Long-Term Human Impact and Vegetation Changes in a Boreal Forest Reserve: Implications for the Use of Protected Areas as Ecological References
- First Online:
- Cite this article as:
- Josefsson, T., Hörnberg, G. & Östlund, L. Ecosystems (2009) 12: 1017. doi:10.1007/s10021-009-9276-y
- 343 Views
Northern boreal forest reserves that display no signs of modern forest exploitation are often regarded as pristine and are frequently used as ecological reference areas for conservation and restoration. However, the long-term effects of human utilization of such forests are rarely investigated. Therefore, using both paleoecological and archaeological methods, we analyzed temporal and spatial gradients of long-term human impact in a large old-growth forest reserve in the far north of Sweden, comparing vegetational changes during the last millennium at three sites with different land use histories. Large parts of the forest displayed no visible signs of past human land use, and in an area with no recognized history of human land use the vegetation composition appears to have been relatively stable throughout the studied period. However, at two locations effects of previous land use could be distinguished extending at least four centuries back in time. Long-term, but low-intensity, human land use, including cultivation, reindeer herding and tree cutting, has clearly generated an open forest structure with altered species composition in the field layer at settlement sites and in the surrounding forest. Our analysis shows that past human land use created a persistent legacy that is still visible in the present forest ecosystem. This study highlights the necessity for ecologists to incorporate a historical approach to discern underlying factors that have caused vegetational changes, including past human activity. It also indicates that the intensity and spatial distribution of human land use within the landscape matrices of any forests should be assessed before using them as ecological references. The nomenclature of vascular plants follows Krok and Almquist (Svensk flora. Fanerogamer och ormbunksväxter, 2001).
Keywordsforest ecosystemdisturbancenature reserveland usenative peoplepollen analysisinterdisciplinary studiesforest history
Understanding the ecological effects of long-term human land use in the past is increasingly recognized as essential in attempts to interpret the structure and function of contemporary ecosystems (Foster and others 2003; Elmore and others 2006; Willis and Birks 2006). For example, pre-industrial human activities in forest landscapes may continue to have profound effects not only on current vegetation composition and forest structure (Foster and others 1992; White and Mladenoff 1994; Motzkin and others 1999a), but also on biodiversity (Mitchell and others 2002), ecological processes (Bradshaw and Hannon 1992; Carpelan and Hicks 1995) and nutrient availability (Currie and Nadelhoffer 2002; Fraterrigo and others 2006). This is generally acknowledged for ecosystems that have been heavily influenced by humans in recent time, but much less so for “pristine” areas that have not been disturbed by recent human activities. However, very few studies have been undertaken to study human influence on forest ecosystems within areas designated as pristine. True examples of such forests, that is, those that have never been directly influenced by human activities, are only found at a few locations around the world. Typical examples include forests in uninhabitable locations such as the remote table mountains of the Guyana region in South America (Rull 2007) and the isolated Foja Mountains on the western side of New Guinea (Diamond 1982; Cyranoski 2006). However, large forested areas in the boreal parts of the northern hemisphere, including North America and the Russian Taiga, are also widely considered to be pristine, in a more limited sense, because they have never been cleared and have not been exploited by modern forestry (Norton 1996). In Europe, there are only a few such areas left, primarily in northerly national parks and nature reserves in Russia and Fennoscandia (Norway, Sweden, Finland, and northwestern Russia). The large forest reserves in northern Fennoscandia are among the best examples of pristine forests (in the limited sense) in Europe and, because they have been largely unaffected by agrarian colonization and modern forest exploitation, they are regarded as representing “base-line” conditions and are frequently used as ecological references and models for conservation and restoration (compare Fries and others 1997; Linder and others 1997; Lähde and others 1999). However, a growing body of research indicates that pre-industrial human activities of indigenous peoples (Lane 2006) have influenced such forest ecosystems in North America (Barret and Arno 1982; Bonnicksen 2000; Östlund and others 2005) and the northernmost parts of Europe (Hicks 1995; Hörnberg and others 2006; Karlsson and others 2007) to a greater extent than previously acknowledged. In the northern parts of Fennoscandia, the long presence of people is manifested by evidence of Mesolithic settlements, along the main watercourses, 8,000 to 10,000 years old (Matiskainen 1996; Bergman and others 2003).
To elucidate the importance of past human influences on forest ecosystems, it is necessary to use historical ecological approaches (Fuller and others 1998; Hörnberg and others 1998; Benitez and Fisher 2004; Schulte and others 2005). However, most previous relevant studies have been carried out in managed or semi-natural forests with recognized human land use. In this study, we analyzed the long-term ecological effects of human activities in a large, remote forest, in the Tjeggelvas Nature Reserve (TNR), to assess the impact that humans may have had on this protected forest in a long-time perspective. The TNR is considered to contain one of very few remnants of pristine boreal forest in northern Fennoscandia. We focused on vegetation changes over the last 1,000 years at three sites using analyses of pollen, organic matter contents (loss on ignition residues), and charcoal fragments in stratigraphic records obtained from forest hollows, in combination with archaeological field surveys. The objectives of the study were to: (1) analyze and compare changes in vegetation composition between an area with no recognized history of human settlements and two peripheral areas in which human activity is known to have occurred (sites of an indigenous Sami settlement and a 19th century agrarian settlement), (2) distinguish the effects of human land use with respect to subsistence strategies, and (3) interpret and discuss temporal and spatial gradients of long-term human impact on the vegetation composition and ecosystem structure in northern boreal forest ecosystems. The timeframe was chosen because it is well covered by both paleoecological and archaeological archives (there are no verified archaeological remains at the study sites dating from before AD 1,000) and during this time substantial shifts occurred in the type, intensity, and extent of human land use in the interior parts of Fennoscandia.
