Key words: Fagus sylvatica, Picea abies, pollen analysis, palaeoecology, immigration, population dynamics, forest continuity, forest history, disturbance, succession, Late Holocene, south Sweden

This thesis is based on four papers and the present synthesis. The four papers are listed below and presented as appendices I–IV. They are referred to in the text according to their respective Roman numeral label.

Till sidans topp/Up

Introduction

Small vs. large palaeoecological sites

Fagus and Picea

Holocene distributional changes for Fagus and Picea in Europe
Fagus
Picea
Climate vs. human influence as causes for distributional changes
Fagus
Picea
Holocene history of Fagus and Picea in southern Sweden
The present distribution of Fagus in Sweden and historical distributional changes
The establishment of Fagus forest in southern Sweden (based on earlier palaeoecological investigations)
The present distribution of Picea in southern Sweden and recent distributional changes
Ecological characteristics of Fagus and Picea relevant for this study
Dispersal biology
Regeneration and establishment
Climatic factors

Selection of sites and field work
Laboratory methods, pollen and charcoal analysis
Chronology

Paper I
Paper II
Paper III
Paper IV

Fagus
Is it possible to detect early occurrences of species by pollen analysis?
The establishment of Fagus in Sweden
The role of disturbance and cultural activities
The pre-Fagus vegetation
Is Fagus still migrating northwards in Sweden?
Picea
The establishment of Picea in southern Sweden
Picea invasion at stand-scale
Competition between Picea and Fagus

Acknowledgements

Svensk sammanfattning (Swedish summary)

References

Appendix I [Abstract]
Appendix II [Abstract]
Appendix III [Abstract]
Appendix IV [Abstract]

Till sidans topp/Up

The composition and structure of Swedish woodlands have changed considerably throughout the Holocene. These changes have been caused by several factors that include the immigration of species, climatic change, competition, succession, soil processes, disturbance regimes, and more recently, cultural activities. In southern Sweden these changes have been highly dynamic and complex, especially since the latter part of the Middle Holocene. During this period two very competitive and potential forest dominants – beech, Fagus sylvatica, and Norway spruce, Picea abies – immigrated into Sweden from different directions. The immigration of these species has contributed to the great shift of woodland composition during the last few thousand years (e.g., Björse et al. 1996). This period also witnessed the development of the present-day cultural landscape. The reasons for these rather recent vegetation changes can be difficult to understand fully as they may have been caused by a combination of natural and cultural factors.

This thesis focuses on local forest history, particularly on the immigration and establishment of Fagus and Picea at the spatial scale of the forest stand in southern Sweden. Many earlier regional-scale pollen studies have established when these species became regionally abundant (e.g., Nilsson 1964; Berglund 1966, 1991; Digerfeldt 1972, 1982; Göransson 1977; Regnéll 1989; Lagerås 1996), but we still lack detailed information about their establishment, particularly in which type of vegetation they became established, and how this came about. This is difficult to interpret with any certainty from regional-scale pollen diagrams because of poor spatial resolution.

However, an alternative way to study this problem is to use small forest hollows, i.e., small sites with organogenic deposits (peat or gyttja) in close proximity to forest stands of interest. Another possibility is to use deposits of mor humus on the forest floor, but such deposits are not available in all forest types. Small sites, especially closed-canopy sites, have the advantage that the majority of the pollen that becomes deposited has not travelled very far (e.g., Andersen 1970; Jacobson & Bradshaw 1981; Heide & Bradshaw 1982; Bradshaw 1988; Chen 1988; Yazvenko 1991; Jackson & Wong 1994; Sugita 1994; Calcote 1995). A pollen diagram from a small site therefore gives a more local picture of the vegetation than diagrams from lakes or bogs. Small sites with peat or gyttja deposits, or mor humus, have been used successfully elsewhere to study local forest history, for instance in Denmark (e.g., Iversen 1964, 1969; Aaby 1983; Andersen 1984, 1988, 1989), Ireland (e.g., Mitchell 1988; Hannon & Bradshaw 1989), and USA (e.g., Bradshaw & Miller 1988; Davis et al. 1992, 1994; Sugita et al. 1994). This type of site has until recently not been used very often in Sweden. Some Swedish studies are however, worth mentioning (e.g., Bradshaw & Zackrisson 1990; Bradshaw & Hannon 1992; Bradshaw 1993; Segerström et al. 1994, 1996; Lindbladh & Bradshaw 1995), as they clearly have shown the potential of local-scale forest studies.

In order to investigate and discuss the immigration and establishment of Fagus and Picea at stand-scale in southern Sweden four sites with small forest hollows were selected for high resolution pollen analysis (cf. Bradshaw 1988). These sites are Siggaboda, Flahult, and Mattarp in Småland, and Bocksten in Halland (Fig. 1). At these sites five small hollows in total were investigated (two adjacent hollows were studied at Bocksten, in order to evaluate stand-scale differences within a small area). All forest hollows selected for this study lie in close proximity to present Fagus stands, or stands where Fagus is abundant. In addition, Picea is more or less common in the vicinity of all selected hollows, and at least single Picea individuals occur near to all hollows. A further reason why these sites were selected is that they are all situated within the area of overlap for the present distributions of Fagus and Picea. Within this area of overlap, Fagus and Picea have had the possibility of competing with each other for some time. Moreover, Fagus and Picea have immigrated into southern Sweden from different directions, and they have both reached their present distribution limits rather recently. The geographical distribution of the selected sites therefore gives a gradient throughout the area with overlap where these species most likely became established at different times.

To avoid confusion I will clarify that I have used the terms forest and woodland interchangeably throughout this synthesis as well as in the appendices, i.e., the term forest has not been used in the same strict sense as suggested by Rackham (1976). Additionally, all dates in this thesis are given in uncalibrated ¹⁴C years BP, unless otherwise stated.


Fig. 1. (A) The forest regions of southern Sweden according to Sjörs (1965). (B) Detailed map of southernmost Sweden showing the distribution limits for Fagus and Picea forests according to Lindquist (1931, 1959). Note that isolated Fagus stands occur north of the northern limits of Fagus-Picea forest as indicated in Fig. 1A. The studied sites are also indicated, as well as other local- and regional-sce pollen sites referred to in this synthesis (A-Y, see also Table 2).
[Back to contents]

This section of the synthesis briefly summarises characteristics of different types of sites that may be useful for palaeoecological reconstruction.

As mentioned in the introduction, small sites have the advantage of having a relatively small pollen source area and they therefore give a more local-scale picture of the vegetation. Investigations of the relationship between pollen assemblages and vegetation have been an important task for paleoecologists, but it was not until rather recently that these relationships were examined in a quantitative manner (Prentice 1985).

Today, the opinion that the size of the sedimentary basin reflects the size of the pollen source area is well established (i.e., a forest hollow has a small source area and a large lake a large one). Many theoretical studies have indicated this relationship (e.g., Tauber 1965; Jacobson & Bradshaw 1981; Prentice 1985, 1988; Sugita 1993, 1994), which has recieved confirmation from studies of pollen/vegetation relationships (e.g., Berglund 1973; Tauber 1977; Bradshaw & Webb 1985; Jackson 1990; Yazvenko 1991; Sugita et al. 1994; Calcote 1995).

However, even if the general relationship is clear, many researchers differ in opinion when they try to define specifically the radius or diameter of a given pollen source area. High precision in this matter is probably unobtainable and arguably not of critical importance. Even if the pollen source area for a small site has a radius of 20, 50, or even 100 m, the main part of the pollen grains are still derived from the local vegetation (i.e., the local pollen signal is strong), and the pollen diagram constructed is valid for stand-scale interpretations. Sugita (1994) has presented a model predicting that the "relevant pollen source area" for a small hollow (radius = 2 m) is around 50–100 m. A small lake (radius = 50 m) has according to the same model a somewhat larger source area around 300–400 m. The "relevant pollen source area" is defined as the distance from a sampling site beyond which the pollen represent a constant background pollen rain and within which differences in plant abundance will be recorded as variance in pollen analyses among sites. Only 30–45% of the total pollen loading comes from within this area, but this is enough for detection of the spatial variability in the vegetation. Calcote (1995) also found that the correlation between pollen and vegetation ceased to improve as the vegetation sampling radius was increased to 50–75 m. His study also implies that pollen grains deposited in small hollows record stand-scale vegetation.

The majority of pollen analytical studies published up to now have been based on studies of lake sediments from relatively large basins, i.e., sites which have a comparatively large pollen source area. Pollen grains are transported to this type of site from many different vegetation types, and these grains are mixed together forming the pollen assemblage. Pollen samples from lakes or bogs with a diameter of c. 1 km normally reflect the vegetation within a large area. Theoretical and empirical studies have indicated that pollen grains are transported to such sites from within a radius of several tens of kilometres (e.g., Jacobson & Bradshaw 1981; Bradshaw & Webb 1985).

One way to improve the spatial resolution in palaeoecological studies is to use small sites where the majority of pollen grains deposited have not travelled very far. Sites useful for such studies are often small basins where organogenic sediments are deposited, particularly when they are situated below a dense canopy. These sites are often referred to as closed-canopy sites (Bradshaw 1988). Closed-canopy sites normally provide both a temporal and spatial resolution that enable the pollen analysis to be a useful tool for stand-scale reconstruction. They can also reveal information about pollen taxa that are rarely found in regional-scale studies, especially for taxa producing few pollen grains or that are not dispersed very well. Far-travelled pollen (<1 km) normally only comprise a minor part of the pollen assemblage deposited at closed-canopy sites. When the canopy is opened up this immediately influences the pollen deposition and the pollen source area will increase. A larger opening (or a larger basin) simply means a larger pollen source area.

Andersen (1970) was among the pioneers investigating the relationship between pollen assemblages deposited on the forest floor and composition of the forest within 20–30 m from the sampling point. His investigation has been repeated and largely confirmed by several other studies (e.g., Bradshaw 1981a; Heide & Bradshaw 1982; Chen 1988; Yazvenko 1991; Jackson & Wong 1994; Calcote 1995), indicating that pollen/vegetation relationships are fairly well understood at this scale. However, the pollen source area for certain pollen types is probably somewhat larger than originally thought (e.g., Jackson & Wong 1994). Different pollen types have different source areas, as pollen grains vary in size and dispersal ability (Prentice 1985, 1988; Sugita 1994). This also affects the pollen source area giving light pollen types, or pollen types with sacci, a larger source area than heavy types.

Several small sites are needed to study stand-scale differences within an area, or differences in vegetation development on different soil types. Ideally, a net-work of small sites in different settings is preferred to reveal such differences. It is however, also possible to connect stand-scale studies with regional-scale ones. One example of such a study is that made by Andersen (1984), which described the forest history of Eldrup Forest in Denmark using pollen analyses based on sediments from small, wet hollows. Andersen was able to detect vegetation changes affecting the forest that were not discernible from a regional pollen diagram from a bog lying outside, but nearby the forest. He was also able to detect stand-scale differences within the forest, from small sites lying not more than 100 m apart.

Sites useful for local-scale studies can be divided into three major types (Table 1) according to Bradshaw (1988). Small wet hollows are small basins where organogenic sediments (peat or gyttja) are deposited. Temporal resolution is normally good (usually extremely good during the last 500 to 1000 years), and the sediments may be continuously deposited over several thousands of years, and sometimes spanning more or less the whole of Holocene (e.g., Sugita et al. 1994). Pollen grains are normally well-preserved in wet hollows, but less favourable preservation conditions may occur if the site dries up periodically.

Preservation of pollen is probably the largest problem when analysing sediments from small sites (Bradshaw 1988). Preservation conditions may vary from excellent to relatively poor within the same profile. Pollen grains are normally more susceptible to oxidation and microbiological activity when deposited in small sites than in lakes or bogs. A high concentration of degraded pollen grains, or pollen types (and spores) that are resistant to degrading, most likely indicate that preservation conditions in the sediment have been poor (Havinga 1971, 1984). However, these problems are mostly confined to soils and shallow, wet hollows that may dry out completely during the summer.


