Introduction

There is increasing evidence to suggest that human and ecological communities do not follow uniform responses to climate-induced stress (Cumming et al. 2008). Studies of cultural and ecological change in the Arctic have uncovered extended periods of disequilibria, where biota and hunter-gatherer communities followed non-linear responses to climate volatility (Normand et al. 2013; Riede and Pedersen 2018). We here evaluate the role of social-ecological disequilibria in shaping the demise of the Medieval Norse settlements in Greenland. Using insights from Niche Construction Theory (NCT), we argue that environmental modification had a significant role in the Norse demise, not because they degraded their environments, as some have argued (i.e., Diamond 2005), but because they were culturally invested in transported “domesticated landscapes” (Terrel et al. 2003: 323) that limited their capacity to adapt to climate change.

Scandinavian Norse settlers introduced agriculture to Greenland in the late tenth century, importing an ‘agricultural niche package’ of non-native domesticated animals and plants tuned to Norwegian and Icelandic environmental conditions (Riede 2011). Early settlers cleared willow and birch scrub to extend grasslands for grazing livestock, creating synanthropic landscapes with impacts on vegetation, soil, and landscape structures (Dugmore et al. 2005). The introduction of agro-pastoralism accelerated soil erosion, but this is far more isolated than often reported and had a limited impact on home-field hay production (Golding et al. 2011).

We focus on the biocultural relationship between Norse farmers, their cultural landscapes, and the changing biophysical conditions in Medieval Greenland. Synergistic changes to climatic stability and European markets operating at regional scales had significant long-term impacts on local-scale resilience (Dugmore et al. 2012). Anthropologists and cultural geographers have explored the detailed relationships between culture and landscapes in different ways, from the cultural-semiotic relationships that form common identities to the interactions shaping social practices and values (Berkes 2017; Brace and Geoghegan 2011). Biological and physical impacts on landscapes often have significant cultural impacts—eroding social institutions, traditional practices, and iconic habitats (Adger et al. 2013). We argue that the translocation of the Norse ‘agricultural niche’ from stable environments in Norway to the different and less predictable ecologies of Greenland was responsible for social-ecological disequilibrium and the decline of Norse settlement.

Historical Ecology, Macroevolution, and Niche Construction Theory

Historical ecology—the study of long-term interacting human and natural processes manifest in landscapes—is a cross-disciplinary field connecting ‘climatic and biotic variability (such as wildlife grazing, browsing, and fire) in the context of human land use and management’ (Armstrong et al. 2017: 8; Crumley 1994). NCT raises questions of how the dispersal of domestic species across the globe has affected landscape and seascape ecologies and how this influences human capacities to adapt to environmental change (cf. Armstrong et al. 2017).

The colonisation of new environments involves a process of ‘landscape learning’ wherein initial settlers acquire ecological information about resource location and timing, limitations of the environment (i.e., carrying capacity) and sustainable social organisation (Rockman 2003). Constructing sustainable social-ecological systems can be constrained by information about long-term eco-dynamic changes and the historically contingent ecological knowledge of new settlers (Dugmore et al. 2009). The human capacity for social learning through language and symbol systems ‘vastly increases the fidelity of information transmission, making it possible to modify and fine-tune these behaviours’ (Zeder 2016: 332). In macroevolutionary theory, the cultural transmission of information is a fast-operating process and enhances adaptability to new environmental conditions by constantly updating knowledge, practice, and belief systems (Berkes 2017). This information can be transferred between generations in art, objects, myth and legend, and ritual performances to give meaning to local environments known as traditional ecological knowledge (TEK) (Riede 2012).

Niche construction (NC) can be defined as behaviours whereby organisms, or agents, modify environments in such a way that ‘selective pressures’ acting on organisms are changed (Odling-Smee et al. 2013). NC is divided into ecological, genetic, and cultural inheritance domains to explain the ontogenetic processes of human environmental modification (‘ecological inheritance’) and gene-culture co-evolution by modified natural selection (‘genetic inheritance’). Knowledge of niche-modifying behaviour is stored in cultural-spiritual practices (‘cultural inheritance’) to enhance the survival of future generations (Smith 2011). Storage of historical information in cultural traditions is essential, especially in agricultural economies where information transfer is vital to making predictions about timing cultivation, harvest, slaughter, and so on. It is also essential for managing domestic resources in periods of resource scarcity when impacts that have not previously been experienced require adaptive strategies equal to the stress upon the agricultural system (Kennett and Marwan 2015).

NC research is primarily concerned with understanding why humans chose to ‘abandon more mobile [hunter-gatherer] strategies’ in favour of ‘well-defined resource catchment territories’ (Zeder 2015a, b: 3196). The transition from mobile hunter-gatherer communities to sedentary agriculture marks a stark transition in human societies’ cultural investment in fixed resources, infrastructure, and socio-political institutions (Zeder 2009). NC frameworks explain how, in various independent geographical contexts (cf. Smith 2011), resource-rich environments, communal living, and longevity collectively secured a predictable resource system predicated on selective plant and animal domestication (Zeder 2015a, b). But landscape modification and knowledge transmission cannot ensure adaptation to the local environmental conditions. Volatile climatic regimes and environmental degradation can undermine information transmission from past experience (Kennett and Marwan 2015) and in new environmental contexts (Rowley-Conwy and Layton 2011).

Disequilibrium and Niche Construction

Research has recorded significant (often multi-decade to century-long) time delays between climate change and the response of vegetation communities (Svenning and Sandel 2013; Normand et al. 2013) that has challenged conventional models in community ecology that assume a uniform response to temperature change (Blonder et al. 2017; Svenning et al. 2015). Human communities can exhibit similar ‘adaptive lags’ when ‘the discrepancy between past and current environments […] produces a mismatch between behaviour and the environment’ (Laland and Brown 2011: 97).

