Journal of the North Atlantic G. Sveinbjarnardóttir 2018 Special Volume 11 1 Introduction “At that time Iceland was covered with woodland between mountain and seashore” (“Í þann tíð vas Ísland viði vaxit miðli fjalls ok fjöru”) (Íslendingabók, 5). This sentence from the Book of Icelanders, Íslendingabók, which is a short history of the earliest habitation in Iceland, written in the 12th century, is often quoted to demonstrate the rapid change that occurred in the vegetation cover of the country at the time of settlement, now generally acknowledged to have taken place in the second half of the 9th century. The text, written some 250 years after the time of settlement, suggests that the author had some knowledge of what the environment may have been like about two and a half centuries earlier. Archaeological and environmental research has proved his statement to have some substance. A pioneer in the study of environmental change was geologist Sigurður Þórarinsson, who laid the foundation for the study of volcanic ash layers (tephra) in Iceland (Þórarinsson 1944), which are extensively used for dating changes in the environment and also in archaeological research. In the latter discipline volcanic ash layers are for example found in the building turf and also sealing structures, showing, respectively, after and before what time they were built. The various ash layers, which are deposited in layers to form stratigraphic markers, have been dated on the basis of written sources, radiocarbon dates and stratigraphy, creating what Þórarinsson called tephrochronology (Þórarinsson 1981, Larsen 1996). In his research Þórarinsson pointed out that there was a tremendous increase in erosion at the point where landnám sets in, marked by a volcanic ash layer which until recently was conveniently dated to AD 871 ± 2 (Grönvold et al. 1995), now revised to AD 877 ± 1 (Schmid et al. 2017). Because of its date, close to when the country is, based on written sources, assumed to have been settled, it is generally called the landnám layer.1 Although it is not disputed that there were drastic environmental changes at landnám, views vary as to the temporal, spatial and social aspects of these processes, which are increasingly being shown not to have been as clear cut as suggested by Þórarinsson and others, or indeed the same everywhere in the country (e.g., Dugmore and Buckland 1991, Gathorne-Hardy et al. 2009, Lawson et al. 2007, Ólafsdóttir et al. 2002, Streeter and Dugmore 2014, Vickers et al. 2011). Estimates of vegetation cover at the time of settlement vary depending on the methods used. Thus, a modelling study based on estimates of dynamics of the vegetation cover in response to temperature change suggests that total vegetation cover of Iceland decreased from around 52% at landnám to 28% in 1990, and that the birch (Betula) woodland cover during the same period decreased from an estimate of 8% to ca. 1% (Ólafsdóttir et al. 2001, Streeter et al. 2015). The estimates based on the present-day spatial distribution of birch, historical sources and geographical information which suggest in which areas birch could never have thrived gave different results, suggesting that woodland covered at least 25% of the country at the beginning of settlement, with the total vegetation cover being ca. 65% (e.g., Arnalds 1987). Today woodland coverage is on the increase due to preservation and plantation work. The indigenous woodland species are birch and willow. Birch can grow to considerable heights under favourable conditions, but much of the woodland cover probably consisted of low-growing bushes or brushwood. Archaeological, Written, and Palaeoecological Evidence for Environmental Change in the Wake of Settlement in Iceland: Two Case Studies Guðrún Sveinbjarnardóttir* Abstract - A 12th century source (Íslendingabók) suggests that Iceland was wooded “between mountain and seashore” when the country was first settled in the ninth century. Today only ca. 1% of the surface of the country is wooded and large areas have suffered heavy erosion. The results of the earliest investigations undertaken of environmental change gave a rather uniform picture, showing an equally drastic escalation in erosion and decline in woodland right at landnám, the time when the country is thought to have been settled. Recent research suggests a more diverse picture. This is for example evident in the two areas, Eyjafjallasveit in the South and Borgarfjörður in the West discussed in this article, for which the historical, archaeological and palaeoecological evidence for changes in the environment in the wake of the earliest settlement is surveyed. It demonstrates that the effect of landnám on the environment is not a uniform event which can be applied equally to all sites. A combination of evidence from different disciplines is more likely to provide a truer picture of developments. Selected Papers from the North Atlantic Archaeology Group 2013 Symposium Journal of the North Atlantic *Department of Scandinavian Studies, University College London, London WC1 E6BT, gudrun.s@ucl.ac.uk. 2018 Special Volume 11:1–12 Journal of the North Atlantic G. Sveinbjarnardóttir 2018 Special Volume 11 2 The favoured explanations for the reduction of the woodland cover have been (i) clearance by the initial settlers to make way for cultivation, for which there is some evidence (e.g., Byock 2001:33, Smith 1995:334, Sveinbjarnardóttir et al. 2011:166, Þórarinsson 1944:192ff), (ii) coppicing for charcoal making, which was used for the extraction of iron from iron-ore, and (iii) sheep grazing, suggesting that the sheep, if allowed to graze in the woodland, would cause woodland regeneration failure (e.g., Einarsson 1962:452, Smith 2005:187). Pollen work carried out at several sites in the south-west of the country shows that the woodland was on the decline already before the deposition of the landnám tephra, and that it declined dramatically after that time, when there is also a dramatic increase in grasses, a consequence of the opening up of former wooded areas for the expanding grassland (e.g., Einarsson 1962, Hallsdóttir 1996). This evidence for the arrival of a human population and its effect on the environment is certainly fairly clear cut and dramatic, and all these early charts tell a very similar story about the development of the birch and the grasses. More recently, pollen work has been carried out in other areas showing certain variations in the development of landscape change, calling for different explanations. One of those areas is the Reykholtsdalur valley in the west of Iceland. Here there was birch woodland present in the valley bottom pre-landnám which was quickly cleared soon after the earliest