Table 1. Some important properties of small sites, slightly modified after Bradshaw (1988).
 

	Small wet hollows	Mor humus deposits	Soils
Length of record	usually >1000 years	100-1000 years	10 years
Continuity of record	usually good	good	poor
Temporal resolution of single sample	10–100 years	1–10 years	?
Sediment mixing	minimal	minimal	major problem
Pollen preservation	usually good	good	usually poor
Site availability	restricted	restricted	abundant

[Back to contents]

This section of the synthesis gives a general background to topics and problems discussed in this thesis. It briefly describes what is known about the present (and historical) distribution of Fagus and Picea, as well as the migration and establishment of these species. It also deals with some ecological characteristics for these species that may be relevant for this study. [Back to contents]

Holocene distributional changes for Fagus and Picea in Europe

It is beyond the scope of this thesis to describe and discuss in detail the migration patterns of Fagus and Picea in Europe as a whole. This has recently been done more thoroughly in Huntley & Birks (1983) and Huntley (1988). However, a brief summary of these changes can be useful as a background to the Late Holocene distributional changes that have affected southern Sweden. It should also be noted that two species are native to Europe for both the genus Fagus (F. sylvatica, F. orientalis) and Picea (P. abies, P. omorika). When Fagus and Picea are discussed or mentioned in this thesis, they refer to F. sylvatica and P. abies. The other two species, F. orientalis and P. omorika, have a restricted occurrence mainly in south-eastern Europe, and they are not considered in this thesis.

Fagus

The distribution changes for Fagus in Europe since the termination of the last Ice Age have been summarised by Huntley & Birks (1983). During the maximum of the Weichselian glaciation Fagus was probably restricted to the mountainous areas of the Balkans, Italy and the southern part of the Carpathians, even if the pollen evidence for this is scant. From Late Glacial time Fagus pollen grains are only found at some scattered sites, for instance in Italy. This distribution picture did not change much until about 9000 BP when high pollen percentages for Fagus are encountered at sites in the southern Balkans and in the Carpathians. During the earliest part of the Holocene, Fagus apparently migrated along the European mountain ranges. Fagus probably grew in mountain forests during this time. The rapid climatic amelioration at the transition to the Holocene most likely initiated the migration of Fagus.

At about 8500 BP Fagus rapidly migrated northwards through Italy and the northern part of the Balkans. This pattern was stable for more than 1000 years until about 7000 BP when eventually Fagus populations in the Carpathians also started to expand northwards. At about 6500 BP Fagus had extended its distribution further to the north and the west. At this time also an outpost area appeared in southern France. Continuing expansion with high rates of population movement of around 250-300 m/year eventually led to the merging together of all separate Fagus areas at about 5000 BP. Fagus was then more or less continuously distributed from the Massif Central in the west to the Carpathians in the east.

At about 4000 BP three, new, main features appeared in the migration patterns for Fagus: expansion into the Pyrenees, into Poland, and a recolonisation of the mountains in the southern Balkans. Continuing expansion at approximately the same rate as before brought Fagus rapidly into northern Germany, south-eastern England, northern Spain, and eventually also into southern Scandinavia.

In summary, the expansion patterns for Fagus in Europe during the Holocene show three main features (Huntley 1988): a comparatively slow expansion during the earliest part of the Holocene, an expansion only in the central and eastern part of Europe during the Middle Holocene, and a rapid expansion in the north and west during the Late Holocene, when Spain, England and Scandinavia finally became invaded. [Back to contents]

Picea

The distribution limits for Picea in Europe have also changed considerably throughout the Holocene. These changes have been summarised by Huntley & Birks (1983). They have clearly shown that Picea occurred within two separate areas during the Late Glacial: one area in western Russia, centred around Moscow, and one in south-eastern Europe, extending from the Austrian Alps in the west to the Carpathians in the east. 

During the beginning of the Holocene Picea populations in western Russia decreased considerably and almost vanished. At approximately the same time Picea expanded comparatively rapidly in the mountains of south-eastern Europe. In this area the expansion continued towards the west along the Alps and in the Carpathians until about 4000 BP. Later on, a slight recession occurred, and the area with Picea forests then became somewhat disrupted. After about 2000 BP a new, but rather small expansion has occurred, especially on the northern Alpine forelands. 

In western Russia Picea became re-established about 7000 BP, and during the period 7000–5000 BP, a vast area with Picea forests, extending from eastern Finland to western Russia, was established. From about 4000 BP until the present day Picea populations have moved rapidly (with rates approaching 500 m/years) towards the west and south-west and expanded its distribution limits (Aario 1965; Aartolahti 1966; Moe 1970; Tallantire 1972a, 1977; Persson 1975; Huntley & Birks 1983; Tolonen 1983; Hafsten 1985, 1992; Huntley 1988). However, Picea individuals occurred in Scandinavia at a few favourable outpost sites as early as during the Middle Holocene (Kullman 1995), but it was not an ecologically important tree until the main expansion westwards. Today almost the whole of Fennoscandia lies within the present distribution limits for Picea, except parts of the extreme north, south and west. [Back to contents]

Climate vs. human influence as causes for distributional changes

Fagus
There has been a long and lively debate trying to determine why Fagus apparently failed to migrate into the lowlands of Northwest Europe during the middle part of the Holocene. The traditional explanation has been that Fagus was unable to invade the pre-existing dense woodlands in northern and central Europe, so long as they were undisturbed (Iversen 1973; Godwin 1975; Behre 1988). During the earliest part of the Holocene Fagus probably invaded rather open forest types in mountainous areas, and these types did not significantly hinder the invasion (Huntley & Birks 1983). When the pre-existing woodlands in Northwest Europe started to become disturbed by human interference, expansion of Fagus seems to have occurred rapidly, and with migration rates as high as during earlier parts of the Holocene. Iversen (1973) proposed that the late establishment of Fagus in, for example Denmark, was probably related to a colonisation of podsolic soils in previously cleared forests. 

Huntley & Birks (1983) argue that the northwards migration of Fagus in north and west Europe after about 5000 BP coincides with the first significant indication of human interference with the vegetation, and that this coincidence probably has significant implications. This coincidence has led to the postulate that Fagus acted as an opportunistic tree species when cleared areas eventually became available. Even if such conditions favour establishment of Fagus, this postulate only rests on a temporal coincidence and does not take into account the possibilities for Fagus to invade undisturbed forests (Huntley 1988). Moreover, the migration of Fagus was apparently more gradual than the expansion of the Neolithic culture, and this migration has continued until the present day. In Switzerland, north of the Alps, the possible influence of human activity, or natural disturbance, on the timing of establishment and expansion of Fagus has also been discussed (Richoz et al. 1994). 

The alternative explanation is that Fagus migrated simply as a response to climatic change. It is possible that a decreasing continentality in climate and milder winters during the latter part of the Holocene, facilitated the migration towards the north. This postulate also gains support from distribution changes for Fagus grandifolia in North America (a species of similar ecological requirements), where this species has migrated to the north and west throughout almost the entire Holocene (Davis 1976, 1981; Bennett 1985; Dexter et al. 1987), at a time when apparently the continentality of the interior parts of the continent decreased. Prehistoric humans in North America are also believed to have had only a minor effect on tree migration. The great resemblance in migration patterns for Fagus in North America and Europe may indicate that Fagus has migrated purely as a response to climatic changes throughout the Holocene (Huntley & Webb 1989; Huntley et al. 1989). In Europe milder winters in the area north of the Alps have been postulated as a cause for the Late Holocene expansion of Fagus (Huntley 1988). A similar change in winter temperatures has also hypothetically been assigned as a cause for the Late Holocene expansion of Fagus grandifolia in eastern North America (Bartlein et al. 1986; Huntley 1988).

Iversen (1973) and other authors have strongly argued against the importance of climate for the migration of Fagus, as the invasion into Northwest Europe occurred after the Holocene climatic optimum. They instead assumed that a northward migration ultimately has to be driven by increasing temperatures. Evidence from other species, e.g., Corylus avellana, has indicated a southwards regression in Scandinavia after 6000 BP. This led to an interpretation that Fagus apparently migrated northwards during a time when temperatures were decreasing. Iversen concluded that the expansion northwards of Fagus evidently not was determined by a climatic response. However, with the benefit of hindsight Iversen1s interpretation is certainly a simplification of the climatic changes that really have occurred. These changes comprise not only changes in temperature, but also alterations in seasonality, as well as independent changes in both summer and winter temperatures, and precipitation, and these changes are ultimately driven primarily by orbital forcing. These complex alterations can cause contradictory changes in distribution limits for different species (Webb 1986; Huntley & Webb 1989). The southwards recession in Corylus distribution can still be explained by a decreasing temperature, but this decrease is only applicable to the summer temperature, while the contemporaneous northwards expansion of Fagus may be explained by milder winters. 

The hypothesis that Fagus was unable to invade woodlands in Northwest Europe until cultural activities created suitable conditions for establishment takes no account of the fact that natural disturbances, such as fires and storms, occur in all forest types. Fagus is dependent on ground disturbance for successful regeneration (Watt 1923; Bjerregaard & Carbonnier 1979; von Röhrig et al. 1978), and natural disturbances can equally well create suitable seed beds. Huntley et al. (1989) therefore conclude that human disturbances alone cannot have been the ultimate cause for the Holocene distributional changes for Fagus in both Europe and North America. [Back to contents] 

Picea
Factors lying behind the expansion of Picea are also uncertain, and varying migration rates in different areas during different times are difficult to explain. The westward migration of Picea along the Central European mountains during the Early Holocene may be explained by a climatic amelioration occurring at the transition to the Holocene, which enabled Picea expansion from Late Glacial refuges (Huntley & Birks 1983).

The distribution changes for Picea in the northern part of its European extension are also difficult to understand. The decrease in area during the Late Glacial and the earliest part of the Holocene may have been caused by a warmer climate. The re-establishment and expansion after about 7000 BP probably reflects a trend towards more suitable conditions for Picea. Its seedlings are rather sensitive to summer drought, and they also require sufficient insulating snow cover during the winter. An alteration in the ratio precipitation/evaporation may be a more important climatic factor than a direct change in temperature (Huntley & Birks 1983).

In several regional pollen diagrams from Fennoscandia the establishment and expansion of Picea coincides with indications of cultural disturbances, for example forest clearance or agriculture (e.g., Almquist-Jacobson 1994). These cultural activities may have facilitated the expansion of Picea. However, as pointed out by Tallantire (1972a), Picea also increases in several pollen diagrams without contemporaneous cultural activities. This may indicate that human activity is not the only cause explaining the migration of Picea in Fennoscandia (Huntley & Birks 1983). Huntley (1988) has postulated three factors explaining the migration of Picea, and these factors do not necessarily involve human interference:

Before the introduction of modern forestry with planting of Picea outside its present distributional limits, and introduction of foreign provenances, Tallantire (1972a, 1972b) argues that the migration patterns for Picea were driven purely by climate. In addition, Picea has probably not been affected during prehistoric times by selective removal or forest grazing, as many other deciduous trees have been. The expansion of Picea may in some areas during later periods have been somewhat controlled by humans, for instance by slash-and-burn activities, or through active removal from forest-meadows, as it was unwanted in this vegetation type (e.g., Weimarck 1953). [Back to contents]

Holocene history of Fagus and Picea in southern Sweden

The present distribution of Fagus in Sweden and historical distributional changes

Fagus has a pronounced southerly distribution in Sweden (Fig. 1). Its main distribution occurs in the nemoral vegetation zone (sensu Sjörs 1965), but it also extends into the southernmost part of the boreo-nemoral zone, where the natural ranges of Fagus and Picea overlap each other. Within this outpost area Fagus generally occurs as a minor component in the forests, except at a few and scattered stands where it can be a local dominant (Lindquist 1931, 1959; SOU 1971). It is normally a dominant forest tree within the nemoral vegetation zone. It thrives well on several soil types as long as they are well-drained (e.g., Rackham 1980), avoiding soils that are too wet, especially peaty soils (e.g., Wibeck 1909).

