Pastures Found… Farming in Greenland (re)introduced

Challenges and solutions Northern Worlds – Report from workshop 2 at the National Museum, 1 November 2011 Edited by Hans Christian Gulløv, Peter Andreas Toft and Caroline Polke Hansgaard Challenges and solutions Report from workshop 2 at the National Museum, 1 November 2011 Copyright © The National Museum and the authors 2012 Edited by Hans Christian Gulløv, Peter Andreas Toft and Caroline Polke Hansgaard Translated and revised by James Manley Layout Anne Marie Brammer Printed in Denmark by Rosendahls – Schultz Grafisk ISBN: 978-87-7602-192-4 Published with financial support from the Augustinus Foundation and the National Museum A digital version of the publication can be found on the home page of the National Museum: http://nordligeverdener.natmus.dk Front cover illustration: Rødøy in Flatøysund, Alstahaug area, Helgeland, South Nordland Photo: Flemming Kaul Northern Worlds – Challenges and solutions Report from workshop 2 at the National Museum, 1 November 2011 Edited by Hans Christian Gulløv, Peter Andreas Toft and Caroline Polke Hansgaard Copenhagen 2012 2 Contents • Challenges and solutions – status of Northern Worlds Hans Christian Gulløv 6 • A sense of snow? Archaeology, weather and the conception of northernness | Bjørnar Olsen 9 Research theme A Climate changes and society: When climate boundaries move 24 • The landscape and climate of the early Mesolithic hunters of Lundby Mose, southern Zealand – The end of the last glacial period and the Preboreal warming | Catherine Jessen 26 • ’Small trees’ from North East Greenland | Claudia Baittinger 30 • Kitchen middens and climate change – what happens if permafrozen archaeological remains thaw? Henning Matthiesen, Jørgen Hollesen and Jan Bruun Jensen 36 • Conservation and drying methods for archaeological materials modified for use in northern areas Martin Nordvig Mortensen, Inger Bojesen-Koefoed, Jan Bruun Jensen, Poul Jensen, Anne Le Boëdec Moesgaard, Natasa Pokubcic, Kristiane Strætkvern, David Gregory, Lars Aasbjerg Jensen, Michelle Taube and Nanna Bjerregaard Pedersen 42 • The Weather War: The German operation ‘Bassgeiger’ on Shannon Island 1943/44 | Tilo Krause and Jens Fog Jensen 46 • Depopulation of the Cape Farewell region | Einar Lund Jensen 55 • The whaler and the ostrich egg – Introduction to a project on life on the North Frisian Islands and whaling in the Arctic Ocean | Christina Folke Ax 58 Farming on the edge: Cultural landscapes of the North 70 • Agricultural landscapes of Arctic Norway | Flemming Kaul 72 • Pioneering farmers cultivating new lands in the North – The expansion of agrarian societies during the Neolithic and Bronze Age in Scandinavia | Lasse Sørensen 87 • Shetland – the Border of Farming 4000-3000 B.C.E. Ditlev L. Mahler 125 • Resources, mobility and cultural identity in Norse Greenland 2005-2010 | Jette Arneborg and Christian Koch Madsen 137 • Pastures Found… Farming in Greenland (re)introduced Christian Koch Madsen 142 • Churches, Christianity and magnate farmers in the Norse Eastern Settlement | Jette Arneborg 167 • Greenland dietary economy | Jette Arneborg 171 • Agriculture on the edge – the first finds of cereals in Norse Greenland | Peter Steen Henriksen 174 Research theme B Northern Worlds Workshop 2 4 Contents Research theme C Networks in the North: Communication, trade and culture markers 178 • A common sea – the Skagerrak and the Kattegat in the Viking Age | Anne Pedersen 180 • Networks in the north – foreign artifacts in the hands of the Vikings | Maria Panum Baastrup 187 • Nørremølle – the largest Viking silver hoard of Bornholm – Interactions in the Baltic Sea | Gitte Tarnow Ingvardson 192 • Greenlandic runic inscriptions | Lisbeth M. Imer 200 • Skin Clothing from the North | Anne Lisbeth Schmidt 205 Close-up of a string of beads on an amaut, a woman’s jacket, combining large worn 18th-century glass beads with unworn seed beads produced in the 19th century. Photo: Peter Andreas Toft. Pinhoulland seen from the north west down towards Voe of Browland. Photo: D. Mahler. • Challenges of cultural and colonial encounters – European commodities in the Historical Thule Culture | Peter Andreas Toft 209 • Timber houses in Greenland – diffusion and innovation Niels Bonde, Thomas S. Bartholin, Claudia Baittinger and Helge Paulsen 228 • Tunit and the birds – echoes of another world Martin Appelt and Mari Hardenberg 231 • Memory of a myth – a unique Late Dorset ritual structure Ulla Odgaard 247 • Pre-Christian Cult Sites – archaeological investigations Josefine Franck Bican, Anna Severine Beck and Susanne Klingenberg 261 • Contributors 269 Northern Worlds Workshop 2 142 Pastures Found… Farming in Greenland (re)introduced Christian Koch Madsen Danish Middle Ages and Renaissance The Norse settlements that were established in Greenland around the turn of the first millennium AD were based on farming practices and cultural traditions reaching back across the North Atlantic to Norway and wider Scandinavia. The PhD project Landscape and Livelihood, Economy and Hierarchy: Pastoral Farming in Norse Greenland aims at outlining the Norse Greenlandic farming communities as they developed from a shared North Atlantic landnám setup through nearly 500 years of existence on the unique and dramatically changing Arctic frontier. With the geographical focus on the central Eastern Settlement, South Greenland, and using recent precision survey data on ruins from this core Norse settlement area, the project explores how traditional archaeological data and research themes may be expanded and nuanced – for example by using critically and theoretically weighed historical evidence – to understand specific and complex human-environment dynamics (fig. 1). As exemplified in this short paper, for instance, the case study of pre-modern Greenlandic farming shows how initial inexperience in farming and herding in the specific Greenlandic setting led to catastrophic declines in livestock numbers; consequences, and subsequent adaptations which may also have been experienced and implemented by the Norse settlers. Farming reintroduced Some 450 years after the Norse population abandoned their settlements in Greenland in 1450 AD, livestock farming was reintroduced to Greenland. Certainly, small-scale gardening and the tending of a few domestic animals had been practiced ever since the earliest colonization in the 18th century (fig. 2) (e.g. Erslev 1877; Jespersen 1916; Kleivan 1983); however, from 1906 onwards, farming increasingly took on serious scope as a full-time professional occupation. This was the fulfilment of a vision sustained 143 Fig. 1. View of modern sheep farm and hayfields in the central Vatnahverfi region, South Greenland. Like all existing Inuit sheep farms, this one at Tasilikulooq lies at the site of a Norse farmstead (E171); the 20th-century reintroduction of sheep farming in Greenland was, effectively, a second landnám developing along lines not too dissimilar to Norse settlement, thus enabling a discussion of aspects of potential overlap. Photo: C.K. Madsen 2006. Fig. 2. Julianehåb Colony, present-day Qaqortoq, around 1830. Visible close to the