Lateglacial to Holocene vegetation development in the Central Rila Mountains,Bulgaria

Abstract

The vegetation history of the Central Rila Mountains for the last 14,000 years was reconstructed by means of pollen analysis and radiocarbon chronology of a core retrieved from Lake Manastirsko-2 (2326 m). The Lateglacial landscape was dominated by open herb vegetation composed of Artemisia, Chenopodiaceae, Poaceae, and other cold-resistant herbs (14,000–11,700 cal. yr BP). Stands of Pinus, Betula, and Juniperus–Ephedra shrubland partly enlarged during the Lateglacial interstadial. Pioneer forests of Betula with Pinus and Juniperus occupied barren soils in the early Holocene (11,700–7900 cal. yr BP), while mixed oak forests with Tilia, Ulmus, Acer, and later on Corylus spread at lower elevations. A coniferous forest belt with Pinus sylvestris, Pinus peuce, and Abies developed after 7900 cal. yr BP in the conditions of milder winters, cooler summers, and increase in precipitation. The late Holocene dynamic vegetation changes were associated with the invasion of Picea abies after 3400 cal. yr BP, while Fagus communities slightly enlarged in the river valleys. Indications of human activities are visible in the pollen diagram since the ‘Late Bronze Age’ (3400–3200 cal. yr BP). The vegetation development in the study area followed a similar pattern when compared with palynological and macrofossil records from other parts of the Rila and the adjacent Northern Pirin Mountains. On a larger geographical scale, the postglacial vegetation history of the Rila Mountains displays common features with sites in the Romanian Carpathians, whereas the differences observed are result of the location of tree refugia, competing abilities, climate changes, and human activities.

Keywords

Holocene Lateglacial pollen analysis radiocarbon chronology Rila Mountains vegetation history

Introduction

Palaeoecological investigations on the Balkan peninsula in relation to the postglacial European vegetation and flora history have always been of great interest to scientists because this region functioned as a glacial refuge area for many plants, particularly for trees, and as one of the starting places for the reforestation of the continent (Lang, 2003). In this aspect, abundant information has been collected during the last decades from different parts of the peninsula in terms of climate changes and vegetation responses, location of plant refugia and migration patterns, tree-line fluctuations, and human impact on the natural environment (Bennett et al., 1991; Beug, 1982; Bozilova et al., 1996; Feurdean et al., 2014; Jalut et al., 2005; Müller et al., 2011; Panagiotopoulos et al., 2013; Sadori et al., 2015; Tzedakis, 2004, 2009; Tzedakis et al., 2002, 2006, 2013; Willis, 1994, etc.). Important premise for such investigations is the availability of continuous long pollen and plant macrofossil records, supplemented by reliable chronologies, which offer perspective opportunities for exploring and understanding the mechanisms behind the postglacial vegetation history in this part of Southeastern Europe (Tonkov et al., 2013).

Within all research areas on the Balkans, two of its highest midland massifs Rila (2925 m) and Pirin (2917 m) claim a particular phytogeographic and palaeoecological interest. The studies conducted in these mountains outlined the basic chronologically defined postglacial vegetation stages with their specific features (Bozilova, 1981; Bozilova et al., 1990; Stefanova and Ammann, 2003; Stefanova and Bozilova, 1995; Stefanova et al., 2006a, 2006b; Tonkov et al., 2002b, 2008, 2013, etc.).

In the last years, most of the palaeoecological investigations in the Rila Mountains were focused on its northwestern part, in the cirque of the Seven Rila Lakes. Long sediment cores from a system of three glacial lakes (Trilistnika, Ribno, and Sedmo Rilsko), located on tiers in the subalpine belt between 2216 and 2095 m, were analyzed for pollen, diatom, and plant macrofossil (stomata) content in an attempt to elucidate the vegetation and climate changes, location of possible tree refugia, tree-line fluctuations, and origin and evolution of the lakes since the beginning of the Lateglacial (Bozilova and Tonkov, 2000; Hofmann and Lotter, 2003; Tonkov et al., 2008, 2013; Velčev et al., 2011). A detailed reconstruction of the dynamic Lateglacial vegetation and climate shifts (16,000–11,700 cal. yr BP), bound to consistent radiocarbon chronologies and comparable to the palaeoclimate signal from Greenland ice-cores (Björck et al., 1998, INTIMATE event stratigraphy), was also presented (Bozilova and Tonkov, 2011; Tonkov et al., 2006, 2011). Of particular interest is the dominance of Lateglacial herb communities composed of Artemisia, Chenopodiaceae, Poaceae, and other cold-resistant herbs and their partial replacement by stands of Pinus, Betula, and shrubland of Juniperus–Ephedra during the Lateglacial interstadial. Regarding the Holocene, the successive afforestation phases triggered by climate shifts, tree establishment, expansion, and decline, and influenced by human activities as well, were chronologically delimited, although local-to-regional correlations remain still indispensable.

In this paper are presented the results from pollen analysis of a sediment core which dates back to 14,000 years from a subalpine lake in the Central Rila Mountains, an area that still remains insufficiently investigated. The current palaeoecological evidence (pollen and plant macrofossils) from this part of the mountains originates from two small neighboring lakes, and their age was determined at c. 6000 cal. yr BP (Tonkov and Marinova, 2005). The new results on the vegetation dynamics span the largest part of the Lateglacial and the entire Holocene, and are compared with palaeoecological data (pollen, plant macrofossils, radiocarbon ages) from other sites investigated in the Rila and the Northern Pirin Mountains (Figure 1). It also appeared reasonable to compare in broad lines the general trends of the postglacial vegetation development and some of its specific features with similar studies from high altitudinal sites in the Romanian Carpathians, focusing predominantly on the dynamic changes recorded during the Lateglacial and the early Holocene (Feurdean et al., 2007, 2011, 2014; Magyari et al., 2012, etc.).

Figure 1.

(a) Location of the Rila Mountains (closed triangle) on the Balkan peninsula. (b) Site of investigation (closed rectangle) – Lake Manastirsko-2. Location of sites (•) in the Rila and Northern Pirin Mountains mentioned in text: II – Northwestern Rila Mountains. 1. Cirque of the Seven Rila Lakes (Lake Sedmo Rilsko (Bozilova and Tonkov, 2000, 2011); Lake Ribno (Tonkov et al., 2011, 2013); Lake Trilistnika (Tonkov et al., 2006, 2008)); 2. Lake Panichishte (Bozilova et al., 2002); III – Central Rila Mountains; 3. Lakes Ostrezki (Tonkov and Marinova, 2005); IV – Southwestern Rila Mountains; 4. Lake Suho Ezero (Bozilova, 1995; Bozilova et al., 1990); Northern Pirin Mountains; 5. Lake Ribno Banderishko (Tonkov et al., 2002b); 6. Peat-bog Mozgovitsa (Marinova and Tonkov, 2012); 7. Peat-bog Praso (Stefanova and Oeggl, 1993); 8. Lake Bezbog; and 9. Lake Kremensko-5 (Stefanova et al., 2006a, 2006b).

Study area

Rila Mountains are the highest massif on the Balkans (Musala peak, 2925 m) located between 41°52′30″–42°21′40″N and 23°01′22″–24°01′E in Southwestern Bulgaria. The massif has a roughly triangular shape (70 km × 60 km × 40 km) with an area of 2396 km². The mountains are bordered on the south and southeast by the Pirin (2914 m) and the Western Rhodopes (2191 m) mountains and on the west and north by the valleys of the rivers Struma, Iskar, and Maritza. Deep river valleys follow the major fault lines across Rila and divide the massif into four parts: I – Eastern (2925 m), II – Northwestern (2731 m), III – Central (2716 m), and IV – Southwestern (2630 m) (Glovnia, 1968; Figure 1).

