Climate in Nordic historical research – a research review and future perspectives

ABSTRACT This article assesses the development and current state of climate history research conducted in the five Nordic countries and Estonia. The possible societal impacts of past climatic changes already interested a handful of Nordic historians in the early twentieth century, but the lack of data on past climate fluctuations constrained scholarship in this field until recently. The data availability has increased fundamentally over the past decades due to the advances of palaeoclimatology. However, these advances have created new challenges, related to the ability to utilize data from the natural sciences in historical research as well as acquiring a basic knowledge on climatology. In many European countries, climate history has established itself as a strong academic subfield and consequently has created approaches as to how to overcome some main pitfalls, like climate determinism, related to the early works in the field. These epistemological advances are just beginning to gain a foothold in Nordic historical research. Thus, the article concludes with ten recommendations to improve future research in Nordic climate history.


Introduction
Ongoing climate change, and related environmental changes, will affect -and have already affected -human societies.A profound understanding of climate-society interactions over long time scales provides an essential perspective to contemporary challenges. 1 Historical research can offer multiple examples of complex pathways and patterns between changes in climate and human societies.These examples, in turn, may help us to anticipate the plurality of the challenges of anthropogenic climate change in the imminent future.The academic subfield of climate history (or historical climatology) explores these past interactions.This field started to develop, in its more modern form, from the 1960s onwards, when works by Emmanuel Le Roy Ladurie and Hubert H. Lamb,  among others, demonstrated what written sources might reveal about past climatic variations and how they might have influenced human well-being and livelihoods in the past. 2 Historians such as Christian Pfister and Astrid E.J. Ogilvie contributed significantly to the field by emphasizing the importance of analysing historical sources for climate reconstruction. 3dvances in palaeoclimatic research in the Nordic countries over the past few decades have produced quantitative climate reconstructions, which also provide potential material for historians to examine the role of climate in influencing the human past.Furthermore, due to their northern location and dependency on climate-sensitive crop cultivation in historical times, the Nordic countries provide a relevant study location for climate history research.However, interest in this research field has been relatively modest.Encyclopaedias and handbooks on the history of the Nordic countries barely include environmental or climatic aspects -and even fewer discussions about the possible influence of climate variability on human history.An exception is the latest encyclopaedia of Swedish history, to which Fredrik Charpentier Ljungqvist contributed short articles on climate history. 4The Danish, Norwegian, Swedish, and Finnish multivolume handbooks on agricultural histories address rather moderately the possible impacts of climate variability and change on crop yields. 5Despite the comparably good documentary source material covering the past ca 500 years, new high-quality palaeoclimate data from natural archives, and the promising environmental context, climate history has remained a comparatively marginal research area in the Nordic countries in the past two decades.During this time the field has experienced a large expansion and has drawn considerable interest in most of the rest of Europe. 6The marginal state of Nordic climate history research is evident in the recent handbook in the field, The Palgrave Handbook of Climate History (2018), in which the Nordic countries and Estonia are treated very briefly.Hitherto, from the Nordic countries, only Norwegian research on climate history has been summarized in a state-of-the-art review.This publication also includes, to a limited extent, an overview of the research in other Nordic countries. 7n the present review, we assess studies of relevance for climate history, from the first pioneering efforts in the early decades of the twentieth century to the most recent works, for all five Nordic countries, as well as Estonia to some degree.We place particular emphasis on studies written in Danish, Estonian, Finnish, Icelandic, Norwegian, and Swedish, rather than on works in English, as they are arguably less accessible for the international research community.Our hope is that this review will act both as an orientation to the state of research in the Nordic countries and serve as an overview for scholars elsewhere to allow them to better include results from the Nordic countries in a larger transnational context.We include studies which (1) are fully or partly conducted by historians and/or (2) rely entirely or mainly on documentary sources.This limits the temporal extent of this literature review to the period back to the thirteenth century.Consequently, the increasing number of high-quality archaeological, palaeoecological, and geographical studies exploring older times are not assessed here.The rationale for this limitation is lack of space, despite our full acknowledgement that these studies can be highly relevant for the understanding of past climate-human links.We concentrate on research exploring the medieval and early modern periods, yet we include some studies focusing on the nineteenth and early twentieth centuries as well.We refrained from addressing nineteenth to twenty-first century climate-society interactions in the Nordic countries in more detail because of the very different nature of these relationships with increasing agrarian modernization, industrialization, urbanization, and globalization.
Nevertheless, these more recent linkages are very important matters and would, no doubt, deserve a review article of their own.

What is climate history?