settlement, while the birch on the slopes lingered on until about AD 1150 when it also began to disappear (Erlendsson et al. 2012). Similar results have been obtained from sites in the north of Iceland (Lawson et al. 2007). There could be different reasons for this later clearance of the woodland in some locations. Some researchers have interpreted it as an attempt to manage the use of the woodland which remained after the initial onslaught in the wake of the earliest settlement (Vésteinsson and Simpson 2004). Others have pointed out that this remaining woodland is further away from the farm and was, for that reason simply used later (Erlendsson et al. 2012). Medieval written sources suggest that the right of access to woodland was an important asset. The largest holders of this right were the church farms and it is in the church charters that the earliest information on this is to be found. For Eyjafjallasveit in the South, one of the areas discussed in this article, the earliest charters date from about AD 1270 (DI 2, 84–85), and for Reykholtsdalur in the West, the second study area, from about AD 1150 (DI 1:279–280, Gunnlaugsson 2018). An indication of how important woodland and its preservation was, is the fact that Grágás, a collection of medieval laws compiled in the latter half of the 13th century, has paragraphs in its Land- Claims Section on how to deal with woodland when land changes hands (Grágás 1992:289), and how to behave in and divide co-owned woodland (Grágás 1992:326–330). Here the uses to which woodland is put are specified as being fuel for cooking, as building material, for making charcoal to sharpen scythes by, and for mending domestic items. The importance of this asset and the punishment for misbehaving with it is made very clear. In the following, some of the literary, archaeological and palaeoecological evidence for changes in the environment, particularly as concerns the woodland cover, in the wake of the earliest settlement is discussed in two areas in Iceland: the western extremity of Eyjafjallasveit in the South and the Reykholtsdalur valley, together with one of its shieling areas, in the West. The present author has carried out research in both areas, including as a member of the Leverhulme Trust funded project Landscapes circum landnám (Edwards et al. 2004) which involved a great amount of environmental work. The study of the archaeological remains in Þórsmörk formed part of an earlier Leverhulme Trust funded project (Sveinbjarnardóttir 1983). The site of Reykholt has been the subject of extensive excavations (Sveinbjarnardóttir 2012, 2016) and its shieling areas were the subject of the Reykholt shieling project (e.g., Sveinbjarnardóttir et al. 2011). This article re-visits all this research, adding information from written sources on the ownership of woodland, with the aim of re-assessing the evidence and testing the findings in the light of present knowledge in the field. Case study I: Eyjafjallasveit Eyjafjallasveit lies just west of the southernmost point of Iceland (Fig. 1). It is bounded by large glacial rivers to the east (Jökulsá á Sólheimasandi), west (Álar, a tributary of Markarfljót) and north (Markarfljót), and crowned by the glacier Eyjafjallajökull at its northern boundary. The area is now virtually devoid of trees, except for the inland area of Þórsmörk and parts of Langanes, the stretch of land running alongside the northern boundary towards Þórsmörk, where there is low-growing brushwood. Þórsmörk was formerly a pasture area for the lowland settlements and an important source of wood, and attempts were made to establish permanent farms there during the earliest phase of settlement. Now parts of it have been fenced off and turned into a National Park. Until the recent past Langanes was also an important source of woodland. Journal of the North Atlantic G. Sveinbjarnardóttir 2018 Special Volume 11 3 There are several place-names in the area which can be translated as “woods” indicating that the sites were covered with woodland in the past. At the eastern boundary there is the farm Skógar, in the middle of the area there is Holt which was a major church seat and at the western boundary there are the Mörk farms. At none of these sites is there woodland today. Þórsmörk, where there is still some woodland, means the woods of the pagan god Þór. Supporting this place-name evidence of past woodland are treetrunks, often of considerable size, which have been found in pre-settlement deposits at Stóramörk (Mairs et al. 2006) and Holt (Buckland et al. 1991). The historical, archaeological and environmental evidence Several early settlement sites are known in Þórsmörk (Sveinbjarnardóttir 1983, 1992; Tómasson 1996). At two of them, Steinfinnsstaðir and Þuríðarstaðir efri, slag was found on the surface and analysed as ironworking slag, both smithing and smelting slag (McDonnell 1983). For both activities charcoal is needed. At a third site a quantity of slag was also found on the surface, but it has not been analysed (Tómasson 1996:130ff). The sites in Þórsmörk, all of which are now heavily eroded (Fig. 2), have not been excavated, but have been dated to the earliest period of settlement on the basis of artefacts found, including a bead and a Borre style buckle, both of which can be dated to the 10th century (Fig. 3) (Callmer 1977:77, 85; Graham-Campbell 2013:70). The presence of pagan graves in the area (Eldjárn 2000:46–49), also suggests an early date for the settlement. The Pagan religion was the dominant one in Iceland until Christianity was adopted in ca. AD 1000 (see e.g., Vésteinsson 2000:50). It may be easily assumed that Þórsmörk was an attractive place to settle being wooded and secluded, at least until erosion set in making parts of it uninhabitable. The woodland is now thriving again in places after being fenced off and protected from sheep and exploitation by humans. Evidence has been obtained of landscape instability in Þórsmörk between the eleventh and thirteenth centuries, after which there is some improvement in the stability (Dugmore et al. 2006). The period of instability coincides with the duration of settlement in the area. According to contemporary written sources the farms were already abandoned by the 13th century (Biskupa sögur I, 291, Sturlunga saga I, 532). Although this may have been as a result of erosion, it has been suggested that the church Figure 1. Map showing the study area in Eyjafjallasveit with some of the place names mentioned in the text. farms which owned the woodland at that time wanted to preserve it for their own use when woodland in the lowland had disappeared. Ownership of woodland in Þórsmörk is reported in the earliest preserved charters for the church at