Rila massif is built by crystalline rocks of Mesozoic and Pre-Mesozoic age. The granitoids form the central and eastern mountain parts, while the Northwestern Rila and the western fringes of Southwestern Rila are built up mainly of metamorphic rocks, gneisses, schists, and amphibolites. A number of authors consider that Rila Mountains were glaciated at least twice in the Pleistocene during the Riss and the Würm glacial stages (Glovnia, 1968; Velčev, 1995, etc.) following the Alpine geochronological scheme of the Quaternary glaciations, and others (Tonkov et al., 2008; Velčev, 1995, 1999) claimed to have found evidence from an earlier glaciation (Mindel). All authors stick to the opinion that the existing visible glacial landforms (cirques, numerous lakes, different types of moraines, and trough valleys) were shaped during the last glaciation.

Recently, with the application of cosmogenic nuclide dating (¹⁰Be) on terminal moraine samples, the extent of the last glaciation in the Rila Mountains was mapped. Most probably, the maximum extent has occurred in two phases: the first one around the beginning (24–23 kyr BP) and the second one around the end of the Last Glacial Maximum (18–16 kyr BP), separated by a retreat phase (Kuhlemann et al., 2013).

The typical mountain climate in Rila above 1000 m is characterized by the decrease in the annual air temperature by 0.5°C with each 100-m increase in altitude. At 1800–1900 m where the present timber-line runs, the mean January temperature is −6°C and the mean August temperature is 11.4°C. As a result of the predominant directions of the moisture transport, the northwestern and northern slopes at altitudes 1000–1700 m receive 1000–1050 mm yr⁻¹, while the high peaks above 2500 m receive about 900 mm yr⁻¹, most of it snow (Velev, 2002).

The modern vegetation and its altitudinal distribution in six vegetation belts in the Rila Mountains have been presented in several publications (Bozilova and Tonkov, 2000; Tonkov et al., 2008, 2011). The study area of the Central Rila Mountains comprises the following vegetation belts: beech forests, coniferous forests, and subalpine and alpine plant communities. The deciduous forests are primarily found along the valley of the Rilska river up to 1400–1500 m. The plant communities of Quercus dalechampii Ten. and Carpinus betulus L. are replaced in the vicinity of the Rila Monastery (1200–1400 m) by single-dominant forests of Fagus sylvatica L. and mixed forests of F. sylvatica with Abies alba Mill. In these forests as an admixture are found Acer campestre L., Acer pseudoplatanus L., Acer heldreichii Orph. Ex Boiss., C. betulus, and Fraxinus excelsior L. The coniferous belt above 1600 m is compact, well developed, composed by the communities of Picea abies (L.) Karst, Pinus sylvestris L., and partly A. alba. In some places are found secondary communities of Betula pendula Roth. and Populus tremula L. In the upper part of the coniferous belt, the participation of the Balkan endemic Pinus peuce Griesb. becomes rather considerable. This species shapes the upper tree-line on many places (1950–2050 m) together with P. abies. The subalpine belt above 2000 m is dominated by the communities of Pinus mugo Turra, together with Juniperus sibirica Burgad., Vaccinium vitis-idaea L., and Vaccinium myrtillus L. On open areas, nearby springs and touristic paths, are found herb communities of Verbascum longifolium Ten, Rumex alpinus L., and Cirsium appendiculatum Griesb. The typical alpine belt shelters a number of herb species such as Nardus stricta L., Sesleria comosa Vel., Carex curvula All., the glacial relics Saxifraga retusa Gouan, Ranunculus crenatus W. et K., the Balkan endemics Dianthus microlepis Boiss., Lilium jankae A. Kern., and Festuca riloensis (Hack et Hay) Margr.-Dann. Traces of human activities are present in all vegetation belts, and at many places the tree-line was artificially lowered in former times.

The study site is Lake Manastirsko-2 situated at 2326 m (42°07′46.78″ N; 23°24′42.16″ E) in the Central Rila Mountains. The lake is 75 m long and 50 m wide, with a surface area of 0.2 ha and a maximum depth of 2 m. The shores of the lake are flat and marshy, surrounded by sparse groups of P. mugo and J. sibirica among herb vegetation. The inlet brings water from a small upper lake and the outlet drains into Vodniza river (Ivanov, 1964) (Figure 2).

Figure 2.

A view of Lake Manastirsko-2 with the coring place (•) (photo by S Tonkov).

Methods

Core and lithology

A sediment core 415 cm long and 2 cm in diameter was obtained from the peaty southeastern part of the lake with a hand-coring equipment (Figure 2). The lithology of the sediments is the following: 0–120 cm peat with sand, 120–385 cm peaty gyttja, and 385–415 cm light-green gyttja.

Pollen analysis

Pollen analysis was conducted at 5 cm (2.5 cm for the Lateglacial section) interval after the application of the standard acetolysis procedure (Faegri and Iversen, 1989). The pollen sum (PS) used for percentage calculations is based on arboreal pollen (AP) + non-arboreal pollen (NAP). In most instances, a PS of c. 500 grains was reached. Spores of mosses, pteridophytes, pollen of aquatics, and Cyperaceae are not included in the PS. Their presence was expressed as percentages of the PS. The determination of spores and pollen was made using the reference collection of the Laboratory of Palynology and the pollen keys in Faegri and Iversen (1989), Moore et al. (1991), and Beug (2004). A percentage pollen diagram was constructed with the program TGView ver. 1.7.16 (Grimm, 2011) (Figure 3). Pollen accumulation rates (PAR; pollen grains cm⁻² yr⁻¹) for the major tree/shrub and herb taxa were also calculated (Figure 4) by addition of three Lycopodium tablets (13,911/11,300 spores each) per sample of 2 cm³ before preparation (Stockmarr, 1971). Unlike percentage data, the PAR value of a given taxon is directly dependent on the abundance of this taxon in the surroundings of the lake and independent of values of other taxa (Davis and Deevey, 1964; Davis et al., 1973; Seppä et al., 2009, etc.). Fossil stomata were occasionally found on pollen slides from several sample depths (75, 185, 245, 265, 315, 325, and 365 cm) and their presence (*) is also shown on the pollen diagram (Figure 3). All stomata were identified as Pinus sp. following Sweeney (2004). For conifers, the stomata found on pollen slides are derived from needles and thus provide a valuable proxy for local presence (Ammann et al., 2014).

Figure 3.

Percentage pollen diagram from Lake Manastirsko-2: (a) pollen curves of selected trees and shrubs; (b) pollen curves of selected herbs, pteridophytes, and aquatics.

Figure 4.

Pollen accumulation rates (PAR; pollen grains cm⁻² yr⁻¹) for the main tree/shrub and herb taxa from Lake Manastirsko-2 plotted against age (cal. yr BP).

Radiocarbon dating and local chronology

AMS ¹⁴C dating on 12 bulk sediment samples was performed at the Angstrom Laboratory, Division of Ion Physics, ¹⁴C-Lab, Uppsala University. The dates have been calibrated to calendar years (±2σ range) with the computer program OxCal v3.10 (Bronk Ramsey, 2005) using the relevant atmospheric data (Reimer et al., 2013). The results are shown in Table 1. In the text, all dates are cited as cal. yr BP. An age/depth sedimentation plot with a linear fit was also constructed (Figure 5).

Table 1.

Results of radiocarbon measurements of Lake Manastirsko-2.

Lab code (UA)	Sample depth (cm)	¹⁴C age (BP)	¹⁴C age (cal. BP, 2σ)	Material dated
UA-47271	26–28	1245 ± 75	980–1300 (1140)	Peat
UA-49249	36–38	1517 ± 31	1330–1530 (1430)	Peat
UA-47272	80–82	1915 ± 38	1730–1950 (1840)	Peat
UA-48421	130–132	2703 ± 41	2750–2880 (2815)	Peaty gyttja
UA-47273	180–182	3380 ± 42	3480–3730 (3605)	Peaty gyttja
UA-48422	240–242	5858 ± 47	6540–6790 (6665)	Peaty gyttja
UA-472714	280–282	7680 ± 93	8310–8660 (8485)	Peaty gyttja
UA-47275	320–322	8972 ± 106	9650–10,400 (10,025)	Peaty gyttja
UA-47423	360–362	9732 ± 75	11,060–11,270 (11,165)	Peaty gyttja
UA-49250	388–390	10,380 ± 85	11,950–12,550 (12,250)	Gyttja
UA-47276	405–407	11,209 ± 168	12,700–13,400 (13,050)	Gyttja
UA-49251	413–415	11,796 ± 174	13,250–14,050 (13,650)	Gyttja

Figure 5.