As Riede (2009: 3) notes, ‘human tool-use and landscape modification are responsible for the […] hand-in-glove fit of human societies to their environment,’ yet even a highly flexible culture can ‘experience limits to its tolerance space, outside which it is unable to behave adaptively’ (Laland and Brown 2006: 98). This is because ecological changes operating outside collective experience of variation can limit the capacity of specialised strategies to yield predictable returns (Zeder 2016). For many complex societies, climate uncertainty has created disequilibria between specialised agricultural systems and the ecologies that are suitable to support them (Kennett and Marwan 2015). The pre-Hispanic Puebloan communities of the US Southwest illustrate how agricultural instability can result from the interplay between technological specialisation and unanticipated high-magnitude events. In the Hohokam, farmers overcame semi-arid environments by constructing a large-scale irrigation system supplied by the Lower Salt River (Howard 1993) that operated for over a millennium. However, the robustness of the system made it vulnerable to abnormal flow (Ingram 2008) and in the late fourteenth century, high flow devastated the system, causing regional depopulation (Nelson et al. 2012).

Agricultural expansion into new environments often signals the success of a particular mode of production, but can also lead to maladaptation. The Norse expansion across the North Atlantic islands in the Viking Age caused wide deforestation, vegetation clearance, soil erosion, and sometimes the extirpation of local species (Dugmore et al. 2005). The Norse showed a lasting resilience to feedbacks associated with environmental degradation. They were less resilient, however, to the impacts of climate variability and politico-economic changes in Europe (Dugmore et al. 2013).

Adapting the Norse Economy in the Sub-Arctic

The settlement of Greenland (c. 985 AD) took place within a wider context of raiding, trading, and exploration known as the Viking Age (c. 793–1066 AD). Recent evidence suggests that the Norse settlers were seeking trade goods to replace ivory and hide lost following the extirpation of walrus in Iceland. In Greenland, walrus, fur bearing animals, and arctic exotica (i.e., narwhal) were abundant (Frei et al. 2015). Suitable lands for farming were located on the southwest coast in two main regions called the Eastern and Western Settlements (Fig. 1). The influence of oceanic and atmospheric circulation and seasonal sea-ice delivery result in average temperatures in Greenland that are ~8 degrees cooler than continental Europe at the same latitude (Dugmore et al. 2005). The inner-fjords of the Eastern and Western settlements support a largely continental climate, receiving less annual precipitation and lower average wind speeds than the more oceanic climate of the outer-fjords (Arneborg 2005).

Fig. 1
figure 1

Showing the Norse Eastern and Western Settlement in Greenland and the Norse ivory and high-status goods export market between the late-10th and early-fifteenth century. Ivory was a high-value, low-bulk commodity used for lay and ecclesiastical adornments throughout Europe. In Greenland the largest concentrations of walrus were to be found in the Northern Hunting Grounds located at Disko Bay. Ivory would have then been refined in the settlements for export to Bergen on the royal vessel. The nature of export to Europe via Bergen represents a loose tie to wider European markets. Icelandic high-bulk, low-value goods conversely maintained diverse market ties to the German Hansa towns, England, Norway and Denmark

Settlement patterns were characteristically dispersed and connected by networks of seasonally occupied shielings, bearing witness to the extensive utilisation of scattered grazing resources (Vésteinsson 2009). Zooarchaeological evidence indicates a landnám assemblage of domestic animals (cattle, caprines (sheep and goats), pigs, horses, dogs, and cats) characteristic of manor farms in contemporaneous Norway indicating the cultural significance of animal husbandry to Norse identity and status (Perdikaris and McGovern 2008). Subsistence strategies were modified to utilise local wild resources as the Norse switched from fishing and waterfowl supplements characteristic of Iceland and the Faroe Islands to hunt reindeer and seasonally abundant migratory seals (Smiarowski et al. 2017). Evidence of drive lines and hunting dogs to catch caribou (Rangifer tarandus) grazing near settlement areas suggests that Norse settlers imported terrestrial hunting techniques developed in Scandinavia (McGovern 1985). In addition, evidence of seabird and seal bones on inland farms indicates a network of exchange and a pooling of labour resources to maximise hunting efficiency. Techniques used in previous environments, such as communal boat drives, netting, and clubbing could have been used to harvest harp (Phoca groenlandicus) and hooded (Cystophora cristata) seals in the spring months (Dugmore et al. 2007a).

Early domestic assemblages were adjusted to suit long winters and constraints on fodder production (McGovern 1991). The quality and extent of pastureland and the availability of household labour would have determined the level of farmed food surplus possible (Vésteinsson et al. 2002). On small- and medium-sized farms, pigs disappeared and the proportion of cattle to caprine species declined (McGovern et al. 2014; Smiarowski et al. 2017). High-status farms appear to retain larger numbers of cattle (a principal symbol of social status). As survey data from Greenland’s Eastern settlement suggest, many low-status farms made radical transitions to hardier goat dominant livestock assemblages (Madsen 2014), however, low ratios of cattle were also maintained (McGovern 1994). Large livestock assemblages, the capacity to grow, harvest, and store large fodder reserves made elite farms more resilient to climate variability and extended periods of agricultural dearth. In similar periods, medium, and small farms would have relied on either supplementing production with wild resources or acquiring surplus production from large farms (Dugmore et al. 2012).

The Seasonal Round: the Economic Year in Greenland

The seasonal round of farming was organised with a notable concentration of activities and peak labour demand in the summer months (Fig. 2). In the winter months livestock were stalled in byres and grazed on stored fodder reserves harvested in the autumn months, and non-agricultural activities such as cloth production and the refinement of walrus tusks were carried out (Østergård 2004; Frei et al. 2015). In late spring, livestock were moved to the outfields and upland pastures to graze on the new growth; dairy production took place in specialised shielings (Ledger et al. 2013; Madsen 2014).

