No. 42 Journal of the North Atlantic 2022 Behind the Ice: The Archaeology of Nunatarsuaq, Southwest Greenland Christian Koch Madsen and Ann Eileen Lennert Journal of the North Atlantic The Journal of the North Atlantic (Online ISSN #1935-1933, Print ISSN #1935-1984), with an international editorial board, is a collaborative publishing effort of the Eagle Hill Institute, PO Box 9, 59 Eagle Hill Road, Steuben, ME 04680- 0009 USA. Phone 207-546-2821, FAX 207-546-3042. E-mail: office@eaglehill.us. Website: www.eaglehill.us/jona. Copyright © 2022, all rights reserved. On-line secure subscription ordering: rate per year is $40 for individuals, $32 for students, $250 for organizations. Authors: Instructions for authors are available at www.eaglehill.us/jona. The Eagle Hill Institute (Federal ID # 010379899) is a tax exempt 501(c)(3) nonprofit corporation of the State o f Maine, USA. The Journal of the North Atlantic (JONA) is a multi-disciplinary, peerreviewed and edited archaeology and environmental history journal focusing on the peoples of the North Atlantic, their expansion into the region over time, and their interactions with their changing environments. The journal—published online in the BioOne.org database and on the JONA website, and indexed in a full range of journal databases—serves as a forum for researchers, and as an information resource for instructors, students, and the intellectually curious who would like to learn about the latest research and study opportunities within the region. The journal publishes a wide diversity of research papers, as well as research summaries and general interest articles in closely related disciplines, which, when considered together, help contribute to a comprehensive multidisciplinary understanding of the historical interplay between cultural and environmental changes in the North Atlantic world. Specifically, the journal’s focus includes paleo-environmental reconstruction and modelling, historical ecology, archaeology, ecology of organisms important to humans, anthropology, human/environment/climate interactions, climate history, ethnography, ethnohistory, historical analyses, discussions of cultural heritage, and placename studies. The journal publishes individual papers on an article-by-article basis. Whenever a manuscript has completed its peer review process and the article galley has been approved by the author, it will be immediately published online in the BioOne database and on the JONA website. This publishing model is also available for special volumes such as conference and symposium proceedings or other collections of papers. In effect, this means that articles are grouped online over time, i.e., the table of contents of volumes will grow as articles are posted online, which has the advantage of rewarding prompt authors, while enabling tardier authors to retain the option of being included in a special volume without delaying its publication. The Journal of the North Atlantic’s publishing format is versatile enough that authors can include supplementary files with their articles. These supplements may include dataset, figure, and table files (e.g., files requiring a larger than normal journal page size, such as large maps), as well as text and protocol files, audio and video files (e.g., for ethnographic studies), and even Powerpoint files. The Journal of the North Atlantic is indexed in the Web of Science (clarivate. com), EBSCO.com, and by way of author entries in Google Scholar and Researchgate. It is included in full-text in BioOne.org and JSTOR.org. Board of Editors Jette Arneborg, Denmark Gerald F. Bigelow, Scotland, UK Rosie Bishop, Stavanger, Norway Colin Breen, Northern Ireland Alison Cathcart, Stirling, UK Mike J. Church, England, UK Jane Downes, Scotland, UK Andrew J. Dugmore, Scotland, UK Mark Gardiner, England, UK Erika Guttmann-Bond, The Netherlands Ramona Harrison, Norway Agnar Helgason, Iceland Joerg-Henner Lotze, USA, Publisher Niels Lynnerup, Denmark Christian Koch Madsen, Greenland Ingrid Mainland, Orkney, UK Meriel McClatchie, Ireland Thomas H. McGovern, USA Natascha Mehler, Germany Jacqui A. Mulville, Wales, UK Anthony Newton, Editor Astrid E.J. Ogilvie, USA Alexandra Sanmark, Orkney, UK J. Edward Schofield, Scotland, UK Niall Sharples, Wales, UK Ian A. Simpson, Scotland, UK Przemyslaw Urbanczyk, Poland Eileen Tisdall, Stirling, UK Orri Vésteinsson, Iceland Alex Woolf, Scotland, UK James Woollett, Canada Cover Photograph: Winter view from Norse site V16/NKAH 1463 – Saqqarsuaq, looking southwest. A more recent cairn is visible to the left, right and below the clearly outlined rooms of the excavated Norse dwelling. Photograph © Christian Koch Madsen. “ Skálholt Map” courtesy of The Royal Library, Copenhagen, Denmark Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 1 Introduction Nuup Kangerlua (the Nuuk fjord system, Figs. 1–3) is archaeologically one of the most well-studied parts of Greenland with hundreds of sites spread across the entire region and representing all of Greenland’s archaeological cultures (ca. 4500 BC– AD 1721, Barlow et al. 1997, Gulløv 1983, 1997, Knudsen and Andreasen 2009, Knuth 1944, McGovern et al. 1996, Pasda 2014, Rosing 1958, Roussell 1936, 1941). Best represented and understood are the sites from the medieval Norse Vestribyggð (Western Settlement), ca. AD 1000–1400 (Arneborg et al. 2012b), and the West Greenlandic Thule culture that took over the fjord system after the disappearance of the Norse (Gulløv 1997:88, 344). Although the Thule culture nominally ended with the Danish-Norwegian colonization of Greenland from AD 1721, the change of lifestyle was gradual and many traditional sea- and land use practices continued largely unaffected well into the 19th century. Accounts from this period explain the concentration of archaeology in Nuup Kangerlua by describing a wealth of natural resources, including marine wildlife and famous caribou hunting areas, as well as Greenland’s richest sources of high-quality steatite (e.g., Bendixen 1921:175ff., Bruun 1908, Cranz 1765:24ff., Egede 1741:32ff., Meldgaard 1986, Rink 1877:330ff., Toft et al. 2010). However, Norse and Inuit strategies for accessing these resources were fundamentally different: the Norse were sedentary agropastoralists whose farms in the inner fjords served as bases for their livestock economy (cattle, sheep, goats, pigs, and horses), transhumance, and medium to longdistance marine and terrestrial hunting (Dugmore et al. 2009, Madsen 2014b, 2019, McGovern 1985, Mc- Govern et al. 1996). In contrast, the Thule culture were mobile hunters whose migrations followed seasonal resource availability, generally switching between outer fjord winter settlements, and inner fjord and hinterland summer hunting camps according to customary use rights (Egede 1741:33ff., Gulløv 1983:200ff., 1997:353ff., Meldgaard et al. 1983). However, in Nunatarsuaq (“big area of land surrounded by ice”) at the head of Nuup Kangerlua (Fig. 1–3), neither Norse, nor Thule Inuit archaeological evidence has appeared to match the historic resource abundance. The region has been little investigated and has been presented as a marginal, inaccessible landscape (Bruun 1908:217, Roussell 1941:16). Here, we present and discuss the archaeology of Nunatarsuaq by combining prior records with the results of field surveys and test excavations carried out in 2012–2013, 2015, and 2019. We attempt to: a) resolve former Norse site registration inconsistencies; b) combine earlier site records with new description, high-precision mapping, and 14C-dates of Norse sites to provide detail settlement pattern analysis; and c) outline the character of Thule culture and later Inuit archaeology and activities in Nunatarsuaq. In conclusion, we draw on the long-term historical ecology built from the combined evidence of Norse and Thule settlement land- and sea-use patterns to evaluate the frequently made assertion that climatic and environmental deterioration was a main driver of Norse settlement decline, perhaps even collapse (e.g., Behind the Ice: The Archaeology of Nunatarsuaq, Southwest Greenland Christian Koch Madsen1* and Ann Eileen Lennert2 Abstract - Here we present the results of archaeological surveys carried out 2012–2019 in Nunatarsuaq, a remote and little investigated region bordered by glaciers and the Kangersuneq ice-fjord at the head of Nuup Kangerlua, Southwest Greenland. We provide a detail analysis of Nunatarsuaq’s medieval Norse sites and settlement patterns, clarify previous site identification inconsistencies, and outline the character of subsequent Thule culture/historic Inuit activities. The longterm historical ecology of Nunatarsuaq and Kangersuneq informed by this evidence contradicts an existing notion of the region’s marginality. In fact, we find that the Norse settlement included three sizable farms practicing transhumance, a set of new 14C-dates implying that activities were part of first colonization (ca. AD 1000) of the Norse Western Settlement, and continued into the 14th century. We find no evidence that Little Ice Age climatic deterioration, possibly setting in as early as AD 1200, had an immediate impact on Norse settlement in Nunatarsuaq. Successful Norse adaptation strategies probably involved heavy reliance on the locally abundant wild marine and terrestrial species that also attracted and sustained the subsequent Thule culture and later Inuit groups. Journal of the North Atlantic 1National Museum of Denmark/Greenland National Museum & Archives. Hans Egedesvej 8, Postbox 145, 3900 Nuuk, Greenland. 2Arctic Sustainability Lab, UiT The Arctic University of Norway, Hansine Hansens veg 18, 9019 Tromsø, Norway. *Corresponding author: christian@natmus.gl. Associate Editor: Mark Gardiner, School Of History And Heritage, University of Lincoln. 2022 42:1–32 Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 2 Figure 1. Topographic map of Nunatarsuaq and Kangersuneq with indication of Norse (red symbols) and Thule culture to pre-1900 Inuit (yellow symbols) site types discussed in the text. The terrain model is based on the ArcticDEM (Porter, et al. 2018), the approximate LIA glacial maximum termini and extent of ‘Saqqarsuaq Lake’ on (Lea, et al. 2014; Weidick, et al. 2012). Present Ice Sheet and glacial margins are digitized form Landsat satellite imagery from 2018-19 (courtesy of the U.S. Geological Survey). Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 3 Barlow et al. 1997, Diamond 2005:211ff., Kuijpers et al. 2014, McGovern et al. 1988, Seidenkrantz et al. 2007). If so, one would expect marginal settlement areas to display the first evidence of stress, disruption, and abandonment. Such a hypothesis, however, is entirely contingent on the perceived notion of regional marginality and inaccessibility, aspects of which we explore for the Nunatarsuaq case study through the framework of human securities. Nunatarsuaq: Landscape, Environment, Resources Nunatarsuaq covers an area of ~675.000 km2 and is located at the head of Nuup Kangerlua between 64°24 and 64°40 (Figs. 1–2). The distance to Greenland’s capital of Nuuk at the mouth of the fjord system is ~90 km as the crow flies, ~120 km if sailing by the winding fjord arms. Nunatarsuaq’s coastline is mostly steep and unapproachable, except for the Narsaq headland on the northern tip of the landmass and the valleys that ascend northeast from Kangersuneq (“fjord with many headlands”) into the mountainous hinterland. This hinterland is traversed by valley systems and interspersed lakes, which mostly lie at 400–700 masl, separated by mountains that reach between 1000–1200 masl. Although Nuup Kangerlua lies within the Low Arctic zone, the climate of the interior fjord is Figure 2. 1860 map of the Nuuk fjord region by S. Kleinschmidt, the first map to show the locations of Norse sites with Greenlandic place names in Nunatarsuaq (Royal Library in Copenhagen). Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 4 Subarctic and Continental, characterized by less precipitation, as well as by colder winter and higher summer temperatures, than the outer fjord. A series of meteorological data from 1939–56 and 1966–69 (Cappelen 2012, Tamstorf 2001) from Kapisillit ~8 km west of Kangersuneq shows January–March to be the coldest months (means ranging from -9.0 to -9.5 °C); May is the first month with mean plus degrees (4.0°C); July has the highest mean temperature (10.2°C); until monthly mean subzero temperatures (-2.6°C) set in again in October. Temperature variability is notable, especially during April and October–November. Yearly mean precipitation at Kapisillit from1939–56 was 255 mm, i.e., low when compared to a mean 756 mm in outer fjord Nuuk. The meteorological data reflects a period after the 1920s when (winter) warming of Southwest Greenland has reached levels of the Medieval Climate Anomaly (MCA, Vinther et al. 2010). Paleoclimatic studies suggest that the relatively mild MCA in Southwest Greenland persisted to the 13th century AD (e.g., D’Andrea et al. 2011, Lasher and Axford 2019, Patterson et al. 2010). The onset of the “Little Ice Age” (LIA, ca. AD 1250–1750) appears to have been characterized by abrupt cooling and significantly increased variability in temperature, precipitation, and storminess (Andersen et al. 2006, Dugmore et al. 2012). Local climate signals from the Ameralik fjord south of Kangersuneq imply that abrupt atmospheric cooling, extensive sea ice (fast ice) cover, decreased meltwater supply and, possibly, decreased local wind stress began shortly after AD 1200 (Kuijpers et al. 2014, Møller et al. 2006, Seidenkrantz et al. 2007). Nunatarsuaq is surrounded by ice on all sides (Fig. 1): To the east by the Greenland Ice Sheet; to the north by Narsap Sermia (NS); to the south by Qamanaarsuup Sermia (QS); and to the west Figure 3. 1908 map of the Nuuk fjord region with indication and numbering of Norse sites (red dots), contemporary and former Inuit winter (black dots) and summer (squares) settlements, as well as areas with plentiful natural resources where: F = bird colony; R = caribou; S = seal; Hv = walrus; a, N, r, H, xxx = different fish. Also note the extent of winter ice, the terminus of the Narsap Sermia (NS) and the locations of Norse sites V12 and V16 (after Bruun 1908). Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 5 by the Kangersuneq ice fjord. When not covered by winter fast ice, the sources of calf ice that often blocks Kangersuneq are Akullersuup Sermia (AS) and Kangiataa Nunaata Sermia (KNS) at the head of the fjord, the most productive tidewater glaciers in Southwest Greenland (Van As et al. 2017). QS and an unnamed glacial outlet just north of it are land-terminating glaciers with a wide outwash plain below. However, during KNS’s advanced stage, an ice-dammed lake—“Saqqarsuaq Lake”—flooded most of this plain (Fig. 1). Evidence of glacial dynamics is visible across all Nunatarsuaq and is essential for evaluating the region’s accessibility and settlement, as well as archaeological site representativity. Several recent studies allow for reconstruction of late Holocene glacial dynamics in considerable detail (cf. Fig. 1): • NS appears to have been relatively stable over the last ca. 250 years (Weidick et al. 2012) until 2004–06, after which episodic rapid retreats (Motyka et al. 2017) have brought the terminus to its present stage ~ 6 km east of the LIA maximum. A minor, rapid advance of NS’s terminus may have occurred at the end of the 19th century (Weidick et al. 2012), destroying several Thule culture and later Inuit hunting camps on the Narsaq coastline (Nordenskjöld 1914:638f.). This late glacial advance is of key importance to the identification of Norse site V12 (see below). • QS is estimated to have reached its LIA maximum by 1880–90, followed by a slow, continuous retreat to its present stage some 2 km to the east (Pearce et al. 2018, Weidick et al. 2012). A continued 19th-century advance of QS fits well with accounts of both Norse and Inuit sites in the valley being “swallowed” by ice or meltwater rivers (Bendixen 1916, Bruun 1917, Jensen 1889, Nordenskjöld 1914). • Ice-dammed “Lake Saqqarsuaq” formed when the LIA advance of KNS closed off the mouth of the outwash plain below QS. Lake Saqqarsuaq appears to have emptied between 1810–1850 (Lea et al. 2014, Weidick et al. 2012). Because of the gently sloping valley floor, the lake was shallow, and even minor changes in water level would have greatly changed the size of the lake (Fig. 1). • KNS reached its maximum LIA terminus in 1761, at this point lying almost across from Norse V15, after which the glacier has retreated to its present position ~ 22 km south (Lea et al. 2014, Pearce et al. 2018). Increased calf-ice production from this retreat appears to have significantly reduced Kangersuneq’s accessibility, which matches early-19th to mid-20th century accounts that the fjord could only be safely navigated in August (Giesecke 1910:151, Jensen 1889:88, Roussell 1932b). However, this notion of Nunatarsuaq’s inaccessibility is, to some extent, biased by the seasonality of the historical observations, i.e., late summer caribou hunting and scientific expeditions. To the permanently settled Norse, the presence and duration of fast ice had different implications: Kangersuneq is one of the few fjord arms in Nuup Kangerlua that freezes over every year with solid fast ice forming in October and lasting until the start of June (Bendixen 1921:189, Bruun 1917:75f.). This ice cover could extend halfway out Nuup Kangerlua (cf. Fig. 3) providing Nunatarsuaq with easy access to nearby Norse settlement areas for up to eight months of the year, possibly longer during the LIA. Kangersuneq is also historically known as one of the best areas in Nuup Kangerlua for (ice-based) winter and spring hunting of common or harbor seal (Phoca vitulina Linnaeus), ringed seal (Phoca hispida Schreber), even harp seal (Pagophilus groenlandicus Erxleben), and beluga whale, or white whale (Delphinapterus leucas Pallas) (see Fig. 3 and Bendixen 1921:251ff., Bruun 1908, Bruun 1917:75f., Lennert and Mikkelsen 2015, Nielsen 1910). Common and harp seal were the most important marine species hunted by the Vestribyggð Norse (Dugmore et al. 2009, McGovern 1985). An abundance in Nunatarsuaq of wild terrestrial species—caribou (Rangifer tarandus Linnaeus), Arctic hare (Lepus arcticus ), seagull (Larus sp.), and ptarmigan (Lagopus mutus Montin)—hunted by both Norse and Inuit is also noted by early observers (Bendixen 1921:251, Bruun 1908, 1917:75f., Giesecke 1910:150ff., Nielsen 1910, Rink 1877:362, Roussell 1932b). A most graphic example of this is a 1956 retelling of old, Inuit hunting stories from Nunatarsuaq describing how hunters had to carry their game through herds of grazing caribou and streams and small lakes so full of urine as to make them undrinkable (Larsen 2017). Interestingly, the accounts of relative caribou abundance in Nunatarsuaq persist through the periodic crashes that characterize West Greenlandic caribou population dynamics (Meldgaard 1986), perhaps also an effect of limited human predation. Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 6 made later in 19th century by naval officers J.A.D. Jensen (1889) and D. Bruun (1908; 1917, Fig. 3). Apart from supplementary field observations of V12 and V13a provided by Danish official O. Bendixen in (1916), none of these surveyors ever visited Nunatarsuaq, but based their maps and descriptions on earlier records and Greenlandic interviewees. Although the latter have proven remarkably accurate site descriptions, there are unsurprisingly inconsistencies between all the early site records and maps. In August 14–22, 1932, Danish archaeologist, Aa. Roussell carried out an archaeological investigation in Nunatarsuaq, which was reported in his own publications and field notes (Roussell 1932a, b; 1941), as well as by his field assistant E. Knuth (Knuth 1934, 1944). Besides making the first (sketch) plans of Norse sites and excavating the dwelling of V16 (Fig. 12), Roussell also partly reorganized the inventory of Norse sites, introducing further inconsistency to earlier records. Subsequent archaeological investigations in Nunatarsuaq have been limited to short surveys or new site registrations based on local interviewees (Gulløv 1983; Kapel 1982, 1989; Vebæk 1951), as well as a recent paleoenvironmental study at V15 (Schofield et al. 2016). The later investigations have not significantly changed our impression of the Norse archaeology but have identified several Thule culture and later Inuit sites in Nunatarsuaq (Fig. 1 and Table 1). In 2012, the authors initiated a collaboration between two research projects to improve our understanding of Nunatarsuaq’s long term historical ecology. The Winter is Coming Project (WiCP), 2016–20, reported in this paper, explored the dynamics of Norse settlement and land use in marginal areas to identify drivers of societal change and decline. The other project, A Millennium of Changing Environments in the Godthåbsfjord, West Greenland: Bridging Cultures of Knowledge, 2012–17, was a study combining knowledge co-production with social and natural scientific approaches to describe the long-term and culturally embedded historical ecology of Nuup Kangerlua, the results of which are published elsewhere (Lennert 2016, 2017; Lennert and Mikkelsen 2015). To facilitate the research aims of both projects, we adopted a survey strategy that mainly used the fjord near Norse sites as nodal and entry points— reached by helicopter or boat—for surveys into the vast Nunatarsuaq hinterland (Fig. 1). This strategy served to ensure adequate survey coverage of previously recorded Norse sites and the unsurveyed entryways to hinterland resource areas by Inuit hunters in later history, potentially revealing any unnoticed geographical overlap between Norse and Thule culture and later Inuit land-use practices. The field An evaluation of Nunatarsuaq’s agro-pastoral resource capacity, a satellite-based vegetation classification study of 25 x 25 m resolution (Tamstorf 2001), shows that the vegetation types usually favored as pastureland for domestic livestock by Norse farmers comprise a minor part of the landmass: 4 % (dwarf) shrub heath, 4 % meadow/fen, 3 % grass-heath and 1 % willow scrubland, i.e., in total only ca. 12 % of Nunatarsuaq’s surface; and that these vegetation types are geographically confined to the lower valleys and mountain slopes facing Kangersuneq and the QS-valley. By comparison, the same types of pastureland vegetation in central areas of the Norse Eastern Settlement cover between ca. 45–70% (Madsen 2014a). A greater part of Nunatarsuaq’s surface is comprised of fell field (46%), lichen heath (15%), and stone or exposed bedrock (12%) and, as noted during our field surveys, the vegetation above ~150 masl rarely reaches above ankle height. The impression of Nunatarsuaq’s generally limited agro-pastoral resource capacity is confirmed by a recent study combining modeled biophysical parameters and infrastructure to identify future areas for farming expansion in Nuup Kangerlua (Westergaard-Nielsen 2015). In this study, Nunatarsuaq ranks as one of the least promising areas for farming expansion, although areas displaying some potential overlap completely with the distribution of Norse farmsteads. Archaeological Investigations in Nunatarsuaq 1808–2019 The first record of archaeological sites in Nunatarsuaq was provided by German mineralogist, C.L. Giesecke, who visited the region between August 12–15, 1808 (Giesecke 1910:150ff.). Giesecke’s concise accounts of two Norse sites located towards the head of Kangersuneq and by QS, respectively, are important because one of these sites—as we argue below— has not since been rediscovered. Unfortunately, no accurate map of Nunatarsuaq or Kangersuneq existed at the time of Giesecke’s visit to allow him to fix his descriptions geographically, a situation not much improved when the first compendium of Norse sites in Greenland was published decades later (GHM III 1845:Tab. XII). In the map annex to this publication, the head of Nuup Kangerlua is barely recognizable, and a single Norse site not since confirmed was mapped on the western side of Kangersuneq. The first usable map of Nunatarsuaq with place names and Norse site locations was published in 1860 by German missionary S.P. Kleinschmidt (Fig. 2). More accurate and systematic surveys and maps, introducing systematic Norse site numbering, were Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 7 after baseline postprocessing, supplemented by tape measurements, detail recording and photographs of all features. Thule culture and later Inuit features were only point surveyed, but described in detail and photographed. For cross-reference between old and new site records, we use the old Norse site number surveys were carried out over 42 days in the years 2012–2013, 2015, and 2019, by two or three-person teams, moving across the landscape in a 20–100 m line to cover as much ground as possible. Norse features were surveyed with a Leica SR 20 differential GPS with a local accuracy of < 20 cm Table 1. List of all currently recorded (Thule) Inuit sites in Nunatarsuaq. The table gives site place names (where available), site type interpretations referred to in the text, summary feature descriptions, and recorded numbers of features on each site. NKAH no. Place name Site type Description No. features 1224 Kissaviarssuup qangattaa Cave site “The falcon’s cave” (Gulløv 1983:164). 1 1287 - Cave site Cave, ca. 6.0 x 7.5 m, with stone and turf built wall at the entrance. Still in use (Kapel 1989:2, authors 2019). 1 1347 Saakalivilik Cave site Very good and old cave for 15 people (Gulløv 1983:173) 1 1348 Iluliartup qangaatarssua Cave site ‘The cave by the lake with much ice’; cave with two parts, in the larger 20 people can sleep next to each other (Gulløv 1983:173) 1 1349 Narsaq Tent site 1-2 tent foundations and 2 tent rings, Thule culture/historic Inuit, meat caches/cairns and a grave (Kapel 1982, authors 2019) 3-4 1350 Narsaq Tent site 1 tent foundation with platform, vegetation suggests of considerable age, Thule culture (Kapel 1982) 1 1351 Narsaq Tent camp 4 tent rings, one older looking with platform, several hearths, Thule culture/historic Inuit (Kapel 1982) 4 1392 Narsaq Tent site 1 tent foundation with platform, stone flagging visible, Thule culture (Kapel 1982) 1 5581 Umiviik Overnight site 1 small horseshoe shaped tent foundation, 19th early 20th century (authors 2012). 1 5582 Naajarsuit Tasiat Tent site 1 large tent ring, historic Inuit/recent (authors 2013) 1 5583 Naajarsuit Tasiat Tent site 2 tent rings/foundations and 2 hunter’s beds, Thule culture (authors 2013) 4 5584 Naajarsuit Tasiat Tent site 2 tent foundations, 1 hunter’s bed or shelter, 1 hearth, Thule culture (authors 2013) 3 5585 Naajarsuit Tasiat Overnight site 1 rock shelter, Thule culture/historic Inuit (authors 2013) 1 5586 - Overnight site 1 hunter’s bed, nearby a cairn, Thule culture/historic Inuit (authors 2015) 1 5587 - Overnight site 2 hunter’s beds, lying ca. 30 m apart, Thule culture (authors 2015) 2 5588 - Overnight site 1 hunter’s bed, Thule culture/historic Inuit (authors 2015) 1 5589 - Overnight site 1 multi-phase hunter’s bed, nearby a shooting blind (authors 2015). 1 5591 Mannissuup Naqinnera Caribou drive 27 cairns, single or composite, forming an 850 m barrier. Three coverts or lookouts are found along the drive (authors 2019) 5650 - Hearth Single hearth, possibly Saqqaq culture (authors 2019) 1 5651 - Meat cache Single meat cache (authors 2012). 1 5652 - Shooting blind Single shooting blind (authors 2013). 1 5653 - Single cairn Single cairn, route marker (authors 2015). 1 5654 - Single cairn Single cairn, route marker (authors 2015). 1 5655 - Single cairn Single cairn, route marker (authors 2015). 1 5656 - Single cairn Single cairn, route marker (authors 2015). 1 5656 - Single cairn Single cairn, route marker (authors 2015) 1 5657 - Open air hearth Single hearth (authors 2015) 1 5658 - Meat cache Single meat cache (authors 2015) 1 5660 - Meat cache Single meat cache (authors 2015) 1 5668 - Meat cache Single meat cache (authors 2019) 1 5669 - Meat cache Single meat cache (authors 2019) 1 5670 - Cairn Single cairn and nearby shooting blind (authors 2019). 2 Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 8 system (V for Vestribyggð + no.) when possible but refer to Nunatta Katersugaasivia Allagaateqarfialu heritage (NKAH) numbers for newly discovered Norse and all Thule culture sites. In 2019, we excavated seven small (< 50 x 50 cm) test units to retrieve datable samples from identifiable archaeological contexts. To minimize the potential of old wood effect, the samples selected for dating were in all but one case (cf. Table 2) small, < 0.5 cm diameter, twigs of local willow or birch with assumed low own age. The test units were excavated by layer with trowel without screening. Test unit stratigraphic sequences are summarized in Table 3, while sample materials, uncalibrated 14C-dates, calibrated ranges (1 and 2 sigma), and median probabilities are listed in Table 2. In the text, all 14C-dates are presented with 95.2% probability (2 sigma). After evaluating archaeological research in Nunatarsuaq 1808–2019, the region must still be considered poorly investigated, and uncertain site representativity remains a serious caveat to settlement pattern interpretations. Survey coverage is biased towards the Narsaq headland, the lower Nunatarsuaq valleys, whereas the mountainous hinterland and fringe of the Ice Sheet is still virtually terra incognita. The probable existence of numerous unrecorded Thule culture and later Inuit sites in interior areas is suggested by cave site 1348 (Fig. 1, Table 1), place names (e.g., Qamanaarsuup, see above and Fig. 1, i.e., “the places where one waits and keeps a lookout”, implicitly “for prey”), historically recorded hunting practices (Lennert 2017, Nielsen 1910), and parallel archaeological evidence from elsewhere in Nuup Kangerlua (e.g., Gulløv 1983, Knudsen and Andreasen 2009, Knuth 1944). We estimate that current Thule culture and later Inuit site representativity in Nunatarsuaq may be as low as 20–40%, including the sites destroyed by glacial advances (see above). Conversely, based on past site records by local interviewees (see above) and previously demonstrated Norse site selection criteria (Albrethsen and Keller 1986, Madsen 2014b), we estimate that up to 90% of the still existing Norse farmsteads in Nunatarsuaq have been identified, but perhaps only 50% of the shieling sites. Intra-site feature representativity presents another issue because poorly preserved and overgrown ruins are easily overlooked. For instance, in our surveys we identified quite many new Norse features, thereby changing our impression of the farmstead’s size and function. In Table 4, the survey intensity percentage provides a rough estimate for intra-site survey coverage of the combined archaeological investigations. Table 2. List of carbon-14 dates from Norse sites in Nunatarsuaq with description of sample materials, archeological contexts, and summary 14C-age statistics. Sample ID Lab Code1 Material/species2 Context3 14C Age BP4 14C cal. AD (1σ) 14C cal. AD (2σ) Median probability cal. AD V13 x01 SUERC-92664 Charcoal, Betula sp. Ruin 1, lower [02] 999 ± 24 995 - 1038 988 - 1148 1025 V13 x02 SUERC-92653 Charcoal, Salix sp. Ruin 1 [06] 1052 ± 24 984 - 1017 901 - 1025 996 V13a x01 SUERC-92654 Charcoal, Betula sp. Midden, upper [02] 858 ± 24 1166 - 1213 1055 - 1250 1189 V13a x02 SUERC-92655 Charcoal, Salix sp. Midden, lower [02] 847 ± 24 1168 - 1219 1158 - 1255 1198 V14 x01 SUERC-92656 Charcoal, Salix sp. Midden [02] 676 ± 24 1281 - 1381 1275 - 1388 1298 V14 x02 SUERC-92660 Charcoal, Ericales sop Midden [05] 660 ± 21 1286 - 1484 1281 - 1390 1349 5576 x01 SUERC-92661 Charcoal, Salix sp. Midden, upper [02] 929 ± 24 1042 - 1154 1032 - 1160 1097 5576 x03 SUERC-92662 Charcoal, Betula sp. Ruin 1, upper [02] 970 ± 21 1022 - 1147 1018 - 1153 1091 5576 x04 SUERC-92663 Charcoal, Salix sp. Ruin 1 [04] 936 ± 21 1040 - 1151 1033 - 1155 1099 V15 UM25 10-12 cm5 UBA-31331 Charcoal and Montia fontana seeds Homefield, upper 767 ± 22 1248 - 1276 1222-1279 1260 V15 UM25 32-34 cm5 UBA-31332 Montia fontana seeds Homefield, lower 674 ± 42 1277 - 1387 1265 - 1395 1314 1Scottish Universities Environmental Research Centre (SUERC). 2All samples speciated by Dr. Susan Ramsay, Falkirk, UK. 3Context numbers refer to Table 3. 4All dates calibrated in Oxcal v4.3.2. Bronk Ramsay (2017) by IntCal13 atmospheric curve (Reimer et al. 2013). 5Dates from Schofield et al. 2016). Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 9 unclear whether this signifies a difference in function, seasonality, chronology, or all three. The above feature types are typical of the Thule culture (ca. AD 1200–1721) but continued in use into the 20th century—indeed some feature types are still used today—and they may rarely be typologically dated more accurately than to the Thule culture/later Inuit. In the following, the term Thule culture (indicated by yellow on maps) refers to all Inuit features that are assumed predate ca. AD 1900, whereas red signifies very recent features. Interpretation of Norse archaeological surface evidence often is more challenging because of the buildings’ heavier construction, longer use and reuse, and therefore more complex collapse remains; their greater functional variety; and limited evidence from archaeological excavation. Although a substantial number of Norse features and ruins have been excavated, most in the first half of the 20th century (e.g., Andreasen 1982, Nørlund and Stenberger 1934, Roussell 1936, Roussell 1941, Vebæk 1991), specialized buildings, such as livestock houses, sheds, barns, storehouses, etc., have rarely yielded direct evidence of their use. Instead, they have been functionally identified from a combination of spatial layout, building materials, landscape setting, and ethnographic parallels. To avoid overinterpretation of the surface archaeological evidence, but at the same time providing a systematic, robust, and verifi- Archaeological Feature and Site Types and Definitions Spatially limited urban, agricultural, and industrial development in Greenland has ensured exceptional surface preservation of archaeology, not at least in a North Atlantic context where continued farming and land use most often superimpose traces of past activities (e.g., Arge et al. 2005, Vésteinsson et al. 2002). The interpretational potential of archaeological surface evidence—including a nearcomplete, fossilized medieval Norse settlement landscape—was early realized and functional classification of features and sites has long been a recurring theme in Greenlandic archaeology, often greatly informed by pre-20th century ethnographic parallels (e.g., Bruun 1928; Gulløv 1983, 1997; Madsen 2014b; Meldgaard et al. 1983; Roussell 1941). Nunatarsuaq’s Inuit archaeology includes feature and site types familiar in all West Greenland: tent rings, cave sites, hunter’s beds, rock shelters, meat caches, open-air hearths, shooting blinds, caribou drive systems, fox traps, and cairns (Table 2). A less familiar feature type is tent foundations, i.e., more substantial tent features where a low turf and stone wall held down the skin canvas. Tent foundations may—as the Inuit winter houses—also have a platform, and they appear to reflect a greater labor investment than the tent rings. However, it is Table 3. List of the excavated test units in Nunatarsuaq with summary description of the stratigraphy from where the samples (Table 2) were retrieved. Test unit ID Test unit size Summary layer description and interpretation V12 / NKAH 1483 - 1 25 x 25 cm Surface to 8 cm [01]: vegetation layer and topsoil, ca. 30 cm [02]: mixed silty sand with painted wood and modern tent peg, [03] stone (flag?). V13 / NKAH 1482 - 1 50 x 50 cm Surface to 15-16 cm [01]: vegetation layer and topsoil, 34-42 cm [02-05]: layers of mixed turf (and stone) collapse with bone fragments and charcoal, 35-38 cm [06]: possible thin floor layer with charcoal, below [04] and [06]: gravelly subsoil [07]. V13a/NKAH 1509 - 1 50 x 50 cm Surface to 13-16 cm [01]: vegetation layer and topsoil, 30-32 cm [02]: mixed midden deposits with poorly preserved bone fragments and charcoal, 34-36 cm [03]: silty sand with no cultural inclusions, below [03] gravelly subsoil [04]. V14/NKAH 1508 - 1 50 x 50 cm Surface to 15-17 cm [01]: vegetation layer and topsoil, 23-25 cm [02]: mixed midden deposit with bones and charcoal, 30-45 cm [03]: midden deposit with many well-preserved bones, charcoal and some artifacts, 31-36 cm [04]:sterile turf horizon with partly decomposed grass, 42-56 cm [05]: midden deposit with many well-preserved bones and charcoal, 43-57 cm [06]: sterile turf horizon with partly decomposed grass, 70-73 [07]: midden deposit with many bones, charcoal and some artifacts, below [07] layers of decomposed peat/turf, natural vegetation horizons. V15/NKAH 1462 2 cm core Surface to 10 cm [01]: vegetation layer and topsoil, 21 cm [02]: series of thin vegetation horizons with specks of charcoal and bone fragments separated by sterile silt layers, 45 cm [03] midden deposits with charcoal and bone fragments, below [03] gravelly subsoil [04]. NKAH 5576 – 1 25 x 25 cm Surface to 18 cm [01]: vegetation layer and topsoil, 21-22 cm [02]: mixed collapse deposit w. a few poorly preserved bone fragments and spots of charcoal, 50-51 cm [03]: mixed midden deposits with poorly preserved bone fragments and charcoal, 52-53 cm [04]: coarse sand layer (erosion event) with no cultural inclusions, 55-57 cm [05]: cultural layer w. charcoal (landnám), below [05] sandy subsoil [06]. NKAH 5576 – 2 25 x 25 cm Surface to 6-7 cm [01]: vegetation layer and topsoil, 20-22 cm [02-03]: turf collapse layers with poorly preserved bone fragments and charcoal, 30-23 cm [04]: turf/stone collapse overlaying floor layer (?). Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 10 ings are too poorly preserved to classify beyond most likely having been roofed. Storehouses stand out from other outbuildings by being purposefully built for ventilation and cooling, either in dry-stone masonry technique or wholly or partially in wood. As the wood is not preserved, the latter type is identifiable as low stone foundations too narrow (<50 cm) for roof-supporting turf walls and with no surrounding collapse material. Storehouses are situated on drained ground (bedrock or gravel) and in wind-exposed settings. Storehouses cover a range of different functions, e.g., food/milk stores, meat-drying houses, warehouses etc., although the evidence for such functional separation is unclear (Arneborg et al. 2012b, Roussell 1941:230ff.). Freestanding walls and dykes are non-roof supporting built barriers, usually forming all or part of an enclosure, most frequently animal pens and folds, hay-yards, or homefield boundary walls. Pens or hay-yards are separable from roofed outbuildings by their often more irregular shape, frequent inclusion of natural features (boulders, cliffs, ravines, water), greater internal width, and good preservation of often more narrow walls with little original turf. Humanmade channels or ditches are self-explanatory but can be difficult to recognize without excavation. The combination of various buildings allows for site classification, where we make a basic distinction able dataset, we here adopt an interpretive model with six functional categories of Norse vernacular architecture: 1) dwellings; 2) insulated outbuildings; 3) uninsulated outbuildings; 4) storehouses; 5) freestanding walls or dykes; and 6) channels or ditches. Dwellings (i.e., farmhouses) are identifiable as low, grassy mounds with scattered collapse stone, often preserving vague outlines of rooms, passages, and walls. They are typically the largest ruin of a roofed building on site and have an associated midden area. Large quantities of turf/stone collapse reflect heavy insulation, extended use, continual modifications and, in centralized dwellings, clustering of several specialized buildings (byres, barns, stores, various craft rooms, etc.) (Andreasen 1982; Nørlund and Roussell 1929; Roussell 1936, 1941). Outbuildings are defined by having been roofed, built either with pure turf, or combined turf and stone, and walls on a stone foundation one to several courses high. We distinguish between insulated versus uninsulated outbuildings, the former defined by having long walls >90 cm wide, the latter by long walls <90 cm wide. This somewhat arbitrary distinction serves to separate outbuildings that more likely facilitated cold-period housing of livestock from outbuildings that more likely served functions where heat preservation was unneccesary (e.g., barns, sheds, stores) or, alternatively, summerperiod housing of people or animals. A few outbuild- Table 4. List of all currently recorded Norse sites in Nunatarsuaq, providing key measurements for each site discussed in the text. Bruun-no.1 NKAH no.2 No. features3 No. of roofed buildings3, 4 Area (m2) of roofed buildings4 No. enclosures Area (m2) enclosures5 Homefield Area (ha) Survey intensity V12 1483 3 (2) 2 (2) 35.5 1 34.0 - 60 % V13 1482 8 (1) 4 (1) 172.5 3 35.1 - 70 % V13a 1509 27 (4) 19 (1) 1720.8 5 372.7 3.160 80 % V14 1508 9 6 870.1 3 248.3 1.008 70 % V15 1462 16 (2) 9 1297.1 4 1042.0 2.897 70 % V16 1463 5 (2) 5 (2) 383.3 0 - - 60 % 5576 8 (1) 6 (1) 268.0 2 83.6 - 60 % 5577 1 1 4.3 0 - - 90 % 5578 6 (2) 3 30.7 1 29.9 - 80 % 5580 1 1 2.4 0 - - 90 % V16_alt - (2) 1? 240.3? ? ? - - V516 1480 16 (2) 8 (1) 1192.3 2 103.2 2.153 90 % V76 1491 11 6 1063.0 1 18.0 3.198 70 % Ø29/29a6 2231/ 2230 34 23 3301.0 6 620.8 12.389 90 % Ø666 4318 32 18 3181.1 8 418.8 7.374 90 % Ø836 4427 18 13 1532.1 2 181.7 6.206 90 % Ø1496 3575 24 13 3107.7 4 139.4 2.507 80 % 1The old site numbering system referring to settlement + site no. (V for Western Settlement, Ø for Eastern Settlement). 2Official Greenland National Museum & Archives heritage site numbers. 3Numbers in brackets indicate uncertain/unverified features. 4Features interpreted as dwellings, insulated and insulated outbuildings, and storehouses. 5Measured inside area of all enclosures except bounded grazing areas; 6Data based on recent precession surveys by C.K. Madsen and stored at Greenland National Museum & Archives. Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 11 between farmstead and shieling. Farmsteads were the main nuclei of domestic habitation, production, accumulation, and consumption, and are defined by the presence of a farmhouse (dwelling) with associated midden, winter housing for domestic livestock (insulated outbuildings), storage/processing facilities (uninsulated outbuildings and storehouses), as well as a hay-fodder production area, i.e., homefield or meadow. Farmsteads may have more or less dispersed or centralized layouts (Arneborg et al. 2012b, Roussell 1941) that are in general so predictable in their combination of features that a lack of any single elements calls for a consideration of missing evidence or alternate site functions (see discussion). If a site is not classified as a farmstead, it is by definition a shieling, here defined broadly as a seasonal, task-specific production or logistic site (Madsen 2019:122). Shieling, or transhumance, is a widely known farming practice in all of the North Atlantic that facilitated use of seasonally available and often geographically dispersed outfield resources, i.e., mainly pasture and fodder resources and continued dairy production (e.g., Bruun 1928, Mahler 2007, Reinton 1969, Sveinbjarnardóttir 1991). Shielings were thus the satellite infrastructure to the farmsteads that facilitated transhumance, and in Norse Greenland, they appear to make up a half to one-third of all recorded sites (Madsen 2014b:213, Vésteinsson 2010:144). Unfortunately, transhumance and shielings have been relatively little investigated in Greenland (Albrethsen and Keller 1986; Berglund 1998; Ledger et al. 2013; Madsen 2014b, 2019; Roussell 1941) and are not a well-understood phenomenon in terms of site types and their development. V12/NKAH 1483 – Narsaq V12 – Narsaq (“plain”) has alternately been mapped on the coastline west and east of the tip of the Narsaq headland (e.g., Fig. 3, Gulløv 1983, Jensen 1889, Kapel 1982), although a position east of the tip, and facing the NS glacier, have generally been accepted from a single eye-witness description and accompanying map: “By Narsaq, between a steep point and the edge of the glacier is a small ruin on the slope immediately above the fjord. Part of the ruin has slid into the water” (au. trans. after Bendixen 1916, Bruun 1917:Tavle I). Around this time, local Greenlanders believed the ruin no longer existed (Bruun 1917:77), and subsequent surveys have not succeeded in locating any Norse features on the Narsaq headland (Kapel 1982, Vebæk 1951). In 2019, we identified a Norse site that may be the “missing” V12. This site is located just southwest of the tip of the Narsaq headland (Figs. 1, 4). There, 1 to 3 Norse ruins lie by a small point facing the mouth of Kangersuneq, sheltered from winds coming off the glacier by the Figure 4. Survey map of V12/NKAH1483 – Narsaq. 1: animal pen, 2-3: possible, uninsulated Norse outbuildings, A-C: Thule culture/later Inuit tent features. Red diamonds are more recent features (map: C.K. Madsen 2020). Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 12 hillside that ascends steeply to 80–90 masl. behind the site before flattening out onto the plain. The vegetation on this western-facing slope is the richest on the Narsaq headland. The tip of the small point is exposed bedrock that slants gently into the fjord, to the south partly sheltering a stony beach, providing one of few places on the jagged coastline, where a small boat can easily be dragged up. A small stream just south of the site is the only notable freshwater source on the tip of the Narsaq headland. The convergence of these favorable landscape features explains the continued use and reuse of the site, also as Thule culture hunting camp NKAH 1349 (Gulløv 1983:173, Kapel 1982:97, features A–D) and continuing to this day. Later activities have greatly disturbed the Norse ruins, which may explain why earlier surveys failed to notice them. In fact, only Ruin 1 is unmistakably Norse, a stone foundation for a turf-built animal pen perched on the edge of a slope that drops off into the fjord just a few meters below (Fig. 4). A stone heap butting against the pen may be a completely collapsed Norse turf/ stone building (Ruin 3), the stones of which have partly slid down the slope, partly been reused in the Inuit features. Ruin 3 is possibly the ruin recorded by Bendixen in 1916 (see above). Feature 2 (D) was interpreted as one of the older Thule culture tent features on site (Kapel 1982), but the rectangular shape of the ~80 cm wide stone foundation, the backside of which is dug some 60 cm into the slope, suggest a Norse origin. However, a 25 x 25 cm test unit excavated in Ruin 2 (D) revealed a shallow stratigraphy (Table 3) producing a modern iron tent peg and painted wood fragments and thus the feature has been reused and disturbed very recently. We interpret V12 as a Norse dairy shieling or, if features 2–3 are not Norse, as a herding station (see discussion) used during seasonal pasturing of animals on the Narsaq headland. There are four arguments for the identification of this site as the “original” V12. First, the setting corresponds well with Bendixen’s concise (1916) description, especially if noting the maximum terminus of the NS glacier (cf. Fig. 3). Second, because Bendixen’s site visit occurred at the time of or slightly after this maximum, the site could not have been located east of this line (cf. Fig. 1). Third, continued use of the site as a summer hunting camp by Thule culture and Figure 5. Survey plan of V13/NKAH 1482 and NKAH 5576. At V13; 1: insulated longhouse, 2-3, 6: uninsulated outbuildings, 4-5, 7: animal pens, A-B: hunter’s beds, C. “wolf trap”. At NKAH 5576; 1: dwelling, 2, 7 animal pens/shelters, 3: uninsulated outbuilding, 4-5: insulated outbuildings and 6: storehouse (wood on stone foundation) (map: C.K. Madsen 2020). Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 13 V13 is situated at the mouth of the valley Mannissuup Naqinnera at ~190 masl and just on the edge of the steep drop-off towards NKAH 6676 and Kangersuneq. A giant, egg-shaped boulder sits on this edge, giving rise to the site’s fitting Greenlandic place name. A stream runs just south of the site and, ~100 m further south, the small river Mannissuup Kuua that meanders through the full length of the valley (Fig. 1). The valley floor is made up of stony glacial till vegetated by rather expansive, but mediocre quality pastureland of low grasses, dwarf shrub, and heathland. Ruin 1 is the most distinct feature on site, probably the ruin referred to above. It is a well-preserved, 1.0–1.2 m wide stone foundation for an insulated turf building measuring 15.8 x 7.0 m. Ruin 1 is unusual for its large size and the greatest internal width (4.8 m) of all the surveyed ruins, as well as its slightly curving long walls. A 50 x 50 cm test unit excavated centrally in the ruin revealed a thin floor layer (Table 3, [06]) 14C-dated to cal. AD 901–1025 (Table 2), the earliest Norse date from Nunatarsuaq. The top of the collapse material (Annex 2, [02]) in the test unit was 14C-dated to cal. AD 988–1148. The size, later Inuit attest to the favorable landscape setting. Fourth, in 2019 we surveyed the entire coastline of the Narsaq headland and the steep, northeastern facing—and therefore shaded—shoreline facing masses of packed calf ice in front of NS presents a very atypical and unfavorable landscape setting for a habitation site. V13/NKAH 1482 – Mannissuaq The geographic location of Norse V13 – Mannissuaq (old spelling Mánigssuaq, “big egg”) has been consistent since the earliest record (cf. Figs. 1–3). However, apart from an aerial inspection (Kapel 1989), no archaeological survey has formerly been carried out at the site, which is described as a single Norse ruin: “(…) in a high lying valley with a steep drop-off to the fjord (…)” (au. trans. after Bendixen 1916). Locals also mentioned: “a small ruin, part of which is under water. Here is only a small habitable site under a steep mountain” (au. trans. after Bruun 1917:77). During our 2019 survey, we found these site descriptions to correspond to two site units spaced ~250 m apart and registered as V13/NKAH 1482 and NKAH 5576 (Fig. 5). Figure 6. Survey plan of V13a/NKAH 1509 – Illorsuit. 1-3, 15, 21-22: uninsulated outbuildings, 4, 12-13, 16, 18: enclosures (animal pens/hay yards), 5, 11, 13: uncharacteristic/unverified outbuildings, 6: dwelling, 7: stable complex, 8-10, 12, 14, 16, 19-20: insulated outbuildings, 23: storehouse, 24-25: possible irrigation channels, 26-27: possibly boundary walls, A-B Thule culture/pre-1900 Inuit tent foundations (map: C.K. Madsen 2020). Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 14 central hallway. The western rooms sit directly on bedrock, while those in the back are dug somewhat into the slope. The ruin is preserved as a two or three course, 1.0–1.2 m wide double row stone foundation for pure turf walls. On the grassy slope below are midden deposits mixed with turf collapse from the dwelling (Table 3). The remainder of NKAH 5576 ruins (Fig. 5) include animal pens and shelters 2 and 7, insulated outbuildings 4–5 (sheep/goat sheds?), uninsulated ruin 3 (a barn or shed?) and a mainly wooden storehouse on stone foundation perched on the bedrock edge of the shelf with the site. While NKAH 5576 displays a layout typical of small Norse farmsteads, the unusual setting on the hillside and lacking homefield, as well as the insubstantial character of the dwelling, leads us to interpret the site as a transitional farmstead/shieling (see discussion). There are three dates from the site, one from the lower midden and two from inside the dwelling (Fig. 5, Table 3), all statistically identical with a median probability around cal. AD 1100 (Table 2). Thus, if NKAH 5576 was a farmstead, the tight range of the dates imply that it was relatively short-lived, perhaps replacing V13’s early landnám settlement. Even if considering V13’s northern cluster of ruins (Fig. 5 nos. 2–5, 7), part of NKAH 5576 does not change the general impression of it as a small site. V13a/NKAH 1509 – Illorsuit Although the location of Norse V13a – Illorsuit (“large houses”) was roughly accurate in Kleinschmidt’s map of Nunatarsuaq (Fig. 3), the site was omitted or slightly misplaced in later maps (cf. Jensen 1889:tavle IX). However, the accounts of Bendixen (1916) and Bruun (1917) leave no doubt that they were describing V13a, which was first sketch mapped by Roussell (1941:62f.). Apart from an aerial survey (Kapel 1989), the site has not been archaeologically investigated until our site visits in 2012 and 2019, during which we added 10–14 new features to the existing site record (cf. Fig. 6 and Roussell 1941:Fig.40). V13a is located by a wide bay one third into the Kangersuneq, where the 23–27 features of Norse farmstead V13a lie at the mouth of a broad valley (Figs. 1, 6). This valley ascends steadily 5 km northeastwards to above 500 masl with a sizable river meandering down its middle. Below ~200 masl there are considerable, but patchy areas with lush vegetation of willow shrub, grassland, and meadow, but generally the drained and boulder-strewn valley floor presents low-quality pastureland. The relict Norse homefield by the fjord stands out distinctly against the surrounding natural vegetation; part of shape, built and date of Ruin 1 suggest that it was a small longhouse, or skáli, belonging to the earliest phase of Norse settlement—in fact, the only third such longhouse ever dated in Greenland (Albrethsen 1982; Albrethsen and Ólafsson 1998; Vebæk 1993). The landscape setting of Ruin 1 is surprising for an early dwelling, but poorly preserved bone fragments in the lower collapse layers (Table 3, [05]) imply human habitation. Perhaps it was a short-lived attempt at setting up a farm that moved to NKAH 5576 (see below) after a hundred years. Based on its appearance and location, Ruin 6 was undoubtedly part of the same settlement, but it is uncertain whether outbuildings 2–3, and animal pens 4–5, 7 belonged to it as well. Alternatively, these ruins could have comprised a dairy shieling or a part of NKAH 5576 (see below) used for seasonal grazing. Ruin 2, perhaps a sheep/goat shed, is disturbed by a later Thule culture hunting camp, traces of which are found all over V13. Unfortunately, we only had time to survey some of the features (Fig. 5) and at least two tent rings of some age were noted, but not surveyed. This Inuit hunting camp lay at the end of the newly discovered caribou drive system NKAH 5591. Feature C appears to be a meat cache reusing the chamber and stones of a large chamber trap built against a boulder. With an internal width of ~70 cm, Feature C is too big to be a normal fox trap and must instead be grouped with the rarer, so-called “wolf traps” that are found in several parts of southwest and southeast Greenland (e.g., Bruun 1895:431f., Holm and Vilhelm 1889:80, Kapel 1982). Next to the trap lay the “trap door”, a flat stone measuring ~55 x 45 cm with a conical hole in its center. NKAH 5576 – Mannissuaq The steep slope below V13/NKAH 1482 is vegetated by dense and 1.0–2.5 m high willow shrub. However, at 90 masl is a narrow shelf on the hillside with a grassy clearing, providing the setting for the seven nucleated Norse ruins of NKAH 576 (Fig. 5). There are no traces of, nor room for, a homefield as every level part of the shelf is occupied by a building. This setting agrees well with Bruun’s (1917:77) description of V13, which also mentions a partly submerged ruin by the fjord. While such a ruin was not observed during our 2019 survey, the former presence of a shed for boat and fishing gear is plausible, although the boulder-strewn beach today makes for a poor landing site. Although not preserved to great height, the ruins of NKAH 5576 are quite distinct and undisturbed by later activities (Fig. 5). Ruin 1 is a dwelling measuring ~17.9 x 15.7 m and made up of at least seven merged rooms (or buildings) split in two parts by a Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 15 name “Qatsigsut” (modern spelling Qatsissut) derived from the earlier site records (Fig. 2, Bendixen 1916, Bruun 1917:78, Jensen 1889:110). We argue, however, that the site originally numbered V14 and named Qatsissoq (Figs. 2–3) is identical to the newly discovered site NKAH 5577 (see below). Fig. 2 (no. 79) shows a Norse site named Iviángiussánguit (“lovely, small breast-like”, implicitly hills) roughly where V14 is recorded today, although we have no explanation for the alternating place names. V14 – Illorsuakasiit was resurveyed (Kapel 1989) before our 2012 and 2019 site visits, during which we added three new Norse features to the existing record. V14 is located by a small, sheltered bay about halfway up the Kangersuneq (Fig. 1, 7). The ruins are nucleated on the lowest, gently sloping part of a small headland sheltered under a steep mountain side. On recent sites maps, V14 has been placed ~300 m to the north, where a sizable river running through the steeply ascending valley drops into the fjord (Gulløv 1983:152, Krogh 1982:252). Patchy but rich willow shrub covers the lower mountain slopes, while the valley system to the northeast offers sizable areas of grassland and meadow up to ~350 masl. Like at V13a and V15, V14’s relict Norse homefield stands out clearly from the surrounding vegetation but appears less drained, perhaps because of a small stream running parallel to the homefield (Fig. 7, A.). This stream is the farmstead’s only stable water source (a second stream B south of the homefield is today completely dried up), which could explain why stream A may have been extended artificially (see NKAH 5578 below). The ruins of V14 are generally well-preserved, a possible effect of their sheltered landscape setting (Fig. 7). They concentrate around Ruin 1, a sizable, but completely collapsed dwelling with a midden in front. Ruin 2 is a heavily insulated building (stable complex with byre?). Ruins 5 and 7 are smaller, insulated outbuildings (sheep/goat sheds?), the former with an enclosure (hay-yard?). Ruins 2 and 4 are separate enclosures (animal pens), ruin 6 a small uninsulated building (barn/shed). Ruin 8 sits just above the fjord and appears a turf and wooden storehouse raised on a heavy, single row stone foundation. Ruin 9 is a homefield boundary wall running downslope towards the fjord. The notably more centralized layout, lower number of buildings and small homefield implies that V14 was a more modest farmstead (see discussion). The 50 x 50 cm test unit excavated in the midden (Fig. 7) revealed a 61 cm deep stratigraphy of cultural deposits on peaty, wet soils, the lower part separated by several horizons of only partly decomposed grass (Table 3). Inexplicably, the 14Cthis homefield, and possibly additional ruins, must have been lost to coastal erosion, although it today appears somewhat stabilized. The majority of V13a’s ruins are spread in and around the homefield area, most preserved only as stone foundations for pure turf walls (Fig. 6). The large dwelling 6 and insulated outbuilding 7 (stable complex?) with a merged midden area occupies a central position. Ruins 8–10, 12, 14, 16–17, and 19–20 are insulated outbuildings (animal sheds and stables?), of which 12 and 16 lie at some distance— ~450 and 320 m, respectively—from the farmstead’s nucleus, both with adjoining enclosures (animal pens or hay yards). These more distant features most likely served as close shielings—herding/milking stations—during spring and fall grazing and herding of livestock on the lower valley, home pastures. Ruins 4, 13, and 18 are also animal pens, while Ruins 1–3, 5, 15, and 21–22 are uninsulated outbuildings (various sheds and barns?), Ruin 23 a storehouse. Ruin 5 stands out by a distinct outer wall enclosing a central heap of collapse stone, i.e., a layout reminiscent of the Norse churches with churchyard (Arneborg 2004:250). However, the outer boundary of Ruin 5 measures only 13.3 x 10.1 m, which would make for a very small, in fact the smallest, of all known Norse churchyards in Greenland (cf. Roussell 1941:32ff., Krogh 1982:27ff., Arneborg 2004:250). Ruins 11 and 13 recorded by Roussell (1941) could not be relocated, but both were described as poorly preserved and indistinct. A distinct homefield boundary wall mentioned by Bendixen (1916) is probably one of two lines of intermittent boulders 26–27 that may have been foundations for turf walls. Features 24–25 are possible irrigation channels, if not old, dried-up stream beds later used as caribou tracks. The homefield sits on a heavily drained, gravelly subsoil, and an irrigation system could have helped mitigate summer droughts. The combination and number of features at V13a show that it was a sizable farmstead, possibly specializing in sheep/goat husbandry (see discussion). Two 14C-dates from a test unit excavated in the midden (Fig. 6) overlap greatly and are suggestive of mixed deposits but cover a range of cal. AD 1055–1250 to 1158–1255 ([02] top and bottom, Tables 2–3). There are also several Thule culture and later Inuit features at V13a, of which we only surveyed the older looking tent foundations, A and B, near the beach and with an associated midden area. V14/NKAH 1508 – Illorsuakasiit The Norse site presently recorded as V14 – Illorsuakasiit (“terrible large houses”) was first described by Roussell (1932, 1941), but by the place Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 16 Figure 7. Survey plan of V14/NKAH 1508 and NKAH 5578. At V14: 1 - dwelling; 3, 5, 7 - insulated outbuildings; 6 - uninsulated outbuilding; 8 - storehouse; 2, 4 - enclosures; A-B streams mentioned in the text. At 5578: 1: uninsulated outbuilding, 2: enclosure, 3-6: confirmed or possible dykes, 5: storehouse (m ap: C.K. Madsen 2020). Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 17 dates from the upper and lower midden ([02] and [05], come out slightly inverse at cal. AD 1275–1388 and 1271–1390 (Table 2), respectively. Although statistically almost identical, this overlap is even more puzzling from the fact that we had to attempt to date the lower [05] from a heather type (Ericales) sample more likely to turn out older dates. The test unit at V14 is the only place in Nunatarsuaq where we have observed excellent bone preservation, the unit producing ca. 250 bones > 1 x 2 cm (not analyzed). NKAH 5578 NKAH 5578 is a new Norse site discovered in 2019 (Figs. 1, 7–8). Following stream A from V14 uphill to just above at 220 masl, one reaches a terminal moraine formed by a small cirque glacier. The “bowl” behind the moraine now shelters a small patch of lush, inundated grassland and willow shrub. Four to 6 Norse ruins lie on both sides and at the southern end of the moraine. Ruin 1 is a small turf and stone building with two, separate rooms nestled against a boulder. Except for the boulder and being slightly smaller (~7 x 5 m), Ruin 1 is very reminiscent of a small V53 ruin (~9 x 6 m) excavated by Roussell (1941:288ff.), which had a fireplace in each room and was interpreted as a shieling. Ruin 2 is an animal pen, Ruin 5 a foundation for a small wooden storehouse placed on the windward side of the moraine. Ruin 3 is a double row and ~1 m wide stone foundation for a wall or dyke lying at right angles to the moraine and ending abruptly in the meadow, Ruins 4 and 6 possible additional stretches of wall or dyke. The combination of a turf/stone (warm-period) habitation building 1, animal pen 2 and storehouse 5, implies that NKAH 5578 was small dairy shieling to V14 used when pasturing livestock in the lower valley, i.e., a set up similar to the distant clusters of outbuildings and pens at V13a and V15 (Figs. 6, 9). However, the walls or dykes, especially No. 3, are puzzling. They could have formed a delimited grazing area using the steep slope and moraine as part of the enclosure or, alternatively, a small dam or reservoir that fed the stream A (Fig. 7). The latter interpretation seems supported Figure 8. Photo, looking SW, of the newly recorded NKAH 5577 ruin 1. Perched on the edge of a small shelf, this probable Norse shelter or lookout enjoys an excellent view of the surrounding valley system. The terminal moraine with new Norse site NKAH 5578 is visible in the lower part of the valley towards Kangersuneq (photo: C.K. Madsen 2019). Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 18 by the unnaturally looking channel or cut through the southern end of the moraine by Ruins 1–2 and which forms the beginning of stream A. NKAH 5577 – Qatsissoq The single archaeological feature at NKAH 5577 – Qatsissoq (“elevated one” or “highland”) was discovered during our 2019 survey. We suggest this site is identical to the Norse V14 described in earlier records (Bendixen 1916, Bruun 1917:78, Jensen 1889:110). These all map V14 as somewhat withdrawn from the fjord and describe a single ruin lying unusually high on the mountain, giving origins to the Greenlandic place name. At least this place name and the single ruin mentioned excludes the possibility that the “original” Qatsissoq can be identical to the current V14– Illorsuakasiit (see description above and Roussell 1941), although NKAH 5578 (see below) could be an alternative for the “original” Qatsissoq. The ruin at NKAH 5577 is perched on the edge of a triangular shelf overlooking the joining of two valley systems above V14 and NKAH 5578 (Figs. 1, 8). At 450 masl, NKAH 5577 is the highest lying Norse site recorded in Nunatarsuaq and fits the place name Qatsissoq perfectly. Well-used caribou tracks leading across the shelf suggest it was a nexus in caribou migrations and would explain why 19th-century Inuit hunters would have known of the site. The ruin itself (Fig. 8) is stone-built, measuring only ~2.3 x 2.1 m, squarish of shape, and possibly had no eastern wall. Although heavily collapsed, the building’s western corner is preserved as a ~60 cm wide wall with six courses of largish, flattish stones irregularly stacked to a height of up to ~60 cm. Nearby collapse stones, especially downslope from the shelf, show that the walls originally stood somewhat higher; it is also possible that the walls originally included interspacing turf layers or turf superstructure. While the build of the ruin does not appear especially Norse, it is also atypical of the Thule culture. Sitting directly on exposed bedrock, it was impossible to carry out test excavation or probing of the feature, which thus remains undated, although lichen growth suggests that it is of considerable age. Placed with an excellent view of the valley system above V14 and NKAH 5578, we suggest NKAH 5577 was a Norse shelter or lookout used during hunting or herding, possibly connected to NKAH 5578. Figure 9. Survey plan of V15/NKAH 1462 where: 1: dwelling, 2-4, 10-11: insulated outbuildings, 6: uninsulated outbuilding, 7: storehouse, 6, 8, 14 enclosures, 9, 13 homefield boundary wall, 15-16: possible irrigation channels (map: C.K. Madsen 2020). Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 19 V15/NKAH 1462 – Umiivik The identification of V15 – Umiivik (“place where you pull the umiaq on land”) has been consistent (Figs. 2–3), the only exception being Giesecke’s (1910:150) first record of the site, where he named it “Illorsuit”. However, three conditions demonstrate that he visited V15: First, only two locations in Nunatarsuaq match his account of a wide valley with many Norse ruins, i.e., V13a and V15 (Figs. 6, 9). Second, landing at V13a would place Giesecke too far from the QS and KNS glaciers for him to visit them on single day’s hike (see below). Third, a round hill overlooking the valley with V15 (Fig. 1) is such a prominent landmark that it can only be the moundshaped hill referred to by Giesecke (1910:150). The definite identification of Giesecke’s landing point is essential to interpret his travel account, the glacial history of the QS and KNS and, as we argue below, the existence of another Norse site in the QS valley. Later site descriptions of V15 (Bruun 1917:78; Kapel 1989; Roussell 1932b, 1941:59ff.) were generally confirmed by our 2012 surveys, although we added at least six new Norse features to the site record (Fig. 9). The 14–16 features of V15 are dispersed in and around a clearly discernable, heavily drained homefield area at the mouth of a broad valley. Coastal erosion today appears moderate, but some of the homefield area, possibly also ruins, must have been lost to erosion. A large river flows into the fjord south of the site, from there winding its way up the middle of the valley, about 800 m from the fjord and at 200 masl, bending around a prominent terminal moraine. Giesecke’s hill rises abruptly on the far edge of the wet plain behind this moraine (Fig. 1). The valley above Umiivik presents fairly rich and expansive pastureland vegetation up to ~300–400 masl, including considerable areas of meadow along and on both sides of the river. Gradually narrowing, the valley continuous up into the mountainous hinterland almost to the edge of the QS valley. The ruins of V15 are mostly well-preserved, including the large dwelling 1 with a midden area in front. Ruins 2–4 and 10–11 are insulated outbuildings, Nos. 3 and 5 still preserve in situ stall stones, suggesting they were stable complexes with byres with a combined capacity for at least fifteen cows. Ruin 6 is an uninsulated outbuilding, perhaps a sheep/goat shed for outfield grazing, whereas Ruin 7 is a storehouse placed on a drained knoll. Ruins 8 and 12 are probably animal pens, whereas the enclosure built against Ruin 6 could be a hay-yard. Figure 10. Photo, looking SW, of the newly recorded Norse NKAH 5580 ruin 1, a small storehouse or herder’s/hunter’s hut. The upper valley above V15 – Umiivik and the lush plain behind a terminal moraine is visible to the left (photo: C.K. Madsen 2015). Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 20 No. 14 is also an enclosure, partly delimited by the steep slope created by the river, and so large that it may be described as a grazing enclosure, a feature type normally associated with cattle herding (Madsen 2014b:118). Ruins 9 and 14 are stretches of homefield boundary walls. On the drained and stony glacial till, the connecting parts of the boundary wall must have been built without a foundation and are therefore not visible today. Features 15–16 are possible irrigation channels, if not old dried-up stream beds reused as caribou tracks. Roussell undertook test excavations on Ruin 2 and some undefined units to locate graves of a Norse church but only found what he interpreted as temporary, kettle-shaped cooking pits (Roussell 1941:61–62). We did not test excavate at V15, but probing in the lower part of the midden with a 2 cm diameter corer revealed shallow cultural deposits with poor organic preservation (Table 3). Samples from a test unit in the homefield retrieved during a recent paleoenvironmental investigation yielded two 14C-dates (Schofield et al. 2016). Inexplicably, but like at V14, the upper and lower 14C-dates are reversed at cal. AD 1223–1278 and 1265–1395, respectively (Arneborg et al. 2012b). This indicates ongoing, later-medieval field amendment at V15, which we interpret as a large farmstead. There are also numerous traces of Thule culture and later Inuit hunting activities at V15, including modern debris pointing to the site’s continued use, which has partially disturbed and obscured earlier features. Unfortunately, we did not have time to detail investigate and survey most of these features, but Umiivik appears to have been a significant summer hunting camp. This corresponds well with Giesecke’s description (1910:152) of several Inuit families tenting at the site in 1808. Several Inuit chambered fox traps, some reusing stones of the Norse ruins (e.g., Fig. 9), are found along the beach and up the valley, showing that the area was later much-used for fox trapping. NKAH 5580 – Umiivik Valley The single Norse ruin of NKAH 5580 is located 2 km up the valley from V15 – Umiivik at about 300 masl. (Figs. 1, 10). The feature was built up against a low, southeastern-facing cliff wall a little up the valley’s side and offers an excellent view of the surrounding landscape. Using boulders split off the cliff face as part of its walls, the ruin is somewhat irregular of shape, measuring 5.2 x 4.9 m externally. Between the Figure 11. Survey plan of V16/NKAH 1463 where: 1 - dwelling; 2 - uninsulated outbuilding; 3 - storehouse; and 4–5 - possible stone foundations for turf buildings or enclosures. Note the proximity of the glacier at the LIA maximum (dashed grey line) (map: C.K. Madsen 2020). Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 21 boulders are dry-stone masonry walls ~60–70 cm wide, preserved in uneven courses up to a height of ~1.2 m. Most conspicuous is a “doorframe” by the building’s northern opening (Fig. 10) showing that it was a roofed building. The dry-stone masonry and setting up the valley’s side implies the ruin could have been a small storehouse, possibly for drying and storage of meadowhay from the plain below. Alternatively, the building was a herder’s/hunter’s hut situated to overlook the surrounding terrain. Either way, it must have been a shieling to V15 only some 1.8 km distant. No excavation or probing was carried out at NKAH 5580. V16/NKAH 1463 – Saqqarsuaq The site today registered as V16 – Saqqarsuaq (“big sunny side”) was located and partly excavated in 1932 (Roussell 1932a, b, 1941:78, 162). The site was later photographed during a helicopter survey (Kapel 1989), although mistaken for an unregistered site, likely because it has previously been misplaced ~2 km to the southeast in official heritage maps (Krogh 1982). V16 has seen no other archaeological investigation until our 2012 and 2016 site visits that identified two new possible features. As we argue below, today’s V16 is probably not identical to a Norse site noted by Figure 12. Excavation plan of V16/NKAH’s dwelling 1, with interpretation of the rooms as: I, VI–VII - passages; II - living room (stofa); III - kitchen (eldhús); IV–V - storerooms (bur); VIII–IX - byres; X–XII - barns (after Roussell 1941:Fig.99). Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 22 Giesecke (1910:151) in 1808 and formerly mapped as V16 some way up the QS-valley (cf. Fig. 3). V16 lies in a southwestern facing, kettle-shaped recess in the mountainside just above the 200 m contour line and ~750 m from the fjord (Fig. 1, 11–12). To the north, the mountainside ascends abruptly to above 900 masl. Ruins 1–2 are situated on a grassy slope that may have been a small homefield. However, the meadow/mire in the bottom of the kettle must have been the main area of fodder growth (meadowhay). The surrounding pastureland is very patchy and poor. The impression of an ostensibly marginal setting is reinforced by the proximity of KNS during its LIA maximum, the limit of which is clearly seen 400–420 m southwest of the site (Fig. 11), although it is unclear if the glacier was already encroaching on V16 while it was inhabited. Indeed, V16’s setting prompts one to speculate—as Roussell also did (1932b)—why the site was not placed further up the QS-valley, which offers some of the richest and most expansive pastureland in Nunatarsuaq. The layout of V16 is strikingly nucleated and was presented as archetypical of the smallest centralized farms in Roussell’s (1941:162) Norse site typology. Sheltered under a low hummock is Dwelling 1 with no visible midden area, below it an uninsulated outbuilding 2 (a barn?). Ruin 3, a dry-stone storehouse, occupies a more wind-exposed setting north of the hummock. In front of Ruin 3 are indistinct foundations, Ruins 4–5, for outbuildings or enclosures. Several fox traps are found around and in the Norse ruins. In a hollow under a boulder that likely served as a lookout (Fig. 11, A), we found a worked scapula with cutmarks and a row of drill holes, a bone parting technique generally associated with the older Thule culture (Gulløv 1997). Reuse of dwelling 1 for Thule culture habitation was also noticed by Roussell during its excavation (1932b). This excavation was carried out over only three days and without screening, but still provides a valuable record as the only excavated Norse ruin in Nunatarsuaq (Fig. 12 and Roussell 1932a, b; Roussell 1941:162ff.). Dwelling 1 consisted of eight rooms or buildings divided on two not interconnected, functional halves: The eastern half served human habitation, the western half as quarters for the livestock, including a byre with stall stones indicating space for 4–6 cows. The dwelling was very coarsely built with irregularly and loosely stacked walls partly made up of natural boulders and floors dug well into the sloping terrain. Phasing of building development was only noticed in one area (room IX). The excavation produced a low number of artifacts (n = 15) and animal bones (n = 39), the latter displaying a mix of domestic and wild species typical of Norse sites: 9 cattle, 2 goats, 12 sheep/goat, 5 Caribou, 5 Harp Seal, 1 Harbor Seal, 4 walrus, 1 hare (Degerbøl 1936, McGovern 1985). Roussell’s field notes suggest the scarcity of finds was real, not only a bias of excavation methodology. In all respects, V16 appears a very modest farmstead and, along with NKAH 5577, belongs to the problematic type of sites we here label transitional farmstead/shieling (see discussion). Giesecke’s V16 During his 1808 Nunatarsuaq visit, Giesecke undertook a one-day, inland excursion during which he found: “… a clear remnant of an old Norse building that protruded above the small shrubs. It formed a square and measured close to 50 × 50 feet (~15.5 x 15.5 meter). In some places the walls reached c. 1 ell (~0.6 m) and a bit more above the moss, grass, and shrubs. The inner walls were difficult to discern, but the shrubs indicated three main rooms. Several other buildings in this rocky valley were probably buried below the glacier long ago” (Giesecke 1910:151 transl. by N. Mehler). The site was later numbered V16 and mapped near the front of QS (Fig. 3), the latter surveys adding the information that there were: “two houses near a river that runs from the Inland Ice” (auth. transl. after Jensen 1889:110) and: “a not particularly large (Norse) ruin, but where there were previously more, which are believed to have been washed away by a river or destroyed by the ice” (auth. transl. after Bruun 1917:78). Although no Norse site near QS has subsequently been identified, there are several reasons to suggest that the Giesecke’s site is not identical to the site registered as V16 today. First, having established that Giesecke embarked from V15 – Umiviik and not V13a – Illorsuit (see above), his survey route can tentatively be reconstructed: he must have travelled up the valley above V15—on the way climbing the prominent, domeshaped hill (Fig. 1)—to where it descends towards the QS, and where he found the Norse ruin. From there, he headed back via the two mountain peaks on the range between the Umiviik and QS valleys, taking note of the geography and “Saqqarsuaq Lake” (see Fig. 1 and above), before finally heading back to his camp at V15. In his diary, Giesecke notes (1910:152) that this excursion was carried out between 2:00 and 22:00 pm and covered ~22.5 km (3 meilen), a distance correlating almost exactly with the reconstructed route—and it precludes that he could have embarked from V13a. Second, while the KNS was in retreat after AD 1761, QS was apparently advancing during the 19th century (see above), thereby supporting the noted destruction of several ruins in the valley and at Giesecke’s V16. The topography and distance of KNS’s LIA margin to today’s V16 (Fig. 11) seems to exclude the possibility that it was ever immediately threatened by the glacier. Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 23 Third, the account of ruins having been washed away by a river also precludes that today’s V16 and Giesecke’s V16 are identical since only minor streams pass by the former, and there are no traces of any nearby, old river outlets. The geographic position of Giesecke’s V16 shown in Fig. 1 is rough and based on georeferencing of Bruun’s site map (cf. Fig. 3). However, Giecke’s V16 must have been found some way up the QS valley and the nearby river referred to by observers was the outlet from the small, unnamed glacier north of QS (Fig. 1). A 1932 search for Giesecke’s V16 failed to locate the site (Knuth 1934, 1944) and neither did we manage to find it despite intense searching along the still visible banks of “Saqqarsuaq Lake” in 2015. Quite possibly, Giesecke’s V16 no longer exists. However, Giesecke’s V16 must have been a relatively small site, probably like NKAH 5576 and V16 – Saqqarsuq, i.e., a small farmstead or summer farm. Norse farmsteads located far up the sheltered, glacial valleys are known in several other parts of the Western Settlement (Andreasen 1982, Knuth 1944, Roussell 1941). Thule Culture and Later Inuit Archaeology in Nunatarsuaq Apart from an uncertain feature resembling a Saqqaq-culture hearth (Fig. 1, NKAH 5650), the archaeological sites in Nunatarsuaq broadly date to the Thule culture. Prior to our project, only eight such sites were recorded in the region, to which we have added 22 new sites (Table 1). These range from single features to multicomponent campsites, but generally display well-known site types in Nuup Kangerlua associated with late summer (caribou) hunting (Gulløv 1983, Knudsen and Andreasen 2009, Odgaard et al. 2008, Pasda 2014). Habitation sites are represented by summer tent camps NKAH 1349–51, 1392, 5583–84 (Fig. 1), as well as by at least one to four tent features found, but not all mapped, at V13, V13a, V14, and V15. Except for V14, all these sites had associated midden areas, implying prolonged or reoccurring use. Although the summer camps vary in size from one to four tent features, typological differences suggest that no more than a maximum two tent features were in simultaneous use at each site, i.e., these sites were seemingly not occupied by more than one or two families at one time. The three tent camps on the Narsaq headland, including NKAH 1349 overlapping V12 (Fig. 4), support the claim that this was the main entryway to the Nunatarsuaq hinterland in late 19th to early 20th century (Lennert 2017, Lennert and Mikkelsen 2015, Nielsen 1910). However, the summer tent camps at V13, V13a, V14, and V15 show that all entranceways to the Nunatarsuaq hinterland Figure 13. Sketch plan of Thule culture/historic Inuit cave site NKAH 1224 showing the extent of the cave, the built-up wall at its mouth and the midden area in front. Red stars indicate the 2019 test units (map redrawn after Kapel 1989: n.p.). Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 24 were used, although the latter perhaps mainly prior to AD 1761 when ice conditions in Kangersuneq were more favorable. The four cave sites found in Nunatarsuaq (NKAH 1224, 1287, 1347–48) were also significant habitation sites and, according to their descriptions (Table 1), could house at least similar-sized groups as the largest tent camps. NKAH 1224, 1347–48 appear to be situated to overlook the surrounding valleys and migrating caribou, although the locations shown in Fig. 1 are only roughly accurate. Only NKAH 1287 has been archaeologically investigated (Kapel 1989) and briefly visited during our 2019 survey (Fig. 13). The cave measures ~6.0 x 7.5 m and displays signs of very recent use. Two small test units excavated in the midden in front of the cave produced only 20thcentury debris, and it is uncertain how far back the use of this or any of the other caves can be dated. Overnight sites (Fig. 1) are hunter’s beds or rock shelters used by one or a couple of hunters on brief excursions from the summer camps (Gulløv 1983, Meldgaard et al. 1983), represented by sites NKAH 5592–93, 5584–88. These rather inconspicuous sites are expectedly underrepresented. However, all appear placed at convenient nodes in the landscape, where the caribou would pass nearby, i.e., narrow isthmuses, mountain passes, narrow valleys. Hunting blinds (at NKAH 5589, 5670) are also associated with hunting excursions, whereas the meat caches (NKAH 5651, 5658, 5660, 5668–69) and cairns (NKAH 5653–5656, and two near 5584) located along the valleys are associated with the movement between campsites and the fjord. The impression of Nunatarsuaq as a lightly used hunting landscape is only contradicted by the discovery of caribou drive system NKAH 5591 a little east of V13 (Fig. 1). This drive consists of a least 27 cairns (inussuk), single or composite, that form a line ~850 m long, possibly extending further up the valley. Such drives needed the collaboration of a substantial number of people to be “activated” and reflect large-scale communal hunting activity (Meldgaard et al. 1983). Although NKAH 5591 has not been dated, parallel evidence suggests that caribou drive hunting was mainly practiced prior to the population and settlement disruption caused by colonization (Meldgaard et al. 1983). Discussion Of the ten or eleven Norse sites in Nunatarsuaq, V13a, V14 and V15 (Figs. 6–7, 9) are clearly farmsteads. V13a and V15 display similar layouts with numerous dispersed buildings, including more dis- Figure 14. Graph showing accumulated building and homefield areas for Nunatarsuaq sites interpreted as farmsteads and church farms from the Norse Western (V51, V7) and Eastern (Ø83, Ø149, Ø66, Ø29&29a) Settlements. Left vertical axis and bars shows the area in sq. meters for the five main building categories: dwellings, insulated outbuildings, other outbuildings (uninsulated outbuildings and storehouses), enclosures and uncertain/unverified buildings, while the line and right vertical axis display the areas of the homefield. Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 25 tant clusters of outbuildings (V13a, nos. 11–13, 16, 22; V15, nos. 6, 13, 15) used during early spring and late fall grazing of lower valley home pastures, i.e., functioning almost as close shielings (dairyand herding stations). NKAH 5578 probably had similar function at the smaller farmstead V14 (Fig. 7). V13a, V14 and V15 display other shared attributes, including possible water management (irrigation) and homefields enclosed by boundary walls. Distinct preservation of the homefield areas even today points to intensive soil amendment strategies (e.g., turf-cutting, clearing of stones, and sustained fertilizing). Comparing the number of features at V13a and V15 (27 versus 16, Table 4), V13a would seem by far the largest farmstead. However, the difference is far less pronounced if the areas of roofed buildings are contrasted (Fig. 14, dwellings, insulated and uninsulated outbuildings, and storehouses), V13a having more, but relatively smaller outbuildings, V15 fewer, but larger outbuildings. We suggest this difference indicates a degree of economic differentiation where V13a was more focused on sheep/goat husbandry, the many small outbuildings serving selective feeding of separately stalled animal groups (ewes, yearlings, rams, wethers, bucks, doelings etc., e.g., Bruun 1928:268f.), whereas V15 was more focused on cattle. This is also substantiated by V15’s byres (Fig. 9, nos. 3, 5) with room for at least fifteen cows, as well as the only example in Nunatarsuaq of a large grazing enclosure (no. 14), a feature type elsewhere associated with wealthy, cattle oriented farms (Madsen 2014b:118; Nørlund and Stenberger 1934:99ff.). Interestingly, the maximum cow capacity at V15 is on par with larger farms in contemporary Norway and Iceland (McGovern et al. 1988; Øye 2004:117). This example demonstrates that accumulated areas of roofed buildings is a more robust measure for a farm ranking than number of features. Roussell (1941:List VII) ranked V13a, V14, and V15 as farms of similar type. However, Fig. 14 shows a clear difference between almost equally sized V13a and V15 and the smaller V14, a difference also exhibited in the homefield areas. Still, Fig. 14 shows V14 to be a larger farmstead than would be assumed from its number of ruins (n = 9, Table 4). Fig. 14 also displays comparable architectural evidence from six Norse church farms in the Vestribyggð (V7, V51) and Eystribyggð (Ø29a, Ø66, Ø83, Ø149). Church farms were the economic and religious centers of Norse Greenland (Arneborg et al. 2012b, Madsen 2014b, McGovern et al. 1988, Roussell 1941), thus providing an upper baseline for identification of the most wealthy farms. By such comparison, Nunatarsuaq farmsteads V13a and V15 do not rival the largest of the Eystribyggð church farms (Ø29a, Ø66, Ø149), but do appear to match church farms V7, V51, and Ø83, a surprising implication considering the agropastoral marginality of Nunatarsuaq. However, interpretation of Fig. 14 requires consideration of several caveats: First, most buildings on the church farms were excavated (e.g., Nørlund and Stenberger 1934, Roussell 1936, 1941, Vebæk 1991), making measurements of building more precise, but also smaller than collapsed ruins at unexcavated sites. Second, there is no chronological resolution to the architectural evidence, which thus potentially represents hundreds of years of accumulated activity. However, consistent site ranking suggests that architecture size—especially dwellings, livestock buildings, and storehouses—does reflect site rank and status to notable extent (Madsen 2014b; McGovern 1985, 1992; Roussell 1941). Third, church farms seemingly had the setup of medieval estates with economic resources spread over several dependent farms (Arneborg et al. 2012b, Madsen 2014b, Vésteinsson 2010), the buildings on the central estate thus only revealing part of their economic wealth. This could explain the very modest appearance of V7, V51, and Ø83 in Fig. 14, albeit survey intensity (Table 4) and erosion of features (McGovern et al. 1996) undoubtedly plays some part. Fourth, farm buildings also reflect to the actual need for winter stalling and feeding of livestock. Although this issue has seen little formal investigation, Roussell (1941:85), for instance, observed more buildings on farms in northern Iceland than in the milder south, while McGovern (1985) suggested that relatively larger barn areas in the Vestribyggð than in the Eystribyggð reflected a greater need for winter stalling of livestock. Despite these concerns, the Nunatarsuaq farmsteads, notably V13a and V15, were sizable and probably independent farms. This is reinforced by their evidence of labor and resource-intensive irrigation and homefield amendment, which is normally attributed large Norse farms (e.g., Arneborg 2005, Krogh 1974, Madsen 2014b), and by matching church farms V7 and V51 that were situated in a similar environment. V13a, V14, and V15 also starkly contrast NKAH 5576 and V16. While the details of the V16 dwelling’s layout (Fig. 12) can certainly be questioned, the terrain surrounding the dwellings at V16 and NKAH 5576 precludes that they were ever substantially larger than documented. They, thus, stand apart not only by their small, nucleated buildings, but also by their higher altitude setting and lack of homefield areas, although the meadow at V16 most likely served as a hay-production area. The similarity of NKAH 5576 and V16 even extends to Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 26 the layout of their centralized dwellings, which both appear of rather insubstantial built and are separated in two main parts, at V16 demonstrably serving human and livestock habitation (Fig. 12). NKAH 5576 and V16 present a problematic Norse site type: While they appear too large and functionally complex to be shielings, they occupy marginal agropastoral settings, with or without a small homefield area, and exhibit highly centralized layouts (cf. Table 4, Fig. 14). Such sites have been interpreted as the smallest of farmsteads (Roussell 1936, 1941:228), “full shielings” (Albrethsen and Keller 1986), “complex shielings” (Madsen 2014b:193), and “summer farms” (Ledger et al. 2013). The last three terms stress seasonal (summer) use, but possibly involving all farmstead activities, albeit on a smaller scale. We suggest sites like V16 and NKAH 5576 be termed “transitional farmsteads/shielings” to reflect the relatively intensive and broad economic nature of onsite activities, but also a likelihood for having switched functional status, perhaps several times, between farmstead and shieling depending on need, population size, labor availability, livestock numbers, etc. Located on each side of V13a and V15 (Fig. 1), NKAH 5576 and V16 appear to have been satellites to the former. How “Giesecke’s V16” fits into this pattern is unclear. If the dates from NKAH 5576 are presented (Table 2), the transitional farmsteads/shielings were established during early settlement, but were abandoned, or changed function, well before by AD 1200. With their few ruins including animal pens and uninsulated housing, we interpret V12, NKAH 5577 (Figs. 4, 7), and possibly the northern cluster of ruins at V13 (Fig. 5), as more simple shielings, most likely summer dairy shielings used when grazing livestock on outfield valley pastures (Fig. 1). Such sites undoubtedly also doubled as gathering pens during seasonal roundups of livestock. Interestingly, the pens at these shielings (V12 no.1, V13 Nos. 3–4, 8, and 5578 no.2) enclose almost identical areas, 34.0, 35.1, and 29.9 sq. meters (Table 4), respectively, implying that they facilitated roughly similar sized flocks. Based on ethnographic parallels (Madsen 2014b), these pens would allow for the gathering of an estimated maximum of 17–83 ewes/she-goats with lambs/kids, a middling number seeming most realistic. Similar to the observed cattle capacity at V15, such herd sizes of sheep/goats are comparable to those kept on middle-sized farms in contemporary Iceland and Norway (McGovern et al. 1988, Øye 2004:117). Single-feature sites NKAH 5578 and 5580 were outstations— shelters and lookouts—used for herding or hunting in the lower valleys. In terms of general settlement patterns, the Norse sites in Nunatarsuq conform to other parts of the medieval settlements in Greenland: Farms were located in the environmental niches that provided the best opportunities for agro-pastoral farming, while seasonally exploiting more distant resource through transhumance with different types of shielings (Albrethsen and Keller 1986, Guldager et al. 2002:15ff., Madsen 2014b:189ff., Roussell 1941:228f.). Unfortunately, the 14C-dates from Nunatarsuaq (Table 2) do not allow for detailed reconstruction of settlement dynamics, except perhaps for the move of V13’s early phase dwelling downslope to NKAH 5576, which was then abandoned before AD 1200. The Nunatarsuaq 14C-dates fall neatly within the accepted occupation range ca. AD 1000–1400 for the Vestribyggð (Arneborg et al. 2012a), although with a clear statistical tendency (Table 2) for a winding down of activities well before AD 1400. However, nothing implies that the above noted, abrupt cooling in Nuup Kangerlua from ca. AD 1200 had an immediate impact on the settlement. The most surprising 14C-dates are from the short-lived dwelling V13 ruin 1 (Fig. 4), which suggests that Nunatarsuaq was settled in the earliest colonization wave. The Thule culture archaeology of Nunatarsuaq is more challenging to interpret because of expected low site representativity and a complete lack of chronology (Table 1). Clearly, late summer-early fall caribou hunting was the main activity during this period. Caribou drive system NKAH 5591 shows that the hunt could be relatively intensive, but in general the sites constitute small hunting camps or cave sites housing only a couple of families at the time. This correlates well with the above noted late 19th- and 20th-century historic accounts of low-intensity Inuit caribou and bird (gull) hunting mainly using a route via the Narsaq headland, east along the valleys and into the hinterland near the Ice Sheet (Lennert 2017, Nielsen 1910). Winter seal hunting and fox trapping were also important economic activities in the same period (Bendixen 1921, Bruun 1917, Nielsen 1910), as evidenced by hundreds of fox traps spread along the shore of Kangersuneq and by the valley mouths. However, the impression of relatively lowintensity later-19th to 20th-century Inuit use of Nunatarsuaq may in part be a later historical artifact considerably influenced by an increased calf ice production of KNS after AD 1761 that made the area more inaccessible (see above and: Lea et al. 2014; Weidick et al. 2012) coupled with regional changes in settlement patterns, a declining population, and decreasing seal numbers following colonization in AD 1721 (Bendixen 1921, Bruun 1908, Gulløv 1997, Rink 1877). In fact, the largest summer camps identified in Nunatarsuaq are found by Norse sites V13, V13a, and V15. The placing and distribution of Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 27 Thule culture sites suggest the use of all entryways to access resources across the entire Nunatarsuaq, although probably by rather small groups of people. How far back this more extensive Thule culture use of Nunatarsuaq can be dated is uncertain, but quite possibly it was associated with several winter settlements situated on Kangersuneq’s western shore across from Nunatarsuaq (Fig. 1 and Gulløv 1983). These sites represent the most inner-fjord, Thule culture winter settlements in Nuup Kangerlua, i.e., a rare break with typical West Greenlandic migration patterns and a strong testimony to former marine resources in Kangersuneq plentiful enough to sustain even winter occupation. One intriguing aspect of activities in Kangersuneq and Nunatarsuaq is that the region is also mentioned in Inuit oral tradition as a place of dense Norse settlement that attracted numerous early Inuit occupation (Rink 1866:205). This first meeting of Norse and Inuit initiated with a period of peaceful