fjord are, left, a parcelled-out garden and, right, three grazing cattle. Small-scale subsidiary household farming had been a typical element of the colonies, mission and trade stations ever since colonization (watercolour by J. M. Mathiesen 1800-1860). Northern Worlds Workshop 2 144 since the earliest decades of Danish colonial rule (Bendixen 1927; Gad 1969), a vision which also seems to have been nourished by the historical awareness of the Norse settlers and by the prevailing disbelief in the shortcomings of their essentially European farming tradition (e.g. Wormskjold 1814). Renewed farming in South Greenland developed from a Danish Government sheep-breeding experiment: in 1906, 19 Faroese sheep were sent to South Greenland, followed in 1915 by another 170 sheep, this time of Icelandic stock, to breed a stable population from which to expand sheep farming locally (Chemnitz 1920; Walsøe 1919). Except for a very few earlier and later imported sheep of varied origin (Austreheim et al. 2008; Johnstrup 1878; Walsøe 1919: 42), the entire present Greenlandic sheep population grew from these first individuals. As shown in fig. 3, by 1966 sheep numbers had reached a peak hitherto of an estimated 47,968 ewes (i.e. not including lambs, rams, and wethers), a population development spanning only 60 years. From an archaeological perspective, this sheep population development is in itself an interesting observation; the aggregate 189 sheep imported in the early 20th century does not seem to significantly exceed the likely number of a Norse founding sheep population. Accordingly, the development from a landnám population platform to near-maximum sheep stocking levels could potentially be achieved in the specific environmental setting of South Greenland in as little as 60 years, i.e. a few generations. Besides providing an estimate of the potential rapidity of Norse livestock development, this covers so brief a timespan that it could problematize the identification of landnám phases in zooarchaeological bone assemblages from Greenland. However, this and the following lines of argumentation all build on the premise that there is considerable overlap between Norse and recent pastoral farming practices, a premise that warrants pause for critical consideration. Renewed farming in Greenland did not develop in the form of the reoccupation of Norse farmsteads by Scandinavian descendants, as originally envisaged by Danish Government officials; instead it was realized by industrious Inuit; with only the most basic technological aids, and often in areas much isolated from established infrastructure, pioneer Inuit, from the 1930s onwards, reclaimed Norse farmsteads and fields (e.g. Christensen 1954; Kristiansen 1955). This was effectively a second landnám. Nonetheless, it was a landnám developing from a different cultural background and aided by mid20th-century agricultural technology and knowledge. Clearly then, it is imperative 145 explicitly to single out both potential overlaps and obvious differences between Norse and recent Inuit farming, a discussion that can only be presented in summary form in this short paper. First, the project sections off the period 1906-1976, designated the period of pre-modern farming, as the most pertinent for a comparative case study. Such a division is obviously somewhat capricious and constructed, since Greenlandic farming developed continuously. However, the preceding phase of colonial farming was characterized by being small-scale, subsidiary household production (Bendixen 1927; Bentzen 1920: 35) tied to the colonial settlement pattern, i.e. the trade and mission stations that were, in turn, tied to Inuit coastal settlement (fig. 2) (Amdrup et al. 1921a, 1921b). Thus, apart from two farming settlements established at the inner fjord sites of Igaliku in 1782 (Bóbe 1915) and Narsaq in the 1880s (Amdrup et al. 1921b), both suggestively also sites of Norse magnate farmsteads, colonial farming had a geographical distribution markedly different from the Norse one, something that would also to a certain extent characterize the early settlement pattern of Inuit farmers. The period of farming developments following 1976, here designated modern farming, is harder to delimit decisively from pre-modern farming; however, it is characterized by being the period in which Greenlandic farming emerged as a fully modernized, mechanized, and commercial occupation. This process of maturing was, naturally, already under way, but 1976 marks the year when the sheepbreeding station at Upernaviarsuk was converted into an independent institution (Heerfordt et al. 1980), future expansions in sheep farming came under detailed and centralized supervision, sheep breeding had been centred on fewer larger farms (fig. 8), and farmers from then on in general received increased government subsidies (Egede et al. 1982; Heerfordt et al. 1980; Laursen & Ørnsholt 1979). These developments were largely a result of the hard winter of 1975/76 (fig. 9), which left the Greenlandic sheep population at its lowest point since 1949; because of the improvements subsequently introduced, i.e. the transition to modernized intensive farming, 1976 also became the last catastrophic winter. To sum up, after 1976 Greenlandic sheep farming had evolved to such modern standards that comparison with Norse farming practices is generally unfeasible. Pre-modern farming outlined The period of pre-modern Inuit farming, then, is the one that seems most informative as regards Norse farming practices, among other reasons because, compared with the preceding period, it is richly Northern Worlds Workshop 2 146 No. sheep in Greenland 1906-2006 50 000 No. sheep 40 000 30 000 20 000 10 000 Year Cattle, horses, and goats in Greenland, 1895-2006 200 No. animals 150 100 50 Year Cattle Goats Horses 147 and accurately documented. The animal husbandry initially practised during the pre-modern farming phase included all the species brought to Greenland by the Norsemen: goats, sheep, cattle, horses, and pigs, listed roughly in the order of their prevalence at the beginning of the period (cf. figs. 3, 4). Pre-modern farming also saw experimental breeding of rabbits, geese, and reindeer, as well as the keeping of quite a considerable number of chickens (Christensen 1946; Hansen 1926: 75; Kampp 1964; Walsøe 1936); however, as none of these do- Fig. 3. Number of officially registered ewes in Greenland in the period 1906-2006; dotted lines indicate years with missing information. The sharp drops in the curve reflect the so-called ‘catastrophic’ winters, i.e. winters with heavy sheep losses, discussed in the text. Notice also the relatively stable population post-1976 after regulation and modernization of Greenlandic sheep farming (data aggregated from: Amdrup et al. 1921; Chemnitz 1920: 30; Hansen 1926: 75; Jensen et al. 1963: bilag 2; Kampp 1964: fig. 1; Walsøe 1919, 1936; Statistisk Årbog 1951-1985; Grønland: Statistisk Årbog 1983-2006). Fig. 4. Number of cattle, horses, and goats in Greenland 1895-2006; dotted lines indicate years with missing information. Figures from 1896 to 1945 are only estimates, because there was no official