Fig. 2
figure 2

The seasonal round in Norse Greenland’s Western Settlement area. Grey shaded areas represent areas of high labour activity (after McGovern 1980)

Core farming activities were balanced against additional hunting and foraging tasks across the wider land- and seascapes of southwest Greenland and long expeditions to the northern walrus hunting grounds (McGovern 1985; Enghoff 2003) and communal hunting drives to harvest migrating seals (McGovern et al. 1988). Zooarchaeological analyses have recorded an increased dietary reliance on migratory harp seals in both the Eastern and Western settlements from AD 1300 (Ogilvie et al. 2009; Smiarowski et al. 2017). Other communal hunting and foraging tasks would have taken place on a seasonal basis such as harvesting nesting birds in the spring and caribou hunting in the late autumn (Dugmore et al. 2007a).

Labour demand would have peaked between June and August, as hunters made the long and dangerous voyage to the Northern Hunting Grounds (Nordurseta) (McGovern 1985). Walrus ivory, hides, and other arctic exotica were high-value commodities on the European market, and were traded in return for essential resources such as iron (Gulløv 2008). Grœnlandie vetus chorographia, a seventeenth century text, describes the voyage to the Nordurseta as a journey that took six-oared boats between ‘15’ and ‘27’ days to reach the Disko Bay region (Frei et al. 2015). Heavy labour demands are likely to have coordinated a gendered division of tasks. In Norse societies, textile production was organised solely by women on the farmstead (Hayeur-Smith 2014). Runestone inscriptions and various written accounts from continental Scandinavia reflect the central role of women in farming tasks when men were away (Roesdahl 2016: 62).

Population estimates based on accounts in the Vinland Saga of the voyage to Greenland, ethnographic data from Iceland, and sustainable population models suggest a small peak around AD 1300. Assuming an initial population of 300–500 individuals, and an exponential growth rate of 0.62%, population is most likely to have peaked at ~1400–2200 by AD 1300 (Lynnerup 2014). A population of this size divided between the Eastern and Western settlements would have made labour management difficult in the summer months. Although absolute population numbers remain uncertain, there is good evidence for exceptional labour demands. Skeletal remains indicate both hard labour and a regular use of teeth as tools (Scott et al. 1992; Scott and Jolie 2008). Despite a protein-rich diet and limited exposure to the diseases endemic in contemporary Europe, life was likely to have been brutally hard for most of the Norse settlers even in periods of comparative prosperity.

Exogenous Change

Archaeological evidence suggests that, from the mid thirteenth century the Eastern settlement became increasingly vulnerable to higher-scale changes across the North Atlantic. Increased climate volatility, changes to market structures and trading systems in continental Europe, and the possibility of hostility with the Thule Inuit culture have gained significant attention in multidisciplinary studies over the last 40–50 years (McGovern1994; Dugmore et al. 2007a; Nelson et al. 2016).

Climate Change

Despite the relative mildness of the Medieval Climate Anomaly (MCA) even early settlers would have experienced harsh winters and summer droughts in Greenland (Dugmore et al. 2007b). Nevertheless, the Norse proved highly adaptive, establishing and expanding settlement range until the twelfth century (Madsen 2014). From the mid-thirteenth and fourteenth centuries multi-decade cooling increased the length and harshness of winters, sea-level rise submerged coastal home-fields, and storminess and summer sea-ice increased (Kuijpers et al. 2014; McGovern et al. 2014).

Recent climate models suggest that North Atlantic cooling was triggered by a series of globally significant volcanic eruptions in the thirteenth century, culminating in the eruption of the Mt. Samalas Volcano in AD 1257–1258; each event reducing solar insolation and increasing the onset of summer sea-ice in the Denmark Strait (Miller et al. 2012; Lavigne et al. 2013). The onset of the generally cooler and more variable Little Ice Age (LIA) stressed farming economies; homefields became less productive and longer winters increased livestock mortality, causing a decline in dairy production (McGovern et al. 2014). Results from isotopic studies of human bone collagen indicate a transition from a largely terrestrial diet supplemented by ~40% marine resources in the eleventh century to a diet topping ~80% marine input in the final phases of settlement (Arneborg et al. 2012) that coincides with climate cooling commencing in the mid-thirteenth century (Fig. 3). Hydrographic proxies also record cooling in the fjord systems of southwest Greenland (Jensen et al. 2004; Kuijpers et al. 2014) and increased summer sea-ice prevalence, which would have affected boat travel, seal migration routes, summer grazing patterns, and fodder availability across the Eastern and Western settlements (Ogilvie et al. 2009; McGovern et al. 2014).

Fig. 3
figure 3

Isotope records (above) collected from human skeletal remains in Norse Greenland indicating a transition from high terrestrial low marine to low terrestrial high marine with radiocarbon error bars and median lines (Arneborg et al. 2012). Multiproxy temperature records (below) for the Arctic-Sub-arctic region (60–90°N) between 900 AD and 1500 AD (McKay and Kaufman 2014). Thick line indicates 5 year moving average to show the multicentury cooling trend commencing from the mid-thirteenth century. There is a negative correlation between temperature anomaly and isotopic signature indicating increasing adoption of marine proteins as the climate cooled

In the Eastern Settlement consumption of non-migratory harbour seals (Phoca vitulina) declined significantly after AD 1300 likely due to sea-ice, displaced harbour seal populations, and increasing dependence on migratory seal species (Ogilvie et al. 2009). Ecological changes such as these would have stressed the application of TEK and increased the ‘degree of adjustment’ required to utilise coastal resource spaces effectively (Berkes 2017).

European Economic Transition

In continental Scandinavia, the impact of the Black Death (AD 1346–48) in Norway and the Kalmar Union of 1397 AD effectively shifted the center of power from Norway to Denmark (Epstein 2009). Danish royal and mercantile interests were focused upon the Baltic and North Seas, with increasing trade with Russia opening-up sources of walrus products in the Barents Sea by 1400 AD. In Europe by the fourteenth century, power and trade had accumulated around a coalition of northern German towns, shifting Scandinavian societies to the periphery of the European world-system (Nedkvitne 2014).