coexistence, followed by growing hostilities between the two populations that led the Norse to abandon the area and gather in the Ameralik Fjord, where they were eventually killed by the Inuit. Although the historical accuracy and details of this oral tradition can of course be queried, it does suggest early Thule culture settlement in Kangersuneq and local knowledge of the Norse settlements in Nunatarsuaq. Exploring Aspects of Settlement Marginality Based on the archaeological evidence presented here, the medieval farms in Nunatarsuaq can hardly be considered marginal and appear to match other parts of the Norse settlements in Greenland, perhaps even middle-sized farms in contemporary Iceland and Norway. However, because of the limited excavation evidence and chronological framework, it is difficult to evaluate if or how marginality—perceived or real—could have been a driving factor of settlement change, and ultimately desertion, in Nunatarsuaq. Several recent studies have highlighted how drawing on the concept of Human Securities can guide archaeologists to query their data in new ways and emphasize experienced livelihood in the past (Hegmon 2013, Hegmon and Peeples 2018, Nelson et al. 2016). Here, we explore Nunatarsuaq’s marginality through aspects of food, environmental and community securities afforded by the above established archaeological evidence and regional, long-term historical ecology. Food security can be described as physical and economic access to basic food, i.e., not only availability of enough food to survive, but also opportunity for accessing it (UNDP 1994). Although we have little direct evidence of the food economy in Nunatarsuaq, evidence from both Norse settlements suggest the economy would greatly, and increasingly, have depended on wild terrestrial and marine resources, in the Vestribyggð especially common seal, harp seal, and bearded seal (Erignathus barbatus Erxleben) (Arneborg et al. 2012a, Dugmore et al. 2009, Hartman et al. 2017, McGovern 1985, McGovern et al. 1996). It has been argued from Norse archaeofauna that seal provisioning strategies mainly took the shape of communal drive hunting in the outer fjords during the seasonal (spring) seal migrations and that increased sea ice conditions and storminess setting in at the onset of the LIA may have critically reduced access to this food resource (Dugmore et al. 2009, Hartman et al. 2017, Ogilvie et al. 2009). While such changes in food accessibility are likely to have impacted the Nunatarsuaq Norse, the above presented archaeological and historical evidence from later periods shows that Kangersuneq offered a uniquely rich niche for ice-based hunting of common seal, ringed seal, harp seal, especially in early spring. A noted increase in the duration and extent of fjord ice cover after ca. AD 1200 could even have had a positive effect on local seal abundance that was, at least, rich enough to sustain Thule culture winter settlement after the disappearance of the Norse. To this must be added the historically affluent caribou populations in Nunatarsuaq, a wild resource also greatly used by the Vestribyggð Norse (McGovern 1985, McGovern et al. 1996). Food preservation would also not have presented an issue with mean subzero temperatures persisting from October through May (see above), possibly longer during the LIA. Thus, while Nunatarsuaq would seem marginal in terms of its agro-pastoral capacity, the region enjoyed prime, perhaps improving, access to marine and terrestrial wild resources; and even if these were only supplemental food sources, food security was probably equal, if not better, than in other parts of the Norse Vestribyggð. Environmental security can be described as the resource stability, and predictability, of the inhabited environment, i.e., the possibility of acquiring needed resources, as well as foreseeing and counteracting any changes in resource availability (UNDP 1994). As noted above, the Norse in the Vestribyggð appears to have faced cooler and more unpredictable environmental conditions after ca. AD 1200, possibly including periodic, prolonged summer droughts from shortly before AD 1200 and into the 13the century (Andersen et al. 2006). Such environmental changes could have significantly impacted the viability of agro-pastoral farming and have also been suggested as a cause for the growing importance of Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 28 wild resources in the Norse subsistence economy (e.g., Arneborg et al. 2012a, Dugmore et al. 2012, Hartman et al. 2017, McGovern 1985). The evidence of homefield amendment and irrigation observed on the Nunatarsuaq farms must be seen as strategies to improve production and mitigate adverse environmental conditions, although it is uncertain to what extent they were successful. A more effective way of coping was probably by adopting a farming strategy that relied heavily on the use of extensive outfield resources, i.e., transhumance with shielings, an adaptive measure also made possible by the small, dispersed Norse population (Albrethsen and Keller 1986; Madsen 2014b). Nunatarsuaq offers a prime example of such very dispersed settlement with transhumance (Fig. 1), where transitional farmsteads/shielings V16, NKAH 5576 and, possibly Giesecke’s V16, played an important part. The meadow at V16 perhaps offered a low hay yield, but was less susceptible to droughts, while NKAH 5676 accessed considerable areas of resilient willow and birch shrub for leaf fodder. In terms of environmental security, Nunatarsuaq appears to have faced no greater challenges than other settlement areas. Community security can be described the security and resilience of an established group membership, i.e., “a family, a community, an organization, a racial or ethnic group that can provide a cultural identity and a reassuring set of values. Such groups also offer practical support” (UNDP 1994:31). In terms of medieval livelihood, this includes the prospect of starting a farm, getting married and establishing a family, of partaking in important seasonal community and religious events, of drawing on communal help and support for labor-intensive or collective activities etc. The Nunatarsuaq community consisted of a maximum six potentially coexisting Norse farms (V13a, V14, V15, V16, NKAH 5576, and “Giesecke’s V16”). Based on average household sizes in contemporary medieval to early modern Norway and Iceland (4.5–6.14 people per household, Pulsiano and Volf 1993, Vésteinsson 2006, Øye 2004), the Nunatarsuaq community thus counted a maximum 27.0–36.8 people. While such a small community would certainly have been vulnerable to demographic adversity—e.g., disease, famines, violence etc.—Nunatarsuaq was essentially part of the wider Vestribyggð community, from which it was only disconnected for some two to three months of the year (June–July), as the Kangersuneq was crossable, by boat or over the ice, between August and May. Connectedness may even have improved with the post-AD 1200 increased ice cover and decreased local storminess and reduced glacial output (see above). The average distance between farms in Nuntarsuaq is 7.4 km, 6.0 km, if including the transitional farmsteads/ shielings, i.e., every farm could reach its neighbor by half a day’s walk, probably less if going over the frozen fjord. The average distance of the Nunatarsuaq farms to the nearest Norse settlement at Kapisillit is 15.2 km, 16.7 km if including the transitional farmsteads/shielings, i.e., reachable within a single day’s travel. In comparison, average farm distances in peripheral settlement areas of late medieval to early modern Iceland range between 3.1 (Mývatnsveit), 3.3 (Hornstrandir), and 5 km (Krókdalur) (Vésteinsson et al. 2014), i.e., significantly less than in Nunatarsuaq. Farm dispersal and community disconnectedness have been argued as a reason for the eventual abandonment of these Icelandic settlement areas, a community insecurity that would appear to apply even more critically to Nunatarsuaq. However, average inter-farm distances in the Eystribyggð have been found to vary from 2.6–5.5 km depending on regional setting (Madsen 2014b; Vésteinsson 2010), a range also greater than the Icelandic parallels, but able to sustain the Norse communities there. With the nearest churches of V7, V23, and V51 located some 20–30 km from the Nunatarsuaq farms, access to religious networks and services could have been an especially adverse aspect of the region’s community security. However, there were many Norse farms located at greater distances from religious centers in other parts of both settlements and, again, there are, on the whole, no reasons to suppose that Nunatarsuaq should have faced significantly greater community insecurity. Conclusion In this paper, we have presented and discussed the archaeological evidence of 11 Norse sites, 4 of them new sites and 2 not previously mapped, as well as outlined the evidence of Thule culture and later Inuit activities in Nunatarsuaq, a remote and little investigated part of Nuup Kangerlua (the Nuuk fjord system), Southwest Greenland. Based on detail recording and classification of 84 individual archaeological features, Norse settlement in the region comprised three fair-sized farms (V13a, V14, V15) with associated satellite dairy-shielings (V12, 13, NKAH 5578), smaller outstations (NKAH 5577, NKAH 5580), as well as two to three sites interpreted as transitional farmsteads/shielings (V16, NKAH 5576 and, possibly, “Giesecke’s V16”), i.e., sites that likely changed function over time and according to need. Nine 14C-dates from excavated test units, as well as two supplementary 14C-dates from an earlier study (Schofield et al. 2016), show that Norse settlement in Nunatarsuaq fall within the accepted range for the Norse Vestribyggð (ca. AD Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 29 1000–1400, Arneborg et al. 2012b), while early 14C-dates from a possible, short-lived lándnam farm at NKAH 5576 suggest that the region was settled during the first colonization wave. Subsequent Thule culture activities are represented by 32 archaeological sites, 23 of them new, found across the entire Nunatarsuaq, most of them related to late summer caribou hunting. We find that low-intensity Inuit caribou hunting noted in the later historical period (post-AD 1850) is contrasted by earlier, larger archaeological sites located at all entryways to the Nunatarsuaq hinterland hunting grounds, including a newly identified caribou drive system (NKAH 5591). Thus, pre-colonial Inuit hunting activities in Nunatarsuaq appear to have been of higher intensity, probably affected negatively by post-colonial disruption of traditional settlement patterns, decreasing population numbers, and increasingly adverse ice conditions in Kangersuneq. Drawing on the long-term historical ecology afforded by archaeological and historical sources, and exploring this evidence through a human securities perspective, we find that the notion of Nunatarsuaq’s marginality or inaccessibility is overstated and partly an artifact of later historic hunting and research activities, i.e., the timing of these events to (late) summer months when Kangersuneq was most impassable. In fact, a former abundance of seal, caribou and other wildlife in Nunatarsuaq and Kangersuneq seems to have offered a uniquely rich inner fjord environmental niche not only able to sustain a small, dispersed Norse community of some 27.0–36.8 people residing on fair-sized, independent farms, but also later Thule culture winter settlements. This result, as well as a limited chronological framework, excludes the possibility of using Nunatarsuaq as a case study for evaluating a research hypothesis testing mainly climate-driven settlement dynamics and abandonment. Changing environmental conditions—including an abrupt change to cooler, drier conditions after ca. AD 1200—most likely forced a shift in Norse economic strategies in Nunatarsuaq similar to elsewhere in Norse Greenland, i.e., a decreasing importance of agriculture practices evermore reliant on wild forage and an increasing dependence on marine resources. However, we observe no direct evidence that Norse settlement in Nunatarsuaq—the land behind the ice—fared any worse than other parts of the Norse Vestribyggð. To the contrary, 14C-dates of the Nunatarsuaq settlement appears to cover the same range as, and two of its farmsteads even rival in size the largest church farms in the rest of the Vestribyggð. In short, we find no evidence to suggest that climatic or environmental change was a main driver of settlement change or abandonment in Nunatarsuaq. Acknowledgements The Winter Is Coming Project (WiCP), 2016–20, was funded by the Danish Council for Independent Research and housed at the National Museum of Denmark. A Millennium of Changing Environments in the Godthåbsfjord, West Greenland-Bridging Cultures of Knowledge, 2012– 17, was funded by the Ministry of Science Innovation and Higher Education of Denmark and housed at Ilimmarfik and Greenland Climate Research Centre. The authors would also like to thank the field survey participants— Kirstine E. Møller, Michael Nielsen, and Frederik Fuuja Larsen—who also translated the Greenlandic place names, as well as the Greenland National Museum & Archives for their support throughout the project. i The changing official number systems for Norse sites in Greenland are published in Nunatta Katersugaasivia Allagaateqarfialu’s (Greenland National Museum & Archives) online heritage database Nunniffiit (http://nunniffiit.natmus.gl/). ii Translated at the request of the authors by Natascha Mehler, Institut für Urgeschichte und Historische Archäologie, Universität Wien. Literature Cited Albrechtsen, S.E. 1982.Træk af den norrøne gårds udvikling på Grønland. Pp. 269–287. In B. Myhre, B. Stoklund, and P. Gjærder (Eds.). Vestnordisk byggeskikk gjennom to tusen år: Tradisjon og forandring fra romertid til det 19. årh.Arkeologisk museum i Stavanger, Stavanger, Norway. 287 pp. Albrechtsen, S.E., C. Keller. 1986. The use of Saeter in Medieval Norse farming in Greenland. Arctic Anthropology 23:91–107. Albrechtsen, S.E., G. Ólafsson.1998. A viking age hall. Pp. 19–30. In J. Arneborg and H.C. Gulløv (Eds.). Man, Culture and Enviroment in Ancient Greenland: Report on a research programme. Danish Polar Center & the Danish National Museum, Viborg, Denmark. 212 pp. Andersen, K.K., P.D. Ditlevsen, S.O. Rasmussen, H.B. Clausen, B.M. Vinther, S.J. Johnsen, and J.P. Steffensen. 2006. Retrieving a common accumulation record from Greenland ice cores for the past 1800 years. Journal of Geophyscial Research 111:1–12. Andreasen, C. 1982. Nipaitsoq og Vesterbygden. tidsskriftet Grønland 5:177–188. Arge, S.V., G. Sveinbjarnardottir, K.J. Edwards, and P.C. Buckland. 2005. Viking and Medieval settlement in the Faroes: People, Place and Environment. Human Ecology 33:597–620. Arneborg, J. 2004. Det Europæiske Landnam–Nordboerne i Grønland. Pp. 219–278. In H.C. Gulløv (Ed.). Grønlands forhistorie. Nordisk Forlag, København, Denmark. 434 pp. Arneborg, J. 2005. Greenland irrigation systems on a west Nordic background. An overview of the evidence of irrigation systems in Norse Greenland c. 980–1450 AD. Pp. 137– 145. In Water management in medieval rural economy. J. Klápste (Ed.). Vol. Památky Archeologicke´–Supplementum 17. Institute of Archaeology, Academy of Sciences of the Czech Republic, Prague, Czech Republic. 275 pp. Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 30 Arneborg, J., N. Lynnerup, and J. Heinemeier. 2012a. Human diet and subsistence oatterns in Norse Greenland AD c.980– AD c.1450: Archaeological interpretations. Journal of the North Atlantic Special Volume 3–Greenland Isotope Project: Diet in Norse Greenland AD 1000–AD 1450:119–133. Arneborg, J., N. Lynnerup, J. Heinemeier, J. Møhl, N. Rud and Á.E. Sveinbjörnsdóttir. 2012b. Norse Greenland dietary economy ca. AD 980-ca. AD 1450: Introduction. Journal of the North Atlantic Special Volume 3 (Greenland Isotope Project: Diet in Norse Greenland AD 1000–AD 1450):1–39. Barlow, L.K., J.P. Sadler, A.E.J. Ogilvie, P.C. Buckland, T. Amorosi, J.H. Ingimundarson, P. Skidmore, A.J. Dugmore, and T.H. McGovern. 