register of livestock (data aggregated from: Amdrup et al. 1921b; Bruun 1895; Hansen 1926; Jespersen 1916; Knuthsen et al. 1906; Walsøe 1936; Statistisk Årbog 1951-1985; Grønland: Statistisk Årbog 1983-2006). mesticates were bred by the Norse settlers, they will not be considered here. Interestingly, although goats were seemingly more numerous than sheep at the transition from colonial to pre-modern farming, they were completely phased out by the mid 1950s (fig. 4), a diachronic development quite opposite to that of Norse animal husbandry (Dugmore et al. 2007; Enghof 2003). In the latter context, the relative increase in goats has been ascribed to their greater resilience and broader food spectrum (McGovern 1992). However, the specifics of keeping goats in early 19th-century Greenland are unfortunately rarely touched upon in the 19th-century sources; one author (Jespersen 1916) notes that goats were in decline because the Inuit had no liking for goats’ milk, thus also pointing to the seemingly arbitrary factors that sometimes influenced developments. Pigs in Greenland were always very few; experimental breeding culminated during and just after World War II, but was abandoned because it was simply too impractical (Christensen 1946; Jespersen 1916). Because of their low numbers and insignificant role in pre-modern farming, goats and pigs will not be discussed here either, leaving only sheep, cattle, and horses to be discussed. During pre-modern farming these livestock were herded extensively, i.e. they Northern Worlds Workshop 2 148 were mostly left free to graze in the mountains and valleys and fend for themselves throughout the year and rounded up merely for culling and shearing; only the cattle would be stalled and fed through the hardest winter months (fig. 5) (Heerfordt et al. 1980; Jensen 1951). In the South Greenlandic environment, with its highly fluctuating and often harsh winter weather, this markedly extensive mode of production was a very risky farming strategy: the result was that livestock were depleted during the hardest winters (fig. 9), sometimes in disastrous proportions, which accounts for these winters being termed ‘catastrophic’ (e.g. Christensen 1950; Jensen 1958). This disadvantage of extensive farming was already realized by early farmers and observers (e.g. Bend- Fig. 5. Abandoned cattle byre at Igaliku, once the seat of the Norse bishop at Garðar (E47); both the stones of the byre and the hayfield in front have been reclaimed from the Norse farmstead. Such re-use of Norse vestiges was common in pre-modern Greenlandic farming (Photo: K. Krogh 1979). ixen 1927; Hansen 1926; Knuthsen et al. 1906; Walsøe 1936), and was indeed known from elsewhere in the North Atlantic; in the Faroes for instance, such catastrophic winters were called felli and were thought to occur on average every 14 years (Brandt 1985). That farming in South Greenland remained strikingly extensive throughout its premodern phase in spite of the known risks was thus not a result of poor judgement or inexperience; it was rather a direct consequence of factors specific to the Greenland setting, i.e. the very factors that characterized pre-modern farming and make it such an interesting case study: first, the pronounced lack of agricultural machinery up until the late 1960s (fig. 6), primarily caused by Greenland’s distance from Denmark, meant that it was exceedingly hard to expand hayfields and thereby increase production of winter fodder (e.g. Christensen 1946, 1955; Jensen 1951; Walsøe 1919). Hay production was mostly limited to the re-use of relict Norse homefield areas, expanses that were quickly occupied (e.g. Christensen 1946: 152; Jensen 1951: 71). The slow increase in cultivated area during the phase of pre-modern farming is strikingly obvious from fig. 7. At the same time, livestock numbers increased dramatically, widening the critical gap between potential fodder resources and numbers of livestock in potential need of 149 Fig. 6. Raking hay at Qassiarsuk; because of Greenland’s logistical remoteness from Denmark and the trade monopoly, agricultural machinery and aids were very few – essentially just a few tractors owned by the sheep-breeding station – until the 1970s. Instead, fields had to be improved and worked by hand or with the aid of horses (Photo: Vebæk 1964). fodder (cf. figs. 3-4, 7), even though Inuit farmers also readily resorted to wildgrowing fodder resources (Amdrup et al. 1921b; Bendixen 1927; Christensen 1955; Hansen 1926; Heerfordt et al. 1980; Jensen 1958; Walsøe 1919, 1936). Apart from the spatial and, especially during the first half of the period, the technological overlap with Norse farming, points of prospective contrast with Norse farming should also be suggested: first, before the modern phase of farming, holdings were scattered among many livestock owners, some of whom were full-time farmers, but the majority of whom were to varying extents farmers/ hunters (fig. 8) (e.g. Jensen et al. 1963: tab. 3). The former group was settled in the inner fjord areas on prime Norse 900 250 800 700 200 600 500 150 400 100 300 200 50 100 0 0 1928 1962 1972 1977 1994 2006 Fig. 7. The aggregate cultivated area in Greenland at intervals in the period 1928-2006. Note the very slow expansion in cultivated area up to the mid-1970s (data aggregated from: Heerfordt et al. 1980; Høegh 2007; Jensen et al. 1963; Walsøe 1936). 1935 1947 1965 1978 1988 2006 Fig. 8. The number of registered sheep owners in Greenland at intervals in the period 1936-2007. Notice the drop in the number of holdings after the catastrophic winter of 1975/76 and the subsequent professionalization of farming in Greenland (after Dietz 1989; Heerfordt et al. 1980; Høegh 2007; Jensen et al. 1963; Walsøe 1936: 162). Northern Worlds Workshop 2 150 farmland, while the latter was found in the middle fjord and coastal areas. The distribution over several, mostly smallscale owners is probably somewhat similar to the Norse farming pattern. However, the middle fjord and coastal situating of the pre-modern holdings is surely linked to traditional Inuit settlement pattern and, according to the initial findings of the project, a trait rather opposite to than the same as Norse settlement; new, not yet published dates suggest that the Norse outer fjord and coastal sites were the latest to be occupied and the earliest to be abandoned. In summary, the Norsemen in Greenland were farmers who increasingly turned to hunting, while the Inuit were hunters who increasingly turned to farming; perhaps Norse farming practices were, as a consequence, more effective than even the 20th century pre-modern ones, for instance in their use of manure and irrigation (cf. e.g. Christensen 1953; Knuthsen et al. 1906; Krogh 1982). In addition, premodern sheep farming steadily became a specialized, commercial venture – increasingly relying on subsidies and imports – while Norse farming remained self-sufficient household production with the classic North Atlantic composition of animal husbandry. A number of such overlaps and contrasts – suggestive rather than conclusive – are considered in the project, but lie beyond the scope of the present