The northern distributional limit of Fagus is a diffuse 50–100 km broad band running from NNW to SSE across the southern part of the boreo-nemoral vegetation zone. Within this band the abundance of Fagus falls off rapidly, and to the north of this area only a few isolated outpost stands occur.

The present distribution of Fagus in Sweden is well-known and has been studied and mapped several times during this century (e.g., Nilsson 1902b; Lindquist 1931, 1959; Hjelmqvist 1940; Lindgren 1970; SOU 1971). Its distribution has traditionally been divided into three regions: a Fagus region (mainly belonging to the nemoral vegetation zone), a Fagus-Picea region, and a marginal or outpost region (cf. Lindquist 1931, 1959). However, Lindgren (1970) has argued for a more strict division, i.e., into a Fagus and an outpost region. These authors' characterisation of the Fagus region are very similar, except that Lindgren assigned the Fagus forests along the Swedish west coast to the outpost region. Lindquist (1931, 1959) instead assigned these forests to the Fagus region.

Several studies of, for example, historical documents and maps have clearly demonstrated that Fagus forests were more abundant and widespread during the Middle Ages and Early Modern Time than during the present day (e.g., Wibeck 1909; Malmström 1937, 1939; Troedsson 1966; Svenningsson 1992; Brunet 1995). This has also been confirmed by regional-scale pollen analyses (Björse et al. 1996). The earliest history of Fagus in Sweden is however, not well-known from historical documents, as Fagus was already present when the oldest preserved documents were established.

Fagus is a masting tree and its nuts have been highly valued, especially for pig breeding. As a consequence of this many historical documents describe Fagus stands and their usage (e.g., Wibeck 1909). The oldest existing documents describing Fagus stands in for example, western Småland date from about the 12th and 13th centuries. These documents clearly demonstrate that Fagus stands were already important for feeding pigs. The legislation affecting stands with Fagus (but also Quercus robur) was strict during the Middle Ages and Early Modern Time. For instance, the forest act of 1647 stated that every individual of Fagus and Quercus that was felled should be replaced with the same species. This legislation was, however, never firmly rooted among farmers, and during the latter part of the 18th century the usage of masting trees became freer, partly as a consequence of a more liberal legislation.

The value of a Fagus stand throughout the Middle Ages and later was determined by the number of pigs that could feed within the stand during years with a large mast. Later on, when the area of Fagus forest decreased, the importance of pig breeding also declined. The breeding of pigs probably had a positive effect on the regeneration of Fagus, as pigs through their subsistence created seed beds that where favourable for Fagus. The canopy of a Fagus stand used for pig breeding was probably also held rather open, to increase the production of nuts. It is very likely that such stands had an open structure with both old and young Fagus individuals intermixed.

The Fagus population has been severely reduced during the medieval and historic periods because of human exploitation. Wibeck (1909) and Malmström (1939) have documented the particularly marked reduction in Fagus forests in south-west Sweden during the 18th and 19th centuries using historical documents and maps. Several factors have contributed to this reduction. Among these factors, tilling and grazing activities were probably the most important, but locally also the production of potash (which required enormous amounts of wood) was significant (e.g., Åhman 1983). These activities made the remaining Fagus stands less resistant to invasion by other forest trees.

The immigration of Picea from the north is probably a further factor that has contributed to the decline of Fagus. Picea has migrated southwards from central Sweden (Moe 1970; Tallantire 1972a, 1977; Persson 1975) during approximately the same period as Fagus has immigrated from the south. Both species can flourish under nearly the same conditions, and consequently there is an intense competition between these two dominant species when they meet in the same stand. [Back to contents]

The establishment of Fagus forest in southern Sweden (based on earlier palaeoecological investigations)

Fagus immigrated into Sweden from the south (e.g., Huntley & Birks 1983; Huntley 1988). It is difficult using regional pollen diagrams alone to determine when Fagus first became established in the pre-existing forests. Single Fagus pollen grains are common in lake sediments, especially in Skåne, well before 4000 BP (e.g., Nilsson 1964; Berglund et al. 1991). It is not possible to deduce with any certainty whether these pollen grains originate from scattered Fagus trees growing in Swedish forests, or if they have been transported to Sweden from the area south of the Baltic – where Fagus forests already existed at that time (Huntley & Birks 1983; Huntley 1988). It is however, possible that immigration of new species can occur with very low population abundances (e.g., Kullman 1995) and such low frequencies cannot be detected with conventional pollen analysis (Bennett 1986). Studies of macrofossils may be a potential tool to reveal such early occurrences.

About 3500 BP Fagus first reached pollen percentages that are sufficiently high to imply it was established as small stands throughout the landscape (Table 2). A first weak expansion occurs around 3500–3200 BP in Skåne (e.g., Nilsson 1964; Regnéll 1989; Berglund et al. 1991). Fagus was probably only of minor importance in southern Sweden until 2200-1500 BP. Fagus attained pollen percentages of 3–5% in regional pollen diagrams around 2200–1500 BP in south and central Skåne (e.g., Nilsson 1964; Regnéll 1989; Berglund et al. 1991), and around 1500 BP in northern Skåne, Halland and Blekinge (Digerfeldt 1974, 1982; Berglund 1966). Fagus pollen percentages reach peak values after 1300–1200 BP in Skåne, Halland and Blekinge (Nilsson 1964; Berglund 1966; Digerfeldt 1974, 1982; Gaillard 1984; Regnéll 1989; Gaillard & Göransson 1991; Göransson 1991). Fagus probably achieved its greatest abundance in southern Sweden during this period.

Recent studies using AMS ¹⁴C dating of terrestrial plant macrofossils instead of conventional ¹⁴C dating of bulk gyttja samples, have shown that bulk gyttja samples may give spuriously old ages even in non-calcareous sediments (e.g., Lagerås 1996). This may be an explanation for unusually old dates from many sites, e.g., Lake Sandsjön (Thelaus 1989). This form of dating bias (and other biases, see Table 2; cf. Fig. 17) makes it difficult to discuss the earliest history of Fagus in southern Sweden.

The history of Fagus in its northern outposts is poorly known. Fagus occurs today in this area in single, dispersed stands, or as single, scattered individuals in other forest types. Many stands located near large farms and manor-houses are believed to be planted (cf. Lindquist 1931), although it is difficult to distinguish natural occurrence from planting purely by examination of the existing stand (e.g., Willis 1993). A rich occurrence of certain indicator species, for example lichens such as Lobaria pulmonaria, may indicate a long stand-continuity, at least for some hundred years (Nilsson et al., unpublished; James et al. 1977).

Regional pollen diagrams from central Småland show that Fagus pollen percentages of 3–5% were not reached before 1200–1000 BP (Table 2), and that these values peaked around 1000 BP and afterwards (Digerfeldt 1972, 1977; Königsson 1989; Svensson 1988). Several recent regional pollen investigations from central northern Småland have found a low Fagus representation. Percentage values of 0.5–1% are not reached before 1500–500 BP (Lagerås 1996; George L. Jacobson, unpublished), but maximum values only slightly exceed 1%. These percentages certainly mean that scattered Fagus individuals grew in the area, and Fagus may have been locally important, at least for as long as 1500 years. Lagerås (1996) argued that single Fagus individuals already occurred in this area before 3000 BP. However, this interpretation is based on the occurrence of single Fagus pollen grains, and these may just as likely be of long-distance origin. An investigation of the bog Dags mosse in Östergötland, which is located close to a nearby Fagus outpost on Omberg, had pollen percentages of 0.5–2% during the last 1000 years (Göransson 1989). [Back to contents]

Table 2. The establishment and expansion of Fagus in southernmost Sweden according to ¹⁴C dated levels in local and regional pollen diagrams (see also Fig. 17). The sites are ordered from approximately the south to the north (cf. Fig. 1). This table comprise dates from various sources, and they are not always strictly comparable. All dates have also been rounded off to the nearest hundred. First, these sites differ greatly in size and pollen source area, which means that local and regional features may be recorded differently. Second, Fagus pollen percentages are based on different pollen sums, in older diagrams mostly on a tree pollen sum, and in newer studies mostly on a total terrestrial pollen sum. Third, datings are made on a variety of materials (Sphagnum peat, fen peat, gyttja, etc.) and these dates may be more or less reliable. Some sites are only dated by pollen-analytical correlations. Fourth, datings have not always been carried out on levels critical for the Fagus pollen curve, which means that the nearest available date to this critical level has been used instead. In some cases a timescale in calibrated calendar years is available. In such cases dates derived from the time scale have been used, but for comparison with other dates in the table they have been converted to years BP. Fifth, Fagus pollen curves are generally very weak in regional pollen diagrams from the northern part of the studied area, which makes these curves difficult to interpret. Sixth, my interpretations of the Fagus pollen curves may differ somewhat from the interpretations suggested by the original authors.

¹ First establishment: means the level where small Fagus stands may have occurred in the area. Single Fagus individuals may have occurred in the area before this date. Fagus pollen percentages around 0,5% in regional pollen diagrams may be indicative for this phase.
² Local/Regional expansion: Local expansion means the level where Fagus stands certainly started to occur in the area. Fagus pollen percentages around 0,5–1% in regional pollen diagrams may be indicative for this phase (cf. Woods & Davis 1989). Regional expansion means the level where Fagus forests expanded regionally. Fagus pollen percentages around 2–5% in regional pollen diagrams may be indicative for this phase. The first date usually refers to the level where Fagus percentages exceed 0.5–1%, and the second date to the level where percentages exceed above 2–5%.
³ Culmination: means the period when Fagus must have been at its peak of abundance in the landscape.
⁴ These pollen diagrams are not ¹⁴C dated. They are instead pollen-analytically correlated with the well-dated diagram from Ageröds mosse (Nilsson 1964).
⁵ The studied sediments cover the time from around 600 AD to present time.
⁶ Bulk gyttja ¹⁴C dates. These dates may be more or less affected by reservoir effects, see for example Lagerås (1996).
⁷ Bulk peat ¹⁴C dates.
⁸ Chronology based on AMS dated terrestrial macrofossils.
⁹ Regional pollen diagram (based on sediments from medium-sized to large sites).
¹⁰ Local pollen diagram (based on sediments from forest hollows or other small sites).

The present distribution of Picea in southern Sweden and recent distributional changes

Picea is today the most abundant tree species in southern Sweden, but this abundance has however, a rather recent origin (Table 3; cf. Fig. 21). The forests in Sweden dominated by Picea belong to the boreo-nemoral and boreal vegetation zones (sensu Sjörs 1965) (Fig. 1). These zones are often referred to as the "southern coniferous belt" and the "northern coniferous belt" in Swedish literature (e.g., Lindquist 1948). The boreo-nemoral vegetation zone includes forest types that have been strongly influenced by humans for several thousands of years (e.g., Bradshaw et al. 1994). This influence has been extremely marked during the last 150 years (e.g., Nilsson 1992), particularly due to land-use changes, which have led to considerable changes in forest composition and structure. These recent changes have favoured Picea and coniferous forest types at the expense of deciduous forests.

It was demonstrated by discovery of macrofossils that Picea was a late immigrant into southern Sweden even before the development of pollen analysis as a tool for reconstructing past vegetation (e.g., Sernander 1892). The pioneering work of von Post (1924), which presented pollen analytical maps showing the distribution changes for tree species throughout the Holocene, demonstrated the late expansion of Picea in southern Sweden. Regional pollen studies published during the last decades have shown that Picea started to expand in the northern outskirts of the Småland Uplands and in southern Östergötland at about 1900–1500 BP (Göransson 1977, 1989, 1995). In central northern Småland Picea started to expand about 1200-900 BP (Lagerås 1996) (Table 3), and apparently at about the same time in central Småland (Digerfeldt 1972, 1977; Svensson 1988).