These transitions on the European continent had contrasting impacts on North Atlantic trade. Unlike Iceland, Greenland lacked the combination of a laboring population, grazing sheep herds, and access to fish stocks to generate surpluses of wool and dried fish for export (Madsen 2014). An absence of suitable ship-building timber in Greenland limited the possibility of local production of large ocean-going boats, making the Norse solely dependent on visiting boats from Europe. With waning Scandinavian royal interests in its distant Atlantic outpost, stiff competition from other suppliers of arctic products (i.e., Sami and Karelian trades), and lack of dried fish or woolen goods to attract commercial interest from the Hanseatic League (whose cog ships were ill-suited to the stormy North Atlantic), the Greenlanders were unable to maintain economic ties with Bergen leading to increased isolation from the early fifteenth century (Dugmore et al. 2007a).

Cultural Contact

Little is known of the possible interactions between either the Dorset Paleo-Eskimo or the later Thule Inuit and Greenland Norse, and hostilities mentioned in the Vinland Sagas and Ivar Bardarson’s accounts remain ambiguous and uncorroborated in the archaeological record (Appelt and Gulløv 2009). After 1300 AD, a written record from the Icelandic Annals suggests at least intermittent hostility with small groups of Norse men and boys reported killed or carried off, perhaps during voyages northwards (Gulløv 2008). However, there is limited evidence in the archaeological record to corroborate such accounts. A recent large-scale genetic study of the modern Greenlandic population found no evidence for any admixture with Norse or Dorset populations and argues for a single, substantial migration event (Moltke et al. 2015). Hostility and estrangement of Inuit cultures is likely to have reduced the possibility of inter-cultural transmission of ecological knowledge.

Discussion

Cultural-Ecological Disequilibria

Preindustrial agrarian states were dependent on predictable climatic conditions for planning utilisation of local-scale resources (Kennett and Marwan 2015) and on accumulated information about the opportunities and limitations of the environment and the capacity of social systems to utilise resources efficiently (Riede 2011, 2012). In farming communities, resources diversification, storage, yield-boosting technologies and social-economic institutions can be used to increase the predictability of local resources (Zeder 2016). But when communities have insufficient knowledge to organise sustainable responses to changing social-ecological feedbacks, the system can become vulnerable to food shortages (Nelson et al. 2016). In many preindustrial societies climate volatility or unanticipated climatic events (i.e., floods or drought) undermined centralised political authority leading to decentralisation and diverse adaptive pathways (Kennett and Marwan 2015; Middleton 2017). This was often associated with path dependent institutions, structures, or behaviours that resulted from historical antecedents of cultural adaptation and environmental change.

In Norse Greenland, the transition to less predictable climate conditions resulted in resource diversification (Smiarowski et al. 2017), the use of storage and irrigation structures (Buckland et al. 2009), and manuring and soil augmentation (Adderley and Simpson 2006) to maintain population stability. Acute climate uncertainty in the fourteenth and fifteenth centuries did not, however, result in reduced political complexity—as in many cases of societal transformation (Tainter 1988; Butzer 2012). In fact, evidence of church building in the fourteenth century suggests a strengthening of politico-religious institutions (Arneborg 2003). Adaptive strategies were largely uniform across the settlement (with the exception of differences in household-scale diet) and aimed at sustaining farming (Dugmore et al. 2013). As we argue, it was a combination of historical-cultural factors and changes operating across different spatial and temporal scales that caused social-ecological disequilibria in Norse Greenland.

Antecedent Landscapes of Norse Greenland – Niche-Construction in Scandinavia

The arrival of agriculture and domestic animals to Scandinavia was part of a complex process of cultural interactions and extension (Rowley-Conwy 2013). From as early as the Late Bronze Age, farm institutions were formed, establishing a spatial continuity for organising the household-barn and farm structure (Myhre 2004). This created conformity to the spatial organisation of the farm and seasonal practices of agriculture in much of Norway, and institutionalised and embedded ideas and practices of niche modification (Øye 2013). By the Medieval period, farms were comprised of a central nucleated cluster of buildings, field systems, and shielings (Fig. 4). Seasonal transhumance was organised between winter byres and close pasturelands and more distant summer pastures, as the cultural niche was expanded into broader the landscape. Differences in climate across Norway, however, dictated the amount of cultivation, stocking capacity, and wild resource use that was possible (Øye 2004).

Fig. 4
figure 4

Conceptual model of the Norwegian farm and legally regulated areas (after Øye 2013)

The transformation of Norway’s natural environment into a network of cultural landscapes and the evolution of integrated agricultural systems, technologies, and institutions by the Middle Ages (Øye 2004) left a “legacy of modified selection pressures” operating over millennia (Laland et al. 1999: 10242). The ‘ecological inheritance’ left by landscape modification processes, which included vegetation clearance and burning, soil augmentation and manuring, irrigation, and managed grazing (Mehl and Hjelle 2017) produced anthropogenic soils and ‘cultural steppe’ landscapes for descendent populations (Dugmore et al. 2005, 2006). This increased predictable returns in the farming niche (Zeder 2016). Technologies and practical knowledge had evolved to effectively manage crop and hay fodder production, seasonal transhumance, and specialised hunting strategies (Perdikaris 1999). Axes, ploughs, tillage and harrowing tools, scythes, and storage technologies were used to intensify local-scale production and store resources through periods low productivity (Øye 2009). Knowledge of local resources and agricultural skills would have been transmitted and updated through social learning, stories, and beliefs leaving an ecological and cultural legacy in Norway that was resilient to the long-term constraints of the environment (Øye 2004; Riede 2012).