Climate history 8 is traditionally considered to have three main objectives: (1) reconstructing past climate variability, (2) studying climate-human relationships, and (3) exploring discourses and written or pictorial representations of climate and weather (Figure 1).As the third objective has barely been studied for the Nordic countries, this review focuses on the first two objectives.The first objective covers the reconstruction of past climate and weather variability from written sources as well as the study of occurrence of extreme events and climate-related natural disasters.The second objective, the study of climatehuman relationships, includes both sides of this relationship: the agency of climate in human history and the agency of humans in adapting to changing climate and coping with extreme weather events.In particular, the dynamics of societal and economic vulnerability to climate variability are at the centre of the investigation. 9ne of the major problems related to climate history concerns climate determinism and reductionism.The concept of climate determinism originates from the early twentieth century, when certain scholars aimed to justify, for example, colonialism or theories on race by climatic circumstances. 10Consequently, considering climate as a 'historical determinant' has received strong objections from historians and especially so since the Second World War. 11Recently, the field has been criticized for practicing climate 'reductionism' -removing climatic consequences from their social and environmental contexts. 12However, the majority of more recent studies presenting such an approach are not conducted by academic historians but either by physical scientists or by nonacademic writers of popular history. 13Commonly, these studies explore the links between climate and human activities by seeking correlations among large datasets, erroneously equating correlation with causation, or focus solely on single dramatic events and ignore the available evidence concerning the complexity of human responses to climate. 14istorians specializing in climate history are well aware of such pitfalls of climate determinism and reductionism and have developed approaches to avoid monocausal interpretations.Some of these approaches are presented later in this article in a Nordic context.
Climate history is an interdisciplinary field.Depending on the objectives of a particular study, conducting research might require knowledge of physical, social, and/or cultural sciences.Information from palaeoclimate reconstructions helps historians to date and quantify past climatic variability and change.This helps us to place studied historical events and phenomena in a climatic context.Yet, the climatic periodization do not always match with a historical one and the used terminology can be misleading.For example, the so-called Medieval Warm Period (MWP) ended in Finland before the historical Middle Ages in this region had begun.Furthermore, these reconstructions reveal that although the so-called Little Ice Age (LIA) was, on average, colder than, for example the eleventh or twentieth centuries, this period also included multi-decadal periods of moderately warm conditions.Above all, the LIA was not a glacial period in geophysical sense.Nevertheless, most of the Nordic reconstructions indicate that the coldest phase of the LIA was between the late sixteenth and the end of seventeenth century (Figure 2).

Basics of historical climate reconstruction
The measuring of different weather parameters, like temperature, precipitation, and air pressure, with meteorological instruments began in the Nordic countries during the first half of the eighteenth century.Collection of these early observations from archival material, and analysing and homogenizing them, constitutes its own subfield of climate history: historical meteorology. 15Another subfield focuses on reconstructing past climate variability from written descriptive observations and proxy records.Proxies are indirect climate data, materials which have been influenced by the climate of the time when they were written or laid down, formed, or grown.These proxies can be found either in the 'archives of nature', like in tree rings or sediment and ice layers, or in the 'archives of societies', which are the written sources. 16hronicles, annals, diaries, and newspapers, among other sources, can contain written descriptions of climate and weather.These include extreme and unseasonal events records, like floods and frosts, as well as descriptions of day-to-day weather.These qualitative materials can be commonly transformed into climate reconstructions with the so-called index method.This method gives the descriptions an ordinal scale value, e.g. from -3, indicating extremely cold, to +3 indicating extremely warm. 17A recent example of an application of this method is a study by Dag Retsö and Lotta Leijonhufvud  (2021), who reconstructed drought in Sweden back into the Middle Ages. 18Besides the index approach, a calibration-verification procedure is another method used in historical climate reconstructions.This latter method commonly employs quantitative materials.Early scholars, private individuals, and state authorities kept records of different events, like the onset dates of harvest or when the harbours were free from ice.As these so-called phenological dates are commonly sensitive to climate variations, it allows them to be transformed into meteorological units, like degrees Celsius.This method requires an overlapping period of meteorological measurement data, preferably divided into two sub-periods, one period that is used to calculate the reconstruction model (calibration) and another that is used as an independent check of the model (verification). 19After the model is validated with this procedure, it can be applied over the whole historical timeseries.For example, Per Øyvind Nordli employed this method when reconstructing central Norway spring-summer temperatures from the first day of grain harvest, compiled from farmers' diaries. 20The index method identifies primarily year-to-year climate variability, whereas the calibration-verification method can also detect slow and gradual, so-called low-frequency, changes.