In southern Småland, Halland and Blekinge Picea expansion occurred late (cf. Fig. 21), and in many areas Picea dominance has only been built up during this century (Berglund 1966; Nilsson 1968; Digerfeldt 1974; Thelaus 1989). According to historical documents the first Picea individuals appeared in northern Skåne during the 17th and 18th centuries (e.g., Hesselman & Schotte 1906; Weimarck 1953; Glimberg 1963; Hallberg 1983), and about the same time in Halland and Blekinge (Malmström 1937, 1939; Björnsson 1946).

Picea obviously reached its present distribution limits rather recently. Its natural distribution limits were mapped during the beginning of this century by Hesselman & Schotte (1906). It is very difficult to deduce where its natural limits occur today, as Picea has been planted outside this limit for more than 100 years, and these plantations have obscured the previous limit. Picea grows well in southernmost Sweden (Langlet 1935), even in areas outside its supposed natural distribution limits. This may indicate that Picea is still migrating southwards, but this migration has also been highly favoured by modern forestry. The present southern distribution limits for Picea are apparently not clearly limited by climate or soil types. Several authors (e.g., Hesselman & Schotte 1906; Malmström 1937; Lindquist 1959) have suggested that the distribution limits for Picea during the recent past have been controlled by cultural factors (such as agriculture and forestry) rather than climate, edaphic or dispersal factors.

Picea easily invades forests dominated by Pinus sylvestris, or other forest types with an open structure, and these forests then quickly become dominated by Picea. Some forestry practices, for example selective cutting, may also have favoured Picea expansion. Small gaps are created when forest stands are used for selective cutting. In these gaps Picea regenerates often more successfully than the tree species originally present in the stand. Forests dominated by Betula sp. have probably also had a significance for Picea invasion. Almost pure Betula forests were common on old slash-and-burn areas (or on abandoned pastures) (e.g., Hesselman & Schotte 1906; Weimarck 1953). Additionally, the cultural landscape with forest-meadows, pastures, and open grazed forests, were important for Picea colonisation and expansion. If such areas are left unmanaged for some time they rapidly become Picea forests, as Picea easily outcompetes Betula, or other deciduous trees.

Hesselman & Schotte (1906) have summarised some important factors that have favoured Picea expansion into southern Sweden, and have made Picea highly competitive compared to many deciduous tree species: 1) Picea is able to invade and establish in most forest types, 2) human interference made previous forest types less resistant to Picea invasion, 3) Picea is able to produce viable seeds at an age of 25–30 years, which is younger than for many other tree species, 4) Picea has wind-dispersed seeds, that are easily spread over large distances.

Picea individuals of a Central European provenance have been introduced into southern Sweden. A more or less organised planting of Picea started around the middle of the 19th century, but occasional plantations may be older. Picea seeds imported from particularly Germany have been sown over vast areas of southernmost Sweden. Wibeck (1912) estimated that during the period between 1888 to 1909 around 35,000 kg of Picea seeds of a Central European provenance were sown in southern Sweden. Almost all Picea seeds sown in Skåne and Halland during this period were imported from Central Europe. [Back to contents]

Table 3. The establishment and expansion of Picea in southernmost Sweden according to ¹⁴C dated levels in local and regional pollen diagrams (see also Fig. 21). The sites are ordered from approximately the south to the north (cf. Fig. 1). This table comprise dates from various sources, and they are not always strictly comparable. All dates have also been rounded off to the nearest hundred. First, these sites differ greatly in size and pollen source area, which means that local and regional features may be recorded differently. Second, Picea pollen percentages are based on different pollen sums, in older diagrams mostly on a tree pollen sum, and in newer studies mostly on a total terrestrial pollen sum. Third, datings are made on a variety of materials (Sphagnum peat, fen peat, gyttja, etc.) and these dates may be more or less reliable. Some sites are only dated by pollen-analytical correlations. Fourth, datings have not always been carried out on levels critical for the Picea pollen curve, which means that the nearest available date to this critical level has been used instead. In some cases a timescale in calibrated calendar years is available. In such cases dates derived from the time scale have been used, but for comparison with other dates in the table they have been converted to years BP. Fifth, Picea pollen curves are generally weak in pollen diagrams from the southernmost part of the studied area, as these sites are situated outside or near to the present distribution limits for Picea forests. Picea has been planted outside its natural limits for at least 150 years. Sixth, my interpretations of the Picea pollen curves may differ somewhat from the interpretations suggested by the original authors.

¹ Local/Regional expansion: means the level where Picea stands certainly started to occur in the area. Picea pollen percentages around 5% in local/regional pollen diagrams may be indicative for this phase (cf. Huntley & Birks 1983). Percentages around 1% may be indicative for presence of single Picea individuals, or alternatively may be caused by long-distance transportation.
² These pollen diagrams are not ¹⁴C dated. They are instead pollen-analytically correlated with the well-dated diagram from Ageröds mosse (Nilsson 1964).
³ The studied sediments cover the time from around 600 AD to present time.
⁴ Bulk gyttja ¹⁴C dates. These dates may be more or less affected by reservoir effects, see for example Lagerås (1996).
⁵ Bulk peat ¹⁴C dates.
⁶ Chronology based on AMS dated terrestrial macrofossils.
⁷ Regional pollen diagram (based on sediments from medium-sized to large sites).
⁸ Local pollen diagram (based on sediments from forest hollows or other small sites).
 

Ecological characteristics of Fagus and Picea relevant for this study

Dispersal biology

A majority of the tree species in Northern Europe have light wind-dispersed seeds. However, an exception are trees within the Fagaceae family (e.g., Quercus sp., Fagus), which have heavy seeds. Without dispersal vectors (animals) these seeds obviously would not be dispersed far outside the stand of origin. For example, the jay (Garrulus glandarius) is probably the main vector for dispersal of Quercus acorns in Northern Europe. The dispersal mechanisms for Fagus are less well-known, despite the fact that during the Holocene it has extended its area of dominance in Europe (and North America) at rates approaching 200–300 m/years (Davis 1976, 1981; Huntley & Birks 1983). Recent studies in North America (e.g., Darley-Hill & Johnson 1981; Johnson & Adkisson 1985; Johnson & Webb 1989) have shown that the blue jay (Cyanocitta cristata) has a significance for long-distance dispersal of seeds from various Fagaceae species. For example, Johnson & Adkisson (1985) measured dispersal of Fagaceae seeds by birds finding that 4 km was far from a maximum dispersal range.

In Northern Europe the jay is probably also important for dispersal of Fagus nuts. Nilsson (1985) has convincingly shown that the jay can act as a vector. Normally the jay prefers acorns, but during years with a restricted Quercus mast, and a normal or rich production of Fagus nuts, the jay also exploits these. During such years the jay transports significant amounts of Fagus nuts, and these can eventually become buried in other forest types.

Picea, by contrast, has light, wind-dispersed seeds. These seeds are normally shed early during spring (February–May), particularly when warm and dry winds prevail. A crusty snow cover (and ice cover on lakes) during this period may also favour long-distance transport of Picea seeds, as they are easily transported even by rather light winds. Picea seeds are therefore normally effectively dispersed to other vegetation types. [Back to contents]

Regeneration and establishment

Fagus has some characteristics that makes it highly competitive compared to most tree species occurring in southern Sweden (Picea being the only other tree species occurring in this area that has similar characteristics). These characteristics are for instance, its ability to withstand shading below a dense canopy of other trees, and that it casts a dense shade itself when it has reached the canopy. These characteristics make it possible for Fagus individuals to grow suppressed for a long time, and wait for better light conditions. Additionally, it also impedes regeneration of light-demanding tree species below its canopy. However, the light conditions at ground level must not be too poor, otherwise its trunk may develop in an almost horizontal direction.

Fagus nuts germinating in stands dominated by other tree species (for example in Picea or Quercus stands) usually seems to have better possibilities to survive as seedlings, than if they have germinated in a pure Fagus stand (Bjerregaard & Carbonnier 1979). The reason for this is unclear but may be an effect of a more intense intraspecific competition within Fagus stands than outside such stands, or specific predators or parasites on Fagus nuts may be absent outside larger Fagus stands (Nilsson 1985). In any case, Fagus nuts germinating below a dense canopy of Fagus have only a small chance of becoming established as seedlings. Fagus nuts germinating on leaf litter also have little chance of survival compared to nuts buried in mineral soil to a depth of a few centimetres (Figs. 2, 3).

Fig. 2. Fagus seedling shortly after germination. When Fagus seeds are buried at a sufficient depth their chances of becoming established as seedlings are better than if they were deposit directly on the ground (Skåne, Häckeberga, 1996-04-27).

Fig. 3. Fagus seeds that are deposited directly on the ground and not covered by mineral soil or leaf litter have little chance of germinating and become established as seedlings (Skåne, Häckeberga, 1996-04-27).

All animals, birds as well as mammals, are important for the regeneration of Fagus if they for some reason bury seeds in the ground (as a food resource for the winter period) or redistribute the top-most layer with leaf litter during their subsistence. Rodents have usually been regarded as having a positive effect on Fagus regeneration, but they probably mostly damage nuts and make them unable to germinate. Birds such as the nuthatch (Sitta europaea) and the marsh tit (Parus palustris) may have a positive effect as they often hide and store seeds for later use. This is also true for the jay, which often hides and buries seeds in places where they can germinate if they remain unused or forgotten. Long-distance dispersal of Fagus nuts by the nuthatch and marsh tit are unlikely as these species have rather small territories (Nilsson 1985).

Fagus is unable to regenerate successfully if gaps exceed a certain size, or if the remaining stand is too open (Bjerregaard & Carbonnier 1979). Then other tree species (e.g., Picea, Betula, Pinus) are able to regenerate successfully depending on the size of the gap. In small gaps chiefly Fagus and Picea are able to regenerate, often in a mixture. In forests dominated by Fagus, its saplings are usually found in small gaps where the ground is covered by leaf litter. Picea saplings occurring in Fagus forests are usually found in larger gaps where the ground is mostly covered by mosses and Vaccinium myrtillus. If regeneration consists of a mixture of both Fagus and Picea saplings, Fagus may finally dominate smaller gaps and Picea larger ones (Nilsson 1902a).

However, Hesselman & Schotte (1906) and Hemberg (1918) suggested that neither Fagus nor Picea had a definite advantage when competing with each other. Hemberg also argued that when the forests were culturally disturbed Picea was usually favoured. In old-growth forests with a mixture of different age-classes and gaps of different size, Fagus instead was favoured. Furthermore, Wibeck (1909) argued when investigating the Fagus forests in western Småland that a condition for successful regeneration of Fagus was an absence of Picea in the local stand. Wibeck also stated: "If one considers, that Fagus in the area sets seed more seldom than Picea, that it has fewer possibilities for seed dispersal, and that its seedlings are much more susceptible to damage by insects, grazing and ground frost than Picea, and that it has an equal competitor in Picea, it is obvious that Fagus has little possibility of winning the long-term competition with Picea."

Picea is also able to withstand dense shading. Separate individuals can survive standing suppressed for decades, and when eventually light conditions are better, as when gaps are created, they quickly use these for regrowth. As a consequence of this factor Picea is a highly competitive tree species in the boreo-nemoral vegetation. However, its shallow root system makes it very sensitive to wind-throw (e.g., Sylvén 1916), and occasional events with strong winds may be an important disturbance factor in stands dominated by Picea (Fig. 4).

Picea regeneration in old-growth forests seems at least in the boreal vegetation zone to be highly dependent on specific microhabitats such as logs, stumps, roots, and elevated hummocks (e.g., Hofgaard 1993; Hörnberg 1995). The major part of the regeneration occur on these habitats even though such microhabitats only cover a small proportion of the forest floor. [Back to contents]

Climatic factors

Fagus forests in Europe are usually found within areas with a precipitation ranging from 450 to 1500 mm/year and a January mean temperature ranging from -5 to 10°C, and a July mean temperature ranging from 10 to 20°C (SOU 1971). Hjelmqvist (1940) has summarised the factors important for the distribution of Fagus. He argues that the northern distribution limit purely is caused by its sensitiveness to frosts, and furthermore, that the distribution limits in the north-west are controlled by low summer temperatures. The eastern distribution limits are controlled by low precipitation (drought) and unfavourable temperature conditions. Additionally, a sufficient chilling period is important for many tree species in inducing budburst in the spring. Sykes et al. (1996) have shown that Fagus is especially sensitive to the length of the chilling period.