The accumulation of knowledge informs expectations of resource acquisition long before they are encountered and so colonisation could have presented an acute challenge to knowledge and practices established in homelands as they mismatched with the environments of new settlements (Rowley-Conwy and Layton 2011). As a result, evolved cultural-ecological (or biocultural) behaviours need to be considered as structures of adaptive knowledge in the North Atlantic.

Contextualising the Norse Cultural Niche in the North Atlantic

Climate-adaptation can be understood through two macroevolutionary processes that allow cultures to become adapted to their environments. The first—landscape learning—explains the exploratory processes whereby individuals and groups gather and share information about the locations, limitations (resource fluctuations and interannual weather extremes), and the organisation of resources within a local catchment (Rockman 2003). The second process—cultural transmission—is closely interwoven with landscape learning. Information is accumulated through observation of local resource systems and recombined into the existing adaptive package (Boyd et al. 2011). Culture plays a critical role in accumulating, transmitting and, at times, limiting human adaptive capacities in new environments.

Accumulating Information – Landscape Learning in Greenland

It has been proposed that early ‘hunters and scouts’ would have explored and assessed the resources in Greenland before the initial settlement in 985 AD (Dugmore et al. 2007a). The abundance of valuable walrus colonies and marine resources within reach of the settlement areas would have required a scoping phase to assess whether a farming niche could be established (Dugmore et al. 2007a; Frei et al. 2015). Suitable landscapes for animal husbandry practices were located in southwest Greenland. Climate data indicate that settlement took place in the comparatively mild and stable conditions of the MCA (Dugmore et al. 2007b; Nelson et al. 2016). In the early phase of settlement, stable climatic conditions would have increased the predictability of selective returns from hay cultivation and the cycles of livestock transhumance. Knowledge of climate variability and its effect on local ecosystems would have been accumulated over the initial settlement period. This allowed existing economic activities to be adjusted to the limitations of the environment. The stable climate helped early settlers predict ecological feedbacks that influence the timing and location of plant growth, marine, and terrestrial mammal migration and domestic livestock reproduction (McGovern 1980).

In agricultural economies familiarity with feedbacks are manifest as economic cues and seasonal structures of labour organisation (Kennett and Marwan 2015). The identification of local resources allowed environmental information to be reincorporated into existing organisational structures—the Norse seasonal round (McGovern 1980; Fig. 2)—by adjusting existing practices, but critically in ways that conformed to the identity of the existing Norse ‘cultural niche.’ The limited availability of standing wood for the construction of longhouse structures, for example, meant wood beam supports were replaced with perpendicular internal walls to support the roof (Høegsberg 2014). Further research, including comprehensive survey work in the Vatnahverfi region of the Eastern Settlement, has recorded a general absence of infield dykes, which are common to contemporaneous Norwegian and Icelandic farms, to separate grazing livestock from cultivated home-fields (Madsen 2014). This absence, combined with evidence of intensive use of shielings, suggests a less intensive type of farming that utilised broader grazing resources (Madsen 2014). Lower primary production in Greenland would have made this a necessary strategy to reduce the chance of erosion on outfield pastureland. Zooarchaeological evidence indicates a gradual adjustment of livestock ratios to favour species that were suited to the longer, harsher winter conditions. This corroborates with isotopic data, indicating an increased reliance on marine mammals, as well as a continuing reliance on caribou hunting (Arneborg et al. 2012; Smiarowski et al. 2017).

By accumulating information about the location and limitations of the environment in southwest Greenland, the Norse were able to suitably adjust their existing agricultural niche to the new resource regime. However, accumulated information in this early phase of settlement merely retrofitted the existing Norse ‘cultural niche’ to conform to mild and predictable conditions – precluding larger-scale adjustments that buffer against environmental stress. In other words, the Norse agricultural system was able to operate under the assumption that resources would remain located in specific places and times within the economic year (McGovern 1994). Because farming societies are information-intensive uncertain conditions can reduce farmers’ capacity to “evaluate the costs and benefits of one strategy or another” (Kennett and Marwan 2015: 2–3). Buffering strategies used in past environments, for example, are unlikely to have operated effectively as the climate deteriorated.

Home-field management was fundamental to temperate European pastoral economies, and irrigation and manuring strategies were widely employed to increase productivity—often to counter declining home-field yields. These imported management strategies could sustain fodder production through cold winters and summer droughts that were characteristic of MCA conditions in Greenland (Dugmore et al. 2007a, 2007b). Evidence of irrigation and manuring is widespread across the Eastern (i.e., Gardar) and Western settlements (Buckland et al. 2009), but the productiveness of this strategy would have declined as multi-decadal cooling enhanced positive feedbacks reducing home-field yields (Golding et al. 2015). In milder climates manuring can boost soil fertility (and yield) but in Greenland an increase of organic material within the soils enhances water retention leading to a build-up of winter ground-ice and a delayed spring thaw, shortening the effective growing season (Adderley and Simpson 2006). Prolonged winters and decreased fodder production would have increased livestock mortality, reinforcing dependence on hunting and foraging strategies to support subsistence (Dugmore et al. 2012). Unanticipated feedbacks such as these would have required the Norse to respond by stretching the application of short-term buffering strategies with the effect of reducing the size of their ‘safe operating space.’

Cultural Disequilibria – Norse Agriculture and Thule Hunting

To understand the level of disequilibrium between the Norse ‘cultural niche’ and the environment, it is necessary to consider how the Norse adaptive package was suited to long-term ecological change in sub-Arctic Greenland, and this can be achieved through a comparison of the Norse and Thule Inuit cultural niches. Thule society, direct ancestors to modern-day Inuit, evolved a highly adaptive ‘cultural niche’ comprising habitat-specific knowledge, complex technologies, and behaviours required to survive in the extreme cold (Gulløv 2008). Archaeological and ethnographic sources record highly varied material cultures and practices comprising different bows, arrows, and harpoon technologies, styles of clothing, and designs of kayak (Park 2005; Mason 2007). This technological assemblage had evolved over millennial timescales, as knowledge of resource-use was accumulated and recombined (Boyd et al. 2011).