Age/depth sedimentation plot for the core of Lake Manastirsko-2. The red (upper) and blue (lower) lines represent the older and younger calibrated ages at the 2σ significance levels, respectively.

Results

Pollen data

On the pollen diagram, five local pollen assemblage zones (LPAZ) were recognized (LM-1 to LM-5) (Figure 3). They reflect successive changes in vegetation development in the study area. Short descriptions are as follows.

LPAZ LM-1: 415–378 cm (Artemisia–Chenopodiaceae–Pinus–Poaceae) (13,700–11,700 cal. yr BP)

The quantity of AP varies between 20% and 45% with two short peaks of 55% at levels of 415 and 405 cm. The main component is Pinus diploxylon-type (10–30%), accompanied by P. peuce (up to 10%), Betula (5–10%,) Juniperus, Quercus, Tilia, Ulmus, Corylus, Alnus (up to 5% each), and some Ephedra (distachya- and fragilis-type). A short-term rise of Abies pollen (5%) is also recorded. Herb pollen is represented by Artemisia (15–40%), Chenopodiaceae (up to 18%), Poaceae (up to 15%), Achillea-type (up to 5%), Cichoriaceae (5–8%), Galium-type, Rumex, Cirsium-type, Apiaceae, Brassicaceae, Dianthus-type, and so on.

LPAZ LM-2: 378–268 cm (Betula–Quercetum mixtum–Corylus–Pinus) (11,700–7900 cal. yr BP)

This zone can be divided into subzone LM-2a (378–312 cm, Quercetum mixtum–Betula–Pinus) and LM-2b (312–268 cm, Corylus–Betula–Quercetum mixtum–Pinus).

In subzone LM-2a (11,700–9800 cal. yr BP), AP rises from 40% to 75% due to P. diploxylon-type (c. 20%), P. peuce (5%), Corylus (up to 10%), Betula (5–8%), Quercus robur-type (up to 10%), Tilia (5–7%), Quercus cerris-type (5%), Alnus, Ulmus, and Tilia. Low pollen frequencies of Acer, F. excelsior, C. betulus, and Abies are found. The NAP is represented by Poaceae (10–15%), Artemisia (up to 8%), Cichoriaceae (5–10%), Chenopodiaceae (below 5%), Rumex (5–7%), Ranunculus-type, Brassicaceae, Apiaceae, Galium-type, each below 3–5%, and so on. Spores of Polypodiaceae show a maximum of 5%.

In the next subzone LM-2b (9800–7900 cal. yr BP), Corylus dominates with 15–25%, together with Betula (up to 20%), Q. robur-type (up to 15%), and Q. cerris-type (5%). The continuous pollen curve of Abies appears, reaching 7%. Pollen grains of Fagus and Picea are regularly determined. Pollen of P. diploxylon-type fluctuates around 20% reaching short-term maxima of 30–40%. The pollen curve of Poaceae declines to 5–8%, of Artemisia is less than 5%.

LPAZ LM-3: 268–225 cm (Pinus–Abies) (7900–5860 cal. yr BP)

The most characteristic features in this zone are the rises of the pollen curves of Abies, P. diploxylon-type, and P. peuce, reaching 25%, 50–60%, and 15%, respectively. Pollen of deciduous trees (Betula, Q. robur-type, Q. cerris-type, Corylus, Tilia, and Ulmus) quickly declines. In this zone begin the continuous pollen curves of Fagus and Picea, although with low frequencies. Changes in the presence of NAP are not recorded except for a decline of Poaceae below 5%.

LPAZ LM-4: 225–102 cm (Pinus–Abies–Picea-Fagus) (5860–2250 cal. yr BP)

Two subzones LM-4a (225–168 cm, Abies–Pinus-Fagus) and LM-4b (168–102 cm, Pinus–Picea-Fagus) are delimited. In subzone LM-4a (5860–3420 cal. yr BP), AP dominates with 80–90%, attributed to P. diploxylon-type (40–60%), maximal frequencies for Abies (25–30%), P. peuce (10–20%), Fagus, and Picea up to 3–5% each. Continuous pollen curves for Betula, Q. robur-type, and Alnus are recorded. The most frequent herb pollen taxa are Poaceae (5%), Artemisia, Cichoriaceae, Rumex, Cirsium-type, Ranunculus-type, Brassicaceae, and so on. Pollen of Scleranthus reappears in this subzone.

In subzone LM-4b (3420–2250 cal. yr BP), the pollen curve of Picea continues to rise up to 7–10%, while Abies declines to c. 10–15%. Still high values are established for P. diploxylon-type (30–40%) and P. peuce (10–15%). A slight increase in pollen of Betula, Corylus, Alnus, and Fagus is recorded. A higher diversity of NAP is present; particularly to mention is the appearance of anthropogenic indicators such as Secale, Hordeum-type, and Plantago lanceolata, accompanied by an increase in Poaceae and Cichoriaceae pollen.

LPAZ LM-5: 102–0 cm (Pinus–Picea–Abies) (2250 cal. yr BP till present)

In the uppermost pollen samples, P. diploxylon-type reaches 60–65%, accompanied by P. peuce (10–15%), Abies (10–15%), and Picea (10–15%). Minor values below 5% are recorded for Fagus, Betula, Alnus, Corylus, and Q. robur-type. Pollen of Juglans is occasionally found. A secondary increase in the pollen curve of Cichoriaceae (5–8%) is established together with Poaceae, Achillea-type, Artemisia, Scleranthus, Brassicaceae, and P. lanceolata.

Discussion

Lateglacial vegetation dynamics

The pollen record from Lake Manastirsko-2, bound to a consistent radiocarbon chronology, reveals the palaeoenvironmental changes and the vegetation dynamics in the Central Rila Mountains for the last 14,000 years. A detailed picture of the Lateglacial vegetation and climate changes in the northwestern part of the mountains has already been presented by evaluating the palaeoecological evidence from the cirque of the Seven Rila Lakes (Tonkov et al., 2006, 2008, 2011, 2013) so that a reliable basis for comparison already existed (Figure 6).

Figure 6.

Local-to-regional comparison of the palynostratigraphy and chronology for sites in the Rila and the Northern Pirin Mountains.

The Lateglacial interval spans the lowermost 35 cm of the sequence (zone LM-1, Figure 3). However, due to the low resolution because of the high values of deposition time (71–45 yr/cm) (Figure 5), a clear distinction between the upper part of the Bölling/Alleröd interstadial complex and the Younger Dryas stadial is hardly possible. In general, the pollen stratigraphy indicates typical glacial conditions in the study area where open herb vegetation dominated by Artemisia–Chenopodiaceae–Poaceae and other cold-resistant species from Achillea, Thalictrum, Centaurea, Rumex, Galium, Apiaceae, and so on was widely distributed. Group/trees such as Pinus, Betula, and shrubland of Juniperus–Ephedra were also found. The minor quantities of pollen of deciduous trees such as Alnus, Quercus, Corylus, Ulmus, and Tilia; the short-term rise of Abies; and a peak of the AP curve at c. 12,700 cal. yr BP could be interpreted as a response to a signal of climate improvement, which correlates with the GRIP δ¹⁸O curve indicating a warmer interval (GI-1a) (Dansgaard et al., 1993). At lower altitudes, deciduous trees gradually started to spread from their local microrefugia where air and soil moisture were sufficient for their growth. The PAR of trees (shrubs) and herbs appeared rather low, c. 500 grains cm⁻² yr⁻¹ for each of both groups throughout the Lateglacial interval. AP influxes are contributed mainly by P. diploxylon-type (150–250 grains cm⁻² yr⁻¹) which confirms the scarce, most likely isolated presence of pines (P. mugo/sylvestris). The abundance of P. peuce and Betula with c. 50 grains cm⁻² yr⁻¹ each can be considered as disparagingly low. Regarding the PAR of the main herbs, a steep decline in their values is observed after c. 12,700 cal. yr BP for both Artemisia and Chenopodiaceae from 400–200 to 70 and from 200 to 50 grains cm⁻² yr⁻¹, respectively. Meanwhile, the pollen influxes for Poaceae gradually increase starting from 100 to c. 200 grains cm⁻² yr⁻¹ at the transition to the Holocene (Figure 4).