Historiography of Nordic climate reconstructions
Nordic research has a long tradition of employing ice-phenological observations in historical climatology.Various written sources offer information on the freezing and ice break-up dates of harbours, rivers, and lakes.Much early work consisted of source compilations of these events, conducted by researchers from the physical sciences.Danish meteorologist Christian J.H. Speerschneider presented in 1915 a catalogue of winter ice conditions around Denmark and in the Baltic Sea back into medieval times, concluding that the average winter severity had not changed much over time. 21Similarly, Icelandic geologist and geographer Þorvaldur Thoroddsen produced in 1916 a compilation of many types of documentary sources concerning weather and sea ice variability in Iceland and concluded that the climate has not changed much since early medieval times. 22This interpretation, however, was hotly contested. 23A number of early Nordic compilations were closely analysed by W.T. Bell and Astrid E.J. Ogilvie in 1978, who were able to show that they contained considerable amount unreliable material. 24The work of Thoroddsen is better than many other similar compilations, but still contains much that is erroneous or at least questionable.Nevertheless, for its time it must be deemed an outstanding pioneering work. 25Later Thoroddsen's compilation was the sole source for two other works.One of these is a book by Lauge Koch entitled The East  Greenland Ice (1945).This includes a reconstruction of sea-ice conditions back to the ninth century (although with many gaps prior to the seventeenth century) from documentary data.From this Koch concluded that a warm period occurred ca 930-1030, a somewhat colder climate ca 1200-1400 and cold conditions ca 1600-1900. 26These conclusions cannot be deemed reliable as they were based on the mix of reliable and unreliable data to be found in Thoroddsen's work and did not have the advantage of many data that were unknown to Thoroddsen. 27In the late 1690s Icelandic climatologist and meteorologist Páll Bergþórsson wrote an article outlining a sea-ice and temperature reconstruction from c. 870 CE (the traditional date for the first settlement of Iceland) that was also based solely on information taken from Thoroddsen. 28Thus, it unfortunately also contains much unreliable information.What was novel concerning Bergþórsson's reconstruction, however, was his conclusion that there is a strong correlation between temperature and the presence of sea ice which has also been confirmed by later researchers. 29onsiderable progress in Nordic research on ice-observations and historical documentary based climate reconstructions were made in the late twentieth and early twenty-first centuries.Having unravelled the textual relationships of the works described above, and added many novel sources unknown to them, Astrid E.J. Ogilvie presented a much more source-critical and updated documentary-based sea-ice reconstructions for Iceland, covering 1500-1800, along with winter/spring thermal indices for Iceland at decadal resolution. 30She also published discontinuous documentary-based climate estimates for Iceland, extending back to the Viking Age. 31 Anders Tarand gathered spring ice breakup dates from the port of Tallinn back to the fourteenth century in 1992, although the time-series are continuous only from 1600 onwards. 32A year later, Juha Kajander compiled an ice break-up date series from the Torne river back to 1693. 33Both publications demonstrated the unique climatic signal embedded within these historical time-series and how, for example, the colder conditions of the Little Ice Age are evident in later ice break-up dates during the early modern period.Other long Nordic ice break-up series include Lake Mälaren in Sweden since 1712 and the Finnish river Aura since 1749.Both series noted later eighteenth and nineteenth century break-up dates compared to the twentieth century dates. 34ther reconstructions have attempted, following the pioneering efforts of Speerschneider, to reconstruct the sea-ice extent rather than its duration.Gerhard Koslowski and Rüdiger Glaser reconstructed the areal ice volume over the period 1701-1993 in the Western Baltic Sea from various documentary sources, finding, in opposition to Speerschneider, considerable multi-decadal variability in winter ice extent.They later extended the sea-ice reconstruction back to 1501, finding even larger multidecadal variability in areal ice volume. 35ince the compilation of these long ice break-up and ice extent series, these data have been used in several quantitative climate reconstructions employing calibration-verification methods.Tarand and Nordli reconstructed the Tallinn winter temperature from the ice break-up dates, and Svetlana Jevrejeva reconstructed the severity of winter seasons in the northern Baltic Sea from documentary data for the ice break-up date at the port of Riga since 1529. 36Both studies found considerable variability across all time-scales in addition to a long-term trend towards milder winters.Dag Retsö assessed the climatological information on winter weather during the early sixteenth century in Swedish correspondence letters, finding a mixture of winter severity. 37Later, Dag Retsö and Johan Söderberg reconstructed winter severity in medieval Sweden from various types of documentary evidence, including ice break-up dates. 38esides direct observations of freezing and break-up, various administrative accounts, especially those documenting different harbour activities, contain indirect information on these dates.Leijonhufvud and others employed these extensive materials for their Stockholm January-April temperature reconstruction extending back to 1502 (Figure 2).The reconstruction identified that the coldest conditions occurred during the late sixteenth and early seventeenth centuries and in the early nineteenth century. 39Beside icephenological observations, Nordic researchers started to implement harvest-related material in historical climate reconstructions in the 2000s.Tarand and Nordli combined the beginning of rye harvest dates with the river Palmse ice break-up dates in Estonia to reconstruct April-July temperature variability. 40Jari Holopainen and others extended the possible sources of phenological proxy data in Nordic reconstructions further by including observations of the flowering dates of various herbs, shrubs and trees in their southwest Finland spring temperature reconstruction. 41Nordli reconstructed spring and summer temperature variability in Trøndelag of Norway back to 1701, and for Vestlandet back to 1734, from peasant diaries.When comparing these Norwegian reconstructions with early instrumental temperature measurement series, he found very close correlations in the eighteenth and nineteenth centuries between the temperature and time of both the sowing and harvest. 42

Norwegian historiography
Within the context of the Nordic countries, Norwegian historians can be considered as the pioneers in the field of climate history.As early as 1913, Norwegian historian Edvard Bull published a short article 'Klima og historie' ['Climate and history'], where he argued that the consideration of past climate variability offered an additional dimension to a materialistic understanding of history.Bull identified several research problems in the study of climate-history relationships, with emphasis on medieval agrarian history in Norway. 43He also anticipated the potential offered by dendroclimatology -a science still in its very infancy -as a key source of annual past climate information to utilize in historical scholarship.In the same year Otto Pettersson triggered a debate that marked Norwegian climate history research for a long time by proposing that the climate in northwestern Europe became more irregular during the fourteenth century, rather than colder and wetter, and increasingly so towards the north.He argued that Norway and Iceland became more severely affected than Denmark and Sweden due to a larger climate sensitivity owing to the more marginal agricultural conditions in the former countries.This, in Pettersson's opinion, could explain why Norway and Iceland were affected more strongly and for a longer time by farm desertion and population decline during the fourteenth and fifteenth centuries, today commonly referred to as the Late Medieval Crisis. 