The Norse ‘cultural niche’ differentiated between a socialised ‘inside’ of fixed dwellings and modified cultural landscapes and a wild, chaotic ‘outside’ (Arneborg 1997; Gulløv 2008). Modified landscapes conformed to European agricultural economies: local intensification, economic division and cyclical organisation of field systems, domesticated animals and human labour (McGovern 2000; Golding et al. 2015). Technologies were highly evolved to support the European farming niche. Axes, scythes, sickles, spades, soapstone storage vessels, and cooking pots are found from the beginning to the end of settlement (Arneborg 2000; Enghoff 2003). Clothing manufacture also corresponded with Medieval European styles and production techniques, with minor differences to design, and would have utilised wool from domestic sheep and goats (Østergård 2004; Hayeur-Smith 2014). These technologies and practices starkly contrasted with the Thule Inuit ‘cultural niche’ (Table 1). Beliefs, practices, and technologies were integrated with the Arctic environment in the Inuit lifeworld—connecting humans and animals, even in the act of hunting, killing, and consuming animal flesh for food and skins for clothing (Ingold 2000). In Inuit cosmologies, the exchange of vital forces between humans and animals is reciprocal, requiring a deep knowledge of animal behaviours and biology (Brody 2001; Berkes 2017). Technologies were also highly evolved: caribou skins were harvested to construct highly insulated and pliable shelters, parkas, stockings, and boots (Meaks and Cartwright 2005).

Table 1 Norse and Inuit Cultural Niches

Norse and Thule Inuit cultural niches supported highly different behaviours. The Norse constructed a cultural niche to support local intensification, while Inuit lifeways observed intimate details of extensive Arctic habitats. Archaeological and genetic evidence, however, suggests contact between the Thule Inuit and the Norse Greenlanders (Gulløv 2008; Raghavan et al. 2014). By the fourteenth century, Thule culture had advanced to the outer-fjords of the Norse settlement areas, most likely for summer trading (Golding et al. 2011). Evidence for exchange of Norse (i.e., chessman – Neo-Eskimo Ruin Island phase) and Thule Inuit (i.e., Neo-Eskimo bird, hunting implement – Norse Eastern Settlement) artefacts has been recovered from high-Arctic Thule and sub-Arctic Norse contexts (Gulløv 2008). But evidence of cultural contact and exchange does not mean ‘social learning’ was possible between the Norse and Inuit cultures. As Gulløv (2008) explains, chess never existed in prehistoric Arctic cultures, and harpoons were never adopted in Norse society; they may have conveyed a (new) symbolic value but use-value is improbable without a compatible knowledge of the other culture.

The Norse and Inuit had vastly different social arrangements and settings for ‘social learning’ practices (Arneborg 1997). This is because the “acquisition and retention of beliefs, values, role expectations and skills” were situated in different cultural-ecological contexts (Kendal 2011: 241). Because the Norse failed to abandon agricultural practices, it is likely that social learning was situated on the farm (McGovern 1991). Cultural practices were essential to the transmission of information vital for enhancing returns from selected livestock. Miniatures of farm animals, swords, and boats are found on Norse sites across Scandinavia and the North Atlantic (Arneborg 2000; Gardela 2012). These can be interpreted as ‘qualifier toys’ that deliver a common sense of identity and operate as a primer for agricultural practices in later life (Riede et al. 2018). In late childhood, children would have undertaken shepherding duties and assisted with other tasks on the farm (Dugmore et al. 2007a). Intergenerational cultural transmission would have delivered accumulated information about local environments and how the Norse ‘cultural niche’ fits within it (Riede 2012). But because learning was situated within the Norse ‘cultural niche’ it is unlikely that innovative information about the wider environment beyond cultural landscapes was easy to acquire and assimilate with existing behaviours.

Social learning in the Thule ‘cultural niche’ is understood through a well-preserved material culture and contiguous ethnographical records (Park 2005; Riede et al. 2018). Ethnographic research in the early twentieth century records children learning complex practices of seal and caribou hunting through play using miniature bows and arrows, harpoons, and kayaks (Park 2005) and carved miniature animal figurines that draw attention to animal behaviours and physiological features. This allowed the accumulation of knowledge about Arctic environments, animal behaviour, and technology craft and use—and technologies could be fine-tuned (“ratcheted-up”) to environmental variations, increasing efficacy (Riede et al. 2018). The level of complexity and technological know-how observed in the Inuit ‘cultural niche’ is the result of the long-term accumulation of adaptive elements to construct an adaptive package for the dynamic Arctic environments of Greenland and Northern Canada.

Critically, while there is evidence of Norse-Inuit artefact exchange, there no evidence of inter-cultural transmission of adaptive behaviours or learning practices. The Norse adopted incremental changes within their existing ‘cultural niche’ (Table 2) but did not adopt new behaviours. Changes to Norse material culture include increasing the thread-count of clothing (Hayeur-Smith 2014), increasing dependence on hunting marine mammals (Arneborg et al. 2012), and adjustments to farmsteads, field systems, and shieling use (Madsen 2014). This strongly suggests a ‘conformist transmission’ of cultural information (Laland and Brown 2011), maintaining the resilience of the Norse agricultural niche at the expense of efficient resource-use in the long-term. The conformity of behaviour is unsurprising as cultural evolution takes place over long periods of time (Bentley and O’Brien 2015) and is reinforced by the institutional setting—which, for the Norse, included local governing (lawmakers and chieftains) and religious (parish churches) institutions (Vésteinsson 2009).