Stóridalur, which date to the 14th century (DI 2, 683, DI 3, 262–263), for the church at Holt dating to the 13th, 14th and 15th centuries (DI 2, 84–85, 681–682, DI 6, 329– 330) and the church at Miðbæli from the 13th century (DI 2, 678–679, DI 3, 260) (see Fig. 1). All these churches also owned woodland in Langanes, the stretch of land between the Mörk farms and Þórsmörk. Church charters, which were mainly compiled until the Reformation, list the most valuable assets belonging to the church at any given time. The comprehensive reporting in them of woodland ownership shows how important it was in the medieval period. The Bishop visited once a year and amended the charter as appropriate. There is no information to be had from them later than the 15th century, but by Journal of the North Atlantic G. Sveinbjarnardóttir 2018 Special Volume 11 4 1709, when the first land survey was carried out in Eyjafjallasveit, the woodland belonging to all these churches was much depleted, and only with difficulty said to be still usable for charcoal making in places (JÁM 1, 48, 62, 92). No charcoal pits have been discovered in Þórsmörk, but several have been found on the eroded slopes of Langanes, the stretch of land leading into Þórsmörk from the west. The use of these pits has been dated with tephrochronology, analysis of soil accumulation rate estimates and radiocarbon dating to between the mid tenth and late thirteenth centuries. The pits are consistently cut through a tephra layer dated to AD 920 and sealed by the AD 1341 Hekla tephra layer with a considerable soil accumulation between the latter and the pit, estimated to represent close to one century (Dugmore et al. 2006).2 One of the pits had two charcoal fills (Dugmore et al. 2006, Table 1, Church et al. 2007, Table 1), suggesting that it was re-used at least once. It is clear that these pits were filled with soil after the coal had been removed, probably to avoid the hazard of domestic animals falling into them. The back-filling of charcoal pits is specified in the lawbook Grágás, where charcoal makers are also required to remove the charcoal from the pit within a specific period of time if they are making coal in the land of someone other than themselves, or face a fine (Grágás 1992:329). Environmental evidence obtained through a stratigraphic study of macrofossil remains and sediment-accumulation rates close to the farm of Stóramörk, with dating based on tephrochronology, suggests the existence of woodland cover close to the farm during its earliest occupation; the same research study suggests its disappearance before AD 920. Between AD 920 and AD 1341 the site suffers heavier erosion than any other settlement in the area, includ- Figure 2. One of the eroded sites in Þórsmörk. Figure 3. Viking age artefacts found in Þórsmörk. Journal of the North Atlantic G. Sveinbjarnardóttir 2018 Special Volume 11 5 ing those in Þórsmörk (Mairs et al. 2006). A charcoal pit was found close to Stóramörk. A sample of charred birch from the pit gave dating estimates that fall between AD 865 and AD 960 (Church et al. 2007). This is an earlier date than those on the pits found further away in Langanes to the east of the farm, suggesting that the latter were in use later. The evidence from Stóramörk is from a single trench and therefore localized, but one interpretation for these results is that they may show how woodland exploitation moved to areas in the interior of Iceland as the lowland woodland, closest to the farms, disappeared (Dugmore et al. 2006:342–343). As the woodland diminished, it would have been important to manage what was left of it in the interior. This may be visualized in the later medieval woodland preservation described above, evident in the environmental data from both Þórsmörk and Þjórsárdalur, a valley somewhat further to the west, where a similar study has been carried out (Dugmore et al. 2007). Case study 2: Reykholtsdalur The second study area is Reykholtsdalur in Borgarfjörður in the West (Fig. 4). It is an inland valley which lies about 25 km from the coast, bordered by rivers to the north (Hvítá) and south (Reykjadalsá). The focus of this study was the Reykholt estate, a church farm and chieftain’s centre where extensive excavations have been carried out (Sveinbjarnardóttir 2012, 2016). The site is best known for its 13th century occupant, the author and chieftain Snorri Sturluson. Shieling sites (summer farms) which belonged to Reykholt from the earliest period of its settlement were also studied, in a parallel lying valley to the north (1 and 2 in Fig. 4) and to the south (3). Reykholt also had a right to resources in a third area (4), in the interior to the east. Today the Reykholtsdalur valley is The historical, archaeological and environmental evidence In addition to Reykholt, only one other site in the valley has been subjected to archaeological investigations. This is the cottage farm Háls which lies some distance up the valley. Smithing and smelting slag was discovered at both sites, as well as pits for charcoal making or storage. All of this activity points to the exploitation of woodland. Some of the evidence has been dated to the earliest period of settlement. Tree trunks have also been found in ditches in pre-historic layers, e.g., at Grímsstaðir, the neighbouring farm to Reykholt down the valley,3 as well as the shieling area belonging to Reykholt in the Kjarardalur valley which lies to the north of the Reykholtsdalur valley (Erlendsson et al. 2018, Fig. 7C). In the earliest preserved charter, now thought to date to about AD 1150 (Gunnlaugsson 2018), the church at Reykholt is said to own shielings in this valley, as well as the right to woodland in that general area, and also in the inland area of Geitland (no. 4 in Fig. 4) (DI 1, 279–280). Ownership of these resources is repeated in thirteenth, fourteenth and 15th century charters (DI 1, 466–480, DI 3, 122–123, 222, DI 6, 173–174). Like in Eyjafjallasveit, the account in the earliest land survey carried out in the area in 1709 shows that woodland in these areas is depleted. Instead, Reykholt relies on woodland on one of the farms it now owns, Hraunsás (see below), as well as driftwood rights in the Western Fjords (JÁM 4, 230–231). At the site Háls, charcoal remains were found that seem to indicate burning for clearance before the site was settled in the 10th century (Smith 1995). The investigations have also revealed remains of Figure 4. Map showing the study area in Reykholtsdalur with some of the place names mentioned in the text. completely devoid of trees, but evidence has been retrieved which suggests that it was wooded before the settlement. This is not supported by any place-names like in Eyjafjallasveit, but by