In the wider interior region of Fennoscandia, human influence on the forests was limited prior to the 16th century, mainly because human population densities were low (<1 person/km2). The earliest known settlements along the upper part of the Pite River valley date from Mesolithic times more than 8,000 years ago and can be linked to a hunter/gatherer population (Bergman and others 2003). These settlements are limited to a few places near the main watercourses (Meschke 1979). However, during the first centuries AD there was a change in land use and settlement patterns related to a shift in subsistence toward herding reindeer (Rangifer tarandus L.), known as caribou in North America (Aronsson 1991; Bergman 2007). Later, reindeer herding came to form an integral part of Sami subsistence strategies and was combined with hunting and fishing (Aronsson 1991; Storli 1993). By the late 16th century, Sami subsistence was characterized by nomadism and seasonal settlements. In the upper part of the Pite River valley, intensive reindeer herding was practiced until the beginning of the early 20th century (Ruong 1945). During the mid-19th century and the first half of the 20th century, small-scale agrarian expansion into the upper part of the Pite River valley resulted in intense land use in certain limited areas (Bylund 1956). In the mid-20th century, however, many of the agrarian settlements were abandoned. Occupants of one of the study sites, the Akkapakte settlement, acquired property rights in the mid-19th century. Here, small-scale cultivation was carried out within the limitations imposed by the harsh climatic conditions, mainly during the first half of the 20th century. However, agrarian practices at Akkapakte were finally abandoned in the 1970s.
Description of the Study Sites
In studies of local vegetational changes based on pollen assemblages, there are clear risks that regional background pollen may mask the changes or distort the results, but these risks can be minimized using closed canopy assemblages from sites that were continuously forested during the study period (Jackson and Wong 1994). This allows local changes in vegetation composition within approximately 100 m radii of sampled sites to be detected and compared (compare Bradshaw 1988; Sugita 1994; Calcote 1995) that may not be possible to identify using archives with larger pollen source areas, for example, lakes, where local changes are obscured due to larger influxes of regional pollen and there are greater risks of overlaps in pollen source areas. Further, to pinpoint local vegetation changes in space and time, and associate them with human activities, it is necessary to establish the pollen sampling sites as close to archaeological finds as possible. Here, therefore, pollen records were analyzed from forest hollows or small ombrotrophic peat mires with a surface area <0.2 ha to detect and compare local changes in vegetation composition at three sites, situated 7–9 km apart northeast of Lake Tjeggelvas (Figure 1), with very similar physical geography and edaphic settings.
The first site (66°37′ N, 17°49′ E) represents an extensive area covering the north and east part of the study area with no known settlements or visible traces of human activity (hereafter referred to as the Low Impact Area). Scots pine with scattered downy birch in dry dwarf-shrub stands, including Empetrum spp., Vaccinium vitis-idaea L., and Cladonia spp. dominate in this area. At this site, a peat core was obtained from a small mire (0.05 ha, 510 m a.s.l.) adjacent to a 0.25-ha pond, located 250 m south of a small lake. Betula nana L., Calluna vulgaris (L.) Hull., Empetrum spp., Rhododendron tomentosum Harmaja, V. myrtillus L., V. uliginosum L., Eriophorum vaginatum L., Trichophorum alpinum (L.) Pers., Carex spp., Rubus chamaemorus L., and Sphagnum spp. characterize the vegetation on the mire. The second site, Munka (66°33′ N, 17°47′ E), is at the site of an old, abandoned, naturally reforested Sami settlement, situated on a small isthmus on the northern side of Lake Munkajaure. Scots pine and downy birch dominate the surrounding forest except close to the lakeshore and on the isthmus, where downy birch is more abundant. A peat mire (0.1 ha, 460 m a.s.l.), located 15 m south of the settlement was selected for pollen analysis. B. nana, Empetrum spp., V. myrtillus, V. uliginosum, V. vitis-idaea, H. splendens, Polytrichum spp., and Sphagnum spp. characterize the vegetation at the mire. The third site, Akkapakte (66°38′ N, 17°41′ E), is situated approximately 150 m from the northern shore of Lake Akkajaure. Here there was a more recent settlement where the main activities were agrarian, predominately hay-making and farming domesticated animals. Open grassland (6.8 ha) and forest, with scattered Scots pine, downy birch, grey alder, aspen (Populus tremula L.), and a few Norway spruce trees characterize the setting. About 100 m west of the settlement, peat was acquired from a mire (0.2 ha) located at 480 m a.s.l. and enclosed within the forest. Typical vegetation on the mire includes A. incana, B. nana, Juniperus communis L., Salix spp., V. vitis-idaea, Andromeda polifolia L., Carex spp., Potentilla palustris (L.) Scop., R. chamaemorus, H. splendens, P. schreberi, and Sphagnum spp.