The young Fagus sapling is very sensitive to frost, and some sort of shelter is usually needed for successful growth (Bjerregaard & Carbonnier 1979). Fagus is much more sensitive to periods with frost during earlier parts of the growing season than for instance, Quercus, as its buds and shoots proliferate comparatively early during the spring. Its florescence may sometimes be severely damaged by frosts, and as a consequence of this the production of nuts is reduced. The production of Fagus nuts is highly dependent on the frequency and duration of frost periods during the spring. Periods with frost may also damage germinating seeds and seedlings. These seedlings may also be sensitive to periods with drought during the spring and summer (SOU 1971).

However, the occurrence of Fagus in southern Europe indicate that it is somewhat more tolerant to drought than many other tree species occurring in northern Europe. In spite of this, Fagus does not grow further to the east than western Russia. Other less drought-tolerant trees, such as Tilia, have a more widespread occurrence to the east than Fagus. This may indicate that drought is not the sole factor restricting its occurrence in the eastern part of Europe, but that low winter temperatures may also influence its distribution. It is possible that Fagus is sensitive to extremely low temperatures that occur in a pronounced continental climate (Huntley 1988). Episodes with extremely low temperatures (below -30°C) during the winter may sometimes create frost-cracks in the wood.

Tallantire (1972a) has summarised the climatic factors that are probably important for the distribution of Picea in Europe. He has particularly discussed phases in its life cycle that may be influenced by climate. Pollination and seed maturation are influenced by temperature. The pollination of Picea is normally completed in May or early June. Its seeds then mature in the cone during the summer, and this maturation needs a mean temperature above 9.5°C. A mean July temperature above 19°C may damage its seeds. Picea seeds are normally shed during the early spring, and they require a temperature of at least 7°C to germinate. To be able to survive as seedlings, and later as saplings, they need at least 200 mm of precipitation during the period from May to July. These saplings also need to be covered by a sufficient snow cover during the winter. The mean temperature during the growing season is also important, particularly to ensure that they become sufficiently resistant to frosts during the following winter. Picea seems to require a period with winter rest lasting about four months, and the mean temperature during this period needs to be below -1°C. Picea rarely grows well in regions with mild winters, such as in maritime climate types. [Back to contents]

This section of the synthesis gives a brief summary of methods used when sampling the selected sites and analysing the recovered peat profiles. For site-specific details, see the respective papers (Appendices I–IV). [Back to contents]

Selection of sites and field work

The general background to why the sites used in this thesis were selected is described in the introduction. Maps and literature have also been used for the selection of areas to search for small, wet hollows. The search for suitable sites has been somewhat biased to Fagus stands. Many interesting forest stands have been visited, particularly in the northern part of the Fagus-Picea and Fagus outpost region, but unfortunately very few of these held small, wet hollows potentially useful for this thesis. Fagus stands, particularly in the northern part of its present distribution in Sweden, are generally located in topographical settings were organogenic sediments rarely accumulate, i.e., on well-drained ground, or slopes.

From the experience obtained through work with five small hollows at four sites, some general conclusions can be drawn as guidelines for future work in this area. These guidelines are based on experiences drawn from the field (through coring and visits at different sites) and the laboratory (through pollen analysis, dating, etc.). The ideal small site is of course a closed-canopy hollow with a diameter smaller than 2 m, but this type of site is unfortunately very rare. Normally one will have to use somewhat larger (and less ideal) sites. Such hollows are relatively common in southern Sweden, particularly in areas with nutrient-poor bedrock or soils.

The thickness of organogenic deposits in the hollow is also important. This property is easy to probe in field and may be a first indication if the site is useful or not. Shallow deposits (10–20 cm) rarely cover periods long enough to reveal events occurring more than some hundred years back in time. A peat thickness exceeding 50 cm is probably needed when trying to detect events occurring more than 500 years back in time. Often a peat deposit around 1 m is sufficient to study long-term dynamics over several thousands of years. However, sediment accumulation is highly variable in small hollows and depends on site-specific factors that are not easy to understand (Table 4). To make estimates of the time covered by a profile in the field is never easy, but the degree of humification may give some hints. Profiles containing poorly-humified peat throughout the whole core never span as long a time period as profiles containing well-humified peat. If the intent is to study vegetation dynamics during the Early Holocene somewhat larger basins are probably needed, as small closed-canopy sites seldom span the entire Holocene. However, there are reports of forest hollows spanning the entire Holocene, and even reaching back into the Late Glacial (e.g., Bradshaw 1981b; Sugita et al. 1994), and such hollows may also be found in southern Sweden. If larger basins are considered, and these have organogenic sediments with a thickness of around 1.5–2 m, they may span the entire Holocene. But these basins then have a considerably larger pollen source area.

Table 4. The approximate ¹⁴C-age at 50 cm and 80 cm respectively, below surface in the studied profiles. These dates have been approximated from levels dated by conventional radiocarbon dating.

All sites have been sampled with a Wardenaar corer (Wardenaar 1987), except site B at Bocksten which was sampled with a standard Russian corer (diameter: 5 cm). All sites have been sampled where the sediments were deepest to obtain profiles that cover sufficiently long time periods. For details and site descriptions, see Papers I–IV. [Back to contents]

Laboratory methods, pollen and charcoal analysis

All the peat profiles recovered have been stored at +8°C before pollen and charcoal analysis and radiocarbon dating. The profiles (monoliths) sampled with the Wardenaar corer were divided into two parts. One smaller part (a 9 cm² section) was reserved for pollen and charcoal analysis, and the larger part for radiocarbon dating. The section selected for pollen and charcoal analysis was frozen at -20°C, and cut into thin subsamples (usually 3–5 mm thick depending on the sediment) using an electric kitchen slicing machine with a serrated rotating blade. Small samples (usually 0.3–1.0 cm³ depending on the pollen concentration) were taken from the centre of these subsamples for pollen analysis. To avoid contamination during cutting, the blade in the slicing machine was carefully cleaned before each new subsample was cut. The surface layers of these thin subsamples were also removed before pollen preparation to avoid contamination during handling and storage of the samples. Lycopodium tablets were added to the subsamples (Stockmarr 1971). The precise volume was measured by water displacement.

The samples were prepared for pollen analysis following standard methods (Berglund & Ralska-Jasiewiczowa 1986). Microscopic slides were prepared from the residue and scored for pollen (at least 1000 pollen in each subsample when this was possible to achieve) and microscopic charcoal (25–250 µm). The pollen key in Moore et al. (1991) was used to determine some critical pollen taxa. Pollen slides in the reference collection at the Laboratory of Palaeoecology, Lund University, were also used to check some pollen types. Macroscopic charcoal was counted from samples that would not pass through a 250 µm mesh during the preparation procedure for pollen analysis. These large charcoal fragments are most likely derived from local fires and have not been transported very far (Patterson et al. 1987; Clark 1988).

The core from Bocksten that was sampled with a Russian corer (site B) has only been used for pollen analysis (the entire core has been used). Thin subsamples (4–5 mm thick) have been cut from this core. Small (0.4–0.7 cm³) samples were taken from the centre of these subsamples for pollen analysis. These samples have been prepared for pollen analysis in a standard manner (see above) except that no Lycopodium tablets were added.

Bulk sections of the remaining peat monoliths (usually 1–2 cm thick) were submitted for conventional radiocarbon dating. Additionally, one AMS dating was made on seeds from one profile (Bocksten A) to check the reliability of the conventional radiocarbon dates, and three AMS-datings were made on bulk peat samples from one profile (Bocksten B). [Back to contents]

Chronology

At each site the chronology was established essentially on the basis of ¹⁴C dates of bulk peat samples. In a few cases improbable dates were rejected (III). However, root penetration can be a serious problem when using bulk samples for dating, particularly when studying peat sections from small sites where trees may have grown nearby. If roots became incorporated in the dated material, bulk samples certainly will provide too young dates. However, comparison of AMS and bulk dates in this project suggest that this problem has not seriously affected the dates used in this study (Appendices I–IV).

Hiati may be another problem when using sections from small hollows that may be very sensitive to hydrological changes. Such changes may affect the deposition rate. Hiati are usually very difficult to detect, but a way to evaluate the possible occurrence of hiati can be the use of total pollen concentration and accumulation rates (cf. II). Rapid changes between consecutive subsamples can be explained by nonlinear age-depth relationships, and therefore, by changes in accumulation rates or the existence of hiati.

The smooth forms of the age-depth curves used for this study (Appendices I–IV) give no cause to suspect the reliability of the dates. In spite of the two major sources of errors, I assume that the problem of root penetration is minor in all studied sites, and that possible hiati were of short duration. [Back to contents]

This section of the synthesis gives a brief summary of the main results reported in Appendices I–IV that may be relevant for the discussion in the next section. Percentage pollen diagrams with all identified taxa were not presented in the original papers. They are included in this thesis as plates (Plates 1–5). [Back to contents]

Paper I

Björkman, L. & Bradshaw, R. 1996: The immigration of Fagus sylvatica L. and Picea abies (L.) Karst. into a natural forest stand in southern Sweden during the last 2000 years. Journal of Biogeography 23, 235–244. (see also Plate 1)

This paper is based on a palaeoecological study of a small nature reserve (Siggaboda nature reserve) situated near Siggaboda in southernmost Småland (Fig. 1). This site is located in the southern part of the area where the present distribution limits of Fagus and Picea overlap each other. Fagus has probably been present in the region at least for 2000 years, as indicated by regional pollen diagrams. According to historical documents, Picea became established in the region during the last 200–300 years.

The studied site is co-dominated by Fagus and Picea, either growing in mixed stands, or in more or less pure groups (Figs. 5, 6). The forest reserve is known for its high biodiversity, the lichen flora (Fig. 7) and beetle fauna are particularly diverse, and typical for undisturbed forests. Moreover, there is minimal evidence of human impact in the reserve. There are no stone mounds or other signs of agricultural activities that are common elsewhere on the landscape. Isolated tree stumps at the edge of the reserve are evidence for some selective felling during this century. The high diversity in the reserve stands in great contrast to that in the forests surrounding the reserve, which are managed and dominated by more or less pure Picea or Pinus plantations. The occurrence of several threatened species within the reserve has led to an opinion among biologists that the present forest type has a long unbroken continuity.

Fig. 5. Pure Fagus stand at Siggaboda nature reserve (Siggaboda, 1996-06-20).

The aim of this study was to date the establishment of Fagus and Picea at this site, and to reveal the origin of the present forest type. Additionally, the aim also was to test the hypothesis that the present forest type had a long unbroken continuity, as suggested by the occurrence of specific lichens and beetles.

The pollen record of a small forest hollow (Fig. 8) located within the reserve indicates that a fire c. 950 BP facilitated the establishment of Fagus. Another fire c. 350 BP finally triggered a succession which led to a co-dominace of Fagus and Quercus in the forest. During the period between 350 and 150 BP these trees dominated the reserve. At about 200 BP the Quercus population started to decline rapidly. No Quercus individuals remain in the present forest. The absence of old survivors may suggest that it was removed by selective felling. At about 200 BP Picea became established in the forest. The oldest present Picea individuals in the forest are about 200 years old, and they probably represent the first Picea regeneration. The present forest composition with co-dominance of Fagus and Picea developed during the last 200 years.
 

