologică (Munţii Maramureşului). ANALELE Universităţii „Ştefan cel Mare” Suceava, Secţiunea Geografie Anul X.

93.

Nagy

Grabherr

Körner

, et al. (eds.) (2003) Alpine Biodiversity in Europe, vol. 167. Berlin: Springer Science & Business Media.

94.

Noroozi

Talebi

Doostmohammadi

, et al. (2018) Hotspots within a global biodiversity hotspot - areas of endemism are associated with high mountain ranges. Scientific Reports 8: 10345.

95.

Novak

Horváthová

(2019) Mountain Pastures Influenced by Sheep Milk Farming in the Carpathians. Nitra: Slovak University of Agriculture.

96.

Oncescu

(1959) Geologia Republicii Populare Romîne. București: Editura Tehnică.

97.

Panait

Hutchinson

Diaconu

, et al. (2019) Disentangling dust and sand deposition using a peat record in CE Europe (northern Romania): A multiproxy approach. Palaeogeography Palaeoclimatology Palaeoecology 532: 109257.

98.

Pauli

Gottfried

Dullinger

, et al. (2012) Recent plant diversity changes on Europe’s mountain summits. Science 336(6079): 353–355.

99.

Perșoiu

Sipos

Rădoane

, et al. (2022) Tectonic, climatic and autogenic controls on the Late Quaternary evolution of the Someș fluvial fan, North-East Pannonian Basin, Central Europe. Geomorphology 417: 108450.

100.

Peters

Friedmann

Stojakowits

, et al. (2020) Holocene vegetation history and environmental change in the Lăpuş Mountains, north-west Romania. Palynology 44(3): 441–452.

101.

Peters

Higuera

(2007) Quantifying the source area of macroscopic charcoal with a particle dispersal model. Quaternary Research 67(2): 304–310.

102.

Pini

Ravazzi

Raiteri

, et al. (2017) From pristine forests to high-altitude pastures: An ecological approach to prehistoric human impact on vegetation and landscapes in the western Italian alps. Journal of Ecology 105(6): 1580–1597.

103.

Plunkett

Swindles

(2008) Determining the Sun’s influence on lateglacial and holocene climates: A focus on climate response to centennial-scale solar forcing at 2800cal.BP. Quaternary Science Reviews 27(1–2): 175–184.

104.

PNPNMM (2016) Management Plan of Maramureș Mountains Natural Park (Planul de Management al Parcului Natural Munții Maramuresului). Anexele nr. 1 si 2 la Ordinul ministrului mediului, apelor si padurilor nr. 1157 din 16 august 2016. https://www.muntiimaramuresului.ro/index.php/ro/administratie/plan-de-management (accessed 15 February 2022).

105.

PNPNMR (2019) Management Plan of Rodna Mountains National Park (Planul de Management al Parcului Național Munții Rodnei). Anexa nr. 1 la Ordinul viceprim-ministrului, ministrului mediului nr. 307 din 01 Aprilie 2019. http://www.parcrodna.ro/fisiere/pagini_fisiere/81/Planul%20de%20management%20PNMR,%20ROSCI0125%20si%20ROSPA0085%20Muntii%20Rodnei.pdf (accessed 15 February 2022).

106.

Rădoane

Chiriloaei

Sava

, et al. (2019) Holocene fluvial history of Romanian Carpathian rivers. Quaternary International 527: 113–129.

107.

Reimer

Austin

WEN

Bard

, et al. (2020) The IntCal20 Northern Hemisphere radiocarbon age calibration curve (0–55 cal kBP). Radiocarbon 62(4): 725–757.

108.

Rothacker

Dosseto

Francke

, et al. (2018) Impact of climate change and human activity on soil landscapes over the past 12,300 years. Scientific Reports 8(1): 247.

109.

Schumacher

Schier

Schütt

(2016) Mid-holocene vegetation development and herding-related interferences in the Carpathian region. Quaternary International 415: 253–267.

110.

Sitko

Troll

(2008) Timberline changes in relation to summer farming in the Western Chornohora (Ukrainian Carpathians). Mountain Research and Development 28(3/4): 263–271.

111.