pollen evidence and tree-trunks which have been discovered in pre-settlement deposits (see below). The Reykholt site and its surrounding area has been the subject of extensive environmental investigations, carried out as part of the project Landscapes circum landnám (e.g., Sveinbjarnardóttir et al. 2007). The Reykholt shielings project also placed importance on environmental investigations (e.g., Sveinbjarnardóttir et al. 2011). Journal of the North Atlantic G. Sveinbjarnardóttir 2018 Special Volume 11 6 ethnographic evidence describes charcoal making pits as varying in size depending on how much coal was to be made. Thus, a pit making enough coal for 4–5 barrels was 1.9 m in diameter and 1 m deep (Þórarinsson 1974:26). Unfortunately, the earliest archaeological remains at Reykholt were too badly preserved to put the charcoal pit found there into a secure context. Its function is a mystery, but one suggestion is that it may simply have served as a storage pit. The results of pollen analysis carried out by Egill Erlendsson on samples from the Reykholtsdalur valley close to the site of Reykholt show how, before settlement, the valley bottom consisted of wetland with drier patches where trees or shrubs of low stature grew. These would have been easy to clear for agriculture, an activity which is evident in a decline of annual deposition of birch pollen. The trees on the valley slopes were sturdier and lingered on until the mid-12th century when a decline becomes evident, and by ca. AD 1300 the woodland up there had disappeared as well (Erlendsson et al. 2012, 2018). The pollen charts show the usual tremendous increase in grasses right at landnám, probably a result of the disappearance of the woodland, which also manifests itself in a drop in soil organic matter (Erlendsson et al. 2018). The species also indicate drainage, a sign of the introduction of agriculture, and there is evidence of barley growing (Hordeum), a clear sign of the appearance of humans, until ca. AD 1300 when it disappears (Erlendsson et al. 2012, Fig. 9). This is in keeping with written evidence of cereal growing at the Reykholt farm which, according to the sources, had ceased by the 14th century (DI 3, 122–123). Some barley grains were also found during the excavation and dated to the earliest period at the site (Sveinbjarnardóttir 2012, Table 2). Further support for the decrease in the woodland is apparent in the results of micromorphological investigations on material from a midden (577) at the farm site which indicate that there was a change in the fuel use at Reykholt after ca. AD 1300 when peat and dung became more prominent than wood (Sveinbjarnardóttir et al. 2007, 198ff). This shift in the type of fuel is not evident in the written sources until 1708 when peat is said to be available for that purpose in the land of Steindórsstaðir (JÁM 4, 230–231), a farm just to the south of Reykholt. The Shielings of Reykholt As part of The Shielings of Reykholt project two sites, now known as Norðtungusel and Reykholtssel (2 and 1 respectively in Fig. 4), and known from written sources to have been owned by the charcoal making and iron-ore-extraction before the site was settled as a farm (Smith 2005). This is interpreted as the sole purpose of the site before it became a farm. The site lies on the boundary between two of the original land takings (landnám) in the district, and investigations show that the area was covered with woodland at the time of settlement. When enough woodland had been cleared so that farming could begin at the site, a farm was established. It does not seem to have been a suitable location for a farm, however; it was abandoned in the 13th century and never re-occupied. The stretch of land to the north of Háls, along the Hvítá river on the north side of the mountain ridge, from the other large landowner and church farm, Stóri Ás, to Reykholt, was wooded in the past. Remnants of this still existed when the land survey was made for this area in 1708 (JÁM 4, 241–243). Most of this land, together with its woodland, will have belonged to Stóri Ás as indicated by its charter of AD 1258 (DI 1, 593–594), although by at least ca. AD 1600 Reykholt held some land there as well (Jarðabréf, 32, 34). In the 15th century Reykholt had also acquired half of the farm Hraunsás, located somewhat further inland on that side of the mountain ridge, which had good woodland (DI 5, 399–400). It has been suggested that it may, indeed have been Reykholt’s main reason for acquiring the farm (Eyþórsson 2008:67). The excavations at Reykholt produced evidence of ironworking in two places at the site: in the form of slag at the farm site itself and in a smithy to the north of the church, from where charred birch was dated with estimates ranging from AD 1030 to AD 1260. Also found at the farm site was a small pit filled with charcoal which, because of its homogeneity is thought to be the product of charcoal-making. Radiocarbon estimates date it to the earliest period of occupation at the site, around AD 1000 (Sveinbjarnardóttir 2012:58, 88, and Table 2). The pit, which is square in shape, is rather smaller (0.55 x 0.53 x 0.32 m) than known charcoal making pits from the medieval period which are also closer to being circular in shape. For comparison, the charcoal pits found in Langanes ranged from ca. 0.95 x 0.5 m (Church et al. 2007, Fig. 2) to ca. 2 x 1 m (Dugmore et al. 2006, Fig. 8), whereas the largest charcoal making pit found at Háls was 3 x 1 m in size (Smith 2005:190), which implies that it served to make a large quantity of coal, almost certainly for commercial rather than domestic use. This is in keeping with the interpretation of the earliest phase of activity at the site being charcoal making and iron-ore extraction (see above). Late-19th century Journal of the North Atlantic G. Sveinbjarnardóttir 2018 Special Volume 11 7 ore and that fuel may have been scarce when it was produced.4 This is supported by the pollen data which show that there was a new period of environmental decline in the valley after ca. AD 1500 which lasted until about AD 1700, interpreted as showing increased grazing pressure. The timing of this event coincides with the operation of shieling (1), according to the historical and archaeological evidence, which is close to the sampling site (Erlendsson et al. 2018). Whether the piece of slag found at site (2) is evidence of iron-ore extraction in the valley or was carried to the site from elsewhere is not known. Its presence at the site is of interest because iron-ore extraction was practiced at Norwegian shielings. Although such activity is known from Icelandic sites, such as Háls discussed above, no such evidence has so far been found at Icelandic shieling