The comparison with the Lateglacial radiocarbon-dated lacustrine records from the area of the Seven Rila Lakes is possible in broad lines for the time window 14,000–11,700/11,600 cal. yr BP. The advantage of these records because of their higher resolution is that they could be compared also with the palaeoclimate isotopic signal from Greenland ice-cores (see Figure 6 in Tonkov et al., 2011). A tendency of climate improvement after c. 16,000 cal. yr BP was established for the Rila Mountains where the dominant cold-resistant herb vegetation composed of Artemisia–Chenopodiaceae–Poaceae species began gradually to retreat. The Bölling/Alleröd interstadial was characterized by the spread of Pinus–Betula stands at high-mid altitudes, while deciduous trees (Quercus, Corylus, C. betulus, etc.) started to emerge from their local microrefugia. The chronological boundaries of the Lateglacial interstadial are clearly defined between 14,800 and 12,800 cal. yr BP in the sediments of Lake Sedmo Rilsko (2095 m) (Bozilova and Tonkov, 2011) and 15,000–12,900 cal. yr BP in the core from Lake Ribno (2184 m) (Tonkov et al., 2011). Even more, the behavior of the total AP/NAP curve from Lake Ribno correlates with the GRIP δ¹⁸O curve and indicates also the colder (GI-1b, GI-1d) and warmer (GI-1a, GI-1e) intervals with shorter duration of the Greenland Interstadial 1. Typical glacial conditions returned during the harshest climatic reversal, the Younger Dryas stadial (12,900–11,600 cal. yr BP), with a re-advance of the herb vegetation and decline of trees which moved down the slopes. The transition to the Holocene at 11,700/11,600 cal. yr BP was marked by a distinct increase in AP, mainly due to deciduous trees and a decline in Artemisia and Chenopodiaceae.

Plant macrofossils of Lateglacial age were found at lower altitude in the sediments of Lake Suho Ezero (1900 m) from the southwestern part of the Rila Mountains. They are represented by a scarce record of several seeds/fruits from herbs (Silene, Carex, Juncus) and a P. peuce needle. Its stratigraphic position was assigned to the Younger Dryas stadial in conformity with the radiocarbon date 10,575 ± 220 ¹⁴C yr BP (12,345 cal. yr BP) (Bozilova, 1995; Bozilova et al., 1990).

The pattern of the Lateglacial vegetation history for the Northern Pirin Mountains appears coherent with the data from the Rila Mountains. One difference recorded is the lower share of Betula together with Pinus in the Pirin Mountains at the onset of the Bölling/Alleröd interstadial complex. The presence of B. pendula (seeds), P. peuce (needles), and J. sibirica (needles) was confirmed for the Younger Dryas stadial (Stefanova et al., 2006b).

A comparison with sites from the northern Balkans, the Romanian Carpathians, revealed that during the Lateglacial this area served as a refuge for Pinus, Betula, Picea, Alnus, Larix, and Juniperus. In the western Southern Carpathians (Retezat Mountains), a sharp increase in Pinus and some Betula marked the transition to the Lateglacial interstadial (Feurdean et al., 2007, 2014). This rapid and widespread reaction of Pinus and Betula is comparable and synchronous in time with the Rila Mountains. The beginning of the Younger Dryas stadial (12,900–12,600 cal. yr BP) witnessed a general reduction of woodland, and pollen percentages of Artemisia, Poaceae, Chenopodiaceae, Pinus, and Juniperus increased at all studied sites (Feurdean et al., 2007). Recent investigations of Lateglacial and early Holocene sediments from two glacial lakes in the same area by combining pollen, stomata, plant macrofossil analyses, and a detailed AMS radiocarbon chronology allowed the reconstruction of the position of the upper tree-line and the response of the vegetation to the climate oscillations (Magyari et al., 2012). Since 13,600 cal. yr BP alongside the ongoing increase in pollen and PAR for Pinus and Betula, thermophilous deciduous taxa (mainly Ulmus, Fraxinus, and Quercus) spread at lower elevations. A characteristic feature according to the palynological, stomata, and macrofossil records was the rise of P. abies around Lake Brazi (1740 m) between 14,300 and 13,800 cal. yr BP. The higher elevation site Lake Gales (1990 m) recorded an increase in tree pollen, but the lack of terrestrial plant macrofossils suggests that the vegetation cover must have been very sparse at around this altitude, similarly with the situation at Lake Manastirsko-2.

Holocene vegetation dynamics

The new pollen record confirms and enriches the palaeoecological information for the main phases in the Holocene vegetation development already established from other parts of the mountains, that is, from the area of the Seven Rila Lakes (Bozilova and Tonkov, 2000; Tonkov et al., 2008, 2013) in the Northwestern Rila Mountains and from Lake Suho Ezero in the Southwestern Rila Mountains (Bozilova, 1995; Bozilova and Smit, 1979; Bozilova et al., 1990). The chronological boundaries of these phases appear in good conformity with the previous results reflecting also the time of the arrival, the establishment, the expansion rates, and decline of the main deciduous and coniferous tree species.

The amelioration of the climate at the onset of the Holocene has resulted in quick afforestation and diminishing of the areas occupied by the mountain herb vegetation dominated by Artemisia, Chenopodiaceae, and other cold-resistant herbs. Initially, during the time interval 11,700–9800 cal. yr BP (subzone LM-2a, Figure 3), forests of Betula with groups of Pinus (P. mugo/sylvestris, P. peuce) and some Juniperus began to colonize areas at mid-higher altitudes on barren soils. The presence of pines in the near vicinity of the lake was justified by the first finds of fossil stomata between 11,000 and 10,000 cal. yr BP, more likely derived from P. mugo, which grows today around the lake. Meanwhile, the birch forests continued to expand, particularly between 9800 and 7900 cal. yr BP, and the participation of Pinus also gradually increased (subzone LM-2b, Figure 3).

The composition of the pollen assemblages indicated the beginning of widespread of Quercus forests with Tilia, Ulmus, Fraxinus, and Acer below the birch zone. This deciduous forest belt reached its maximal distribution c. 10,000 cal. yr ВР at the time when Corylus started to expand in the study area (subzone LM-2b, Figure 3). Although already present at the beginning of the Holocene, hazel at sites in the mountains of Romania and Bulgaria displayed a different pattern compared to Northwest Europe as its culmination was confined to younger ages c. 9000–8000 cal. yr BP (Giesecke et al., 2011).

The arboreal PAR for the time window 11,700–7900 cal. yr BP increased gradually from c. 500 to nearly 2000 grains cm⁻² yr⁻¹, contributed significantly by P. diploxylon-type (500–1000), Betula (up to 300), Corylus (400–450), and Q. robur-type (up to 200). Such values being several times higher than the Lateglacial ones confirm the afforestation process which was enlarged after 9700 cal. yr BP in the study area. Moreover, the short-term steep decline of the arboreal PAR to c. 500 grains cm⁻² yr⁻¹ at 8230 cal. yr BP (level 275 cm), particularly manifested by P. diploxylon-type and less by Betula, can be considered as a vegetation response to the 8.2-kyr climatic event (Alley and Ágústsdóttir, 2005) at higher altitudes. For the first time, evidence of this abrupt cooling was identified in a pollen diagram from the Rila Mountains (Figure 4).