Table 2 Incremental adjustments made to Norse 'Cultural Niche' - an overview of the North Atlantic

Hierarchy and Conformist Adaptations

Collective behaviours are likely to conform to successful or trusted institutions (Boyd et al. 2011; Thompson 2013). The Grágás and Jónsbók law books record legal codes on resource access, governance, and disputes in Iceland, but also likely reflect modes of governance and organisation in Norse Greenland (Sandvik and Sigurdsson 2005). These codes would have regulated the exchange of goods, resource access rights, and the organisation of agricultural land (Madsen 2014). Settlement data also indicate a hierarchical structure. The transition from small annex church structures of the eleventh century to centralised parish churches between the twelfth and early-fourteenth centuries, suggests a two-tier division of society between a church-lord class controlling rich pasturelands with surplus production capacity and a homogenous lower-class struggling to maintain farm production (Vésteinsson 2009).

Differences in private resource access would have reinforced the dependence of lower class farms on larger farms that controlled the distribution of resources (Dugmore et al. 2012). As Thompson (2013) explains, hierarchical institutions are likely to assimilate the impacts of environmental change to stabilise safe limits. By controlling the redistribution of surplus resources to smaller farms, elite farms could maintain a uniform dependence on standard modes of production, limiting capacity to diversify adaptive strategies.

Adaptive Strategies Created New Vulnerabilities – Changing Settlement Patterns

Relocation has long been considered an adaptive response to risk but can also produce unforeseen risks and trade-offs (Halstead and O'Shea 1989). Changing environmental conditions from the mid-thirteenth century are thought to have created significant vulnerabilities among farms established in the late-tenth and eleventh century settlement phase (McGovern 1994). Survey data collected from Greenland’s Eastern settlement suggests a reorganisation of settlement from distinctly dispersed patterns across the inner and outer fjords to more concentrated settlement on the inner fjords from the thirteenth century (Vésteinsson 2009; Madsen 2014). Farm abandonment on outer fjord and upland environments likely reflects changing environmental gradients as pastureland became less productive for harvesting winter fodder (Madsen 2014). Recent palynological studies suggest decreased grazing pressure and abandonment of Saqqaa (c. AD 1350–1400), Lake Vatnahverfi (c. AD 1290–1400), and Saqqaa Tasia (c. AD 1220–1380) of the Vatnahverfi region between the late-fourteenth and early-fifteenth centuries (Ledger et al. 2014a, 2014b). The abandonment of farms across the Vatnahverfi region support Christian Madsen’s (2014) hypothesis that settlement became increasingly concentrated on farms on the inner and middle fjords, as distant and less productive upland and coastal regions were gradually abandoned (Fig. 5).

Fig. 5
figure 5

Abandonment of outer-fjord and upland areas of Norse Eastern Settlement (cf. Madsen 2014)

This refocus of settlements across the inner and middle fjords may have been an attempt to reduce distance between farms sharing labour for communal hunting and to re-settle in areas with more productive infields (Vésteinsson et al. 2002; Madsen 2014). Mid-fjord church sites at Hvalsey and Narsarsuaq appear to thrive from the fourteenth century at a time when farms in the Vatnahverfi area were abandoned (Madsen 2014), suggesting that resettlement concentration strategies were serving their purpose. But while this would have reduced inter-farm isolation it would also have entailed sustaining investment in fixed infrastructure on the inner fjord, which would have enhanced vulnerability to wild food shortages by increasing the distance to and from migratory seal populations beyond the outer fjord (Ogilvie et al. 2009). A shorter active season, less familiar conditions for hunting, and increased sea-ice—making boat transportation difficult and dangerous—would have compounded risks to food shortage.

Mixed Resource Use Reduced Energetic Returns

NCT takes us beyond simplified theories of human behaviour and human-environment interaction to explain the role culture plays in modifying selective environments and the organisms within it (O’Brien and Laland 2012). Humans are seen to have the agency to consciously enhance their environments by selecting and modifying plants and animals to increase their biological and physical fitness (Zeder 2016). As discussed earlier, culture holds an important role accumulating and transmitting this information for future generations in a complex adaptive repertoire (Boyd et al. 2011). Culture operates at the group scale as a system of shared values, beliefs, and symbolic practices. As explained above, for the in Norse and Thule Inuit (Table 1), their systems of beliefs and practices were central to the regulation of their respective ‘cultural niches.’ Culture, and more specifically identity, had a significant role structuring knowledge of the local environment, its resources, and how to make use of them. The Norse identity has been elucidated, in part, from the material assemblages recovered from farms and church burials in Greenland’s Eastern and Western Settlements. Game pieces, toy boats, religious iconography, wooden tools, woollen clothing and metal jewellery reveal economic networks with the Scandinavian homelands and the North Atlantic islands and explain how the Norse viewed themselves as European farmers (Arneborg 2000).

The environment of southwest Greenland was, for the Norse, identified for modification into cultural landscapes. Reference to land is common to Norse written culture as a symbol of status and power (Jesch 2015). The Kings Mirror and Erik the Red’s Saga both draw attention to rich pastureland in Greenland’s settlement areas. Subsistence would have been organised according to seasonal activities, such as the growing/grazing season, harvest, and overwinter livestock stalling (McGovern 1980). This structured seasonal round was iteratively adjusted to different ecologies across the North Atlantic (Dugmore et al. 2005, 2007a). Levi-Strauss (1966) referred to cultural and symbolic structures as ‘totemic operators’—translating cognitive cultural structures into structured interactions with the environment. Cultural structures can be understood as practical ‘schemata’ that filter experience and legitimise action (Dugmore et al. 2012). Taking this point further, Adger et al. (2009) recognise that cultural identity provides not only legitimacy to act, but also directs the goals of adaptation. The socialised agricultural niche can be understood as an important component sustaining the Norse identity in Greenland (cf. Adger et al. 2013). Adaptive strategies would have been planned with the goal of sustaining agricultural production, which in turn limited their capacity to enhance energetic returns from hunting and foraging.