sites. It is possible that the woodland in Kjarardalur was never of good enough quality for charcoal making. In the earliest source for the shieling activity in the valley it is specifically mentioned that wood to be used at the shieling is available from Þverárhlíð, an area to the north and west of the valley (DI I, 279–280). The name of the valley may also shed light on the nature of the woodland. Kjarr means brushwood. Discussion This article begins with a quote from a 12th century written source, suggesting that the woodland cover at settlement in Iceland had by that time diminished drastically. When research in the area of environmental studies began, main emphasis was put on interpreting the evidence as showing the overwhelming effect the appearance of people and their grazing animals had on the environment (Þórarinsson 1961, Einarsson 1962). At the point in the stratigraphy close to the earliest settlement, landnám, now dated by a volcanic ash layer to AD 877 ± 1 (Schmid et al. 2016), drastic changes are apparent in the earliest pollen graphs, including a dramatic decline in birch woodland, increase in grasses as a result of this and an escalation in erosion (Einarsson 1962, Hallsdóttir 1987, Hallsdóttir 1996). These pollen graphs were all from sites in a limited area in the south-west of the country and produced very similar results. Although it is not disputed that there were noticeable environmental changes right at landnám, more recent work, carried out in different parts of the country, has refined the picture and made it more nuanced (e.g., Dugmore 2014; Erlendsson et al. 2012, 2018; Lawson et al. 2007; Ólafsdóttir et al. 2002; Reykholt farm (DI I, 279–280), were investigated in the Kjarardalur valley which lies parallel to and ca. 12.5 km in a direct line to the north of the Reykholtsdalur valley. As far as we know, there has never been permanent habitation in this valley, but several shieling sites are known. Limited excavations were carried out at both sites where ruins are visible on the surface. Radiocarbon dates obtained from site (2) which is lower down in the valley suggest that it was used already in the tenth or early 11th century, followed by site (1) which is further inland where the earliest cultural activity is dated to the 14th century (Sveinbjarnardóttir et al. 2011; Eyþórsson et al. 2018, Table 1). These results are supported by the written sources which suggest that Reykholt moved its shieling activity further inland at about this time (Eyþórsson 2008:142–143). Site (2) was still being used at that time, but now rented out. Samples were analysed from a location to the east of site (1). The results showed that the area was completely covered with birch before landnám and that it declined shortly after the deposition of the landnám tephra. Since the sample site is some distance from where shieling activity seems to have been taking place at the time, a possible explanation is that the woodland was cleared to create better grazing land generally in the area (Erlendsson et al. 2018). Alternatively, the grazing sheep might themselves be responsible for the disappearance of the woodland which seems to have been of the lower stature and therefore easily got at by the sheep. With the decrease of the birch pollen comes an increase in microscopic charcoal (Erlendsson et al. 2018). Whether this is a sign of some human activity, such as land clearance, charcoal production or other use of fire in the vicinity of the sampling site, or created through natural causes is uncertain. It is tempting to suggest the former in view of evidence obtained from site (2) where a layer of charcoal was found inside one of the buildings, imbedded in the natural subsoil (13 in Fig. 5). There is considerable soil accumulation between the charcoal layer and the first floor layer (10 in Fig. 5) in the building which supports the suggestion that it was created as a result of land clearance before the structure was built. The excavation area was, however, too small to be able to establish this interpretation with any certainty. A piece of clean iron smelting slag was found in the floor (10 in Fig. 5) which was radiocarbon dated to the fifteenth or 16th century (Sveinbjarnardóttir and Dahle 2009; Eyþórsson et al. 2018, Table 1). An analysis established that it derives from bog Journal of the North Atlantic G. Sveinbjarnardóttir 2018 Special Volume 11 8 Streeter and Vickers et al. 2011). It is clear in the pollen graphs that deterioration in the vegetation cover did not only take place as a result of the appearance of people and their grazing animals at landnám. Although the disappearance of the birch woodland was rapid at that point, it had already begun to recede somewhat before that time, probably because of a change in the climate. This has also been shown to be true for the onset of erosion (Geirsdóttir et al. 2009). Other studies (e.g., Dugmore et al. 2006, 2007) suggest that woodland management may have played a part in the survival of woodland for longer in some areas. As demonstrated before with reference to medieval church charters, access to woodland was an important asset from early on, and the extent to which its treatment is covered in medieval laws indicates that woodland management was regarded as important. The two study areas discussed in this article provide ample evidence, written, archaeological and environmental, of changes in the vegetation cover at the time of settlement. In Eyjafjallasveit the woodland cover in the lowland coastal area, indicated both by place-name and palaeoecological evidence, disappeared right at landnám (Buckland et al. 1991). This is where most of the earliest farms were placed. The demise of the birch woodland is probably largely a result of land clearance to make room for the farms and their agricultural activities. In the more inland and less densely settled areas the woodland closest to the farms seems to have disappeared first, followed by exploitation further inland and some distance away from the farms (Church et al. 2007, Dugmore et al. 2006). Archaeological evidence shows that a wooded area in the interior, Þórsmörk, separated from the nearest settlement by almost 20 km and a ferocious river, attracted settlers during the earliest period, but that the settlement did not last for a long period of time. Environmental evidence of landscape instability coincides with the archaeologically dated duration of the settlement. This suggests that erosion may have played a part in the abandonment of the farms. The influence of the church farms which were, according to written sources, the dominant owners of woodland in the area, at least in the later medieval period, may also have played a