Meanwhile, Abies also steadily established in the vegetation cover and gradually started to increase prior to its quick enlargement after c. 7900 cal. yr BP. A foregoing short-term peak of Abies pollen, registered at 9300 cal. yr BP, coincides with a similar one already recorded from the northwestern part of the mountains (Tonkov et al., 2008, 2013).

The pollen records from the Northwestern (Bozilova and Tonkov, 2000; Tonkov et al., 2008, 2013) and Southwestern Rila Mountains (Bozilova et al., 1990) revealed the existence of the same initial birch phase which lasted for nearly 4000 years after the beginning of the Holocene. In the area of Lake Ribno, the first fossil stomata of Pinus sp. were also identified at c. 10,000 cal. yr BP, pointing most probably to a local presence for P. mugo (Tonkov et al., 2013). The plant macrofossil record from Lake Suho Ezero showed abundance of fruits/bud-scales originating from B. pendula and few needles of Pinus sp., indicating that the tree-line by that time was placed at c. 1900 m (Bozilova, 1995).

Similarly further south, the palaeoecological record from the peat-bog Mozgovitsa (1800 m) on the Northern Pirin Mountains revealed an abundance of Betula fruits with high pollen values of birch (up to 20%), particularly for the time interval 7800–7300 cal. yr BP (Marinova and Tonkov, 2012). In the subalpine area of the Northern Pirin Mountains, the sediments from the glacial lakes Kremensko-5 (2124 m) and Lake Bezbog (2250 m) contained macrofossils from Betula, Juniperus, and occasional needles of P. peuce between 10,500 and 8000 cal. yr BP. It could be concluded that the early Holocene tree-line in the Northern Pirin Mountains was running higher compared to the Rila Mountains (Stefanova et al., 2006a).

In the Carpathian Mountains, the vegetation response to rapid warming at the onset of the Holocene was immediate. For example, the stomata and macrofossil records from the glacial lakes in the Retezat Mountains suggest that Pinus cembra, P. mugo, Larix decidua, and P. abies spread coincidently and formed mixed open boreal woodland. The local appearance of A. alba between 10,600 and 10,300 cal. yr BP, similar to the situation in the Rila Mountains although a thousand years earlier, suggested that summer mean temperatures were higher than today (Magyari et al., 2012). Pinus (mostly P. sylvestris and P. mugo) was an abundant constituent of the boreal forest that prevailed at the end of Lateglacial and in the early Holocene (12,000 till 10,500 cal. yr BP) and showed great resilience to climate variability and fire during this interval. However, Pinus had poor competitive abilities against more advanced successional tree taxa such as P. abies, Ulmus, Quercus, Tilia, and Fraxinus when the climate conditions became warmer at c. 10,500 cal. yr BP (Feurdean et al., 2011). At low and middle elevations at c. 11,700 cal. yr BP, a large-scale increase in temperate forests dominated by Ulmus, Quercus, Tilia, Acer, F. excelsior, and Corylus started, and this composition was preserved at least until 8000 cal. yr BP representing a larger extension of temperate forest than today (Feurdean et al., 2014).

The forest composition and the altitudinal vegetation zonation started to change after 7900 cal. yr BP. The conifers Abies, Pinus, and P. peuce started to enlarge their areas replacing at many places the birch forests and pushing the mixed oak forests and hazel downslope. This dynamic transformation lasted for c. 2000 years (zone LM-3, Figure 3) and was also supported by the presence of Pinus sp. stomata, most probably originating at this altitude from P. mugo. The PAR of P. diploxylon-type and particularly for P. peuce and Abies doubled which testified for growing population density of the coniferous forests (Figure 4). The presence of C. betulus, Fagus, and P. abies in the transitional zone between the deciduous and coniferous forests remained still restricted. This important change in the composition of the vegetation cover was triggered by a climate shift to cooler summers and warmer winters with increased precipitation in the Northern Mediterranean region during the second half of the Atlantic period (Davis et al., 2003). The 8.2-kyr climatic event (Alley and Ágústsdóttir, 2005) has also probably influenced this change as the various aspects of this shift may have spread over a period of up to 600 years (Rohling and Palike, 2005).

The palynological information from the area of the Seven Rila Lakes presents a synchronous pattern of vegetation changes, characterized by the expansion of Pinus and Abies. Fossil stomata of Pinus sp. were also determined (Tonkov et al., 2008, 2013). Stands of birch continued to be found in the coniferous forests or above the timber-line as shown by fossil fruits of Betula in the sediments of Lake Sedmo Rilsko, radiocarbon dated at 7360 cal. yr BP (Bozilova and Tonkov, 2000).

The plant macrofossil record from Lake Suho Ezero manifested a sharp transition from abundance of fruits/bud-scales of Betula to needles and bud-scales from Pinus sp. and P. peuce. The absence of macrofossils of Abies, compared with 10–15% presence of pollen, suggests that fir has not reached the vicinity of the lake, its communities being predominantly localized in the lower part of the coniferous belt where the ecological conditions were most suitable for their growth (Bozilova, 1995).

Pollen and plant macrofossil data for the Northern Pirin Mountains after c. 7900–7800 cal. yr BP also proved the formation of a coniferous vegetation belt. In the sediments of peat-bog Mozgovitsa needles of P. sylvestris/mugo and P. peuce, seeds of Pinus sp. and P. peuce, as well as wood of Pinus sp. were determined at c. 7300 cal. yr BP (Marinova and Tonkov, 2012). Needles of P. peuce, Pinus sp., and bud-scales of Pinus sp. were radiocarbon dated for the time window 8300–6500 cal. yr BP in the sediments of Lake Ribno Banderishko (2190 m) (Tonkov et al., 2002b). Plant macrofossil evidence from other sites on the Northern Pirin Mountains after 7500 cal. yr. BP comprised needles of P. peuce, bud-scales, and microsporophylls of Pinus sp. in the surroundings of Lake Popovo Ezero-6 (2185 m) (Stefanova and Bozilova, 1995). Needles and seeds of P. peuce, needles of A. alba, and bud-scales and microsporophylls of Pinus sp. were identified at lower elevation in the sediments of Praso peat-bog (1900 m) (Stefanova and Oeggl, 1993). It could be concluded that at c. 8000–7000 cal. BP, the tree-line on the Northern Pirin Mountains has reached its maximum height compared to the present-day situation.

The next stage in the vegetation development between 5860 and 3420 cal. yr BP (subzone LM-4a, Figure 3) was characterized by the wide distribution of Abies, together with Pinus and P. peuce. In the study area, spruce began to gradually penetrate into the coniferous belt, while the presence of beech remained still isolated. The PAR for the main coniferous species started sharply to rise, reaching maximal values at 3400 cal. yr BP (P. diploxylon-type, 3000; P. peuce, 1000; and Abies, 1250 grains cm⁻² yr⁻¹), which indicated denser forest cover and that the timber-line was placed higher than in present times (Figure 4).

The palaeoecological information from Lakes Ostrezki (2320–2340 m) in the Central Rila Mountains has documented a high presence of Abies as one of the major constituents in the coniferous belt with 25–35% pollen, particularly for the time interval 6000–2800 cal. yr BP. The macrofossil finds of Abies comprise only few needles which were evidently air-transported upslope at this altitude. Meanwhile, the distribution of P. sylvestris and P. mugo has not changed, whereas a gradual decline of P. peuce has started after 4500 cal. yr BP (Tonkov and Marinova, 2005).

Indications of human presence are visible in the pollen diagram of Lake Manastirsko-2 after c. 3400 cal. yr BP when pollen grains of Hordeum-type and Secale were determined pointing to agricultural activities in the foothills of the mountain. The increase in both pollen and influx values for Cichoriaceae and Poaceae and the appearance of P. lanceolata suggest an enlargement of the open area around the lake used for seasonal pastures (Figures 3 and 4). The evidence for the first signs of anthropogenic impact derived from the area of the Seven Rila Lakes is comparable with the appearance of anthropogenic indicators such as P. lanceolata and Rumex in the fossil records after 3400–3200 cal. yr BP, coinciding with the starting point of increase in Fagus and P. abies (Bozilova and Tonkov, 2000; Tonkov et al., 2008, 2013).