44he scholarly debate during the 1920s, which nevertheless did not result in actual empirical studies on the topic, determined the viewpoints and lines of argument concerning the role played by climate for the Late Medieval Crisis in Norway, which have persisted into the twenty-first century.Both natural scientists and historians were active in this debate.Norwegian geologist and geographer Werner Werenskiold acknowledged the existence of late medieval climatic cooling but expressed scepticism that it played a major role in the crisis. 45Conversely, meteorologist Anton Røstad found it plausible that the fourteenth century crisis was initiated by cooling, which caused poorer grain yields and decreased pasture productivity, and which was later amplified with the arrival of the Black Death. 46Agrarian historian Sigvald Hasund questioned the evidence for a causal relationship between cooling and the crisis, despite acknowledging the presence of climate-related crisis symptoms elsewhere in Europe at the time.He expressed a view, which was firmly entrenched in later scholarship, that the demographic and economic crisis in Norway was entirely a result of the Black Death.Hasund's main argument was that the temperature decrease was very small and, as such, incapable of explaining the magnitude of the crisis. 47Røstad and Bull instead argued that the Black Death alone hardly could suffice to explain the extraordinary magnitude and the long duration of the crisis in Norway, finding it very plausible that climatic cooling was a contributing factor, considering Norway's northern marginal location for agriculture. 48owever, the lack of palaeoclimate data at this time precluded testing these arguments with empirical studies.In this regard, historian Oscar Albert Johansen explicitly argued against invoking climate as an explanatory factor on the grounds that available climate data from medieval times were inadequate.Future progress within the natural sciences might change this, he recognized, but until then historians had better to avoid invoking climatic change. 49o summarize the debate in Norway during the 1920s, three principal objections were raised against climate being a significant contributing factor to the Late Medieval Crisis.First of all, it was acknowledged that the lack of quantitative climate data hindered the critical investigation of the hypothesis at the time.Second, the late medieval cooling was nonetheless considered to be too marginal to play a significant role for agriculture and pastoral farming.Third was the hesitation to seek explanations for societal change beyond human agency.These views were further reinforced by Halvdan Koht, as well as by a number of further works by Sigvald Hasund, which emphasized the plague as the sole cause of the crisis. 50Nevertheless, some later historians, such as Andreas Holmsen and Johan Schreiner, maintained that the role of climatic changes was one of the most pressing issues in Norwegian medieval history, and they advocated that advances within the natural sciences, especially in dendroclimatology, would allow for resolving the dispute at some point.
The possible role of climatic change for the Late Medieval Crisis in Norway was extensively revisited in the Scandinavian Research Project on Deserted Farms and Villages. 51However, the results were as inconclusive as before, with a tendency to downplay the possible role of climatic change in favour of emphasizing the role of the Black Death and subsequent recurrent plague outbreaks.The lack of reliable palaeoclimate data at that time for Norway in particular, and Scandinavia in general, was again stressed.Jørn Sandnes and Andreas Salvesen concluded, when synthesizing the main project results for Norway, that no new findings had been made with regard to the possible role played by climate. 52n a more recent piece of Norwegian historical scholarship, Salvesen argued that climate could have been a contributing factor, especially acting as a catalyst, to societal change, whereas Kåre Lunden concluded that long-term climatic changes during the past millennium were too small to have had an influence on agriculture or pastoralism in most farm locations in Norway. 53Audun Dybdahl concluded in several publications that climatic change was unlikely to have played any major role in demographic development or settlement structure over the long run, while at the same time recognizing that historians have underestimated the importance of adverse climate conditions for harvest failures, food shortages, famines, and high mortality epidemic events.Dybdahl emphasized the spatio-temporal variations in societal climate sensitivity.Furthermore, he stressed the role of clusters of consecutive years of adverse climate conditions for agriculture that contributed directly or indirectly to mortality peaks in pre-modern Norway. 54Moreover, by employing climate sensitive tree-ring material, he found a good correspondence between narrow tree-ring growths and documentary reports of poor harvests and high mortality in epidemics -although his methods are more qualitative than quantitative in their nature. 55Dybdahl concluded that consecutive late springs and/ or wet and cold summers were the major challenge for agriculture in Norway, frequently resulting in widespread epidemics among an undernourished population. 56

Icelandic historiography
An early twentieth century debate between Icelandic geologist and geographer Þorvaldur Thoroddsen and Norwegian historian Edvard Bull encapsulates some of the key aspects of contemporary debates on climate history: how reliable is the climatic information that written sources provide, what was the amplitude of past climatic changes, and were these changes of an amplitude big enough to influence human history? 57Thoroddsen took frequent fodder failures and cattle losses mentioned in written sources as proof that the periods of Icelandic settlement and the early Middle Ages were not significantly warmer than the centuries thereafter. 58Bull, in contrast, maintained that the medieval period must have been considerably warmer than later centuries because the sources mention grain cultivation being practised across the island until the fourteenth century. 59Both scholars based their argumentation mainly on the saga literature.Today, however, we have much more detailed knowledge of past climate variability on Iceland from natural proxy data in combination with written sources.Climate history researchers in Iceland, being a small scholarly community, have utilized interdisciplinary approaches by integrating archaeological, natural, and documentary sources for much longer than the other Nordic countries. 60lthough in popular media the colonization and end of crop cultivation in Iceland are commonly brought up as anecdotal evidence of medieval warmth and the cooling LIA, respectively, scholars have demonstrated that in reality the causality between climate conditions and Icelandic society was more complex. 61Astrid E.J. Ogilvie, in particular, has been a leading scholar in Icelandic climate history research since the 1990s. 62It has been demonstrated that the presence of sea ice close to the coasts of Iceland had an adverse effect on pasture productivity and hay yields, and that sea ice had a generally adverse effect on cod fisheries. 63Lower ocean water temperatures, typically associated with more sea ice, around Iceland contributed to reducing cod fish stocks, thereby causing economic hardships.Furthermore, the presence of thick sea ice at times prevented Icelandic fishermen from reaching the cod banks off the coast. 64The prevailing view has been that climate variability had a great effect on human livelihood in Iceland, with cold conditions inducing hardships, along with other environmental hazards like the severe erosion on the island, volcanic eruptions and earthquakes. 65However, at least one scholar, economic historian Gísli Gunnarsson, has questioned both the amplitude and the importance of past climate variability, and has instead stressed socio-economic factors for hardships on Iceland. 66However, his publications lacked any quantitative assessment of the possible climatic influences -or the lack of thereof -among the historical Icelandic population.