Materials and Methods
Archaeological Field Survey
Land-based archaeological field surveys were carried out to detect signs of past presence of Sami people within the studied part of the TNR. Locations where occasional Sami activities are expected to have taken place during the last millennium (hearths of árran-type; Sommerseth 2004) are situated near mires, fens, and watercourses (compare Bergman 1995; Liedgren and others 2007). Consequently, the forests close to Bläckajaure, Hålkåsjaure, Munkajaure, and some several other smaller lakes were systematically examined for ancient remains such as hearths and earth ovens (Bergman 1995), bark-peeled trees (Zackrisson and others 2000) and other signs of past human land use, providing a good, although not complete, insight regarding the spatial distribution of past human activity in the area.
Sampling and Radiometric Dating
Peat samples were collected from a small mire at each of the three study sites, using a Russian sampler (Ø 4.5 cm, length 50 cm) as described by Jowsey (1966). In addition, an intact sample of the uppermost 30 cm of the peat profile from each of the three sites was obtained by digging up a small monolith of peat with a spade. To examine vegetation dynamics over the last millennium, the uppermost 45 cm of each of the peat samples was used. All of the peat samples were carefully described in the field to facilitate correlation between core samples and spade samples, then wrapped in plastic film and aluminum foil, transported to the laboratory and stored in freezers at −18°C. For accelerator mass spectrometry (AMS) dating (Bowman 1990), samples from each site, representing a level at which there were notable changes in the pollen profile, were obtained from the peat profiles used for the pollen analysis. When possible, macrofossils of Sphagnum spp. were used because this material is considered to give reliable dates (Nilsson and others 2001). Bulk peat samples constituted the remaining samples that were dated. To avoid contamination by roots in the samples, the bulk peat samples were sieved through a 250-μm mesh. The samples were dried at 105°C for 24 h and at least 6 mg of organic material was used for dating. AMS-datings, calibrated using CALIB Rev. 5.1 software (Stuiver and Reimer 1993), were performed at the Ångström Laboratory at the University of Uppsala, Sweden and peat accumulation rates were estimated using TILIA 2.0 software (Grimm (1991).
Paleoecological Analysis and Vegetation Reconstruction
Identified Anthropogenic Pollen Indicators
Asteraceae undiff. (Asteroidae)c
The following results of the vegetation and land use history analyses at the three studied sites were based on the land-based archaeological field surveys and the percentage pollen diagrams, supplemented by the estimated peat accumulation rates and a quantitative comparison of pollen types between the study sites.
Along the shores of Lakes Bläckajaure, Hålkåsjaure, and Munkajaure, remains of 106 Sami hearths used for occasional activities were detected (Figure 1). Other signs of past human activity, for example bark-peeled Scots pine trees, were principally concentrated along the main watercourses. In addition, the remains of old wooden constructions revealed the locations of two abandoned Sami settlements, one on the northern shore of Lake Munkajaure and the other on the western shore of Lake Bläckajaure. Signs of past human activities were least frequent in the forest in the Low Impact Area. No traces of human activities were detected within a 500-m radius of the mire used for peat sampling in the Low Impact Area.