The pollen record shows high tree pollen percentages throughout the whole period covered by the sampled profile (around 2800 years). These high values are evidence for a long continuity of forest cover and forest processes, and may explain the present high biodiversity. However, this continuity does not imply stability as the site has experienced at least two fires and undergone a revolution as regards forest composition. The present forest type, looking virtually undisturbed and showing old-growth characteristics, apparently has a recent origin. [Back to contents]

Paper II

Björkman, L. 1996: Long-term population dynamics of Fagus sylvatica at the northern limits of its distribution in southern Sweden: a palaeoecological study. The Holocene 6, 225–234. (see also Plate 2)

This paper is based on a palaeoecological study of a small outpost Fagus stand (Mattarp) in northern central Småland (Fig. 1). This site is located close to Fagus' northern distributional limits in southern Sweden. Regional pollen diagrams recently published from this outpost area show a low Fagus representation (maximum Fagus pollen percentages rarely exceed 1%). These diagrams may indicate that Fagus has been present in the region for as long as 1000–1500 years, but it has probably always been rare, except for some small and scattered stands. Many of these outlying stands have been regarded as planted, as some of them are located close to old manor-houses or large farms. The conclusion that can be drawn from the literature and from regional-scale pollen studies is that the history of Fagus in these outlying stands is not well-known.

The aim of this study was primarily to date the establishment and expansion of Fagus and Picea, but also, if possible, to deduce if the present Fagus stand was expanding or not. The studied site lies within a climatically less favourable area called the Småland Uplands, an elevated area in southern Sweden where climatic conditions resemble those further to the north (Sveriges Nationalatlas 1995; see also Lagerås 1996). If Fagus grows well in this area, climate is certainly not a limiting factor for its distribution.

Fagus stand (Fig. 9) indicates that Fagus establishment at this site. The complex cultural landscape in southern Sweden may have favoured The pollen record of a small wet hollow closely connected to the present Fagus became established at about 400 BP. The local establishment coincides with a phase of woodland clearance (starting about 400 BP) and a former system of land-use practised in the area (Fig. 10). Local factors seem to have controlled the immigration and establishment of Fagus at stand-scale at this site. Regional climate had probably been favourable for Fagus expansion for a considerable period, and cannot be regarded as the direct cause for establishment at this site. Furthermore, a phase with woodland clearance occurred at 1300 BP, but Fagus did not invade the site at this time. Presumably dispersal factors (and chance) influenced the exact timing of Fagus establishment in many areas, particularly when there had for some reason been a temporary reduction in cultural activities, following periods of stronger human influence.

The pollen record also indicates that Picea became established locally around 800 BP, but an expansion did not occur until about 400 BP, coinciding with the establishment of Fagus. Picea expansion seems to have been favoured by human influence at this site. The local establishment of Picea is largely in accordance with recently published dates from regional-scale sites around 15 km to the north of Mattarp, where the Picea expansion has been dated to around 900–1200 BP.

Fagus pollen influx values have increased more or less steadily since the time of establishment (cf. Fig. 20). It seems likely that the Fagus stand has continually increased in size, and probably would continue to do so if this was not prevented by human activities. [Back to contents]

Paper III

Björkman, L. 1997: The history of Fagus forest in southwestern Sweden during the last 1500 years. The Holocene 7, 419–432. (see also Plate 3)

This paper is based on a palaeoecological study of a forested area near Bocksten in central Halland (Fig. 1). This area is located in the "Central Halland Fagus Forest Area," an area which presently has abundant Fagus populations (Lindgren 1969; SOU 1971). It lies in the north-western part of the Fagus forest area of southern Sweden, just outside the historical western natural limits of Picea. However, Picea plantations are today common in this area. The present vegetation in the Bocksten area is dominated by intensively managed Picea and Fagus stands.

The aim of this study was to investigate the pre-Fagus vegetation and reveal the origin of the present Fagus forests (and why these consist of almost pure Fagus stands). Furthermore, the aim was also to detect stand-scale differences between nearby stands. In the studied area many suitable small wet hollows occur, which means that studies of stand-scale differences are possible. To achieve this last aim two adjacent hollows (lying about 50 m apart) were selected for palynological studies.

The pollen records of two small adjacent hollows (Figs. 11, 12) indicate that the pre-Fagus forests were dominated by a rich nemoral type with many woody taxa present. Quercus, Tilia, Alnus, and Corylus were important components of this vegetation type. Fagus became established in this area around 1500–1400 BP, and rapidly became a forest dominant. The rapid expansion was probably facilitated by a slight human influence. The rather pure Fagus stands found today in the area (Fig. 13) are likely a product of recent human activities.

Picea become established late at this site, probably during the latter part of the 19th or during the beginning of this century. This expansion was most likely a consequence of land-use changes occurring when a settlement east of the studied site was abandoned. Picea was planted in this area, or self-sown from nearby plantations.

The high pollen percentages for Tilia at site A until about 200 BP is remarkable. In most regional-scale pollen diagrams Tilia shows a gradually decreasing trend throughout the last 3000–4000 years, and this decrease is also indicated in regional pollen diagrams from Halland. It is obvious that an abundant Tilia population occurred locally in the Bocksten area until 200 BP. Regional pollen diagrams may not be sensitive enough to detect such populations as Tilia pollen grains normally are not well-dispersed outside Tilia stands. There is also a significant difference in Tilia representation between the studied profiles. In profile B the Tilia curve is more classical and shows a gradual decrease beginning before the establishment of Fagus. The presence of Tilia rapidly declined in this stand when Fagus became established. The situation was however, different at site A, where Tilia seemingly grew in a mixture with Fagus, Quercus, and Corylus, until this stand was cleared about 200 BP. The explanation for this strong difference between two nearby stands may be differences in cultural influence. Cultural activities may have been stronger at site B (Fig. 14) as this site was located closer to a settlement than site A.

The two profiles studied also revealed stand-scale differences in composition of vegetation and human influence. At site A Fagus grew in a mixed stand together with Tilia, Quercus, and Corylus, for nearly 1200 years, until the stand was suddenly cleared about 200 BP. At site B Fagus became a dominant shortly after its initial expansion, most likely as an effect of human influence. A pre-requisite for mixed Fagus stands probably is a low or absent human influence.

The forest type we today find in the studied area has little resemblance to the type that existed during earlier periods in this area. A marked human influence during the 18th and 19th centuries, and forestry management during this century, have contributed to the shift of forest type. Today a boreo-nemoral vegetation type occurs in this area, but geographically, as well as climatically, a nemoral type could easily grew there. [Back to contents]

Paper IV

Björkman, L. 1997: The role of human disturbance in the local Late Holocene establishment of Fagus and Picea forests at Flahult, western Småland, southern Sweden. Vegetation History and Archaeobotany 6, 79–90. (see also Plate 4)

This paper is based on a palaeoecological study of a small Fagus stand at Flahult in western central Småland (Fig. 1). This site is located in the northern part of the area where the present distribution limits of Fagus and Picea overlap each other. Today Fagus stands are rather common in the Flahult area, but these stands are mostly young, as individuals with small diameters are numerous (Fig. 15). Moreover, these stands also consist of almost pure Fagus vegetation. Historical documents (e.g., maps) indicate that Fagus was more abundant in this area in the recent past. For instance, a map dating from 1799 indicates a large Fagus area some kilometres to the south of the studied site. Later on, this Fagus area was converted to heathy grazing land. Wibeck (1909) has published a map depicting the past and present distribution of Fagus in western Småland. The studied site at Flahult is probably marked on this map, indicating that this site at least was a Fagus stand during the beginning of this century.

The aim of this study was to date the local establishment of Fagus (and Picea) and to investigate the origin of the present composition of the stand. The pollen record of a small alder carr located centrally within the present Fagus stand (Fig. 16) revealed that Fagus became established in a cultural landscape about 900 BP. This establishment was probably connected to a phase with considerably increased human influence, i.e., an expansion of pastures creating a semi-open landscape. Apparently this change in land-use favoured Fagus invasion of this site. Fagus may have been present in the region for some hundred years before the local establishment at Flahult. Local-scale factors seem to have been important for the timing of establishment at this site.

The Fagus dominance in the present stand seems to have a rather recent origin, as Fagus pollen percentages and influx values increase considerably only during the last 50–100 years. The modern composition and stand-structure are probably an effect of reorganisation of land-ownership and declining human activities at the end of the last century. Fagus could respond positively to these changes, and establish an almost pure stand during this century. However, selective removal of trees has occurred during this century, and this probably also facilitated the development of the present almost pure Fagus stand.

There is no indication in the pollen diagram that Fagus prior to the present century was more abundant in the Flahult area than today. Pollen influx data strongly suggest that it was only during this century that it could expand and eventually begin to dominate the local stand. The age structure of the majority of the present stands in the Flahult area also indicates that these stands originated relatively recently.

Picea may have became established in the area c. 100 BP, but it did not gain any importance in the local vegetation. Today only scattered Picea individuals occur in the Fagus stand surrounding the sampling point, and Picea was certainly not more important in the recent past. [Back to contents]

In this section selected results and problems relevant for the scope of this thesis are discussed in a broader sense. More detailed discussions concerning site-specific issues and problems are found in the respective papers. [Back to contents]

Fagus

Is it possible to detect early occurences of species by pollen analysis

This question is a general problem when trying to reveal the earliest history of any species in a region. If low pollen percentages around 0.3–0.5% are difficult to interpret, single pollen grains are even more delicate to evaluate. To distinguish between a rare but local occurrence, and a more abundant but regional presence, is a classical palynological problem, as these two situations may be identically reflected in the pollen record. It will probably never be possible to convincingly prove an early presence of a taxon with pollen analysis alone. Conventional pollen analysis can only record when a taxon starts to be ecologically important in an area (Bennett 1986). It can rarely tell when a taxon starts to occur with a low population abundance, as is most likely when a species migrates into a new area.

This discussion is also applicable to the question of what single pollen grains of Fagus really signify. Single Fagus pollen grains are normally found in profiles before the expansion phase of Fagus. The first single Fagus pollen grains may occur in southern Sweden as early as around 6000 BP. It is unlikely that Fagus at that time even occurred as single individuals scattered in the forests, as it had hardly started to invade the lowland areas of Northwest Europe (Huntley & Birks 1983; Huntley 1988). Single grains occurring around 5000–4000 BP are more difficult to interpret, as it is possible that single Fagus individuals then occurred in southernmost Sweden. Lagerås (1996) argues that single Fagus individuals had already appeared in central northern Småland before 3000 BP. By this time Fagus probably had begun to increase in southernmost Skåne (Berglund et al. 1991), about 350 km south of central northern Småland. Regional pollen diagrams only show a scant Fagus representation in northern Småland, and these percentages rarely exceed 1% even when the Fagus curve is at its maximum. These diagrams may indicate that Fagus started to appear in this area around 1500–1000 BP, but if it was present earlier is questionable.

Studies of terrestrial plant macrofossils might help to shed some light on this general problem (e.g., Kullman 1995), but the dispersal, deposition, and preservation of plant macrofossils are probably even more influenced by chance than pollen grains. At Flahult (IV) a small alder carr is located centrally within a Fagus stand. The pollen record of this site shows comparatively low Fagus pollen percentages throughout a long period from the local establishment of Fagus at about 900 BP until c. 50 BP, when percentages began to exceed 10%. These low percentages for around 850 years indicate a likely local presence of Fagus, but it is difficult to discuss if it was abundant or not during this period. However, macroremains of Fagus (part of a nut and bud scales) have been found in one sample from the profile dating from about 500 BP (G. Hannon, pers. comm.). These macroremains certainly prove that Fagus was present in the local vegetation even if the pollen percentages were comparatively low. [Back to contents]

The establishment of Fagus in Sweden

Viewed on a continental scale the migration pattern of Fagus shows a striking correspondence with modelled climatic change over the millennia (e.g., Huntley 1988; Huntley & Webb 1989; Huntley et al. 1989). However, at finer spatial and temporal scales this correspondence is not obvious. All migration means invasion of species into a pre-existing vegetation, and this invasion takes place at stand-scale. At this scale there are obviously factors other than climate that are crucial for the establishment of a species (e.g., disturbance, seed dispersal, human activities, etc.), but climate may of course also affect these factors.