Smith

AJE

Smith

(2004) The Moss Flora of Britain and Ireland. Cambridge: Cambridge University Press.

112.

Stein

Gerstner

Kreft

(2014) Environmental heterogeneity as a universal driver of species richness across taxa, biomes and spatial scales. Ecology Letters 17: 866–880.

113.

Tanţău

Feurdean

de Beaulieu

, et al. (2011) Holocene vegetation history in the upper forest belt of the Eastern Romanian Carpathians. Palaeogeography Palaeoclimatology Palaeoecology 309: 281–290.

114.

Tanţău

Feurdean

De Beaulieu

, et al. (2014a) Vegetation sensitivity to climate changes and human impact in the Harghita Mountains (Eastern Romanian Carpathians) over the past 15 000 years. Journal of Quaternary Science 29(2): 141–152.

115.

Tanţău

Geantă

Feurdean

, et al. (2014b) Pollen analysis from a high altitude site in Rodna Mountains (Romania). Carpathian Journal of Earth and Environmental Sciences 9(2): 23–30.

116.

Tapody

Sümegi

Molnár

, et al. (2021) Sedimentological-geochemical data based reconstruction of climate changes and human impacts from the peat sequence of Round Lake in the western Foothill area of the Eastern Carpathians, Romania. Quaternary 4(2): 18.

117.

Tasser

Mader

Tappeiner

(2003) Effects of land use in alpine grasslands on the probability of landslides. Basic and Applied Ecology 4(3): 271–280.

118.

Ter Braak

Smilauer

(2002) CANOCO reference manual and CanoDraw for Windows user’s guide: software for canonical community ordination (version 4.5). https://library.wur.nl/WebQuery/wurpubs/fulltext/405659 (accessed 21 February 2022).

119.

Thompson

Oldfield

Thompson

, et al. (1986) Mineral magnetic studies of lake sediments. Environmental Magnetism. Springer London, pp.101–123.

120.

Tobolski

(2000) Przewodnik Do Oznaczania Torfów i Osadów Jeziornych (A Guide to Identification of Peat and Lake Sediments). Warszawa: PWN, p.508 (in Polish).

121.

Vandewalle

Purschke

de Bello

, et al. (2014) Functional responses of plant communities to management, landscape and historical factors in semi-natural grasslands. Journal of Vegetation Science 25(3): 750–759.

122.

van Geel

Buurman

Brinkkemper

, et al. (2003) Environmental reconstruction of a Roman period settlement site in Uitgeest (The Netherlands), with special reference to coprophilous fungi. Journal of Archaeological Science 30(7): 873–883.

123.

Velichkevich

Zastawniak

(2006) Atlas of Pleistocene Vascular Plant Macroremains of Central and Eastern Europe, Part I–Pteridophytes and Monocotyledons. Cracow: W. Szafer Institute of Botany, Polish Academy of Sciences, pp.1–224.

124.

Vincze

Orbán

Birks

, et al. (2017) Holocene treeline and timberline changes in the South Carpathians (Romania): Climatic and anthropogenic drivers on the southern slopes of the Retezat Mountains. Holocene 27(11): 1613–1630.

125.

Vinther

Buchardt

Clausen

, et al. (2009) Holocene thinning of the Greenland ice sheet. Nature 461(7262): 385–388.

126.

Weng

Hooghiemstra

Duivenvoorden

(2006) Challenges in estimating past plant diversity from fossil pollen data: statistical assessment, problems, and possible solutions. Diversity and Distributions 12(3): 310–318.

127.

Werners

, et al. (2016) Climate change adaptation in the Carpathian mountain region. In: Salzmann

Huggel

Nussbaumer

(eds) Climate Change Adaptation Strategies – An Upstream-Downstream Perspective. Cham: Springer, pp.79–99.

128.

Whitlock

Larsen

(2001) Charcoal as a fire proxy. In: Smol

Birks

HJB

Last

(eds) Tracking Environmental Change Using Lake Sediments, Volume 3: Terrestrial, Algal, and Siliceous Indicators. Springer Dordrecht, pp.75–97.

129.

Zuur

Ieno

Elphick

(2010) A protocol for data exploration to avoid common statistical problems. Methods in Ecology and Evolution 1(1): 3–14.