Danish historiography
The influence of climate and weather has been addressed only to a very small extent in Danish historical research.Gunnar Olsen provided an overview of climatology and its relevance for agricultural history in 1939, but offered no insights into the possible impact of climate variability on Danish history. 67The first scholar to actually invoke climate as an explanatory factor in Danish history was Axel Steenberg in 1951, who argued that the Late Medieval Crisis in Denmark had already begun by ca 1300.He based his argument on pollen analysis and archaeological evidence of deserted farmsteads, which he linked to an increased wetness. 68Nevertheless, this research was later largely overlooked and not followed up.A short article from 1978 by Svend Gissel notes the limited attention drawn to climate, and knowledge thereof, among medieval historians in Denmark. 69he first longer treatment of climate by a Danish historian, since that of Olsen in 1939, was a review article by Flemming Mikkelsen in 1983. 70He maintained that the lack of interest in climate variability among historians was related to the erroneous presumption that the climate has been more or less stable for centuries -if not for millennia.Furthermore, he showed that climate had not fit into the theoretical and methodological frameworks of Danish historians.Finally, Mikkelsen argued that increasing interdisciplinary collaboration was necessary for research in climate history.
In more recent times, Nils Hybel has been perhaps the only Danish historian who has dealt more extensively with climate history.He first addressed climate in 1997 when he argued, on the basis of information from few documentary sources, that the weathertriggered Great Famine (1315-1322) presumably hit Denmark as it did elsewhere in northwestern Europe. 71Hybel subsequently explored the possible impacts of climate on medieval famines in Denmark and concluded that climatic factors likely had more severe societal consequences in the high than in the late medieval times.Furthermore, he claimed that the timing of the Medieval Warm Period varied spatially across Europe and was rather modest in comparison to the abrupt twentieth-century warming.Consequently, he proposed that the medieval warmth likely played a minor role in medieval history but argues instead that short-term climate extremes were more important triggers of past events. 72

Swedish historiography
The Swedish ethnologist and agrarian historian Åke Campbell maintained that assessing the relationship between climate and agricultural production belonged to the natural sciences and not to the humanities or social sciences. 73This remained, more or less, the prevailing view for a long time.However, economic historian Gustaf Utterström made an exception in 1954 in the Swedish scholarship, when he linked changing climatic conditions to harvest variations and demographic fluctuations in eighteenth-century Sweden.The next year, in 1955, Utterström published his internationally famous, and still frequently cited, article 'Climatic fluctuations and population problems in early modern history', which can be considered as a manifesto for invoking climate variability in historical research. 74His key point was that the coldest decades, and especially the coldest years, appeared to correspond to most of the large-scale harvest failures.In particular, Utterström emphasized how cold springs and summers triggered agricultural crises, which were in turn capable of inflicting significant economic and demographic consequences.Furthermore, he suggested that a cold climate might have contributed to the crisis of the seventeenth century.Nevertheless, he simultaneously pointed out the importance of socio-political factors in explaining the past population fluctuations. 75owever, presumably because of the lack of palaeoclimate data during his time, Utterström never followed up his hypotheses with empirical studies.Overall, the inclusion of climatic aspects in late twentieth-century Swedish historical research is rather marginal.For example, Erik Husberg mentioned the possible role of colder, and also wetter, climate conditions in the decreasing production of wax during the last three decades of the sixteenth century. 76In his agrarian history of Sweden, Janken Myrdal more generally consider the impact of climate change and variability. 77he late 1990s saw the re-emergence of climate in historical research, especially within the field of agrarian history.Lennart Andersson Palm found that the climatic signal in eighteenth-and nineteenth-century southwestern Swedish yield ratio data was mixed but, importantly, the farmers were shown to have adjusted the timing of agricultural work with changes in temperature and precipitation. 78Sven Lilja investigated the associations between climate, harvest and demographic crises in the early modern Swedish Realm in three articles published between 2006 and 2012. 79He concluded that unfavourable climatic conditions, causing poor harvests, were a recurrent problem in early modern Sweden.These problems, in turn, occasionally triggered mortality crises.Furthermore, Lilja anticipated that multi-annual periods of cold climate anomalies and related harvest losses could be considered one reason why the Swedish Realm had such difficulties in upholding its status as a great power. 80However, these studies lacked statistical assessments of the degree to which climate variability actually explained harvest fluctuations.