part. By forcing the occupants of the settlements to relocate, the woodland could be preserved and managed for their own use. There is evidence in the environmental record of some revival of the woodland after the abandonment which may support this suggestion, although the disappearance of humans and their grazing animals may by themselves be reason enough. Recent research carried out in the north of Iceland (Vésteinsson et al. 2014) can be used as another example where with additional data an alternative explanation for the abandonment of an inland marginal area can be suggested. Several sites were identified, with occupation dated to between the tenth and twelfth centuries. This is an area which was investigated by Sigurður Þórarinsson in the 1970s (Þórarinsson 1977). His conclusion was that the area was abandoned as a result of erosion which was already in progress by AD 1104. A reassessment of one of his soil profiles and further investigations in the area gave a different result, showing that the environmental degradation was less severe. Taking this into account and combining it with archaeological evidence from a neighbouring area it was suggested that the settlement of this marginal area was a deliberate temporary decision Figure 5. Section through main building complex at Norðtungusel showing charcoal lens (13). Journal of the North Atlantic G. Sveinbjarnardóttir 2018 Special Volume 11 9 had to be placed on peat and dung for that purpose, supporting the notion that woodland for charcoal making was being depleted. Conclusion Research into environmental change has benefitted greatly from advances in methodology, producing results which throw a more refined and different light on the issue. An increased number of study areas in different parts of the country is also helping in this respect. It is no longer a foregone conclusion that all areas suffered drastic environmental degradation immediately at the point of settlement. This paves the way for different interpretations for example of the fate of inland settlements in marginal areas, with the aid of information obtained from other sources, such as written sources and archaeological findings. The central theme in the two areas discussed in this article is the fate of the woodland, the importance of which is so evident in the medieval written sources and the demise of which is apparent e.g., in the place-name, archaeological and palaeoecological record. Trees are vulnerable to over-exploitation and in delicate landscapes, like Iceland has, complete disappearance is likely without the undertaking of some measures to manage it. The fate of the woodland in the lowland is similar in both study areas: it disappears right at landnám, probably the result of land clearance to make room for agriculture. The development in the more inland and upland areas varies, but also seems to be conducted by human interference to various degrees. Here, economic factors may have played a part. In the Reykholtsdalur valley, where the high status farm Reykholt became a big landowner from early on and had access to woodland in its resource areas away from the homeland, it was initially feasible to preserve the woodland on the slopes closer to home until also that begins to disappear in the 12th century. In Eyjafjallasveit the evidence shows that exploitation is also rather brutal and continuous until ca. the 13th century when preservation becomes apparent, perhaps resulting from measures taken by the rich landowners. The archaeological record supports this change at that time. Iron-ore extraction which requires a great quantity of charcoal, ceases, and peat and dung become more prominent than wood as fuel. The effect of humans and their grazing animals on the environment at landnám is still a dominant one, but by combining different strands of evidence from a variety of disciplines a more accurate picture is gradually emerging. in an attempt by the settlers to improve their social standing (Vésteinsson et al. 2014). There is no evidence in this area of woodland cover and hence its preservation through management being the reason for the abandonment, but is suggested that a decision may have been made at the point of abandonment to use the area differently. This interpretation is purely hypothetical, as acknowledged by the authors, which is impossible to prove on the basis of the available data. Further research in other parts of the country might help in throwing more light on this issue. In the Reykholtsdalur valley the picture is similar to that in Eyjafjallasveit in that woodland cover in the valley bottom close to Reykholt, which was the main farming area, was cleared right at landnám while the woods on the slopes lingered on into the medieval period. This may be because woodland on the slopes was preserved and managed and only exploited when woodland in other areas further away from the farm had become scarce. Reasons for other changes in the vegetation cover could be linked to changes in the operation of the farm. An example is Kjarardalur, one of the shieling areas belonging to the Reykholt farm, where there is evidence in the environmental record, dated to the 14th century, of temporary revival of the woodland, suggesting that there was less activity, that is less grazing pressure in the shieling area at that time. This change coincides with the move of the Reykholt shieling to the site placed further inland in the valley. Information in 14th century church charters for the Reykholt church shows a temporary drop by almost a half in the number of sheep owned by the church compared with what it was in the 12th century (Eyþórsson 2008, Tafla 8). It is possible that there is a connection between the reduction in the number of sheep, the move to the new site and the relief in grazing pressure in the shieling area, but it cannot be established. There is no available data for the stock levels in the 13th century. It is possible that the reduction in numbers of sheep occurred earlier than the 14th century. Charcoal production was one of the most important uses of the birch wood during the earliest period of settlement. It was used for iron-ore extraction, evidence of which has been found at most early sites (see e.g., Sveinbjarnardóttir 1983, Vésteinsson et al. 2014). The importance of ironore extraction diminished when the import of iron became more feasible in the 15th century (Jóhannesson 1943:56), This would have eased the demand for charcoal somewhat, although access to it continued to be important to use as fuel for cooking and heating, as is evident from the written sources. These sources also indicate that increased reliance Journal of the North Atlantic G. Sveinbjarnardóttir 2018 Special Volume 11 10 Edwards, K.J., P.C. Buckland, A.J. Dugmore, T.M. McGovern, I.A. Simpson, and G. Sveinbjarnardóttir. 