Radiocarbon and Calibrated Ages of Plant Remains from Peat Profiles
Type of material
Reported age (14C years BP)
Calibrated age (AD/BC 2σ)
Calibrated age (BP 2σ)
Low Impact Area
65 ± 25
Low Impact Area
630 ± 40
Low Impact Area
2040 ± 30
160 BC–AD 50
225 ± 35
305 ± 30
2910 ± 30
400 ± 50
1755 ± 30
2590 ± 40
Vegetation History of the Study Sites
Major Changes in Vegetation Composition in the Local PAZs in the Low Impact Area and the Munka and Akkapakte Settlements
Low Impact Area
c. 2000–1200 BP
AP 85% (Betula 30%, Pinus 50%, Alnus < 2%, rise in Picea 2% to 4%). Ericaceae high throughout (15%) until upper levels (5%). Salix < 1%, Poaceae < 1% at lower and upper levels, herbs low. Few charred particles at 43 cm depth, LOI residue low-moderate < 10%
c. 1200–550 BP
AP 85–90% (temporary rise in Betula to 60% then decline to 40%, notable decline in Pinus to 30% then rise to 40%, Alnus 1%, Picea 4%, QM < 1%). Ericaceae 3-8%, Poaceae and Cyperaceae stable <1%, herbs low. Few charred particles, LOI low and stable c. 5%
c. 550–150 BP
AP 70–85% (decline in Betula to 20%, rise in Pinus to 60%, Alnus 1–2%, Picea stable 3%). Fluctuating Ericaceae 1–10%, Juniperus and Salix intermittently <1%. Rise in Poaceae to 2% then decline to <1%, Cyperaceae stable 2%, rise in herbs. Small charred particles occur continuously (highest at 29 cm depth), LOI residue low <5% except at 29 cm depth (13%)
AP 85–90% (further decline in Betula to 15% temporarily increasing to 25% at 5 cm depth, rise in Pinus from 65 to 80%, slow decline in Alnus to 0, Picea stable 2%). Rise in Ericaceae and Calluna in the uppermost levels. Poaceae stable 1%, decline in Cyperaceae to 1%, decline in herbs. Charred particles intermittently, LOI residue very low <3%
c. 4200–2800 BP
Rise in AP 80–90% (rise in Betula 40–55%, Pinus 30–45%, decline in Alnus 5–3%, QM < 1%). Decline in Ericaceae 10–5%, Poaceae < 1%, herbs low. Few charred particles of both size classes, LOI residue low-moderate (6–10%)
c. 2800–220 BP
AP 75–90% (Betula 40–65% and Pinus variable 25–55%, Alnus 1–2%, Picea 2–4%, QM < 1%). Ericaceae 5% higher at 34–24 cm depth (10–15%), Juniperus sporadic and Salix < 1% at 28–21 cm depth. Slow rise in Poaceae to 2%, Cyperaceae 1% at 32–21 cm depth. Cerealia at 28 and 22 cm depth, rise in herbs. Charred particles 28–21 cm depth and high at 38 cm depth, LOI residue low <7%
c. 220–50 BP
AP 70–85% (temporary rise in Betula to 90%, decline in Pinus to >10%, Alnus >1% and slow decline in Picea 1–2%, Populus intermittently <1%). Slow decline in Ericaceae to <1%, Calluna <1%, and Salix intermittently <1–8%. Rise in Poaceae to 7% at 16 cm depth then decline to 2%, Cyperaceae 1–2%. Cerealia at 16, 14, and 12 cm depth, herbs rise further (for example, Rumex rise to 10% at 16 cm depth). Charred particles high throughout, especially at 20–18 cm depth, LOI residue low <4%
c. 50 BP–present
Rise in AP 65–95% (Betula 40% and rise to 75% in upper levels, rise in Pinus to 55% then decline to 15%, Alnus and Picea initially 2% then decline to 0). Juniperus temporarily >1%, rise in Ericaceae to 6%. Decline in Poaceae to 1% and Cyperaceae to 0, Cerealia at 8 cm depth, decline in Rumex to 0 at 6 cm depth. Small charred particles in upper levels, LOI residue low <3%
c. 2700–2500 BP
Decline in AP 95% to 65% (decline in Betula 40% to 15%, rise in Pinus 55–65%, Alnus 5%, Picea 1–2%, QM < 1%). Ericaceae < 1%, fluctuating Cyperaceae 1–30%, rise in Poaceae from 0 to 1%. Few small charred particles, LOI residue moderate 9–12%
c. 2500–350 BP
AP 85–95% (fluctuating Betula 15–30%, Pinus variable 65–75% but decrease to 50% (16–11 cm depth), Alnus variable 1–7%, Picea stable 3–5% but increase to 7% in upper levels, QM intermittently <1%). Ericaceae initially low <1% but rise to 5% (22 cm depth) then decline to 1%, Salix <1%, and Poaceae <1% intermittently, Cyperaceae variable 1–10%, herbs low. Charred particles recur throughout and higher at 34, 20, 18, and 12 cm depth, LOI residue moderate and stable (8–12%)
c. 350 BP–present
AP 70% rise to 85% (Betula 15–20% but rise to 30–40% in upper levels, Pinus 40–55%, Alnus < 1%, Picea decline to 1%). Ericaceae recur throughout <1%, Juniperus intermittently < 1%, Salix decline to 0, Poaceae higher 1–2%, rise in Cyperaceae to 30% then decline to 10%. Rise in herbs. Charred particles recur, rise in LOI residue (19% at 6 cm depth) then decline to 5%
Changes in Pollen Composition During the Last 1000 Years
The Low Impact Area pollen diagram covers approximately 2,000 years, divided into four zones (LI I–IV), the upper three of which (LI II–IV) span the last 1,000 years. Changes in pollen percentages in these zones are summarized in Table 3 and illustrated in Figures 4 and 7. The percentage of AP is high throughout these zones, about 70–90%, and pollen from trees such as Alnus, Betula, Picea, and Pinus occurs continuously. The percentage of Pinus pollen increases from 45% to 75% throughout the lithostratigraphy whereas Betula pollen values decrease from almost 40% to less than 20%. However, at approximately 700 cal years BP Pinus pollen temporarily decreases and Betula increases (LI II). The pollen values of shrubs (Juniperus communis and Salix) and dwarf-shrubs (Calluna and Ericaceae) are generally low. Ericaceae dominate the non-arboreal pollen (NAP) together with Poaceae, Cyperaceae, and Rubus chamaemorus. Pollen types from herbs occur sporadically, at very low percentages. Spores of Lycopodiumannotinum and Polypodiaceae types occur at low proportions whereas the percentage of Sphagnum spores fluctuates strongly. LOI residue values are low to moderate (<10%) throughout the zones, except (14%) at about 450–500 cal years BP. Charred particles occur intermittently.