paper. Centrally, four main points singled out from the history of pre-modern farming make for interesting comparisons with Norse farming: 1) the South Greenland landscapes can, in terms of available summer fodder biomass, naturally support quite a large number of livestock; agricultural surveys have through time projected sustainable sheep numbers between 55,000 and 900,000 (Egede et al. 1982; Kampp 1964; Laursen & Ørnsholt 1979); however, 2) summer fodder biomass is not the key factor regulating fluctuations in overall Greenlandic livestock populations; this factor is rather the availability of winter fodder, either occurring naturally or provided by humans; 3) because pre-modern farmers predominantly re-used Norse homefields or naturally occurring resources for winter fodder hay production, we can fairly assume that fodder production in neither period can have greatly exceeded fodder production in the other; 4) as this amount of fodder could winter-feed only a fraction of the total number of livestock, Inuit farmers relied on extensive grazing; the Norse settlers, likely also to try and maximize their animal husbandry, would have had to rely to some de- 151 gree on extensive farming practices to compensate for the insufficient winter fodder production. The environmentally regulated fluctuations in livestock numbers observed historically can therefore be taken as suggestive of similar fluctuations during the Norse occupation of Greenland. ‘Catastrophic’ winters The inner part of the South Greenlandic fjords, where both Norse and pre-modern Inuit sheep farmers primarily set up their farmsteads, is an environmental niche characterized by a subarctic subcontinental climate; i.e. mean summer temperatures lie above 10°C and the growing season is somewhat longer than in the low-arctic outer fjords (Born & Böcher 2001; Cappelen et al. 2001; Feilberg 1984). By contrast, winters are colder and somewhat drier in the inner than in the outer fjord. More significantly, however, winter weather varies greatly in terms of temperature and precipitation depending on the wider meteorological conditions, especially the North Atlantic Oscillation (NAO), anomalies in this system leading to either cooler arctic winter weather or milder, wetter, and more unstable North Atlantic winter weather (ibid.). A further influence on South Greenlandic winter weather is the equally variable winds, among which the katabatic winds com- ing off the ice cap – especially the warm Foehn winds – are highly influential on the appearance of the South Greenland winter landscape. ‘Catastrophic winters’ occur when certain of these weather factors coincide. However, it must be stressed that ‘catastrophic winters’ is not a meteorological term; rather, it is a ‘hands-on’ term tied specifically to farmer-livestock-environment dynamics, i.e. it is a farmer’s perception of the severity of a specific winter in terms of the loss of livestock, more specifically sheep. Clearly, severity is not an objectively fixed magnitude, but a relative observation. Still, severity is central to the identification and description of this winter phenomenon, because some winters with minor livestock losses are not referred to as catastrophic. If we compare the historical use of the term (Christensen 1950; Jensen 1958; Heerfordt et al. 1980) with actual losses (fig. 9), ‘catastrophic winters’ can tentatively be defined as winters when 10% or more of the total number of ewes were lost. As shown by fig. 9, there were ten such recorded ‘catastrophic winters’ in the period 1906-1976, plus an additional four years with population declines not recorded as ‘catastrophic’. It should be noted that some yearly declines do not appear as grave as they really were; for instance, in 1932/33 and 1933/34, as many as 3040% of the ewes were actually lost, but Northern Worlds Workshop 2 152 60 % Relative Sheep Declines (%) 1906-1976 50 % 40 % 30 % 20 % 10 % 0% Fig. 9. Recorded relative sheep declines (% of the previous year’s population) in the period 19061976. Historically termed ‘catastrophic winters’ were those when 10% or more of the total stock was lost during a single winter. because of a subsequent mild spring a larger proportion of lambs survived and were added to and counted among the stock the subsequent year. Finally, although fig. 9 covers the entire pre-modern phase of farming, population declines in 19061925 should be viewed with some caution, because livestock records from the early years are somewhat uncertain. A characteristic of the ‘catastrophic winters’ of pre-modern farming was exceptionally cold winter/early spring temperatures over prolonged periods. Additionally, they came in two varieties: On the one hand, there were the winters when the prolonged cold was accompanied by deep snow cover that was not occasionally thawed by Foehn winds or breaks of mild temperatures; this resulted in farmers running short of fodder, a deficit they could not remedy by grazing the animals, because of the deep snow cover (Christensen 1950). However, equally disastrous were winters when Foehn winds or generally mild weather thawed only part of the snow cover, but subsequent quick freezing led to the formation of an ice shield that the sheep could not dig through in their search for forage (Born & Böcher 2001; Heerfordt et al. 1980). In the historical records of pre-modern farming (fig. 3) such ‘catastrophic winters’ occurred on average every seventh 153 year, i.e. with double the frequency expected for instance in the 18th-century Faroes (Brandt 1985). However, fig. 9 reveals that the severity of such catastrophic winters was quite variable, with losses ranging from c.10% to 50%. Also shown in fig. 9 is the circumstance that in 7 of the 10 instances, the ‘catastrophic winters’ occurred in 2 or 3 successive years; these years must surely have been extra hard on the farmers. Following ‘catastrophic winters’, repopulation on average occurred within 4.1 years (calculation based only on those years where repopulation actually occurred), although it varies from 2 to 8 years depending on the severity of population decline. isons of the records of sheep numbers and meteorologically observed mean annual temperature in the inner fjord area in 1906-1976, e.g. at Narsarsuaq (Cappelen 2007), do show some correlation; however, the correlation is not strong, clearly signalling that while low temperatures are part of the explanation for ‘catastrophic winters’, they are not the whole explanation. Inclusion of meteorological data on winter precipitation in the same period (ibid.) results in much stronger correlation, emphasizing snow cover as an important factor; unfortunately, detailed winter precipitation records are only available for the period after 1960 and therefore provide little statistical verification. While such crude statistics are certainly in themselves suggestive for farming strategies employable by Norse farmers, they also need to be formulated in terms of climatic data in order to be truly convincing; after all, Norse settlement spanned a period of quite striking climatic and environmental change, i.e. the transition from the medieval warmth period (MWP) to the little ice age (LIA) (e.g. Ogilvie & Jónsson 2001; Patterson et al. 2010; Xoplaki et al. 2011). Comparing ‘catastrophic winter’ records and different meteorological and climatic data is still an ongoing process, the results of which can only be summarily reported here. But as an example compar- Somewhat surprisingly, the strongest positive correlation found so far between ‘catastrophic winter’ sheep declines and climate data overlapping the period of pre-modern farming comes from ice-core records from the Greenland ice cap, especially the Dye-3 ice core (Dahl-Jensen et al. 1998; Vinther et al. 2010). Simply explained, these data sets provide a relative temperature record back through time which can be divided bi-seasonally – winter and summer season – on the basis of their 16O/18O stable isotope content, henceforth referred to simply as 18O data. Significantly, it has recently been demonstrated that winter-season Dye-3 18O data mainly ties in with the Northern Worlds Workshop 2 154 NAO signal and display high correlations with southwestern Greenland temperatures, whereas summer-season Dye-3 18O data generally tie in with temperature records from Greenland’s east coast and NW Iceland (Vinter et al. 2010). In short, the Dye-3 18O data reflect temperatures in the vital winter season and mainly in the geographical region that is of interest to the present study; this, in all likelihood accounts for the correlation with recorded sheep declines (for discussion, see below). Besides the considerable degree of correlation with sheep-number fluctuations, the 18O data is useful because, unlike the meteorological records, it can be extrapolated back in time to periods of Norse farming. Sheep population and 18O data Fig. 10 shows the sheep population in 1906-1976 and corresponding seasonally separated 18O data; what are immediately visible are correlative ‘dips’ in the curves. However, equally visible are instances of correlation with only one set of 18O seasonal data, correlation shifted seasonally to either side, or very little correlation at all (cf. figs. 9&10). Clearly, both data sets need to be carefully evaluated for potential biases before an actual model for 18O data/sheep-decline dynamics can be formulated. As expected, seasonal 18O data alone do not explain all of the ‘catastrophic winter’ sheep declines. Part of the upcoming project work is to refine the model, adding new data such as ice-core accumulation rates (Andersen et al. 2006) and determining the exact prevailing weather of the specific ‘catastrophic winters’; i.e. when exactly did the ‘killing’ cold and snow set in? But even working simply with some basic statistics has a number of interesting implications. To begin with, the years prior to 1926 must be left out of consideration simply because of poor data quality, i.e. very insecure population counts and such low sheep numbers that decline percentages are too easily exaggerated. Thus, the potential ‘catastrophic winter’ of 1922/23 must be omitted, leaving only the nine winters of 1932/33, 1933/34, 1936/37, 1937/38, 1948/49, 1956/57, 1966/67, 1971/72, and 1975/76 in consideration (note, however, that 1936/37 is not mentioned historically as a ‘catastrophic winter’). Using the 18O data means of these recorded ‘catastrophic winters’ as a critical mean (dotted line fig.10, 13a-c) – henceforth termed Winter Critical Mean (WCM) and Summer Critical Mean (SCM) – for when sheep-decline/18O data correlation can be expected (i.e. a catastrophic winter will occur when winter and summer seasonal 18O data < WCM/SCM), seven instances occur in the period 1926-1976. Of these, three instances correspond with actual recorded ‘catastrophic winters’, three cor- 155 No. ewes + winter/summer 18O dye 3 1906-1976 -21 45 000 -23 40 000 35 000 -25 30 000 -27 25 000 20 000 -29 15 000 10 000 -31 5 000 -33 Sheep Winter 18O (‰) Fig. 10. Graph showing correspondences between number of ewes (black line) and seasonal winter (dark grey line) and summer (red line) 18O data (after Vinther et al. 2010); broken dark grey and red lines – termed winter critical mean (WCM) and summer critical mean (SCM) – show mean values for the 9 recorded catastrophic winters 1926-1976 used in the inferential extrapolation to the Norse period. respond with winters with smaller sheep declines, and one does not correlate at all (cf. figs. 9 & 10). While yet imperfect, the model does thus seem to account most predominantly for years with sheep losses, although not necessarily of ‘catastrophic’ proportions. Summer 18O (‰) WCM SCM Horse population and 18O data In the graph (fig. 4) of recorded numbers of horses in Greenland from the colonial phase of farming until today, two things are immediately evident: first, that the population seems to have fluctuated markedly, an interpretation that is however partially due to the generally low number of horses (between 56 and 216) and the consequent exaggerated dips when this is shown in the chosen graphic resolution; the second thing that is evident is the lack of detailed population counts prior to 1945, and in the interval 1958-1965; this critically limits the comparative statistical value of the data set. Northern Worlds Workshop 2 156 Recorded declines in the horse population during the pre-modern phase of farming are shown in fig. 11. If we also accept the >10% decline as a limit for ‘catastrophic winter’ horse losses; there were only four such winters: 1948/49, 1950/51, 1966/67 and 1971/72. Of these four declines, the 1950/51 one was the most distinctive, with a c. 42% (51 horses) loss; it is furthermore the only catastrophic decline to occur in a year without a parallel catastrophic sheep population decline. This could indicate that the 1950/51 decline had other causes, e.g. purposeful culling or disease. Of the five recorded <10% horse population declines, two occurred in years with parallel sheep losses, while three did not; however, these years are extremely problematical because the yearly differFig. 11. Recorded relative horse population declines (% of the previous year’s population) in the period 1945-1976. 