2004. Landscapes circum landnám: Viking settlement in the North Atlantic and its human and ecological consequences: A major new research programme. Pp. 260–271, In R.A. Housley and G. Coles (Eds.). Atlantic Connections and Adaptations. Symposia of the Association for Environmental Archaeology No. 21. Oxbow Books, Oxford, UK. 271 pp. Einarsson, Þ. 1962. Vitnisburður frjógreiningar um gróður, veðurfar og landnám á Íslandi. Saga 24:442–469. Eldjárn, K. 2000. Kuml og haugfé úr heiðnum sið á Íslandi. 2nd rev. A. Friðriksson (Ed.). Mál og menning, Þjóðminjasafn Íslands, Fornleifastofnun Íslands, Reykjavík, Iceland. 615 pp. Erlendsson, E., K. Vickers, F. Gathorn-Hardy, J. Bending, B. Gunnarsdóttir, G. Gísladóttir, and K.J. Edwards. 2012. Late-Holocene Environmental History of the Reykholt Area, Borgarfjörður, Western Iceland. Pp. 17–47, In H. Þorláksson and Þ.B. Sigurðardóttir (Eds.). From Nature to Script. Reykholt, Environment, Centre and Manuscript Making. Snorrastofa, Cultural and Medieval Centre, Reykholt, Iceland. 294 pp. Erlendsson, E., K.J. Edwards, K. Vickers, G. Sveinbjarnardóttir, and G. Gísladóttir 2018. The Palaeoecology and Cultural Landscapes Associated with Reykholt. Pp. 161–203, In G. Sveinbjarnardóttir and H. Þorláksson (Eds.). Snorri Sturluson and Reykholt. The Author, his Life, Works and Environment at Reykholt in Borgarfjörður, Western Iceland. Museum Tusculanum Press, in collaboration with Institute of History, University of Iceland and Snorrastofa at Reykholt, Copenhagen, Denmark. Eyþórsson, B. 2008. Búskapur og rekstur staðar í Reykholti 1200–1900. Sagnfræðistofnun Háskóla Íslands, Reykjavík, Iceland. 165 pp. Eyþórsson, B., E. Erlendsson, G. Gísladóttir, and G. Sveinbjarnardóttir 2018. Natural Resources. Access and Exploitation. Pp. 205–234, In G. Sveinbjarnardóttir and H. Þorláksson (Eds.). Snorri Sturluson and Reykholt. The Author, His Life, Works and Environment at Reykholt in Borgarfjörður, Western Iceland. Museum Tusculanum Press, in collaboration with Institute of History, University of Iceland and Snorrastofa at Reykholt, Copenhagen, Denmark. Gathorne-Hardy, F.J., E. Erlendsson, P.G. Langdon, and K.J. Edwards. 2009. Lake sediment evidence for late Holocene climate change and landscape erosion in western Iceland. Journal of Paleolimnology 42:413–426. Geirsdóttir, Á., G.H. Miller, Th. Thordarson, and K.B. Ólafsdóttir. 2009. A 2000 year record of climate variations reconstructed from Haukadalsvatn, West Iceland. Journal of Paleolimnology 41:95–115. Graham-Campbell, J. 2003. Viking Art. Thames & Hudson, London, UK. 208 pp. Grágás. 1992. Lagasafn íslenska þjóðveldisins. G. Karlsson, K. Sveinsson, and M. Árnason (Eds.). Mál & menning, Reykjavík, Iceland. Acknowledgements Thanks are issued to the Leverhulme Trust, the Icelandic Centre for Research (Rannís, grant no. 080655021), Kultur - og kirkjedepartementet in Norway, the bank Sparisjóður Mýrarsýslu, the Culture and Medieval Centre Snorrastofa, the Ministry of Agriculture in Iceland and the Archaeology Fund in Iceland, all of which supported different parts of the research discussed in the article. I also thank all my co-workers in the different projects whose work I have used and quoted, especially Egill Erlendsson, who provided all the pollen data for the Reykholt study areas, and Stuart Brooks for assistance with finalizing the maps. Literature Cited Arnalds, A. 1987. Ecosystem disturbance in Iceland. Arctic and Alpine Research 19(4):508–513. Biskupa Sögur. 1858–1878. G. Vigfússon (Ed.). 2 vols. S.L. Möller and Hið íslenzka bókmenntafélag, Copenhagen, Denmark and Reykjavík, Iceland. Buckland, P.C., A.J. Dugmore, D.W Perry and G. Sveinbjarnardóttir. 1991. Holt in Eyjafjallasveit, Iceland. A paleoecological study of the impact of landnám. Acta Archaeologica 61:252–271. Byock, J. 2001. Viking Age Iceland. Penguin Books, London, UK. 448 pp. Callmer, J. 1977. Trade Beads and Bead Trade in Scandinavia ca. 800–1000 A.D. Acta Archaeologica Lundensia, Series in 4o Nr. 11. Lund, Sweden. 229 pp. Church, M.J, A.J Dugmore, K.A Mairs, A.R Millard, G.T Cook, G. Sveinbjarnardóttir, P.A Ascough, and K.H Roucoux. 2007. Charcoal Production during the Norse and Early Medieval Periods in Eyjafjallahreppur, Southern Iceland. Pp. 659–672, In C. Bronk Ramsey and T.F.G. Higham (Eds.). Proceedings of the 19th International 14C Conference. Radiocarbon 49(2). DI = Diplomatarium Islandicum. 1857–1952. Íslenzkt fornbréfasafn. 16 vols. of Icelandic documents. S.L. Möller and Hið íslenzka bókmenntafélag, Copenhagen, Denmark and Reykjavík, Iceland. Dugmore, A.J., and P.C. Buckland. 1991. Tephrochronology and late Holocene soil erosion in South Iceland. Pp. 147–159, In J. Maizels and C. Caseldine (Eds.). Environmental Change in Iceland, Kluwer, Dordrecht, Netherlands. 332 pp. Dugmore, A.J., M.J. Church, K.-A. Mairs, T.H. McGovern, A.J. Newton and G. Sveinbjarnardóttir. 2006. An overoptimistic pioneer fringe? Environmental Perspectives on Medieval Settlement Abandonment in Þórsmörk, south Iceland. Pp. 335–346, In J. Arneborg and B. Grønnow (Eds.). Dynamics of Northern Societies. Proceedings of the SILA/NABO Conference on Arctic Anthropology, Copenhagen May 10th–14th, 2004. PNM, Copenhagen, Denmark. 414 pp. Dugmore, A.J., M.J. Church, K.-A. Mairs, T.H. McGovern, S. Perdikaris, and O. Vésteinsson. 2007. Abandoned Farms, Volcanic Impacts, and Woodland Management: Revisiting Þjórsárdalur, the “Pompeii Of Iceland.” Arctic Anthropology 44(1):1–11. Journal of the North Atlantic G. Sveinbjarnardóttir 2018 Special Volume 11 11 Grönvold, K, N. Óskarsson, S.J. Johnsen, H.B. Clausen, C.U. Hammer, G. Bond, and E. Bar. 1995. Ash layers from Iceland in the Greenland GRIP ice core correlated with oceanic and land sediments. Earth and Planetary Science Letters 135:149–155. Gunnlaugsson, G.M. 2018. The deed of the church of Reykholt. Pp. 279–289, In G. Sveinbjarnardóttir and H. Þorláksson (Eds). Snorri Sturluson and Reykholt. The Author, his Life, Works and Environment at Reykholt in Borgarfjörður, Western Iceland. Museum Tusculanum Press, in collaboration with Institute of History, University of Iceland and Snorrastofa at Reykholt, Copenhagen, Denmark. Hallsdóttir, M. 1987. Pollen Analytical Studies of Human Influence on Vegetation in Relation to the Landnám Tephra Layer in Southwest Iceland. Lundqua Thesis 18, Lund University, Lund, Sweden. 45 pp. Hallsdóttir, M. 1996. Frjógreining. Frjókorn sem heimild um landnám. Pp. 123–134, In G.Á. Grímsdóttir (Ed.). Um landnám á Íslandi. Fjórtán erindi. Vísindafélag Íslendinga. Ráðstefnurit V, Reykjavík, Iceland. 200 pp. Íslendingabók. Landnámabók. 