The unlogged forest within the study area is currently dominated by very late-successional, multi-aged Scots pine forest, with scattered downy birch trees and grey alder in wet locations. As described by Aronsson (1991) and Segerström (1990), this tree-species assemblage—and the presence of grasses and herbs such as Chenopodiaceae, Epilobium spp., and Melampyrum spp.—is characteristic of rather open forests. Furthermore, the continuing dominance of the field layer by dwarf-shrubs, mainly ericaceous species, and comparatively low amounts of herbs and ferns, confirm that the soils are nutrient-poor, which are typical features of northern boreal pine forest ecosystems (Sjörs 1965). Thus, the forests of the TNR could be viewed as archetypal pristine boreal pine forests of northern Europe. However, the pollen assemblages observed in the peat cores indicate that changes in vegetation composition have occurred in several periods during the last millennium. Climatic variables heavily influence vegetation composition, and the resulting changes are often difficult to separate from vegetational changes caused by past human land use (Fuller and others 1998; McLachlan and others 2000; Motta and others 2002; Hotchkiss and others 2007). However, because the sites examined are all in such close proximity to one another, they are all under the same climatic influence. Therefore, it is unlikely that past changes in climate made major contributions to the differences in vegetation history among the sites. Rather, it is posited that past human action may have contributed to the unique vegetation histories observed at the study sites.
Inferred Vegetation History in the Low Impact Area
Analyses of anthropogenic indicators in pollen records in northern Fennoscandia have generally focused on regional vegetation history and the introduction of agriculture, including farming of domesticated animals or slash and burn activities (Vorren 1986; Hicks 1988; Segerström 1990). This implies that new interpretations of the pollen record are required in studies of ecosystems with historically low levels of human presence, and that an interdisciplinary approach (including paleoecological, archaeological and ecological analyses) should be applied. Thus, in this study indicators of past human activity have been treated with great caution because the traditional interpretation of many such indicators is based on human land use in rather different contexts (Vorren 1986; Behre 1988; Hicks 1988; Bos and Janssen 1996). During the last 1,000 years, the forest ecosystem in the Low Impact Area, initially characterized by an open Scots pine forest with a species-poor field layer, has undergone only slight changes (Figure 4). Although the tree layer seems to have become more dominated by Scots pine, the field layer appears to have been more stable and persistently dominated by shrubs and dwarf shrubs (Figures 4 and 7) characteristic of undisturbed northern Scots pine forests (compare Aronsson 1991; Räsänen 2001).
Summary of Detected Types of Disturbances or Ecological Processes and Associated Indicators
Type of disturbance/ecological process
Indicator of disturbance
Low Impact Area
1000 cal years BP–present
Reduction in soil fertility
Scots pine +
Downy birch –
600 cal years BP
Charred particles +
Downy birch +
Scots pine −
350–50 cal years BP
Downy birch +
Herbs, incl. apophytes +
Scots pine –
Dwarf shrubs –
Cutting for fire-wood and wooden constructions
Scots pine –
(Downy birch +)
Cereal species +
Rumex acetosa/acetosella +
Other herbs +
200 cal years BP
Charred particles +
Downy birch +
Salix species +
Scots pine −
50 cal years BP–present
Regeneration following abandonment of the settlement
400–300 cal years BP
Clearing land and possibly grazing by reindeer
300 cal years BP
Hydrological change at the mire and increased erosion
300–50 cal years BP
Downy birch +
Herbs, incl. apophytes +
Scots pine −
Cutting for fire-wood and wooden constructions
Scots pine −
(Downy birch +)
Clearing land and possibly grazing by domesticated animals
Furthermore, the increase in Scots pine during the timeframe for this study coincided with a steady decline in downy birch (Figure 4). These vegetational developments may have been related to reductions in soil fertility and long-term absence of disturbances. As discussed by Iversen (1958), changes in the soil and vegetation in the Northern Hemisphere have been closely linked to the glacial–interglacial climate cycles. During their final successional phases, northern boreal ecosystems gradually become more nutrient-poor, which favors conifers such as Scots pine on coarse till (Iversen 1958; Andersen 1969; Birks 1986). Moreover, in a study of ecosystems in the western United States MacKenzie and others (2006) found that soil fertility, notably nitrogen availability, is directly dependent on fire history and successional stage, that is, that long-term absence of large-scale disturbances tends to result in reduced soil fertility. There may be other explanations for the detected vegetational changes, for example hydrological or climate changes. However, because the Low Impact Area soils are well-drained course tills, major hydrological changes are less likely explanations. Furthermore, a climate change (for example the Little Ice Age) causing the vegetational development seen at the Low Impact Area would have also been detected in the peat-cores sampled at Munka and Akkapakte. However, no similar changes in vegetation composition were identified at either of the other two study sites, both of which have been subjected to disturbance by human land use.