The first large-scale quantitative climate historical assessment in Sweden was published by Rodney Edvinsson and co-authors in 2009, in which they assessed the influence of monthly and seasonal temperature and precipitation variability on the output and quality of grain harvest in 1724-1955.They found cold summers dictating harvest success only in the northern parts of Sweden, whereas winter-spring temperatures and summer precipitation variability were more decisive for grain yields in the main crop cultivation regions of present-day Sweden. 81

Finnish and Estonian historiography
The Finnish studies on climate-human interactions have a strong focus on the links between weather, climate, and food crises during the early modern period.This connection has been widely recognized in Finnish agricultural history since the latter half of the nineteenth century. 82The climatic conditions in Finland were understood as the main limiting factors for agriculture, but the weather events causing crop failures were considered as occurring haphazardly.The possible connection between the frequency of weather events causing the crop failures and climatic variability was not addressed until the 1970s.Instead, prior to that, climate was essentially seen as rather stable throughout time.
Eino Jutikkala was in 1976 the first Finnish historian to criticize his predecessors for ignoring long-term climate fluctuations.Furthermore, he highlighted the importance of historians employing climate data from the natural sciences to better understand past climate fluctuations, such as the 'Little Ice Age' (LIA).However, although he recognized the link between cooler temperatures and declining harvest output in the late medieval and early modern times in the Nordic countries, he emphasized that past population fluctuations were not direct consequences of climate-driven harvest failures, but the mortality crises were always mediated by the spread of hunger-related infectious diseases. 83These views were recurrent also in his later publications and have strongly dictated the perceptions of the Finnish scholarship on the role of climate in human history. 84lthough Jutikkala brought the question of climate into the Finnish discussion, the pioneering quantitative analyses within the field were done by Matleena Tornberg, who explored the possible impacts of the LIA on grain yields in early modern south-western Finland. 85Also Susanne Lindgrén contributed to the early development of the field by examining, for example, the relationship between climate and warfare-related sea-ice crossings over the Baltic Sea and the weather conditions contributing to the 1695-1697 famine. 86Other early works on Finnish climate history include articles by Roger Kvist, who explored the link between the cooling temperatures and the declining seal hunt in Ostrobothnia, and Jorma Keränen, who linked the LIA with the increased frequency of crop failures in the seventeenth and eighteenth centuries Finland. 87Since then, the LIA has widely been connected to early modern Finnish famines, especially those in the seventeenth century and in 1867. 88Timo Myllyntaus further developed this perception by arguing that the cooling climatic conditions, in addition to poor soils and primitive cultivation methods, handicapped early modern agricultural production, which explains the increasing frequency of the seventeenth-and eighteenth-century famines in Finland. 89Most recently, Heli Huhtamaa and others have quantified the relationships between weather and crop yields in nineteenth-century Finland, studied long termagricultural adaptation to changing climate, and further conceptualized understanding the role of the LIA within historical research in general. 90Finnish crop yields were found to be sensitive especially to spring and summer temperature variability, which dictated the length, onset, and thermal conditions of the growing season, although the relationship weakened towards southernmost parts of the country. 91ven though the great majority of Finnish historians associate the LIA's colder climatic conditions and the related shortening of the growing season with the early modern famines, one contrary opinion exists.Economic historian Miikka Voutilainen argued in his monograph on famines that 'including the Little Ice Age in historical research is a kind of a fifth-wheel, irrelevant part of the cause-effect relationship, [just] a rhetorical backstory'. 92owever, Huhtamaa challenged this argument by reasoning that the mean temperatures, especially in Finland, were indisputably colder during the LIA than in the centuries before or after.The growing season began later and was colder and shorter than during warmer climatic phases.Consequently, the agricultural production was more vulnerable to disturbances, such as societal distress (like raids and hording of staple crops) or weather events (like summer frosts).Thus, she argued that the prevailing climatic conditions, such as the average conditions during the LIA, are not irrelevant.Instead, climate gives a context to the studied phenomena, as to years of food shortage in this case. 93imilarly to historians in Finland, Estonian historians have mainly focused on the links between climate, agriculture, and famines.Priit Raudkivi, in particular, has explored the climate-and weather-driven subsistence crises in Livonia and Estonia, from the 1310s famine to the nineteenth-century crises. 94In his publications, Raudkivi specifically emphasizes the importance of studying socio-environmental vulnerability within the field of climate history.In their thorough book on Estonian climate, Tarand and others also included a rigorous weather catalogue spanning from the eleventh to the end of the nineteenth century.Although focusing mostly on past meteorological and climatological events collected from written sources, the catalogue also contains records of weatherrelated human affairs, such as occurrences of crop failures, famines, and disease epidemics. 95esides the history of subsistence crises, Finnish and Estonian historians have barely addressed the agency of climate in their research.Instead, meteorologists, dendroclimatologists and geographers have been prominent in the field.For example, Reijo Solantie has published widely on the role of climate with respect to the settlement history, livelihoods and population in historical Finland and Jari Holopainen and others have explored the impacts of climate variability on grain yields and prices in early modern and nineteenth century Finland and Sweden. 96Furthermore, Holopainen and others have proposed the possible impact of the LIA cooling climate as an explanation for why so many stone churches were left unfinished in late medieval Finland. 97In recent years, Finnish historians and natural scientists have started to co-operate, and for example Huhtamaa and dendroclimatologist Samuli Helama have utilized tree-ring material to study long-term agricultural adaptation to climate. 98

Inter-Nordic comparison
Nordic climate history research has two main emphases (1) reconstructing past climate variability from written sources and ( 2) investigating past climate-agriculture relationships and their societal implications.The former research is mainly conducted by physical scientists (with the exception of some Swedish-speaking historians), and the latter by mainly historians.At the moment, only a handful of Nordic and Estonian scholars conduct research on climate history. 99Even fewer historians provide university-level teaching in the field.Because of the lack of teaching of climate history only a small number of master's and doctoral theses focusing on the topic have been written.Consequently, this hampers the emergence of a new generation of researchers in climate history and, in the long term, progress in the field.