In the pollen and plant macrofossil diagrams from Lakes Ostrezki, the continuous pollen curves of P. lanceolata, Rumex, Scleranthus, and Urtica, considered as indicators of human disturbance in mountainous areas according to Bozilova and Tonkov (1990), appeared at 3770 cal. yr BP. Part of Poaceae, Artemisia, and Cichoriaceae pollen could be also included in this group. The abundance of charcoal fragments just before 3770 cal. yr BP and c. 2800 cal. yr BP presumes that some of the forest fires were caused by the local people to enlarge the areas for high mountain-pasture land. A decrease in the macrofossil record from conifers (P. sylvestris, P. peuce), alongside the finds of heliophilous herbs such as Potentilla sp., Dianthus sp., Silene sp., and Doronicum austriacum, indicated openings in the vegetation cover and lowering of the tree-line (Tonkov and Marinova, 2005).

The archaeological information from the Struma valley points to an increase in the number of the permanent settlements in the foothills of the surrounding mountains during the ‘Late Bronze Age’ (3400–3200 cal. yr BP) after a period with a shift to increasing pastoralism (Marinova et al., 2012). For instance, in the Middle Struma valley, the existence of about 33 settlements was reported, compared to only 3 for the preceding period (Grebska-Kulowa and Kulow, 2007).

The invasion of P. abies in the coniferous belt has started after 3400 cal. yr BP, reaching the first maximal distribution at c. 2700 cal. yr BP (subzone LM-4b, Figures 3 and 4), in the conditions of a changing climate with lower average temperatures and increase in precipitation (Van Geel et al., 1998). The time period after 2800 cal. yr BP for the Eastern Mediterranean was characterized by an increase in humidity (Bar-Matthews et al., 1999). Fagus also slightly enlarged at lower altitudes in the river valleys.

The expansion of P. abies, alongside the decline of Abies which started after c. 4000 cal. yr BP, was recognized in the majority of the pollen diagrams published from different parts of the Rila Mountains (Bozilova, 1981; Bozilova and Tonkov, 2000; Bozilova et al., 1990, 2002; Tonkov et al., 2008, 2013). Spruce was the latest tree to arrive and its immigration caused a re-distribution in the composition of the coniferous belt, resulting in the retreat of Abies to lower altitudes. This change in the composition of the coniferous belt, convincingly illustrated by the abundance of needles and seeds from P. abies after 2800 cal. yr BP in the sediment record of Lake Suho Ezero, proved that spruce has reached altitudes of c. 1900 m (Bozilova, 1995). In the non-glaciated mountains of Southwestern Bulgaria, Fagus appeared as a competitive tree replacing fir in the majority of habitats, and this phenomenon was accelerated by human interference with the natural forest cover (Bozilova and Tonkov, 1994).

In the Northern Pirin Mountains, P. abies started to spread in the coniferous belt at c. 4400 cal. yr BP, and after 3400 cal. yr BP it invaded new areas (Tonkov et al., 2002b). By that time, plant macrofossils of P. abies (needles, charred wood) already dominated in the fossil record of the peat-bog Mozgovitsa (Marinova and Tonkov, 2012), while the sequence from Lake Popovo Ezero-6 lacked any macroremains of spruce (Stefanova and Bozilova, 1995). This fact presumes that P. abies hardly ever reached altitudes above 2000–2100 m on the Northern Pirin Mountains compared to its good pollen capabilities for dispersal in the subalpine zone as demonstrated by the analysis of modern surface moss samples (Tonkov et al., 2002a).

In the mountains of Romania, P. abies began to expand already from about 10,500 cal. yr BP, at the time when dense forests of boreal and temperate taxa have prevailed, suggesting that spruce was a strong competitor capable of invading dense forest. However, it appeared that fire has also facilitated P. abies establishment by creating openings in these forests. Till present, P. abies has been able to persist abundantly at elevations above 1000 m, proving continuous, strong competition ability. From about 4000 years ago, the populations of spruce below 1000 m were replaced by F. sylvatica when summers became cooler and winters milder. The recent small return of Pinus (mainly P. sylvestris) and P. abies was the consequence of forest management, that is, forest clearance followed by plantations (Feurdean et al., 2011).

The vegetation development during the last two millennia (zone LM-5, Figure 3) was characterized by an intensification of the anthropogenic impact, proved by the increase in pollen for Cichoriaceae, P. lanceolata, and Scleranthus, indicative of livestock-grazing in the mountain meadows. After c. 1000 cal. yr BP at many places, the tree-line was artificially lowered in order to extend the mountain-pasture land. The find of Juglans pollen pointed to its cultivation in the foothills of the mountain.

The last enlargement of P. abies, and partly of Fagus, was reached at c. 1700 cal. yr BP, and afterwards both trees started to decline. Larger areas were occupied by Pinus, mostly P. mugo in the subalpine belt, and P. sylvestris. In general, the total PAR of trees and shrubs substantially decreased after c. 1500 cal. yr BP, particularly noticeable for P. diploxylon-type, P. abies, Abies, and Fagus. In the sediments of Lakes Ostrezki, the macrofossil record after 1240 cal. yr BP was rather poor, comprising only scarce needles of P. mugo, J. sibirica, and other occasional finds, which reflected the characteristics of the present-day vegetation in the subalpine belt influenced by humans (Tonkov and Marinova, 2005).

From Lake Panichishte (1345 m) in the Northwestern Rila Mountains, a replacement of Abies and Fagus at lower altitudes by P. sylvestris and P. abies was recorded after 1220 cal. yr BP, while stands of Betula, Corylus, and Juniperus occupied open areas (Bozilova et al., 2002). Moreover, the last enlargement of spruce in the altitudinal zone 1300–1600 m on north-facing slopes has taken place after 500 cal. yr BP, coinciding with the duration of the ‘Little Ice Age’ (1550–1850 AD). It should be stated that such a peak of Picea pollen curve was not recorded in the pollen diagrams from the cirque of the Seven Rila Lakes (Bozilova and Tonkov, 2000; Tonkov et al., 2008, 2013).

In the Northern Pirin Mountains, P. peuce and P. abies shaped the tree-line after 2200 cal. yr BP. The rise of P. diploxylon-type pollen and the find of needles of P. sylvestris/mugo confirmed the present-day widespread of dwarf-pine in the subalpine belt (Tonkov et al., 2002b).

Conclusion

The continuous pollen record from Lake Manastirsko-2 provides for the first time information on the vegetation history in the Central Rila Mountains for the last 14,000 years. The results are correlated with palynological and plant macrofossil evidence from other parts of the mountains and with sites from the adjacent Northern Pirin Mountains. On a broad scale, some of the basic features in the postglacial vegetation development and tree species distribution are compared with data from sites in the Romanian Carpathians. Several main conclusions from this study can be outlined:

The Lateglacial landscape at high-mid altitudes for the time interval 14,000–11,700 cal. yr BP was dominated by open herb vegetation composed of Artemisia, Chenopodiaceae, Poaceae, and other cold-resistant herbs. The sparse stands of Pinus and Betula and the shrubland of Juniperus–Ephedra partly enlarged during the Lateglacial interstadial.

In the early Holocene (11,700–7900 cal. yr BP), pioneer forests of Betula with groups of Pinus and Junipers spread on barren soils and below them developed mixed deciduous oak forests with abundant Tilia, Ulmus, Acer, and later on Corylus.

The short-term steep decline of the AP influx values at 8230 cal. yr BP, particularly manifested by P. diploxylon-type and Betula, can be explained as a vegetation response to the 8.2-kyr cooling event at higher altitudes. This climatic oscillation is identified for the first time in a pollen diagram from the Rila Mountains.

In the course of two millennia (7900–5800 cal. yr BP), a compact coniferous belt composed of Pinus (P. sylvestris, P. peuce) and Abies developed following a climate shift to milder winters, cooler summers, and increase in precipitation.