Simplified explanatory frameworks for human behaviour, such as Optimum Foraging Theory (OFT; cf. Codding and Bird 2015), are thus insufficient to explain the goal-oriented behaviour of the Norse farmers. In temperate climates, such as Scandinavia, the majority of activities operated within the farming core, but in Greenland, dietary records indicate a shifting dependence from core to peripheral locations, such as the outer fjord and uplands (Fig. 6). Intense scheduling in the summer months would have stressed the capacity of a small population to hunt resources efficiently from farms on the inner reaches of the fjord. As discussed, the settlement became increasingly concentrated on the inner fjord (Fig. 5). Increasing the distance between settlements and resource spaces would have caused energetic returns from hunting to decline, as hunters were required to travel greater distances to locate, catch, refine and return with seals from outer-fjord environments. (Zeder 2016: 329) summarised the potential trade-offs of investing in cultural niche construction:

“Investment in the management of a [domestic] species depends on the returns and demands of other potential resources, technological capacities, distribution of humans and resources across the landscape, as well as the physiological capacity to utilise these resources.”

In the last 40 years, archaeological studies across the North Atlantic islands have uncovered significant evidence to suggest the Norse were flexible to the impacts of climate change on their subsistence. This was also the case in Greenland. However, here strategies remained oriented towards supporting an unstable farming niche.

Fig. 6
figure 6

A theoretical model of distributed labour intensity across space over months of the year. Combining McGovern’s (1980) seasonal round of subsistence tasks with Øye’s (2013) standardised model of Norse farming, this model distinguishes the spatial intensity of labour between the seasons. This demonstrates the associative intensity of activities between the relatively inactive winter months and the laborious summer months. The months from November to April would have been largely inactive months other than small-scale seal hunting on the outer-fjord and attendance to overwintering livestock in byres and the infield. May to October were highly active involving process of transhumance whereby livestock were transported from byre and infield to outfield and mountain pastures as the summer progressed. Caribou hunting would also have taken place on across mountain pastures between settlements on the inner-fjord. Seal hunting took place primarily between the months of May and July. From July to late-August hunters would have taken to the northern hunting grounds in the Disko Bay region. This intense scheduling throughout the summer months is likely to have increasingly stressed the small Norse population throughout settlement. As the pie charts (below) show loss of a single month, possibly as a result of extended winters would have drastically disrupted the seasonal round and intensified task orientation

Conjuncture: Synergisms between Economy and Climate

Mixed agriculture and hunting made the Norse subsistence system flexible to interannual variations in fodder production, but also limited their capacity to adopt more specialised strategies (Dugmore et al. 2012). Because the Norse were able to make incremental changes to their ‘cultural niche’ without compromising their identity, they can be said to have exhibited resilience. However, as Halstead and O’Shea (1989: 1) stipulated, “Culture endows man with exceptional flexibility in coping with his surroundings … [but] should not mask the fact that an effective strategy must match, in both capacity and scale, the variability with which it is to cope.” As the PAGES 2k climate model has shown, from the late-thirteenth century the North Atlantic entered a protracted cold period (PAGES 2k Consortium 2013). This had the effect of reducing homefield production and increasing livestock mortality on Norse farms (McGovern et al. 2014). Dietary proxies show the Norse adjusting to declining homefield yields by substituting farming shortfalls with increased marine mammal hunting (Arneborg et al. 2012; Smiarowski et al. 2017).

GISP2 data indicate the onset of more volatile and unpredictable climatic conditions from the early-fifteenth century (Dugmore et al. 2007b). The onset of climatic conditions that deviated from average cycles of variations made predictions about timing of farm activities and spring seal hunting increasingly difficult. The onset of acute climate variability also coincided with Greenland’s isolation from European markets (Dugmore et al. 2012), which underwent rapid transformation in the thirteenth and fourteenth centuries. The marginalisation of Norwegian political-economic power and the concentration of the stockfish and cloth trade in English and Northern German Hansa port towns resulted in the stagnation and decline of the North Atlantic ivory trade (Frei et al. 2015; Barrett 2018). The absence of alternative commodities for export to Europe isolated the Norse Greenlanders from trading partners, and the iron, clothing, and status imports they had long supplied (Dugmore et al. 2013). The synergisms between the impacts of climate uncertainty on subsistence and economic isolation from Europe would have undermined attempts to sustain agricultural economies as management of home-field production, livestock herding, and resources vital to technologies for tool making and manufacture of clothing became all but impossible.

Conclusions

The Norse adapted to the environments of southwest Greenland through a combination of landscape learning and culturally transmitted knowledge. During the settlement period information about resource location, timing, and sustainable limits would have been accumulated to refine livestock assemblages and organise seal and caribou hunting (Rockman 2009). The Norse proved flexible to changing climates—refining livestock ratios and increasing the proportion of their diet from non-migratory seals (Arneborg et al. 2012; Fig. 3). This strategy was sustained until the mid-fifteenth century, when settlement came to an end.

Agricultural instability and climate uncertainty is a likely cause of social-ecological disequilibrium. The Norse made incremental changes to their diet, tools, buildings, clothing, and field systems without discernible impacts on their identity as Scandinavian-European farmers. These changes were sufficient to sustain subsistence in the mild conditions of the MCA (c. AD 950–1250). The onset of cooler climatic conditions from the late-thirteenth century would have increased stress on homefield production, and subsequently stretched the application of hunting strategies. The construction of a farming niche would have distinguished between a socialised inside and a hostile outside (Arneborg 1997). Culturally transmitted information, accumulated in temperate Norway and other North Atlantic islands, was ill-equipped to respond to unpredictable interannual climate variation. Economic behaviours were dependent on suitable conditions for home-field production and animal rearing. Norse behaviours—associated with cultural and ecological reliance on sedentary animal husbandry—thus reached their adaptive limits in the late-Medieval climates of Southwest Greenland. The Norse agricultural systems were thus mismatched with the scale of climate variation.