One way to evaluate whether climatic change was important for the establishment of Fagus in southern Sweden is to investigate if its establishment shows a regional coherence or not. If dates for its establishment within a region are very similar, climatic change was probably a controlling factor. The existence of a regional coherence or not for Fagus has been tested by plotting all available data (cf. Table 2) for its establishment in southern Sweden (Fig. 17). From this figure no clear regional coherence in establishment date is apparent. This may imply that climatic change was not the limiting factor for the recent establishment of Fagus in southern Sweden.


Fig. 17. The establishment and expansion of Fagus at local and regional pollen sites in southern Sweden. Sites A-Y are described in Table 2. The first date refers to the establishment and the second to the expansion (see footnote 1 and 2 in Table 2 for explanation). The abbreviations bef. and aft. mean before and after.

A study of a Fagus outpost stand (II) lying close to its northern distributional limits revealed that climate was certainly not limiting for the establishment at this site (cf. Fig. 17). This may also be true for all other sites studied within this project (I, III, IV). A comparison of the pollen curves for Fagus from each studied site with that of the nearest available regional pollen sites shows that Fagus occurred regionally in these areas before it became locally established (Fig. 18). The difference in time between regional and local establishment seems to be greatest in the southern part of the studied area, where Fagus has been present for as long as 1500 years before the establishment at the studied sites (Table 5). In the northern part the difference is smaller, but Fagus may have been present in this area for as long as 1000 years before its local establishment at Mattarp (II, cf. Table 2). However, if the local establishment is compared with the expansion in regional diagrams, i.e., the time when Fagus probably starts to be an ecologically important tree regionally, the difference is smaller. In this case the difference is only c. 300–500 years, and at one site (III) the regional expansion largely seems to coincide with the local establishment. In conclusion, factors other than climate must have controlled the local establishment at the studied sites.

Davis (1987) has presented two alternative models for the migration of tree species: A) a broad continuous front, and B) a discontinuous front with outlying isolated populations. The present day distribution of Fagus in southern Sweden (Lindquist 1931, 1959; SOU 1971) suggests a type B migration, and this model probably applies to its past distribution. This type of migration means that the landscape becomes infilled by dispersal from small isolated outposts that act as seed sources. The timing of stand-scale establishment is then certainly influenced by site-specific factors and chance. In this case, the difference in time between the establishment at a particular stand and the first regional presence may be large. If the migration front had been of type A, the difference between local and regional establishment would have been insignificant.

The dispersal of Fagus nuts is probably highly influenced by chance, particularly long-distance dispersal. Within the stand and its nearest surroundings dispersal of nuts is probably effective, simply as a consequence of diffusion laws. At greater distances dispersal of Fagus nuts then becomes more influenced by chance, and this may explain why Fagus outposts are rare and isolated in the landscape (cf. Fig. 1). [Back to contents]

The role of disturbance and cultural activities

In southern Sweden humans have influenced the vegetation to some degree for over 5000 years (e.g., Berglund 1991), but this influence has been highly variable temporally, as well as spatially, creating the complex cultural landscape we still observe in this area. Very few, if any, areas in southern Sweden are untouched by cultural activities, especially in areas where conditions for agriculture have been suitable. A first expansion of open and semi-open vegetation began in southernmost Skåne c. 2800 BP (e.g., Regnéll 1989; Gaillard et al. 1991b), but later in central northern Småland (Lagerås 1996). The degree of openness varied considerably between regions. For instance, in southern Skåne an open landscape was already created during the Bronze Age, and the landscape has been continuously open since then. In central northern Småland forests have always been present, even during periods experiencing a more intensive land-use. These differences between regions were certainly also important for the invasion of Fagus.

The type of vegetation being invaded by a migrating species is important, and a species may react differently depending on the type involved. Clearly, there must be a significant difference between a species invading an unaffected forest with natural processes prevailing, and one entering a semi-open landscape with cultural activities controlling the vegetation. Fagus seeds are highly dependent on ground disturbance for successful establishment (e.g., Watt 1923; von Röhrig et al. 1978; Bjerregaard & Carbonnier 1979), and an undisturbed forest would consequently be able to resist Fagus invasion for some time. However, disturbance also occurs in natural forests untouched by humans. Natural disturbances such as fires, wind-throw, pests, etc., may create gaps and seed beds suitable for Fagus regeneration. A semi-open cultural landscape may provide even better possibilities for Fagus to become established, as cultural activities may create conditions particularly suitable for its regeneration. In least at two of the studied sites (II, IV) cultural activities obviously created conditions favourable for Fagus establishment. Cultural activities may also have played a role for its establishment at the other studied sites (I, III). At Siggaboda (I) fires probably facilitated Fagus establishment and expansion. The cause of these fires is not known, but they may have been intentional, or unintentional, effects of cultural activities.

To evaluate the importance of cultural activities for the establishment of Fagus at the studied sites a comparison was made between the Fagus pollen curve and curves for selected human impact indicators (Fig. 19). From this figure it is suggestive that cultural activities played a major role in the establishment of Fagus at Mattarp (II) and Flahult (IV). At these sites the curves for meadows, pastures, and cultivated land all increase significantly at about the same time as Fagus became established. At Siggaboda (I) and Bocksten A (III) cultural activities may have favoured Fagus establishment, but the evidence for this is weaker than for Mattarp and Flahult.

Fig. 19. Comparisons between Fagus and Picea curves and curves for selected human impact indicators for each studied site. The curve for meadows has been constructed by adding the pollen curves for Poaceae and Plantago lanceolata. This curve is believed to reflect the area with fresh meadows and pastures that were grazed and/or mowed. The curve for pastures is constructed by adding the pollen curves for Juniperus and Calluna. This curve is believed to reflect the area with dry pastures. The curve for cultivated land has been constructed by adding the pollen curves for Poaceae >40 µm, Secale, Cannabis-type, Centaurea cyanus, and Rumex acetosa/R. acetosella. This curve is believed to reflect the area with cultivated land and associated ruderal land. This grouping of human impact indicators is largely comparable to that used within the Ystad project (Berglund 1991), and it takes into account the results of studies of pollen/land-use/vegetation relationships in modern analogues of traditional cultural landscapes in southern Sweden (Gaillard et al. 1992, 1994).

There are today an overwhelming number of studies published which indicate a strong relationship between cultural activities and Fagus expansion. Recent studies, for instance in Denmark and Sweden, have shown how Fagus populations could expand more or less quickly in areas previously disturbed by human activities (e.g., Iversen 1969, 1973; Aaby 1983, 1986, 1988; Andersen et al. 1983; Andersen 1984, 1988; Gaillard 1984; Regnéll 1989; Berglund et al. 1991; Gaillard & Göransson 1991; Gaillard et al. 1991a; Lindbladh & Bradshaw 1995; II, IV). Moreover, there are also several studies indicating that Fagus had difficulties with establishment in areas with undisturbed forests (Iversen 1973; Aaby 1986; I), or in areas with a strong cultural influence (Aaby 1986, 1988; Bartholin & Berglund 1992; Odgaard 1994). The complex, shifting mosaic patterns in the vegetation of the south Swedish cultural landscape probably created many opportunities for Fagus to invade into new areas, particularly when there has been a reduction or change in human influence following a period with stronger impact (e.g., Gaillard et al. 1991a). [Back to contents]

The pre-Fagus vegetation

At Mattarp (II) and Flahult (IV) Fagus apparently became established in a cultural landscape that had already lost its natural composition and structure. At Mattarp the vegetation preceding the clearance event at about 400 BP that probably facilitated Fagus invasion at this site, consisted of a mixture of Betula,Corylus, Quercus, and Picea, and this type was certainly controlled, or at least largely influenced by cultural activities, most likely by grazing. At about 400 BP this forest was locally cleared and replaced with open arable land and pastures. Fagus (and Picea) began a significant population increase in association with this phase. The first indication of cultural activities in this area is the regular occurrence of single pollen grains of Plantago lanceolata dating to c. 3000 BP, but these grains do not necessarily indicate activities in the local forest stand. The seemingly untouched forest prior to this date was dominated by deciduous trees, in which Quercus, Ulmus, Tilia, and Corylus were important. The composition changed later on, and Betula in particular increased, while Quercus and Tilia declined. This change in composition was likely influenced, or controlled, by human activity (e.g., grazing).

At Flahult (IV) Fagus also became established in a cultural landscape. The unaffected forest was in this area dominated by a mixture of Quercus, Tilia, and Corylus, but this forest type had obviously little significance for Fagus invasion, as it had largely disappeared when Fagus became established. Fagus establishment in this area probably took place on pastures, or in stands dominated by Betula, Quercus, and Corylus, and this vegetation was certainly controlled by cultural activities. At Siggaboda (I) and Bocksten (III) the local stands seems to have been more or less unaffected by cultural activities prior to the establishment of Fagus, at least no semi-open vegetation was present locally when Fagus became established. At Siggaboda Fagus became established about 950 BP, and its local establishment was probably facilitated by a fire (natural?). The local stand was on well-drained ground prior to this event dominated by Quercus, but Corylus and Tilia were also present. At Bocksten the pre-Fagus vegetation was dominated by a rich nemoral type dominated by Quercus, Tilia, Alnus, and Corylus.

At most sites Tilia was once abundant on the landscape, but its abundance diminished regionally, as well as locally, well before the immigration of Fagus. However, at Bocksten, especially at site A, a significant Tilia population survived until 200 BP (III). This late presence of a large Tilia population is remarkable and stands in striking contrast to other regions. In most regional-scale pollen diagrams one often finds a gradually decreasing trend for Tilia. This decrease often starts around 3000–2500 BP, but may sometimes begin much earlier (e.g., Huntley & Birks 1983; Aaby 1986). In some areas this decrease even started before 4000 BP (e.g., Lagerås 1996). It has been argued that climatic change played a role in this decrease, but human interference is probably more plausible (e.g., Turner 1962).

Even if Tilia decreased regionally in Halland, as indicated by Digerfeldt (1982), it obviously still had relatively large populations locally, particularly in areas with little human interference (III). Conditions for Tilia regeneration were probably not optimal during the period it co-existed with Fagus at Bocksten, but it was however, able to maintain its abundance thanks to its effective vegetative regrowth. The grazing pressure was probably not intense at this site, otherwise Tilia would have gradually declined.

Tilia individuals can survive for centuries in areas with unfavourable climate (Pigott & Huntley 1978). Individual Tilia trees can reach a notable age. Around 300 years is perhaps not a maximum age for a single stem, and an age of 500–1000 years is possible for single individuals (Pigott 1989). It can also be difficult to get rid of Tilia trees. Felling alone is not enough as it easily regenerates vegetatively by sprouting from the remaining stool. Felling in combination with intense grazing is probably needed to locally kill off this species (Aaby 1986; Pigott 1989; Rackham 1980). Human disturbance of the forests during the Bronze Age and Early Iron Age was certainly a major reason why Tilia forests became fragmented.

The composition and structure of the pre-Fagus vegetation apparently varied between the studied sites. On some sites Fagus invaded a vegetation that was semi-open and obviously controlled by cultural activities (II, IV), but on other sites it invaded a vegetation that had a more or less natural composition until an eventual disturbance event (natural or human induced) triggered the establishment of Fagus (I, III). These apparently unaffected forests were dominated by Quercus, Tilia, and Corylus, but the specific structure and composition probably varied from stand to stand dependent on climate, soils, competition, disturbance, etc. It is tempting to postulate that this forest type once covered vast areas of southern Sweden before Fagus (and Picea) immigrated (Björse et al. 1996). These forests may have been somewhat influenced by human activity, e.g., grazing, but this influence is difficult to detect or prove. [Back to contents]

Is Fagus still migrating northwards in Sweden?