1968. J. Benediktsson (Ed.). Íslenzk fornrit I. Reykjavík. Hið íslenzka fornritafélag. 525 pp. JÁM = Jarðabók Árna Magnússonar og Páls Vídalín. Vol. 1 1913–1917, Vol. 4 1925–1927. S.L. Möller, Copenhagen, Denmark. Jarðabréf frá 16. og 17. öld: útdrættir. 1993. G.F. Guðmundsson (Ed). Hið íslenzka fræðafélag, Copenhagen, Denmark. 359 pp. Jóhannesson, Þ. 1943. Járngerð. Pp. 40–58, In G. Finnbogason (Ed.) Iðnsaga Íslands. Vol 2. Iðnaðarmannafélagið í Reykjavík, Reykjavík, Iceland. 424 pp. Larsen G. 1996. Gjóskutímatal og gjóskulög frá tíma norræns landnáms á Íslandi. Pp. 81–106, In G.Á. Grímsdóttir (Ed.). Um landnám á Íslandi. Societas scientiarum Islandica, Ráðstefnurit V, Reykjavík, Iceland. 200 pp. Lawson, I.T., F.J. Gathorne-Hardy, M.J. Church, A.J. Newton, K.J. Edwards, A.J. Dugmore, and Á. Einarsson. 2007. Environmental impacts of the Norse settlement: Palaeoenvironmental data from Mývatnssveit, northern Iceland. Boreas 36:1–19. McDonnell, J. G. 1983. Ironworking slags from Steinfinnsstaðir and Þuríðarstaðir efri, Þórsmörk. Appendix II. Pp. 59–61, In G. Sveinbjarnardóttir. Byggðaleifar á Þórsmörk, Árbók Hins íslenzka fornleifafélags 1982. Mairs, K.-A., M.J. Church, A.J. Dugmore, and G. Sveinbjarnardóttir. 2006. Degrees of success: Evaluating the environmental impacts of successful longterm settlement in south Iceland. Pp. 365–374, In J. Arneborg and B. Grønnow (Eds.). Dynamics of Northern Societies. SILA/NABO conference Copenhagen May 2004. 415 pp. Ólafsdóttir, R., P. Schlyter, H.V. Haraldsson. 2001. Simulating Icelandic vegetation cover during the Holocene. Implications for long-term degradation. Geografiska Annaler 83:203–215. Ólafsdóttir, R., and H. Guðmundsson. 2002. Holocene land degradation and climatic change in northeastern Iceland. The Holocene 12:159–167. Schmid, M.M.E., A.J. Dugmore, O. Vésteinsson, and A.J. Newton. 2017. Tephra isochrons and chronologies of colonisation. Quaternary Geochronology 40:56–66. Smith, K.P. 1995. Landnám: The settlement of Iceland in archaeological and historical perspective. World Archaeology 26(3):319–347. Smith K. 2005. Ore, Fire, Hammer, Sickle: Iron Production in Viking Age and Early medieval Iceland. Pp. 183–206, In R. Bork (Ed.). De Re Metallica: The Use of Metal in the Middle Ages, AVISTA Studies in the History of Medieval Technology, Science and Art, Vol. 4. Ashgate Publishing, Aldershot, UK. Streeter R.T., and A.J. Dugmore. 2014. Late-Holocene land surface change in a coupled social-ecological system, southern Iceland: A cross-scale tephrochronology approach. Quaternary Science Reviews 86:99–114. Streeter, R.T., A.J. Dugmore, I.T. Lawson, E. Erlendsson, and K.J. Edwards 2015. The onset of the palaeoanthropocene in Iceland: Changes in complex natural systems. The Holocene 25(10):1662–1675. Sveinbjarnardóttir, G. 1983. Byggðaleifar á Þórsmörk. Árbók Hins íslenzka fornleifafélags 1982:20–61. Sveinbjarnardóttir, G. 1992. Farm Abandonment in Medieval and Post-Medieval Iceland: An Interdisciplinary Study. Oxbow Monograph 17. Oxbow Books, Oxford, UK 192 pp. Sveinbjarnardóttir, G. 2012. Reykholt. Archaeological Investigations at a High Status farm in Western Iceland. Snorrastofa and the National Museum of Iceland, Reykjavík, Iceland. 367 pp. Sveinbjarnardóttir, G. 2016. Reykholt: The Church Excavations. The National Museum of Iceland, in collaboration with Snorrastofa and University of Iceland Press, Reykjavík, Iceland. 310 pp. Sveinbjarnardóttir, G., E. Erlendsson, K. Vickers, T.H. McGovern, K.B. Milek, K.J. Edwards, I.A. Simpson, and G. Cook. 2007. The palaeoecology of a high status Icelandic farm. Environmental Archaeology 12(2):197–216. Sveinbjarnardóttir, G. and K. Dahle. 2009. The Shielings of Reykholt. Archaeological Investigations in 2009. Unpublished report. 25 pp. Sveinbjarnardóttir, G., K. Dahle, E. Erlendsson, G. Gísladóttir, and K. Vickers. 2011. The Reykholt shieling project: Some preliminary results. Pp. 164– 177, In S. Sigmundsson (Ed.). Viking Settlements and Viking Society. Papers from the Proceedings of the Sixteenth Viking Congress, Iceland, 17–23 August 2009. Hið íslenzka fornleifafélag. University of Iceland Press. Reykjavík, Iceland. 511 pp. Sturlunga saga. 1946. J. Jóhannesson, M. Finnbogason, and K. Eldjárn (Eds.). 2 vols. Sturlungaútgáfan, Reykjavík, Iceland. 1110 pp. Tómasson, Þ. 1996. Þórsmörk. Land og saga. Land og mynd, Reykjavík, Iceland. 304 pp. Journal of the North Atlantic G. Sveinbjarnardóttir 2018 Special Volume 11 12 Vésteinsson, O. 2000. The Christianization of Iceland. Priests, Power and Social Change 1000–1300. Oxford University Press, Oxford, UK. 280 pp. Vésteinsson, O., and I.A. Simpson. 2004. Fuel utilisation in Pre-Industrial Iceland: A Micro-Morphological and Historical Analysis. Pp. 181–187, In G. Guðmundsson (Ed.). Current Issues in Nordic Archaeology. Proceedings of the 21st Conference of Nordic Archaeologists. Society of Icelandic Archaeologists, Reykjavík, Iceland. 214 pp. Vésteinsson, O., M. Church, A. Dugmore, T. McGovern, and A. Newton. 2014. Expensive errors or rational choices: The pioneer fringe in Late Viking Age Iceland. European Journal of Post-Classical Archaeologies 4:39–68. Vickers, K., E. Erlendsson, M.J. Church, K.J. Edwards, and J. Bending. 2011. 1000 years of environmental change and human impact at Stóra-Mörk in southern Iceland: A multiproxy study of a dynamic and vulnerable landscape. The Holocene 21(6):979–995. Þórarinsson, S. 1944. Tefrokronologiska studier på Island. Þjórsárdalur och dess förödelse. Ejnar Munksgaard, Köbenhavn, Denmark. 217 pp. Þórarinsson, S. 1961. Uppblástur á Íslandi í ljósi öskulagarannsókna. Ársrit Skógræktarfélags Íslands 1961:17–54. Þórarinsson, S. 1977. Gjóskulög og grónar rústir. Brot út íslenskri byggðasögu. Árbók Hins íslenzka fornleifafélags 1976:5–38. Þórarinsson, S. 1981. The application of tephrochronology in Iceland. Pp. 109–134, In S. Self and R.S.J. Sparks (Eds.). Tephra Studies. Reidel, Dordrecht, Holland. 481 pp. Þórarinsson, Þ. 1974. Þjóðin lifði en skógurinn dó. Ársrit Skógræktarfélags Íslands. 1974:16–29. Endnotes 1Some evidence of human activity before the deposition of the landnám layer has been discovered, but so far the evidence shows that the bulk of the settlement took place after that time. 2Since this was written, charcoal pits have been found in eroded sections in Langanes, which are above the Hekla 1341 tephra layer, and even above the Katla 1500 tephra layer (Richard Streeter, pers. comm. January 2018). There are, indeed, accounts of some woodland and its exploitation in the area as late as the 18th century (Tómasson 1996:158-161). 3Guðmundur Kristinsson, farmer at Grímsstaðir, pers. comm. 2005. 4The piece was analysed by Professor Thilo Rehren at University College London. Unpublished data.