The main debate within the early documentary-based reconstruction research was centred on whether the climate in the past had distinctive colder or warmer phases.Along with the re-emergence of historical climatology in the 1990s, new reconstructions showed colder climatic phases overlapping with the late medieval and early modern period in the Nordic countries, in agreement with the results from palaeoclimatological research utilizing tree-ring measurements and other natural proxy data. 100On the other hand, these document-based reconstructions have not provided much evidence for the so-called Medieval Warm Period due to the scarcity of pre-fifteenth century documentary sources for the Nordic countries.Consequently, reconstructions derived from natural proxy data remain the only material for assessing climate variability throughout the Middle Ages and earlier periods in the Nordic countries.
Although agricultural production has been a connecting theme in Nordic historical research about climate-society interactions, Danish, Icelandic, and Norwegian research has primarily assessed the possible contribution of climatic cooling in connection with the Late Medieval Crisis and associated farm desertion, while Estonian, Finnish, and Swedish research has focused primarily on early modern famine crises along with quantitative assessments of the relationships between various weather/climate parameters and harvest yields.Overall, the spatial focus is commonly restricted by current national borders and inter-Nordic research on climate history is, surprisingly, extremely rare.Nordic cooperation might provide a more comprehensive and versatile understanding of the past, considering the inter-linked history of many Nordic countries and the fact that climatic regimes do not follow present-day national borders.Furthermore, the Nordic research has focused only on the causal linkages between climate, harvest failures, food shortages, and population crises.Questions considering whether climate variability has played a role in the outbreak of violent conflicts, in the development of societal institutions, or in migration fluctuations, for instance, are greatly absent in the research. 101he Nordic research on climate-society interactions has been hindered until the last couple of decades by ( 1) a lack of quantitative data on past climatic variability, (2) a presumption that climatic changes have been too small to play a role for agriculture, and (3) the belief that explanations for historical change should be sought in factors related to human agency alone.Since the 1990s, the increasing number of high-resolution climate reconstructions provide answers to the first challenge and helps in assessing the second.Consequently, recent historical research has been able to detect many Nordic famines and population crises coinciding with climatic anomalies, which has shed new light on the third challenge.However, research conducted over the past decades has had its own shortcomings.

Bridging history and climate sciences
Conducting climate history research requires at least basic understanding of (palaeo)climatology.Climate reconstructions on annual-to-seasonal resolution are today adequately available for the Nordic countries, at least for winter severity and growing season temperatures and hydroclimatic conditions (like precipitation and droughts).However, this type of material can pose new challenges as historians can 'struggle with evaluating [these] highly technical scientific analyses and methods, sometimes failing to understand them at an elementary level.' 102 In Nordic studies on climate history, this 'struggle' materializes, for example, by integrating unsuitable palaeoclimate time-series or presenting them in an improper way in the analysis.For instance, some of the studies reviewed above plot different climate-sensitive series without any scale besides an annual timeline. 103Such a practice makes the climate data incomparable with any other studies and leaves the reader puzzled regarding whether the climate variations presented in the graphs are of a small or big amplitude.Furthermore, many studies have failed to clearly define what weather or climate variables and processes the data indicate.In the worst case, the reader does not even know if the graphs include raw measurement proxy data or if the time-series have already been transformed into climate estimates presenting, for example, variations in degrees Celsius or millimetres of precipitation. 104Consequently, although the results might be interesting, their validity is easy to question.Another common shortcoming is the inclusion of irrelevant climate data considering the topic and geographical focus of a particular historical study.For example, Jutikkala used an icecore-based temperature reconstruction from Greenland to demonstrate a link between temperature variability and crop failures in Finland. 105However, employing Greenland ice-core time-series is questionable when studying short-term temperature variability in Finland as it has been demonstrated that in years when the temperatures in Greenland are considerably higher than normal, frequently the temperatures in northern Europe are lower, and vice versa. 106he practice of rigorous critical source assessment is the fundamental cornerstone of historical research.This criticism should be practiced also with (palaeo)climate data if used in historical research, yet only a few studies reviewed here assessed whether the climate data used were relevant, reliable, and valid for the studied area and questions.At least the spatial domain and temporal precision of the climate reconstructions used should be relevant for the object of study. 107For example, studying annual yield responses to climate at a specific location with the help of hemispheric-scale climate data on decadal resolution makes little sense.As critically assessing the appropriateness of physical science data can be far beyond the expertise of a historian, an easy way to avoid using unsuitable climate data is co-operation with (palaeo)climatologists.