40 % 30 % 20 % 10 % 0% ence has a range of only 1-13 horses, i.e. so few that declines can by no means be convincingly attributed to environmental factors. The overlap of years with both horse and sheep population declines, also in some instances correlating with 18O data (cf. figs. 12&10) suggests what has already been mentioned historically: i.e. that horses were herded extensively during pre-modern Greenlandic farming and were liable to equal fluctuations. However, the correlation is not strong, especially considering that horses are more susceptible to harsh winter conditions than sheep and that horse-population declines would accordingly be expected to follow 18O data even more closely. They do not, suggesting perhaps that horses were, in fact, stalled and fed somewhat differently from sheep. How- 157 No. horses + winter/summer 18O dye 3 temp 1945-1976 -23 -24 200 -25 -26 150 -27 100 -28 -29 50 -30 0 -21 No. Horse Winter 18O (‰) Fig. 12. Graph showing correspondences between number of horses (black line) and seasonal winter (dark grey line) and summer (red line) 18O data (after Vinther et al. 2010). Note especially both winter and summer seasonal 18O data values for 1967, corresponding to a significant decline in the horse population as well as the sheep population (fig. 10). ever, a comparison for the year 1966/67 in figs. 9-12 merits special attention because it appears to have been a year of equal severity for both sheep and horses; in the 18O data too this year stands out (cf. figs. 10 & 12) with extremely low values for both winter and summer: -29.57 and -28.24‰ respectively. It is therefore not unreasonable to say that such low seasonal 18O data values strongly suggest extremely critical winters. Summer 18O (‰) Cattle population and 18O data As displayed in fig. 4, accurate yearly cattle counts did not become customary until after the Second World War, and the period 1958-1965 also lacks published cattle counts. Cattle farming was temporarily abandoned in 1973, the year after the ‘catastrophic winter’ of 1971/72, until revived in 2002. Fig. 4 also shows that cattle farming in Greenland never involved many animals; the 98 registered in 1948 being the highest number recorded. The cattle data set is therefore of even poorer quality and comparative value than the one for the horse population, for which reason graphs reproduced for the other livestock have been omitted for cattle. Northern Worlds Workshop 2 158 Nonetheless, fig. 4 offers some significant information; in spite of missing data: in the period from 1895 until around the end of the 1930s, the Greenland cattle population seems to have been pretty stable and to have begun to expand markedly only as sheep farming did through the 1940s, a development that accords well with the historical record (e.g. Christensen 1946). The high point in cattle numbers, 1948, coincided with a peak in sheep numbers; a preliminary optimum in livestock expansion without corresponding expansion in fodder production that, as mentioned above, resulted in grave losses of cattle, horses, and sheep during the ‘catastrophic winter’ of 1948/49. Unlike sheep and horses, cattle were stalled and fed during the hardest winter months, but in spite of this measure, cattle were still depleted by 38% during the 1948/49 winter. This is of special interest because the Norse settlers also stalled and fed their cattle during the hardest winter months (e.g. Berglund 2001; Nørlund & Stenberger 1934); the potential identification of such exceptionally hard winters, equally affecting extensively grazed livestock and stalled and fed livestock, would evidently be highly significant. However, the seasonal 18O data for 1948/49 shows no corresponding exceptionally low values, revealing that icecore temperatures alone cannot demonstrate all of the ‘catastrophic winters’. The catastrophic winter of 1948/49 is described as having freezing temperatures, but little snow from early October, then in January heavy snow and cold, followed by thawing, which then suddenly turned into week-long icy storms that created an ice shield over the vegetation that the animals could not dig through (Christensen 1950). After a relatively mild February and March, hard winter weather set in again in late April and lasted well into June. From this description, one would indeed expect 18O values for both winter and summer that were exceptionally low; but they were not; nor do mean temperature records from the measuring station at Narsarsuaq – situated in the inner fjord core of the sheep-farming area – show any remarkable lows for 1948/49 (Cappelen 2007). The 1948/49 winter does, however, stand out as a low, as do almost all the recorded ‘catastrophic winters’ in the PC1 time series that consists of average aggregate data from seven Greenland ice cores (Vinther et al. 2003). This would seem to suggest that correlating different ice-core records could potentially Figs. 13a-c. Dye-3 winter and summer 18O values (full dark grey and red curves) (after Vinther et al. 2010) with indication of winter critical means (WCM, broken dark grey lines), summer critical means (SCM, broken red lines), extreme critical winter (ECW, dotted dark grey lines), and extreme critical summers (ECS, dotted red line). Dye 3 18O data (‰) A.D 985-1149 -24 -25 -26 -27 -28 -29 -30 -31 -32 Winter 18O (‰) Summer 18O (‰) WCM SCM ECW Dye 3 18O data (‰) A.D 1150-1350 ECS -24 -25 -26 -27 -28 -29 -30 -31 -32 Winter 18O (‰) Summer 18O (‰) WCM SCM ECW Dye 3 18O data (‰) A.D 1350-1500 ECS -24 -25 -26 -27 -28 -29 -30 -31 -32 Winter 18O (‰) Summer 18O (‰) WCM SCM ECW ECS Northern Worlds Workshop 2 160 make the livestock-climate model more precise. Again, however, temporal overlap with the relevant comparative periods is problematical. 18O data and Norse ‘catastrophic winters’ The main reason for trying find signals corresponding to ‘catastrophic winters’ in historical pre-modern farming and 18O values is, as mentioned, because the latter data set can be extrapolated back to the Norse occupation of Greenland and thereby substantiate potential frequencies of occurrences of such winters during the Middle Ages, assuming that animals were herded somewhat similarly, i.e. predominantly in an extensive farming strategy. Figs. 13a-c show the seasonal Dye-3 18O data with indications of the WCM/SCM calculated on the basis of the records of pre-modern Inuit farming. This simplified, preliminary model must be interpreted with certain reservations; the WCM/SCM limit did not, when used on 18O data from the pre-modern period of farming, single out only truly ‘catastrophic winters’; with equal frequency it indicated winters with <10% sheep declines, and one instance of non-correlation (see above). The observations from values extrapolated back into the Norse settlement period can therefore, for now, only be read in terms of relative proba- bility: observed 18O values falling below both the WCM/SCM during the period of Norse settlement have a high probability of reflecting the occurrence of a harsh year for extensively grazed livestock with resulting population decline, although not necessarily of ‘catastrophic’ proportions, and with a slight chance of no population decline at all. Generally, there were 13 years with sheep population declines in the period 19261976, i.e. a sheep decline resulting from harsh winters occurred every 3.8 years on average (fig. 9). For the preliminary identification of truly catastrophic winters, the extremely critical summer and winter 18O values (ECW/ECS) of 1966/67 are suggested (figs. 13a-c). When we use this model to identify winters that were probably hard on livestock during the Norse period (figs. 13a-c), their relative frequency seems to have varied greatly (tab. 1). The frequency, of course, depends somewhat on how the settlement period is divided up temporally, especially because harsh winters clearly have a tendency to cluster, but here we have chosen to work mostly with 50-year units, because this is the timespan (1926-1976) upon which the model was also based. For the Norse landnám period AD 985-1049 harsh winters seem to have occurred considerable less often than during pre-modern farming, while the subsequent 100 years until Year