Huntley & Birks (1983) stated that Fagus 'may not even now have reached its present potential northern limit, except perhaps in Sweden.' One way to answer the question whether or not Fagus still is migrating is to study outlying stands. Mattarp (II) is such a site. At this site the Fagus stand has increased more or less continually in area since the time of establishment (at 400 BP) (Fig. 20). Climate does not obviously hinder Fagus regeneration at this site. It is tempting to postulate that the northern distribution limits of Fagus in Sweden still represent an active front, and that outlying stands are acting as 'infection centres.' It therefore seems as if present day forest management and land-use is the limiting factor for Fagus expansion in its outpost area in northern Europe. [Back to contents]

Fig. 20. Influx values for Fagus have increased more or less steadily at Mattarp since it became established at c. 400 BP. Periods are described in Appendix II.

Picea

The establishment of Picea in southern Sweden

Picea invaded southern Sweden from the north during a period when the cultural landscape had already been evolving for some time (cf. Fig. 21), and the contemporary vegetation was largely influenced by cultural activities. This circumstance certainly affected its migration and establishment.

When Picea entered northern Småland it found a semi-open landscape where the original forests had largely been destroyed and replaced with pastures and stands dominated by successions of Betula, Quercus, and Corylus. Lagerås (1996) has for instance, found that Picea invaded central northern Småland c. 1200–900 BP, and that it rapidly became an important tree in the landscape. He also argues that the Picea invasion of northern Småland was not triggered by any local change in land-use, but rather by its regional migrational speed. Picea is a dominant species with an effective seed dispersal, and the relatively open and probably grazed Betula-Quercus forests in the area were not particularly resistant to Picea invasion. If Picea instead had entered an area with forests unaffected by human activity, these would probably have proved more resistant to invasion. 

Picea is also very shade-tolerant which means that it easily invades stands dominated by other tree species. An intensive grazing regime may not affect Picea, as grazing animals are usually very selective and normally avoid Picea (Jacobsen 1973; Ahlén 1975). However, it is somewhat sensitive to disturbance, particularly to fires and storms. A disturbance regime with frequently returning fires may affect and control Picea expansion (Bradshaw & Hannon 1992; Bradshaw 1993). [Back to contents]

Picea invasion at stand-scale

The timing of the local Picea establishment (cf. Figs. 18, 21; Table 3) mostly seems to be controlled by its migration rate, i.e., it became established locally when its front reached the studied sites. Picea invaded Mattarp (II) in central northern Småland at c. 800 BP. It was probably relatively rare in the local forest stand as its pollen percentages are comparatively low. An expansion in the local stand did not occur until 400 BP, coinciding with the local establishment of Fagus and a phase with increased land-use and a clearance of the local forest stand (Fig. 19). The open conditions that existed during this period appear to have been suitable for Picea expansion.

Fig. 21. The expansion of Picea at local and regional pollen sites in southern Sweden. Sites A–Y are described in Table 3. Generally c. 2–5% Picea pollen is reached at the indicated date. Picea probably started to become ecologically important at this time, but single individuals may have been present earlier. Abbreviations Rec. and Rec. plant. mean recently and recent plantations. Recently means in this respect that the expansion most likely occurred during this century.

Picea invaded Siggaboda (I) in southernmost Småland at about 200 BP. Its rapid increase in pollen percentages began at c. 150–175 BP. The oldest living Picea individual in the forest stand is about 200 years of age, which indicates that there is a lag between the local establishment and the start of pollen production. The population increase for Picea was very rapid and within c. 50 years it virtually co-dominated the local forest stand together with Fagus. The local expansion of Picea coincides with a period of apparent increase in land-use, as reflected by an increase in the curve for cultivated land (Fig. 19). However, this curve probably reflects a regional development, as the study site was certainly not suitable for agriculture. Slash-and-burn cultivation was common in this region, and this practice seems to have culminated at about this time (Larsson 1980; Weimarck 1953). The curve for cultivated land is mainly dominated by Secale and Rumex acetosa/R. acetosella, and these pollen types may have been dispersed to the study site from nearby cultivated areas.

The pre-Picea vegetation at Siggaboda was dominated by Fagus and Quercus, and some occasional Pinus individuals. This vegetation type did not apparently resist Picea invasion, as its increase was very rapid. However, the local Quercus population went into a rapid decline at about the same time as Picea increased rapidly. The reason why Quercus declined locally is not known, but may rather be an effect of selective felling, than of competition with Picea. No single Quercus individuals remain in the present stand, and this may indicate that the local Quercus population was destroyed by felling. This felling may have favoured the local expansion of Picea.

The establishment of Picea at the other studied sites was late (Figs. 18, 21). At Flahult (IV) Picea became established in the local forest stand c. 100 BP, but it has until now not gained any importance. Today, only scattered Picea individuals occur in the local stand dominated by Fagus. Picea has today started to become an important tree in the area, but this expansion has probably mostly occurred during this century and been facilitated by modern forestry.

At Bocksten (III) Picea became established during the latter part of the 19th or during this century. This area is located just outside the distribution limit that Picea had reached naturally during the beginning of this century (Hesselman & Schotte 1906). This may indicate that Picea was planted at Bocksten, or self-sown from nearby plantations. [Back to contents]

Competition between Picea and Fagus

Picea and Fagus have only been present together in the local forest stands at the studied sites for a short period. This means that they have not had many possibilities to compete with each other (Figs. 18, 19). At Mattarp (II) Picea and Fagus have been present locally since 400 BP. However, Picea has today a restricted occurrence in the local forest stand, although it was probably more abundant in the past, particularly some time after 400 BP, when local land-use seems to have favoured both Picea and Fagus. It is difficult to evaluate if these species really competed with each other during this period, as it is possible that they grew in different parts of the local stand, and not really in close contact.

At Flahult (IV) the local Picea population is still restricted, and it may not yet have started to expand. At Bocksten (III) the rapid expansion of Picea coincides with the decrease of pastures and cultivated land (cf. Fig. 19), and this may imply that Picea was planted on abandoned land. The present forest in this area is dominated by rather pure Fagus and Picea stands, and the composition and structure of these are controlled by forest management.

The only studied site where competition between Picea and Fagus really occurs today is Siggaboda (I). The present forest type at this site is however, relatively young as it originated about 200 BP, when Picea invaded this site. This forest is today dominated by more or less pure stands with Picea or Fagus, or by stands with a mixture of these species. In these mixed stands an intensive competition occurs. In the pollen diagram a rapid increase of Fagus percentages coincides with a rapid increase of Picea (cf. Fig. 19). A rapid increase of both these species at the same time may imply that they expanded in different stands, i.e., no competition occurred. However, Fagus percentages reach a peak at c. 100 BP, but decline thereafter. During this decline Picea percentages are still increasing, and this may indicate that Fagus and Picea had started to compete, and that Picea was favoured in this competition. An alternative explanation may be that Picea had gained importance in the vegetation surrounding the sampled hollow, and finally started to invade the peatland itself, where Betula and Alnus earlier were dominant. This expansion of Picea is also seen in the present day vegetation at the sampled site, where Betula and Alnus today are almost out-competed by Picea. Picea is able to grow on peaty soils, and this ability may explain why Fagus percentages have decreased steadily since 100 BP.

In conclusion, the long-term outcome of the competition between Picea and Fagus in southern Sweden is still difficult to evaluate, at least from the data available in this study. [Back to contents]

(1) Viewed on a continental scale the migration pattern of Fagus shows a striking correspondence to climatic change over the millennia, but at finer scales this resemblance is not that obvious. At stand-scale there are probably factors other than climate that are crucial for the establishment of Fagus (e.g., disturbance, seed dispersal, human activities, etc.).

The timing of establishment and expansion of Fagus does not show a regional coherence in southern Sweden, and this may imply that climate was not particularly important for its establishment in this area.

(2) The present day distribution of Fagus in southern Sweden suggests a migration with a discontinuous front with outlying isolated populations, and this model probably applies to its past distribution. This type of migration means that the landscape becomes infilled by dispersal from outpost stands. The timing of stand-scale establishment is then largely influenced by site-specific factors and chance. This may explain the large difference in timing between regional and local establishment at the studied sites. 

(3) The type of vegetation that becomes invaded by a migrating species is important, and a species may react differently depending on the type invaded. There is a significant difference between a species invading an unaffected forest with natural processes prevailing, and one invading a semi-open landscape with cultural activities controlling the vegetation.

Fagus seeds are highly dependent on ground disturbance for successful establishment, and an undisturbed forest then consequently would be able to resist Fagus invasion for some time. A semi-open cultural landscape may be optimal for Fagus establishment, as cultural activities may create conditions particularly suitable for its regeneration.

(4) At two of the studied sites cultural activities seems to have created conditions suitable for Fagus establishment (Mattarp at 400 BP, Flahult at 900 BP). At these sites the pollen curves for meadows, pastures, and cultivated land all increase significantly at about the same time as Fagus became established. At Siggaboda and Bocksten cultural activities may have favoured Fagus establishment as well (Siggaboda at 950 BP, Bocksten at 1450 BP), but human activity was weaker at these sites. At Siggaboda fires obviously facilitated the establishment and expansion of Fagus. These fires may have been human-induced (intentional or not).

(5) At Mattarp and Flahult Fagus apparently became established in a semi-open, cultural landscape that had already lost its natural composition and structure. At Mattarp the vegetation preceding the clearance event at c. 400 BP largely consisted of a mixture of Betula, Corylus, Quercus, and Picea. At Flahult Fagus became established in open pastures or in stands dominated by Betula, Quercus, and Corylus. It is most probable that the vegetation types preceding Fagus establishment at Mattarp and Flahult were controlled by cultural activities.

At Siggaboda and Bocksten the local forest stands seem to have been more or less unaffected by cultural activities prior to the establishment of Fagus, at least no semi-open vegetation was present locally. At Siggaboda the pre-Fagus vegetation on well-drained ground was dominated by Quercus, but Corylus and Tilia were also present. At Bocksten the pre-Fagus vegetation was dominated by Quercus, Tilia, Alnus, and Corylus.

At all sites studied within this project Tilia once was abundant, but its abundance generally decreased well before the immigration of Fagus. At Bocksten (site A) a significant Tilia population occurred until 200 BP. This late presence of a large Tilia population is remarkable, and may imply that the human influence on this stand was weak.

(6) Fagus may still be migrating northwards in Sweden. It grows well in its outpost area, and it seems that present day land-use, not climate, is the limiting factor for local Fagus expansion. The northern distributional limits of Fagus in Sweden probably still represent an active front, and outlying stands act as "infection centres."

(7) Picea invaded southern Sweden from the north during a period when the cultural landscape had already been evolving for some time, and the contemporary vegetation was largely influenced by cultural activities. Picea is a dominant tree with an effective seed dispersal, and the relatively open and probably grazed forests in the area were not particularly resistant to Picea invasion. An intensive grazing regime may not affect Picea, as grazing animals normally avoid Picea. If it instead had entered an area with unaffected forests these would probably have proved more resistant to invasion.

(8) The timing of local Picea establishment seems mostly to be controlled by its migration, i.e., it became established when its front reached the studied sites. It invaded Mattarp c. 800 BP, but did not expand much locally until 400 BP, in connection with the establishment of Fagus. The open conditions that existed during this period appear to have favoured Picea. It invaded Siggaboda at about 200 BP. The population increase for Picea was very rapid and within c. 50 years it co-dominated the local forest stand together with Fagus. The establishment of Picea at the other studied sites (Bocksten, Flahult) was late. At Flahult Picea became established in the local forest stand at c. 100 BP, but it has until now not gained any importance. At Bocksten Picea was probably planted, or self-sown from nearby plantations.

(9) Picea and Fagus have only been present together at the studied sites for a short period. This means that they have not yet had many possibilities to compete with each other. At Mattarp Picea and Fagus have been present locally since 400 BP, but they have probably not competed significantly with each other, as they may have grown in different parts of the local forest stand. The only studied site where competition between these species really occurs today is Siggaboda. The present day forest type originated at about 200 BP when Picea invaded this site. Picea has out-competed Fagus in the vegetation surrounding the sampled hollow, but this increase for Picea may be explained by its ability to grow on peaty soils. The long-term outcome of the competition between Picea and Fagus in southern Sweden is probably not possible to evaluate from the available data for the studied sites. [Back to contents]

Till sidans topp/Up