Inferred Vegetation History at Munka
Around 1,000 cal years BP a forest with Scots pine and downy birch, and a field layer dominated by dwarf shrubs and few herbs characterized the vegetation at Munka (Figure 5). From about 350 cal years BP until present there were marked changes in the tree layer and forest floor vegetation (Table 4). During this time tree cover fluctuated, as indicated by oscillating tree pollen values (Figure 7). Furthermore, there were steady increases in the abundance of apophytes and other herbs, together with a reduction in the abundance of dwarf-shrubs up to about 150 cal years BP (Figures 5 and 7). The main contributors to the increase in apophytes were grasses, accompanied by various herbs (Figure 5)—including Asteraceae, Cannabis spp., Melampyrum spp., Ranunculus spp., and especially Rumex acetosa/acetosella—that are associated with anthropogenic land use activities in forested parts of northern Fennoscandia (Hicks 1993; Aronsson 1994; Hörnberg and others 1999; Räsänen 2001). This vegetation assemblage was most pronounced between 200 and 75 cal years BP. During this period, the forest at the Munka settlement was characterized by a reversed relationship between Scots pine and downy birch, with dominance of deciduous tree species (Figure 5). As discussed by Hicks (1995), a decrease in pine pollen may reflect extensive use of Scots pine trees at and near the settlement site, and eventually downy birch may have spread into the open area (higher abundance of Betula pollen). In addition, the presence of anthropocores, that is, cereal pollen, between about 350 and 50 cal years BP, is especially noteworthy, because it strongly indicates local cultivation at the settlement site (Table 4). The changes in tree-species composition, the presence of cereal pollen and the distinct increases in apophytes, together with an extension of the herb species assemblage, are strongly indicative of human land use activities at and near the sample site.
The findings suggest that the settlement at Munka was established in the early 17th century and utilized more permanently during the 19th and 20th centuries. The rapid change in peat decomposition at around 400–350 cal years BP, which is also reflected in the lithostratigraphy (Figures 3 and 5), may have been due to changes in hydrological conditions at the mire associated with the establishment of the settlement. The groundwater level may have increased as the area was cleared and the peat may have been compacted due to trampling, or a thin peat layer may have been removed, for example, for roof and wall insulation of the hut. There is, however, no clear indication of a hiatus in the peat profile. Further changes in conditions at the mire are apparent from around 200 cal years BP, reflected in reductions in the abundance of Sphagnum. Other noteworthy features during this time period are the pronounced presence of grasses and Rumex acetosa/acetosella and the occurrence of local fires at around 200 cal years BP (Table 4, Figure 5). As suggested by Aronsson (1991, 1994), both settlement and grazing/trampling in northern boreal forest ecosystems favor grasses and some herbs, including for example Rumex acetosa/acetosella and Melampyrum spp. According to Räsänen (2001), settlement also promotes a greater variety of herbs, reflected in the presence of Rumex acetosa/acetosella, Epilobium species and some Achillea and Solidago species of the Asteraceae. Accordingly, reindeers were almost certainly herded near the settlement site (Table 4).
Although it is generally believed that the Sami did not use fire to modify their local environment, Hörnberg and others (1999) have suggested that deliberate burning may have been carried out at certain forest sites by Sami nomads. The use of fires in forests by indigenous people in other parts of the Northern Hemisphere has also been described, for example, in northeastern USA, where indigenous people burned the forest to improve hunting opportunities and clear land for agricultural fields (Russel 1983; Black and others 2006). However, whether the forest fires near Munka were of natural origin or caused by humans cannot be distinguished from the present data (compare Carcaillet 1998). Furthermore, the occurrence of cereal pollen indicates more intensive human land use, including small-scale cultivation and possibly permanent use of the settlement. Interestingly, this appears to be the first time that evidence of small-scale cultivation of cereals has been documented within the context of a Mountain Sami economy. Finally, a distinct increase in birch and accompanying reductions in grasses and herbs in the uppermost part of this zone are interpreted as effects of successional regeneration following the abandonment of the Sami settlement (Table 4). The pronounced appearance of Melampyrum spp. at approximately 50 cal years BP may reflect the early regeneration phase because changes in the abundance of this species can indicate sudden changes in the local environment (Aronsson 1991).