1997. Interdisciplinary investigations of the end of the Norse Western Settlement in Greenland. The Holocene 7:489–499. Bendixen, O. 1916. Noter angående Nordboruiner i Godthaabsfjorden 1916. Unpublished field report on file at Antikvarisk Topografisk Arkiv, National Museum of Denmark, Copenhagen, Denmark. Bendixen, O. 1921. Godthaab District. Pp. 176–269. In G.C. Amdrup, L. Bobé, A.S. Jensen, and H.P. Steensby (Eds.). Grønland i Tohundredaaret for Hans Egedes landing. Meddelelser om Grønland Vol. 61. Bianco Lunos Bogtrykkeri, København, Denmark. 795 pp. Berglund, M.H. 1998. Red sandstone and Greenlandic wool. Two diagnostic artefacts in the interpretation of a newly discovered Saeter in southern Greenland. Acta Borealia 15:153–174. Bruun, D. 1895. Arkæologiske Undersøgelser i Julianehaabs Distrikt. Meddelelser om Grønland 16:173–461. Bruun, D. 1908. Godthaabsegnen og den gamle Vesterbygd. I tilknytning til et ved samarbejde med grønlænderne tilvejebragt kort over nævnte egn. Milo’ske Bogtrykkeri, Odense, Denmark. 15 pp. Bruun, D. 1917. Oversigt over Nordboruiner i Godthaabog Frederikshaab-Distrikter. Meddelelser om Grønland LVI:57–147. Bruun, D. 1928. Fortidsminder og Nutidshjem paa Island, Nordisk Forlag København, Denmark. 416 pp. Cappelen, J. (Ed.). 2012. Weather and climate data from Greenland 1958–2011–Observation data with description. Technical Report 12–15. Danish Meteorological Institute, Copenhagen, Denmark. 23 pp. Cranz, D. 1765. Historie von Grönland. Erster Teil, Leipzig, Germany. 512 pp. D‘Andrea, W.J., Y. Huang, S.C. Fritz, N.J. Anderson. 2011. Abrupt Holocene climate change as an important factor for human migration in West Greenland. PNAS 108:9765–9769. Degerbøl, M. 1936. Animal remains from the West Settlement in Greenland w. Special Reference to Livestock. Meddelelser om Grønland 88:1–54. Diamond, J. 2005. Collapse: How Societies Choose to Fail or Succeed. Penguin Group, New York, NY, USA. 571 pp. Dugmore, A.J., C. Keller, T.H. McGovern, A.F. Caseley, and K. Smiarowski. 2009. Norse Greenland settlement and limits to adaptation. In N. Adger, I. Lorenzoni, and K. O’Brien, Eds. Adapting to Climate Change: Thresholds, Values, Governance. Cambridge University Press, Cambridge, UK. 514 pp. Dugmore, A.J., T.H. McGovern, O. Vésteinsson, J. Arneborg, R. Streeter, and C. Keller. 2012. Cultural adaptation, compounding vulnerabilities and conjunctures in Norse Greenland. PNAS 109:3658–3663. Egede, H. 1741. Det Gamle Grønlands nye Perlustration, København, Denmark. 131 pp. GHM III. 1845. Grønlands Historiske Mindesmærker. Volume 3. Det kongelige nordiske Oldskrift-selskab, København, Denmark. 950 pp. Giesecke, K.L. 1910. Mineralogisches Reisejournal über Grönland. Meddelelser om Grønland 35:1–478. Guldager, O., S.S. Hansen, and S. Gleie. 2002. Medieval farmsteads in Greenland. The Brattahlid Region 1999– 2000. Volume 9. Danish Polar Center, Copenhagen, Denmark. 142 pp. Gulløv, H.C. 1983. Nuup kommuneani qangarnitsanik eqqaassutit inuit-kulturip nunaqarfii - Fortidsminder i Nuuk kommune - inuit-kulturens bopladser. Kalaallit Nunaata Katersugaasivia & Nationalmuseet, Nuuk, Greenland. 238 pp. Gulløv, H.C. 1997. From Middle Ages to Colonial Times: Archaeological and Ethnohistorical Studies of the Thule Culture in South West Greenland 1300-1800 AD. Meddelelser om Grønland: Man & Society bd. 23. Commission for Scientific Research in Greenland, Copenhagen, Denmark. 501 pp. Hartman, S., A.E.J. Ogilvie, J.H. Ingimundarson, A.J. Dugmore, G. Hambrecth, and T.H. McGovern. 2017. Medieval Iceland, Greenland, and the new human condition: A case study in integrated environmental humanities. Global and Planetary Change 156:123–139. Hegmon, M. 2013. The archaeology of the human experience. The SAA Archaeological Record 13:16–19. Hegmon, M., M.A. Peeples. 2018. The human experience of social transformation: Insights from comparative archaeology. PLoS One 13:1–29. Holm, G., G. Vilhelm. 1889. Beretning om Konebaads- Ekspeditionen til Grønlands Østkyst 1883–85. Meddelelser om Grønland 9:53–143. Jensen, J.A.D. 1889. Undersøgelse af Grønlands Vestkyst fra 64o til 67o N.B. Meddelelser om Grønland 8:33–122. Kapel, H. 1982. Feltundersøgelser sommeren 1982 i Godthaabsfjorden: Udgravninger i Ujaragssuit, 64V2- IV-515. Unpublished field report (No. 33-014) on file at Greenland National Museum & Archives, Nuuk, Greenland. 187 pp. Kapel, H. 1989. Rekognoscering af den indre del af Godthåbsfjorden d. 10/6-1989. Unpublished field report (no. 33–148) on file at Greenland National Museum & Archives, Nuuk, Greenland. 7 pp. Knudsen, P., C. Andreasen. 2009. Culture historical significance on areas Tasersiaq and Tarsartuup Tasersua in West Greenland & suggestions for salvage archaeology and documentation in case of damming lakes. Unpublished field report on file at Greenland National Museum & Archives, Nuuk, Greenland. 18 pp. Knuth, E. 1934. Beretning om Udflugter, foretagne af Eigil Knuth under ophold i Vesterbygden somrene 1932 og 1934. Unpublished field report on file at Antikvarisk Topografisk Arkiv, National Museum of Denmark, Copenhagen, Denmark. 3 pp. Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 31 Knuth, E. 1944. Bidrag til Vesterbygdens Topografi. Det grønlandske Selskabs Årsskrift 1944:81–124. Krogh, K.J. 1974. Kunstvanding - hemmeligheden bag Grønlandsbispens hundrede køer. Nationalmuseets Arbejdsmark 1974:71–81. Krogh, K.J. 1982. Erik den Rødes Grønland. Nationalmuseets Forlag, Herning, Denmark. 245 pp. Kuijpers, A., N. Mikkelsen, Ribeiro, S., and M.S. Seidenkrantz. 2014. Impact of Medieval fjord hydrography and climate on the Western and Eastern Settlements in Norse Greenland. Journal of the North Atlantic Special Volume 6 (In the Footsteps of Vebæk-Vatnahverfi Studies 2005–2011):1–13. Larsen, F. 2017. Rensdyrets og rensdyrjagtens betydning for befolkningen i Vestgrønland - belyst ud fra den materielle og immaterielle kulturarv. Master’s thesis. University of Greenland, Institute of Culture, Language and History. University of Greenland, Nuuk, Greenland. 110 pp. Lasher, G.E., Y. Axford. 2019. Medieval warmth confirmed at the Norse Eastern Settlement in Greenland. Geology of Greenland Survey Bulletin 47:1–4. Lea, J.M., D.W.F. Mair, F.M. Nick, and B.R. Rea. 2014. Terminus-driven retreat of a major southwest Greenland tidewater glacier during the early 19th century: Insights from glacier reconstructions and numerical modelling. Journal of Glaciology 60:333–344. Ledger, P.M., K.J. Edwards, and E.J. Schofield. 2013. Shieling activity in the Norse Eastern Settlement: Palaeoenvironment of the “Mountain Farm,” Vatnahverfi, Greenland. The Holocene 23: 810–822. Lennert, A.E. 2016. What happens when the ice melts? Belugas, contaminants, ecosystems and human communities in the complexity of global change. Marine pollution bulletin 107:7–14. Lennert, A.E. 2017. A millennium of changing environments in the Godthåbsfjord, West Greenland–bridging cultures of knowledge. Ph.D. dissertation, University of Greenland and GEUS–Geological Survey of Denmark and Greenland, Nuuk, Greenland. 224 pp. Lennert, A.E., N. Mikkelsen. 2015. Fangsthistorier oplevet og fortalt ud fra kulturlandskaberne: En rejse gennem tid og klimavariationer. Tidsskriftet Grønland 3:148–156. Madsen, C.K. 2014a. Norse pastoral farming and settlement in the Vatnahverfi Peninsula. Pp. 95–114. In Northern Worlds–landscapes, interactions and dynamics Research at the National Museum of Denmark. Proceedings of the Northern Worlds Conference, Copenhagen 28–30 November 2012. H.C. Gulløv (Ed.). National Museum of Denmark, Copenhagen, Denmark. Number of 95. 114 pp. Madsen, C.K. 2014b. Pastoral settlement, farming, and hierarchy in Norse Vatnahverfi, South Greenland. Ph.D. dissertation, University of Copenhagen, Copenhagen, Denmark. 442 pp. Madsen, C.K. 2019. Marine shielings in Medieval Norse Greenland. Arctic Anthropology 56:119–159. Mahler, D.L. 2007. Sæteren ved Argisbrekka. Fróðskapur - Faroe University Press, Tórshavn, Faroe Islands. 525 pp. McGovern, T.H. 1985. Contributions to the Paleoeconomy of Norse Greenland. Acta Archaeologica 54:73–122. McGovern, T.H. 1992. Bones, buildings, and boundaries: Palaeoeconomic approaches to Norse Greenland. In C.D. Morris and D.J. Rackham, (Eds). Norse and Later Settlement and Subsistence in the North Atlantic. Pp. 193–230. Occasional Paper Series. Department of Archaeology, University of Glasgow, Glasgow, Scotland. Number of 193. 230 pp. McGovern, T.H., T. Amorosi, S. Perdikaris, and J. Woollett. 1996. Vertebrate zooarchaeology of Sandnes V51: Economic change at a chieftain’s farm in West Greenland. Arctic Anthropology 33: 94–121. McGovern, T.H., G. Bigelow, T. Amorosi, and D. Russell. 1988. Northern islands, human error, and environmental degradation: A view of social and ecological change in the Medieval North Atlantic. Human Ecology 16:225–270. Meldgaard, M. 1986. The Greenland caribou: Zoogeography, taxonomy, and population dynamics. meddelelser om Grønland, Bioscience 20:1–88. Meldgaard, M., J.B. Nielsen, and B. Grønnow. 1983. Aasivissuit: The great summer camp. Archaeological, ethnograhical and zoo-archaeological studies of a caribou-hunting site in West Greenland. Monographs on Greenland: Man & Society 230. 96 pp. Motyka, R.J., R. Casotto, M. Truffer, K. Kjeldsen. 2017. Asynchronous behavior of outlet glaciers feeding Godthåbsfjord (Nuup Kangerlua) and the triggering of Narsap Sermia’s retreat in SW Greenland. Journal of Glaciology 63: 288–308. Møller, H.S., K.G. Jensen, and A. Kuijpers. 2006. Late- Holocene environment and climatic changes in Ameralik Fjord, southwest Greenland: Evidence from the sedimentary record. The Holocene 16: 685–695. Nelson, M.C., S.E. Ingram, A.J. Dugmore, R. Street, M.A. Peeples , T.H. McGovern, M. Hegmon, J. Arneborg, K.W. Kintigh, S. Brewingt, K.A. Spielmann, I.A. Simpson, C. Strawhacker, L.E.L. Comeau, A. Torvinen, C.K. Madsen, G. Hambrecht, and K. Smiarowskiet. 2016. Climate challenges, vulnerabilities, and food security. Proceedings of the National Academy of Sciences 113: 298–303. Nielsen, A. 1910. Breve fra Grønlændere (samlede og oversatte ved Knud Rasmussen). Atlanten Medlemsblad Bind II: 8–27. Nordenskjöld, O. 1914. Einige Züge der physischen Geographie und der Entwickelungsgeschichte Süd-Grönlands. Geographische Zeitschrift 20: 628–641. Nørlund, P., A. Roussell. 1929. Norse ruins at Gardar: The Episcopal seat of Medieval Greenland. Meddelelser om Grønland 76:1–170. Nørlund, P., M. Stenberger. 1934. Brattahlid. Meddelelser om Grønland 88:1–161. Odgaard, U. (Ed.). 2008. An Archaeological Survey in the West Greenland Inland, summer 2007, in Advance of Proposed Development of Hydroelectric Power. Report prepared for ALCOA, April 2008. Unpublished field report on file at Greenland National Museum & Archives, Nuuk. 187 pp. Ogilvie, A.E.J., J. Wollett, K. Smiarowski, and T.H. Mc- Govern. 2009. Seals and Sea Ice in Medieval Greenland. Journal of the North Atlantic 2: 60–80. Journal of the North Atlantic C.K. Madsen and A.E. Lennert 2022 No. 42 32 Pasda, C. 2014. Regional variation in Thule and Colonial caribou hunting in West Greenland. Arctic Anthropology 51:41–76. Patterson, W.P., K.A. Dietrich, C. Holmden, and J.T. Andrews. 2010. Two millennia of North Atlantic seasonality and implications for Norse colonies. PNAS 107:5306–5310. Pearce, D.M., D.W.F. Mair, B.R. Rea, J.M. Lea, J.E. Schofield, N. Kamenos, and K. Schoenrock. 2018. The glacial geomorphology of upper Godthåbsfjord (Nuup Kangerlua) in southwest Greenland. Journal of Maps 14:45–55. Porter, C., P. Morin, I. Howat; N. Myoung-Jon; B. Bates, K. Peterman, S. Keesey, M. Schlenk, J. Gardiner, K. Tomko, M. Willis, C. Kelleher, M. Cloutier, E. Husby, S. Foga, H. Nakamura, M. Platson, M. Jr. Wethington, C. Williamson, G. Bauer, J. Enos, G. Arnold, W. Kramer, P. Becker, A. Doshi, C. D’Souza, P. Cummens, F. Laurier, and M. Bojesen. 2018. ArcticDEM. Harvard Dataverse. Available online at https://www.pgc.umn. edu/data/arcticdem/. Pulsiano, P., K. Volf (Eds.). 1993. Medieval Scandinavia: An Encyclopedia. Garland, New York, NY, USA. 768 pp. Reinton, L. 1969. Til seters: Norsk seterbruk og seterstell. Det Norske Samlaget, Oslo, Norway. 146 pp. Rink, H. 1866. Eskimoiske Eventyr og Sagn oversatte efter de indfødte Fortælleres Opskrifter og Meddelelser. C.A. Reitzels Boghandel, København, Denmark. 259 pp. Rink. H. 1877 Danish Greenland: Its People and Its Products. Henry S. King & Co., London, UK. 468 pp. Rosing, J. 1958. Tanker ved et “viftested.” Tidsskriftet Grønland 10:385–392. Roussell, Aa. 1932a. Beretning fra arkitekt Aa. Roussell vedrørende nordbo-arkæologiske undersøgelser i Godthaab Distrikt i sommeren 1932. Unpublished field report on file at Antikvarisk Topografisk Arkiv, National Museum of Denmark, Copenhagen, Denmark. 7 pp. Roussel, Aa. 1932b. Kangersuneq 1932, Unpublished field report on file at Antikvarisk Topografisk Arkiv, National Museum of Denmark, Copenhagen, Denmark. 30 pp. Roussell, Aa. 1936. Sandnes and the neighboring farms. Meddelelser om Grønland 88:5–232. Roussell, Aa. 1941. Farms and churches in the Medieval Norse Settlements of Greenland. Meddelelser om Grønland 89:1–354. Schofield, E.J., J.M. Lea, D.W.F. Mair, D. Pearce, and B.R. Rea. 2016. Palaeoenvironmental investigations at two Norse farms near the margin of the Greenland Ice Sheet. Report to the Greenland National Museum and Archives on fieldwork undertaken in the Norse Western Settlement during August and September 2015. Unpublished field report on file at Greenland National Museum & Archives, Nuuk, Greenland. 17 pp. Seidenkrantz, M.S., S. Aagaard-Sørensen, H. Sulsbrück, A. Kuijpers, K.G. Jensen, and H. Kunzendorf. 2007. Hydrography and climate of the last 4400 years in a SW Greenland fjord: Implications for Labrador Sea palaeoceanography. The Holocene 17:387–401. Sveinbjarnardóttir, G. 1991. Shieling in Iceland: An Archaeological and Historical Survey. Acta Archaeologica 61:73–96. Tamstorf, M.P. 2001. Satellitbaseret vegetationskortlægning i Vestgrønland, Københavns Universitet, Copenhagen, Demnark. 50 pp. Toft, P.A., M. Appelt, and H.C. Gulløv. 2010. Nuuks ”Grå guld.” Nationalmuseets Arbejdsmark: 69–83. UNDP. 1994. Human Development Report, 1994. U.N.D. Programme, Ed. United Nations Development Programme (UNDP), New York, NY. 226 pp. Van As, D. et al. 2017. Increasing meltwater discharge from the Nuuk region of the Greenland ice sheet and implications for mass balance (1960–2012). Journal of Glaciology 60:314–322. Vebæk, C.L. 1951. Vesterbygden. August 1951. Unpublished field report (no. 17) on file at Antikvarisk Topografisk Arkiv, National Museum of Denmark, Copenhagen, Denmark. Vebæk, C.L.1991. The church topography of the Eastern Settlement and the xxcavation of the Benedictine Convent at Narsarsuaq in the Uunartoq Fjord. Meddelelser om Grønland, Man & Society 14:1–81. Vebæk, C.L. 1993. Narsaq: A Norse landnáma farm. Meddelelser om Grønland, Man & Society 18:1–47. Vésteinsson, O. 2006. Communities of dispersed settlements: Social organization at the ground level in tenth to thirteenth century Iceland. Pp. 87–113. In W. Davies, G. Halsall, and A. Reynolds (Eds.). People and Spance in the Middle Ages, 300–1300. Studies in the early Middle Ages. Brepols, Turnhout, Belgium. 366 pp. Vésteinsson, O. 2010. Parishes and communities in Norse Greenland. Journal of the North Atlantic Special Volume 2 (Norse Greenland: Selected Papers from the Hvalsey Conference 2008):138–150. Vésteinsson, O., M. Church, A.J. Dugmore, T.H, 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. Vésteinsson, O., T.H. McGovern, and C. Keller. 2002. Enduring impacts: Social and environmental aspects of Viking age settlement in Iceland and Greenland. Archaeologia Islandica 2:98–136. Vinther, B.M., P. D. Jones, A.K.R. Briffa, H. B. Clausen, Katrine Krogh Andersen, D. Dahl-Jensen, and S. J. Johnsen. 2010. Climatic signals in multiple highly resolved stable isotope records from Greenland. Quaternary Science Reviews 29:522–538. Weidick, A., O. Bennike, M. Citterio and N. Nørgaard- Pedersen. 2012. Neoglacial and historical glacier changes around Kangersuneq fjord in southern West Greenland. Geology of Greenland Survey Bulletin 17:5–65. Westergaard-Nielsen, A. A.B. Bjørnsson, M.R. Jepsen, M. Stendel, B.U. Hansen, and B. Elberling. 2015. Greenlandic sheep farming controlled by vegetation response today and at the end of the 21st Century. The Science of the Total Environment 512–513: 672–681. Øye, I. 2004 Agricultural conditions and rural societies ca. 800–1350: An introduction. Pp. 79–140. In R. Almås (Ed.). Norwegian Agricultural History. Tapir, Trondheim, Norway. 365 pp.