The invasion of P. abies in the coniferous belt with several subsequent expansion phases started after 3400 cal. yr BP alongside a slight enlargement of Fagus communities in the river valleys.

The first indications of stock-breeding and other human activities in the high parts of the Central Rila Mountains were recorded since the ‘Late Bronze Age’ (3400–3200 cal. yr BP).

The vegetation development followed a similar pattern as established from pollen and plant macrofossil studies for other parts of the Rila Mountains and the Northern Pirin Mountains with some site-specific features.

On a larger geographical scale, the Lateglacial and Holocene vegetation history of the Rila Mountains displays common features with evidence from high altitudinal sites in the Romanian Carpathians. The differences observed are primarily a result of the location of tree refugia, migration rates, competing abilities, climate changes, and human activities.

Footnotes

Acknowledgements

The field work at Lake Manastirsko-2 was performed with the assistance of colleagues from the Nature Park Rila Monastery, Rila town. Two anonymous reviewers and the editor Professor F Chambers provided critical comments and suggestions to improve the manuscript.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-profit sectors.

References

Alley

Ágústsdóttir

(2005) The 8 k event: Cause and consequences of a major Holocene abrupt climate change. Quaternary Science Reviews 24: 1123–1149.

Ammann

van der Knaap

Lang

. (2014) The potential of stomata analysis in conifers to estimate presence of conifer trees: Examples from the Alps. Vegetation History and Archaeobotany 23: 249–264.

Bar-Matthews

Ayalon

Kaufman

. (1999) The Eastern Mediterranean paleoclimate as a reflection of regional events: Soreq cave, Israel. Earth and Planetary Science Letters 166: 85–95.

Bennett

Tzedakis

Willis

(1991) Quaternary refugia of north European trees. Journal of Biogeography 18: 103–115.

Beug

H-J

(1982) Vegetation history and climatic changes in central and southern Europe. In: Harding

(ed.) Climatic Changes in Later Prehistory. Edinburgh: Edinburgh University Press, pp. 85–102.

Beug

H-J

(2004) Leitfaden der Pollenbestimmung für Mitteleuropa und angrenzende Gebiete. München: Verlag Dr. Friedrich Pfeil.

Björck

Walker

MJC

Cwynar

. (1998) An event stratigraphy for the Last Termination in the North Atlantic region based on the Greenland ice-core record: A proposal by the INTIMATE group. Journal of Quaternary Science 13: 283–292.

Bozilova

(1981) Vegetation changes in the Rila Mountain during the last 12000 years. Annual of University of Sofia, Faculty Biology 71(2): 37–44 (in Bulgarian).

Bozilova

(1995) The upper forest limit in the Rila Mountains in Postglacial time – Palaeoecological evidence from pollen analysis, macrofossil plant remains and ¹⁴C dating. In: Bozilova

Tonkov

(eds) Advances in Holocene Palaeoecology in Bulgaria. Sofia: Pensoft Publishers, pp. 1–8.

10.

Bozilova

Smit

(1979) Palynology of Lake Suho Ezero in South Rila Mountain (Bulgaria). Fitologija 11: 54–67.

11.

Bozilova

Tonkov

(1990) The impact of Man on the natural vegetation in Bulgaria from the Neolithic to the Middle Ages. In: Bottema

Entjes-Nieborg

Van Zeist

(eds) Man’s Role in the Shaping of the Eastern Mediterranean Landscape. Rotterdam: Balkema, pp.327–332.

12.

Bozilova

Tonkov

(1994) The postglacial distribution patterns of Abies in Bulgaria. In: Lotter

Ammann

(eds) Dissertaciones Botanicae 234. Berlin: J. Cramer, pp. 215–223.

13.

Bozilova

Tonkov

(2000) Pollen from Lake Sedmo Rilsko reveals southeast European postglacial vegetation in the highest mountain area of the Balkans. New Phytologist 148: 315–325.

14.

Bozilova

Tonkov

(2011) 14. Lake Sedmo Rilsko (Bulgaria): Lateglacial vegetation history. Grana 50: 232–234.

15.

Bozilova

Tonkov

Pavlova

(1990) Pollen and plant macrofossil analyses of the Lake Suho Ezero in the South Rila Mountains. Annual of University of Sofia, Faculty Biology 80(2): 48–57.

16.

Bozilova

Filipova

Filipovich

. (1996) Bulgaria. In: Berglund

Birks

Ralska-Jasiewiczowa

. (eds) Palaeoecological Events during the Last 15000 Years: Regional Syntheses of Palaeoecological Studies of Lakes and Mires in Europe. Chichester: John Wiley & Sons, pp. 701–728.

17.

Bozilova

Tonkov

Marinova

. (2002) Pollen and plant macrofossil analyses of Late Holocene sediments from Lake Panichishte in Northwestern Rila Mountains. Razprave IV. Razreda SAZU (Ljubljana) 43(2): 37–51.

18.

Bronk Ramsey

(2005) OxCal Program, version 3.10. Oxford: Radiocarbon Accelerator Unit, University of Oxford.

19.

Dansgaard

Johnsen

Clausen

. (1993) Evidence for general instability of past climate from a 250-kyr ice-core record. Nature 364: 218–220.

20.

Davis

BAS

Brewer

Stevenson

. (2003) The temperature of Europe during the Holocene reconstructed from pollen data. Quaternary Science Reviews 22: 1701–1716.

21.

Davis

Deevey

(1964) Pollen accumulation rates: Estimates from late-glacial sediment of Rogers Lake. Science 145: 1293–1295.

22.

Davis

Brubaker

Webb

III (1973) Calibration of absolute pollen influx. In: Birks

HJB

West

(eds) Quaternary Plant Ecology. Oxford:Blackwell Scientific Publications, pp. 9–25.

23.

Faegri

Iversen

(1989) Textbook of Pollen Analysis. 4th Edition. Chichester: John Wiley & Sons.

24.

Feurdean

Tantau

Farcas

(2011) Holocene variability in the range distribution and abundance of Pinus, Picea abies, and Quercus in Romania; implications for their current status. Quaternary Science Reviews 30: 3060–3075.

25.

Feurdean

Persoiu

Tantau

. (2014) Climate variability and associated vegetation response throughout Central and Eastern Europe (CEE) between 60 and 8 ka. Quaternary Science Reviews 106: 206–224.

26.

Feurdean

Wohlfarth

Björkman

. (2007) The influence of refugial population on Lateglacial and early Holocene vegetational changes in Romania. Review of Palaeobotany and Palynology 145: 305–320.

27.

Giesecke

Bennett

Birks

. (2011) The pace of Holocene vegetation change – Testing for synchronous developments. Quaternary Science Reviews 30: 2805–2814.

28.

Glovnia

(1968) Glacial and periglacial relief in the southern part of Central Rila Mountain. Annual of University of Sofia, Faculty Geology and Geography 61(2): 37–66 (in Bulgarian).

29.

Grebska-Kulowa

Kulow

(2007) Prehistorical sites in the middle Struma River Valley between the end of the 7th Millennium BC and the beginning of the 5th Millennium BC. In: Stefanovic

Todorova

(eds)In the Steps of James Harvey Gaul 2. Sofia: James Harvey Gaul Foundation, pp. 279–296.

30.

Grimm

(2011) TGView, version 1.7.16. Springfield, IL: Illinois State Museum.

31.

Hofmann

Lotter

(2003) The development of the late-glacial diatom flora in Lake Sedmo Rilsko (Rila Mountains, Bulgaria). In: Tonkov

(ed.) Aspects of Palynology and Palaeoecology. Sofia:Pensoft Publishers, pp. 171–184.

32.

Ivanov

(1964) The Lakes in Bulgaria. Sofia: Nauka i Izkustvo (in Bulgarian).

33.