Inferred Vegetation History at Akkapakte
From about 1,000 to 500 cal years BP, the forest composition at the Akkapakte settlement site remained fairly stable, dominated by Scots pine and a field layer consisting mainly of shrubs and dwarf-shrubs (Figures 6 and 7) typical of undisturbed northern Scots pine forests. However, during the following centuries a change in forest structure to more open conditions occurred, indicated by a distinct decrease in Scots pine and temporary increases in downy birch (Figure 6). About 300 cal years BP, there were increases in apophytes, especially sedges (Figure 7), accompanied by increases in juniper, grasses, Filipendula spp. and Melampyrum spp. indicating more open conditions. Several species that are closely related to human presence appeared, for example Brassicaceae, Cannabis spp., Caryophyllaceae, Chenopodiaceae, Epilobium spp., Filipendula spp., Hypericum spp., Melampyrum spp. and Polygonum aviculare L. (Aronsson 1991, 1994; Hicks 1993; Räsänen 2001; Figure 6). The vegetational changes during the last 400 years are interpreted as the result of different types of human land use. The marked decline in Scots pine is believed to reflect disturbance events, attributed to clearing of the settlement site and possibly local forest fires, as indicated by the presence of charred particles throughout the studied time period (Table 4, Figure 6). A more intensive use of the settlement site is indicated by the increase in grasses and the almost simultaneous appearance of herb species at approximately 300–250 cal years BP (Table 4). Furthermore, there was a sudden increase in sedges and decline in ferns at approximately 300 cal years BP, probably due to changes in hydrological conditions at the mire and increases in erosion at the mire or its surroundings, as indicated by increases in LOI residue values (Figure 6). The increase in species diversity, including grasses and herbs, is interpreted as a consequence of settlement whereas the increases in grasses and sedges are believed to be results of other land uses (Table 4). As proposed by Räsänen (2001), trampled sites are usually characterized by grasses but also, to a smaller extent, by sedges. In addition, a strong increase in sedges also implies hay-making at or near the mire (Aronsson 1991). Hay-making was commonly practiced because sedges were the preferred sources of fodder for domesticated animals in these areas (Campbell 1948; Bylund 1956; Segerström and Emanuelsson 2002).
Most likely the Akkapakte settlement was established about the same time as the Munka settlement and initially used temporarily during the summer by Sami passing through the area while herding reindeers. However, there are no strong indications of intensive reindeer herding at the site. Instead, there are indications of farming of domesticated animals, even before the inhabitants at Akkapakte acquired property rights in the mid-19th century, indicating that the settlement may have been used permanently for more than 150 years. Interestingly, however, no traces of cultivation were detected in the pollen record. The most likely explanation for their absence is the location of the sample site in relation to the former arable land, because the sample site is located about 100 m west of the settlement and the former cultivation site may be even further away. All of the available evidence, including observations made by Laestadius (1831) in the early 1800s, indicates that the people responsible for the early human activity at Akkapakte included Sami. However, there were no clear changes in vegetation composition indicating a shift from reindeer herding to agrarian land use. Contrary to findings by, for example, Hicks (1988, 1993) and Räsänen (2001), differences in ecosystem structure at the Akkapakte and Munka sites could not be attributed solely to single subsistence strategies adopted by people occupying the sites. The present vegetation composition at Akkapakte may instead reflect the composite effects of several centuries of land use by both traditional forms of Sami land use and more recent agrarian activities. Furthermore, in contrast to the Low Impact Area there are no signs of ongoing progress toward more nutrient-poor soils at either Akkapakte or Munka. Instead, long-term human land use appears to have restrained or delayed progression toward more nutrient-poor ecosystems at these sites.
Effects of Long-Term Human Activities on Northern Boreal Forest Ecosystems
Although the large TNR contains one of the most remote and inaccessible forests in Europe, people have used the forests in the area for a long time. Furthermore, the human activities here and throughout the northernmost parts of North America and the Russian Taiga differed from those in more central and southern parts of the northern hemisphere in that the land use patterns were typically low-intensity but spatially and temporally complex, over vast areas and extensive periods of time (compare Cronon 1983; Hicks 1995; Östlund and others 2003; Delcourt and Delcourt 2004). Because many indigenous people, including the Sami in northern Fennoscandia, were nomadic and present at low population densities in prehistoric times, their effects on vegetation composition and forest structure have often been considered insignificant. However, this view is simplistic and misleading because the accumulated effects of limited human land use over extensive time periods can change the environment considerably. As shown in this study, long-term land use by indigenous people can substantially alter forest floor vegetation and tree-species composition at settlement sites and in the surrounding forest.