A.D. Harsh Winters Harsh Winter Average 985-1049 13 4,92 1 1050-1099 15 3,33 2 Critical Winters 1100-1149 13 3,85 1 1150-1199 23 2,17 6 1200-1249 8 6,25 1 1250-1299 16 3,13 2 1300-1349 14 3,57 4 1350-1399 14 3,57 1 1400-1449 20 2,50 3 1450-1499 16 3,13 2 AD 1149 were roughly comparable. Extremely critical years occurred only once or twice every 50 years. Overall, this would seem to indicate that during the initial settlement of Greenland, and especially during the landnám, reliance on an extensive farming and herding strategy was quite feasible and sustainable. During the period AD 1150-1199, harsh winters suddenly seem to have appeared with much greater frequency – in fact at almost double the rate of the preceding period – as did critical winters, which all fell towards the end of the period. This would have put a serious strain on one or two generations of Norse farmers, perhaps prompting changes in livestock composition, farming strategies, and general subsistence economy. At this time, such changes cannot be confirmed directly in the archaeological material, because of the lack of finely stratified bone collections or construction chronologies from the Norse Eastern Settlement. It does, however, correspond rather well to the overall shift in 13C stable isotope values in human bones after c. AD1250, signalling a shift towards a significantly more 161 Tab. 1. The modelled averaged frequencies of harsh winters, i.e. winters with probable livestock declines, and the number of extremely critical winters, based on the 1966/67 seasonal 18O data, separated only by intervals of 50 years (note that the first landnám interval is 64 years). marine diet (Arneborg et al. 2008). Yet, because the extremely harsh half century was followed by the seemingly most favourable 50 years for extensive farming during the entire Norse settlement period, the effects of the harsh years may have been cancelled out and extensive farming strategies may even have been reinforced. The period AD 1250-1400 saw years comparable to, or only slight worse than, those in the pre-modern phase of farming; that is, extensive herding was risky, but probably generally sustainable, partly because seriously critical winters were still infrequent and livestock had time to recuperate before a new catastrophic depletion. From AD 1400-1500 onwards, however, harsh winters for extensively herded livestock would seem to have occurred with increased frequency, probably making the risks of extensive farming too great to sustain this mode of farming. If, however, extensive herding had been a significant element in Norse farming throughout most of the settlement period, a possibility for which the 18O data would seem to allow, this change could have been difficult and would have reNorthern Worlds Workshop 2 162 quired redirection of a small labour force, which at the time was heavily directed towards the exploitation of marine resources (Arneborg et al. 2008). Conclusion One aim of the PhD project Landscape and Livelihood, Economy and Hierarchy: Pastoral Farming in Norse Greenland is to incorporate historical evidence of 1920th century Inuit farming in Greenland. Such evidence serves to identify key practical and environmental issues likely to face anyone trying to farm on the edge of the Arctic; it allows the archaeologist or scientist to view the landscape, so to speak, through the eyes of Greenlandic farmers and to approach a more dynamic ‘dwelling’ perspective (cf. Ingold 2011): the Norse settlers must have existed as opportunistic and adaptive farmer/hunters finding their livelihood in landscapes that were intricately familiar and structured, but also landscapes that were constantly changing, posing to some aspects of their livelihood, while creating opportunities in others. Through the case study of the period of pre-modern Inuit farming, the paper gives a preliminary example of how the historical evidence can be used to explore scenarios of environment-farming dynamics, not only for the historical period, but, using seasonal ice core data, potentially also for the Norse settlers. The proposed model can clearly be refined and corrected to generate even more statistically sound and graduated findings, and the model still needs testing against other archaeological and environmental evidence. However, the initial results suggest that extensive farming, i.e. the labour-effective mode of farming where the livestock are left to graze mostly unsupervised throughout the year, was a risky, but feasible and sustainable farming strategy throughout most of Norse settlement in Greenland. Only after AD 1400 does extensive farming seem to have become an extremely precarious strategy; at this point, however, farmers might have had a hard time redirecting a sparse labour force into farm work and changing their basic subsistence economy, which was by then heavily dependent on marine resources. If Norse farmers were unwilling or unable to break with the continued reliance on an extensive farming strategy, it could have influenced settlement economy, and ultimately led to abandonment, as environmental conditions for extensive farming grew increasingly worse. As mentioned, the proposed model, and inferences drawn from it, assumes that Norse farming remained markedly extensive throughout the settlement period. This assumption is clearly unsatisfactory; following decades of harsh winter weather and livestock declines, the Norsemen 163 would certainly be expected to react within their capability to improve their situation and reduce livestock losses. Given the lack of complementary archaeological evidence, the model can therefore only be used to explore ‘what-if’ scenarios, rather than ‘so-that’ scenarios. However, given the historical example of the Faroes, even the most catastrophic livestock declines need not always result in adaptive response; until the 18th century, sheep farming was the mainstay of the farming economy, yet severe sheep declines resulting from hard winters are abundantly recorded without this ever leading to a large-scale move to an intensive herding strategy (Brandt 1985; Joensen 1979; Kampp 1950). Besides being suggestive for the above study of livestock-environment dynamics, the historical Faroese example also cautions us that farming strategy is just one element of a complex farming system that works as a totality, i.e. changes are determined by a range of different factors; and outlining and describing other such factors in the Norse Greenlandic farming settlements is the next step in the PhD project. References Amdrup, G. C., Bóbe, L., Jensen, A. D. & Steensby, H. P. (eds.) 1921a & b. Grønland i Tohundredaaret for Hans Egedes Landing, Vol. I & II. Meddelelser om Grønland 60 & 61. Andersen, K.K., Ditlevsen, P.D., Rasmussen, S.O., Clausen, H.B., Vinther, B.M., Johnsen, S.J. & Steffensen, J.P. 2006. Retrieving a common accumulation record from Greenland ice cores for the past 1800 years. Journal of Geophysical Research 11: 148-227. Arneborg, J., Heinemeier, J. & Lynnerup, N. 2008. “… Husk at folk lever af flere ting end bare brød”. De norrøne grønlænderes kost. Nationalmuseets Arbejdsmark 2008, pp. 