Jalut

Carrion

David

. (2005) The vegetation around the Mediterranean basin during the last glacial maximum and the Holocene climatic optimum. In: Petit-Mair

Vrielinck

(eds) The Mediterranean Basin: The Last Two Climatic Extremes. Explanatory Notes of the Maps (scale 1/7 000 000). Aix-en-Provence: MMSH and ANDRA, pp. 37–57.

34.

Kuhlemann

Gachev

Gikov

. (2013) Glaciation in the Rila Mountains (Bulgaria) during the Last Glacial Maximum. Quaternary International 293: 51–62.

35.

Lang

(2003) Immigration of Tilia in Europe since the last Glacial. In: Tonkov

(ed.) Aspects of Palynology and Palaeoecology. Sofia-Moscow: Pensoft Publishers, pp. 21–42.

36.

Magyari

Jakab

Bálint

. (2012) Rapid vegetation response to Lateglacial and early Holocene climatic fluctuation in the South Carpathian Mountains (Romania). Quaternary Science Reviews 35: 116–130.

37.

Marinova

Tonkov

(2012) Holocene vegetation history of the Northwestern Pirin Mountain (Bulgaria). Plant fossil record from peat-bog Mozgovitsa. Comptes rendus de l’Académie bulgare des Sciences 65(8): 1087–1094.

38.

Marinova

Tonkov

Bozilova

. (2012) Holocene anthropogenic landscapes in the Balkans: The palaeobotanical evidence from southwestern Bulgaria. Vegetation History and Archaeobotany 21: 413–427.

39.

Moore

Webb

Collinson

(1991) Pollen Analysis. 2nd Edition. Oxford: Blackwell Scientific Publications.

40.

Müller

Pross

Tzedakis

. (2011) The role of climate in the spread of modern humans into Europe. Quaternary Science Reviews 30: 273–279.

41.

Panagiotopoulos

Aufgebauer

Schäbitz

. (2013) Vegetation and climate history of the Lake Prespa region since the Lateglacial. Quaternary International 293: 157–169.

42.

Reimer

Bard

Bayliss

. (2013) IntCal13 and Marine13 radiocarbon age calibration curves 0–50,000 years cal BP. Radiocarbon 55(4): 1869–1887.

43.

Rohling

Palike

(2005) Centennial-scale climate cooling with a sudden cold event around 8,200 years ago. Nature 434: 975–979.

44.

Sadori

Giardini

Gliozzi

. (2015) Vegetation, climate and environmental history of the last 4500 years at lake Shkodra (Albania/Montenegro). Holocene 25(3): 435–444.

45.

Seppä

Alenius

Muukkonen

. (2009) Calibrated pollen accumulation rates as a basis for quantitative tree biomass reconstructions. Holocene 19(2): 209–220.

46.

Stefanova

Ammann

(2003) Lateglacial and Holocene vegetation belts in the Pirin Mountains (southwestern Bulgaria). Holocene 13: 97–107.

47.

Stefanova

Bozilova

(1995) Studies on the Holocene history of vegetation in the Northern Pirin Mts. (Southwestern Bulgaria). In: Bozilova

Tonkov

(eds) Advances in Holocene Palaeoecology in Bulgaria. Sofia: Pensoft Publishers, pp. 9–31.

48.

Stefanova

Oeggl

(1993) Zur holozanen Vegetationsgeschichte SW-Bulgariens: Das Moor Praso (1900 m) im Pirin Gebirge. Berichte Naturwissenschaftlich-medizinischer Verein Innsbruck 80: 69–80.

49.

Stefanova

Lazarova

Wright

(2006a) Elevational gradients during the Late-Glacial/Holocene vegetational transition in southern Bulgaria. Vegetation History and Archaeobotany 15: 333–343.

50.

Stefanova

Atanassova

Delcheva

. (2006b) Chronological framework for the Late Glacial pollen and macrofossil sequence in the Pirin Mountains, Bulgaria: Lake Besbog and Lake Kremensko-5. Holocene 16(6): 877–892.

51.

Stockmarr

(1971) Tablets with spores used in absolute pollen analysis. Pollen et Spores 13: 615–621.

52.

Sweeney

(2004) A key for the identification of stomata of the native conifers of Scandinavia. Review of Palaeobotany and Palynology 128: 281–290.

53.

Tonkov

Marinova

(2005) Pollen and plant macrofossil analyses of mid-Holocene radiocarbon dated profiles from two subalpine lakes in Rila Mountains, Bulgaria. Holocene 15: 663–671.

54.

Tonkov

Bozilova

Possnert

(2013) Postglacial vegetation history as recorded from the subalpine Lake Ribno (NW Rila Mts), Bulgaria. Central European Journal of Biology 8(1): 64–77.

55.

Tonkov

Possnert

Bozilova

(2006) Lateglacial vegetation and radiocarbon dating of Lake Trilistnika, Rila Mountains (Bulgaria). Vegetation History and Archaeobotany 16: 15–22.

56.

Tonkov

Possnert

Bozilova

(2011) The Lateglacial in the Rila Mountains (Bulgaria) revisited: The pollen record of Lake Ribno (2184 m). Review of Palaeobotany and Palynology 166: 1–11.

57.

Tonkov

Bozilova

Dimitrov

. (2002a) Surface pollen samples from the north-western Pirin mountain (south-western Bulgaria). Annual of University of Sofia, Faculty Biology 90(2): 11–21.

58.

Tonkov

Bozilova

Possnert

. (2008) A contribution to the postglacial vegetation history of the Rila Mountains, Bulgaria: The pollen record of Lake Trilistnika. Quaternary International 190: 58–70.

59.

Tonkov

Panovska

Possnert

. (2002b) The Holocene vegetation history of Northern Pirin Mountain, southwestern Bulgaria: Pollen analysis and radiocarbon dating of a core from Lake Ribno Ban derishko. Holocene 15: 201–210.

60.

Tzedakis

(2004) The Balkans as prime glacial refugial territory of European temperate trees. In: Griffiths

Krystufek

Reed

(eds) Balkan Biodiversity: Pattern and Process in the European Hotspot. Dordrecht: Kluwer Academic Publishers, pp. 49–68.

61.

Tzedakis

(2009) Cenozoic climate and vegetation change. In: Woodward

(ed.) The Physical Geography of the Mediterranean. Oxford: Oxford University Press, pp. 89–137.

62.

Tzedakis

Emerson

Hewitt

(2013) Cryptic or mystic? Glacial tree refugia in northern Europe. Trends in Ecology & Evolution 28(12): 696–704.

63.

Tzedakis

Hooghiemstra

Pälike

(2006) The last 1.35 million years at Tenaghi Philippon: Revised chronostratigraphy and long-term vegetation trends. Quaternary Science Reviews 25: 3416–3430.

64.

Tzedakis

Lawson

Frogley

. (2002) Buffered tree population changes in a Quaternary refugium: Evolutionary implications. Science 297: 2044–2047.

65.

Van Geel

Raspopov

van der Plicht

. (1998) Solar forcing of abrupt climate change around 850 calendar years BC. In: Peiser

(ed.) Natural Catastrophes during Bronze Age Civilization (BAR International Series, 728). Oxford: Archaeopress, pp. 162–168.

66.

Velčev

(1995) The Pleistocene glaciations in the Bulgarian mountains. Annual University of Sofia, Faculty Geology and Geography 87(2): 53–65 (in Bulgarian).

67.

Velčev

(1999) Glacial and cryogenic relief in part of Musala share of Rila. Annual University of Sofia, Faculty Geology and Geography 89(2): 7–21 (in Bulgarian).

68.

Velčev

Tonkov

Bozilova

(2011) Palaeogeomorphological and palaeoecological development of the Northwestern Rila Mountains during the Pleistocene and Holocene. Problems of Geography 3–4: 35–47 (in Bulgarian).

69.

Velev

(2002)Climatic division. In: Kopralev

(ed.) Geography of Bulgaria. Sofia: ForKom, pp. 155–156 (in Bulgarian).

70.

Willis

(1994) The vegetational history of the Balkans. Quaternary Science Reviews 13: 769–788.