Hörnberg and others (1999) suggested that some exceptional forest types in northernmost Fennoscandia may have been the result of human activities in combination with harsh climate conditions, and Motzkin and others (1999b) found that historical human land use activities in central New England, USA, may have influenced tree-species composition for several centuries after the disturbances had ceased. Sami land use included felling of trees for fire-wood and wooden constructions, gathering various natural resources and herding of reindeer. These were activities that extended up to a kilometer from the actual settlement site, and sometimes even further. It seems reasonable to believe that certain activities, for example, reindeer herding and cutting of Scots pine, carried out for many centuries have resulted in more extensive effects on the forest ecosystem. For instance, cutting of standing dead pine trees may have had considerable effects on dead wood characteristics in areas surrounding the settlement sites. Furthermore, grazing, trampling, and fertilization by reindeer and caribou strongly influence vegetation in northern Fennoscandia, Canada and Alaska (Manseau and others 1996; Suominen and Olofsson 2000; van der Wal 2006). Although the foraging behavior of reindeer and caribou differs from the behavior of other domesticated animals in that they wander over large grazing grounds, their effect on the vegetation can be substantial when driven together or kept fenced (Aronsson 1991; Väre and others 1996; Suominen and Olofsson 2000; Holt and others 2008). For example, extensive grazing and trampling can affect soil processes such as soil respiration and mineralization (Stark and others 2000; 2003), as well as the composition of the field layer vegetation (Väre and others 1996; Stark and others 2000; den Herder and others 2003) and lichen and bryophyte communities (Väre and others 1995; Holt and others 2008). The relationships between reindeer herding and tree volume or other site productivity factors are often less clear. However, studies on the effects of reindeer and caribou on ecosystem properties seldom cover time periods longer than 30 years. Consequently, the effects of reindeer on the structure of the forest community near the Munka settlement may have been substantial, affecting productivity and the re-establishment of trees, as well as soil processes and the field layer vegetation, because reindeer herding was probably practiced locally for at least 300 years. As discussed by den Herder and others (2003), the overall effect of reindeer grazing and trampling may impair the natural regeneration of Scots pine. On the other hand, the establishment of birch may be improved by disturbance of ground vegetation through grazing and trampling (Väre 2001). Thus, reindeer herding and other Sami activities may also have caused cascade effects on the diversity of epiphytic invertebrates, bryophytes, macrofungi and lichens as the tree-species composition and, presumably, abundance of dead pine wood changed over time.
The influence of human land use on vegetation composition also varies at different spatial scales. In a study covering the state of Massachusetts, USA, Hall and others (2002) showed that major differences in past human land use had a profound impact on forest composition and structure locally, whereas climatic and geological factors influenced vegetation at broader scales. As shown in this study and by Vale (1998), among others, effects of pre-industrial human activity can be seen across wide scales and intensities, from small areas in which there is clear evidence of anthropogenic effects on the composition of the local vegetation to large parts of the forest reserve in which the human impact appears to have been weak with few visual effects of land use. Consequently, at local scales, the long-term human impact can result in clear and sometimes persistent effects on subsequent vegetation composition and possibly ecosystem processes, whereas at larger scales the effects of human activity are often less readily apparent, but they can result in temporary vegetation changes. Furthermore, Foster and others (1998) suggest that the effects of cumulative human land use may alter regional patterns between the vegetation and the physical environment resulting, for instance, in homogenization and the formation of new species assemblages. Consequently, as illustrated by Foster (1992) and Josefsson and others (2005), current forest structure can mask the extent of past human land use activities.
This study challenges the current view that protected forests in northern Europe are pristine ecosystems that have not been significantly affected by human activities in the past. As shown in this study, low-intensity land use over long times by indigenous people can substantially alter the forest ecosystem at semi-nomadic settlement sites and in the surrounding forest. In this case, the land use has promoted open forests, characterized by downy birch and scattered ancient Scots pine trees with distinct elements of herbs and graminaceous species that otherwise are very rare in this ecosystem and are indicative of more nutrient-rich conditions. This particular combination of forest structure and ground vegetation is a persistent legacy of human land use, clearly distinguishable today. We suggest that the human impact on these forest ecosystems can be discerned from pollen records in increased values of apophytes, because settlements favor grasses and certain assemblages of herbs whereas grazing and trampling favor graminaceous species and herbs such as Rumex spp.
The results have important implications for ecological research in protected areas. Supposedly pristine forests (in remote areas with no recent management) cannot be used indiscriminately to represent reference conditions. Disturbances induced by indigenous peoples must be considered even in low-productivity boreal ecosystems, because slight but persistent human effects can continue to influence ecosystem properties for extended periods of time. Consequently, a landscape matrix of the intensity and spatial distribution of past human land use should be analyzed before areas such as the TNR are used as ecological references. The results also have consequences for both nature conservation strategies and landscape planning, not only in Europe but also in North America where vast forest landscapes in which human interference has been minor still remain (compare Foster and others 1996; Östlund and others 2005). Furthermore, the effects of anthropogenic disturbances are often difficult to separate from those of natural disturbances and climatic changes. An ecological understanding of contemporary forest ecosystems is not sufficient to comprehend patterns and processes, including land use history, that have shaped these ecosystems. We argue that the intricate relationships between human activity and vegetation composition are best elucidated through an interdisciplinary approach, because the impact and extent of past human land use can only be assessed by combining ecological research with methods and models used in several disciplines, for example, paleoecology, history, and archaeology. This is essential for understanding the complexity of human activities and their interactions with forest ecosystems (Ewel 2001; Burgi and Turner 2002; Gillson and Willis 2004; Briggs and others 2006).
We wish to thank Ingela Bergman and Lars Liedgren for archaeological field surveys, Anna Berg for laboratory assistance, John Blackwell for linguistic corrections, two anonymous reviewers, Hanna Karlsson and David Wardle for valuable comments on earlier versions of the manuscript, Henning Rankvist for information regarding the Akkapakte settlement and Björn Helamb at Arctic Air for unsurpassed helicopter transportation to the inaccessible study sites. This study was financially supported by FORMAS.