149-160. Austreheim, G., Asheim, L., Bjarnason, G., Feilberg, J., Fosaa, A. M., Holand, Ø, Høegh, K., Jónsdóttir, I. S., Magnússon, B., Mortensen, L. E., Mysterud, A., Olsen, E., Skonhoft, A., Steinheim, G. & Thórhallsdóttir, A. G. 2008. Sheep grazing in the North-Atlantic region – A long term perspective on management, resource economy and ecology. Rapport zoologisk serie 2008-3, Norges Teknisks-Naturvitenskabelige Universitet, Vitenskabsmuseet, Oslo. Bendixen, O. 1927. Grønland som Nybyggerland. Dansk Forlag, Copenhagen Bentzen, B. 1920. Faareavl og Kohold i Grønland, II. Det Grønlandske Selskabs Aarsskrift 1919, pp. 32-36. Berglund, J. 2001. Omkring dagliglivet på Gården Under Sandet. Tidsskriftet Grønland 7: 267-278. Northern Worlds Workshop 2 164 Bóbe, L. 1915. Anders Olsen og hans Slægt ved Igaliku. Det Grønlandske Selskabs Aarsskrift 1915, pp. 100-112. Born, E. W. & Böcher, J. 2001. The Ecology of Greenland. Eng. trans. by D. EibyJaconsen, Ministry of Environment and Natural Resources, Nuuk. Brandt, J. 1985. En regional analyse af bæreevnens udvikling i de færøske hauger. Fróðskaparrit 33. Bók, Annales Societas Scientarium Færoensis 1985, pp. 19-41. Bruun, D. 1895. Arkæologiske Undersøgelser i Julianehaabs Distrikt. Meddelelser om Grønland 16. Cappelen, J. 2007. DMI Annual Climate Data Collection 1873-2006, Denmark, The Faroe Islands and Greenland – with graphics and Danish summary. Danish Meteorological Institute Technical Report 07-05, Copenhagen. Cappelen, J., Jørgensen, B. V., Laursen, E. V., Stannius, L. S. & Thomsen, R. S. 2001. The Observed Climate of Greenland, 1958-1999 – with Climatological Standard Normals, 1961-1990. Danish Meteorological Institute Technical Report 00-18, Copenhagen. Chemnitz, J. 1920. Faareavl og Kohold i Grønland, I. Det Grønlandske Selskabs Aarsskrift 1919, pp. 30-32. Christensen, K. N. 1946. Faareavlen og Landbruget i Grønland under Krigen. Det Grønlandske Selskabs Aarsskrift 1946, pp. 146-154. -- 1950. Tabet af får i Sydgrønland vinteren 1948-49. Det Grønlandske Selskabs Aarsskrift 1950, pp. 109-118. -- 1953. Grønlands jord under kultur. Tidsskriftet Grønland 4: 407-109. -- 1954. Ødemarken, som fik liv. Tidsskriftet Grønland 6: 201-208. -- 1955. Græsavl i Grønland. Tidsskriftet Grønland 4: 147-154. Dahl-Jensen, D., Mosegaard, K., Gundestrup, N., Clow, G. D., Johsen, S. J., Hansen, A. W. & Balling, N. 1998. Past Temperatures Directly from the Greenland Ice Sheet. Science 282: 268-271. Dietz, R. 1989. Nuværende anvendelse af områder i og tilgrænsende Narsaq kommune, 1988. Grønlands Miljøundersøgelser 1989, 71 pp. Dugmore, A. J., Keller, C. & McGovern, T. H. 2007. Norse Greenland Settlement: Reflections on Climate Change, Trade, and the Contrasting Fates of Human Settlements in the North Atlantic Islands. Arctic Anthropology 44(1): 12-36. Egede, K., Thorsteinsson, I., Køller-Jørgensen, L. & Krogh, K. 1982. Detailplan for fåreavlen i Sydgrønland. Rapport udarbejdet af Fåreavlskonsulent-tjenesten for 165 De samvirkende Fåreholder-foreninger i Grønland og Grønlandsministeriet. Enghof, I. B. 2003. Hunting, fishing and animal husbandry at The Farm Beneath The Sand, Western Greenland. Meddelelser om Grønland, Man & Society 28. Jensen, L. 1951. Det grønlandske landbrugs udviklingsmuligheder. Det Grønlandske Selskabs Aarsskrift 1951, pp. 67-77. -- 1958. Fåreavl på Grønland. Den Kongelige Grønlandske Handel, Copenhagen. Feilberg, J. 1984. A phytogeographical study of South Greenland. Vascular plants. Meddelelser om Grønland, Bioscience 15. Jensen, L., Jørgensen, C. A., Nilsen, J., Mosegaard, H., Sjelle-Hansen, H., Lund, K. B., Andersen, A., Hansen, A., Schledermann-Larsen, C. V., Christiansen, H. C. & Jensen, M. A. 1963. Redegørelse vedr. fåreavl i Grønland. Afgivet af det jordbrugsfaglige udvalg for statens forsøgsvirksomhed i Grønland. Gad, F. 1969. Grønlands Historie II, 17001782, Andet halvbind. Nyt Nordisk Forlag Arnold Busck, Copenhagen. Jespersen, E. 1916. Lidt om Kvægavl i Julianehaabs Distrikt. Det Grønlandske Selskabs Aarsskrift 1915, pp. 74-82. Hansen, K. 1926. Om Mulighed for Landbrug i Grønland. Det Grønlandske Selskabs Aarsskrift 1925-26, pp. 70-88. Joensen, R. 1979. Fåreavl på Færøerne. Færoensia, Vol.XII, C.A. Reitzels Boghandel A/S, Copenhagen. Heerfordt, L., Holmquist, B.K. & Wille, F. 1980. Fåreavl i Grønland. Project i Arktisk Biologi. Københavns Universitet. Johnstrup, F. (ed.) 1878. Gieseckes Mineralogiske Rejse i Grønland ved F. Johnstrup, med et Tillæg om de Grønlandske Stednavnes Retskrivning og Etymologi af Da. H. Rink, Bianco Lunos Bogtrykkeri, Copenhagen. Erslev, Ed. 1897. Om Havedyrkning i nordlige Egne. Geografisk Tidsskrift 1: 87-95. Høegh, K. (ed.) 2007. Årsberetning 2006/2007 fra Konsulenttjenesten for Landbrug, august 2007. Qaqortoq. Ingold, T. 1993. The Temporality of Landscape. World Archaeology 25(2): 152-174. Kampp, Aa. H. 1950. Færøerne. Folk og Erhverv. Det Danske Forlag, Copenhagen. -- 1964. Fåreavl i Grønland. Geografisk Tidsskrift 63: 82-98. Northern Worlds Workshop 2 166 Kleivan, I. 1983. Herrnhuterne eller brødremenigheden i Grønland 1733-1900. Tidsskriftet Grønland 8: 221-235. Knuthsen, H. J., Müller, R., Brummerstedt, C. & Meldorf, G. 1906. Husdyrhold i Grønland. Det Grønlandske Selskabs Årsskrift 1905, pp. 42-46. Kristiansen, A. 1955. Fra fangererhverv til fåreavl. Tidsskriftet Grønland 12: 464-470. Krogh, K. 1982. Erik den Rødes Grønland, Nationalmuseets Forlag, Copenhagen. Laursen, O. & Ørnsholt, H. 1979. Naturmæssige betingelser for Fåreavl i Qagssiarssuk-området (Sydgrønland). Forskningsrapport nr. 6, Publikationer fra Institut for Georgrafi, Samfundsanalyse og Datalogi. Roskilde. McGovern, T. H. 1992. The zooarchaeology of the Vatnahverfi. In Vebæk, C.L., Vatnahverfi – An inland District of the Eastern Settlement in Greenland, pp. 93-107. Meddelelser om Grønland, Man & Society 17. Nørlund, P. & Stenberger, M. 1934. Brattahlid. Meddelelser om Grønland 88(1). Ogilvie, A. E. J. & Jónsson, T. 2001. “Little Ice Age” research: A Perspective from Iceland. Climatic Change 48: 9-52. Patterson, W. P., Dietrich, K. A., Holmden, C. & Andrews, J. T. 2010. Two millennia of North Atlantic seasonality and implications for Norse colonies. PNAS 107(12): 5306-5310. Vinther, B. M., Johnsen, S. J., Andersen, K. K., Clausen, H. B. & Hansen, A. W. 2003. NAO signal recorded in the stable isotopes of Greenland ice cores. Geophysical Research Letters 30(7): 1387-1390. Vinther, B. M., Jones, P. D., Briffa, K. R., Clausen, H. B., Andersen, K. K., DahlJensen, D., Johnsen, S. J. 2010. Climatic signals in multiple highly resolved stable isotope records from Greenland. Quaternary Science Reviews 29: 522-538. Walsøe, L. 1919. Om Faareavl i Grønland. Det Grønlandske Selskabs Aarsskrift 1918, pp. 41-46. -- 1936. Den grønlandske Faareavls nuværende Standpunkt. Det Grønlandske Selskabs Årsskrift 1936, pp. 160-171. Wormskjold, M. 1814. Gammelt og Nyt om Grønlands, Vinlands og nogle fleere af Forfædrene kiendte Landes formeentlige Beliggenhed. Det skandinaviske Litteraturselskabs Skrifter, 10. Årgang, pp. 298-403. Xoplaki, E., Fleitmann, D., Diaz, H., Lucien, von G. & Kiefer, T. (eds.) 2011. Medieval Climate Anomaly. Pages News 19(1).