Climate, weather, and weather damages
Since the pioneering works of Le Roy Ladurie in the 1960s, most studies in climate history have agreed that the triggers of past human calamities were short-term climatic extremes and weather events, rather than long-term climatic regimes like the LIA. 108Therefore, understanding the difference between climate and weather is extremely important in climate history research.Additionally, it is essential to remember that the triggering weather events, like frosts or heavy rains, may or may not be typical for the prevailing climate, as one would expect.If one is proposing a causality between the prevailing climate and occurrence of extreme weather events in the past, this relationship needs to be detected and attributed.
For example, Utterström proposed already in 1961 that due to the severe winters of the early modern period, the onset of the growing season was delayed and consequently the ripening of the crops was postponed to the period when the first autumn frosts commonly occur.A similar argument was made later by Jutikkala, who associated the cold winters with the typical conditions of the LIA. 109Both Utterström and Jutikkala associated the cold winters with long-term climatic fluctuations -but treated the yield-damaging frosts as disassociated weather phenomena.Nevertheless, some Finnish historians have claimed that late summer night-frosts began to occur more frequently from the beginning of the seventeenth century because of the cooling climate. 110However, this interpretation failed to make a distinction between frost damage on yields and the frost phenomenon itself.In many parts of the Nordic countries, the first night-frosts by the end of the summer are commonly radiation frosts.The occurrence of these night-frosts is dictated by daily micro-weather.Therefore, instead of large-scale climate fluctuations, the elevation, proximity of water bodies, soil type, and vegetation cover influence the date of the first frosts. 111For example the known dates of night-frosts in Finland during the seventeenth and nineteenth centuries fell between late August and early September. 112hese dates, in fact, do not differ considerably from the first frost dates of the warmer midtwentieth century. 113Thus, there is no evidence that frosts occurred, on average, more frequently or earlier during the LIA.Nevertheless, we do know that frost damage on grain yields occurred more frequently during the LIA than during the following warmer period.However, one should further attribute the causality of this connection.In the case of crop failures and night-frost in the Nordic countries, the typical LIA climatic conditions explain the link.Average LIA growing seasons in the region (1) were colder, (2) started later, and/ or (3) were shorter than during the warmer climatic phases.These conditions often delayed the ripening of the crops to a period when the first night-frosts commonly occurred (Figure 3). 114evertheless, the geography of the Nordic countries is versatile.Thus, the relationships between climate and agriculture also vary from region to region, as demonstrated in Finnish and Swedish research on eighteenth-to early twentieth-century meteorological and crop-yield data. 115For example in Denmark, which is located further south and influenced by a marine climate, the relative shortening of the growing period due to cooler and/or wetter conditions of the LIA might not have caused any major frost-related problems during the late medieval and early modern periods. 116Consequently, more research is needed on assessing the relationships between historical agricultural production and weather/climate on a regional or local scale across the Nordic countries.Due to numerous technical developments in agriculture over the past century, as well as because of changes in the genetic plant material, the climate sensitivity of medieval and early modern agriculture might have differed from today's.Thus, such assessments need to be undertaken with pre-industrial agricultural source material. 117

Human agency in climate history
Although the Nordic research has proven that crop failures were more common during colder climatic phases in the past, this does not automatically mean that the crop failures culminated in subsistence crises.Nevertheless, Myllyntaus, for example, has interpreted records of crop failure inevitably meaning the presence of famine in early modern Finland. 118This perception can be criticized, however, as the role of individuals, institutions, local environments, politics, and multiple other factors influence how climatic and weather events can materialize on a societal level. 119For example, much of Europe was struggling with the weather-triggered Great Famine from 1315 onward.Raudkivi demonstrated that this was likely also the case in Livonia.In this region environmental factors, like the soil type, and socio-economic systems, like the dependency on the Hanseatic League relations, increased social vulnerability to food crises. 120Consequently, the cold and rainy weather extremes in the mid-1310s triggered a chain of events, which escalated into the famine crisis.Yet, the situation was likely very different north and east of medieval Livonia.Huhtamaa has shown that, although the harvest was presumably badly damaged in 1314 in Finland and north-west Russia (Novgorod), here man-made factors increased the resilience for coping with a bad harvest.In Finland, multiple livelihood options, and in north-west Russia food system diversity and strength in trade and economy decreased the regions' vulnerability to subsistence crises.Thus, as the written sources from the region do not mention a severe famine taking place in the 1310s, Huhtamaa concluded that it is unlikely that the crisis would have extended beyond Livonia to the north-easternmost corner of Europe. 121In international climate history research, the concepts of adaptation, vulnerability, and resilience are important tools to understand the complex and dynamic relationships between climate variability and human societies. 122Rather than the prevalence or amplitude of unfavourable weather and climatic conditions, the human capacity to cope and respond to these anomalies explains whether they may or may not influence the society. 123Hybel has concluded that the 1315-1317 Great Famine must have also occurred in Denmark, although contemporary sources do not record the subsistence crisis in the country. 124Indeed, several sources from the surrounding regions document that heavy rains in 1314-1316 destroyed the crops, and tree-ring-based reconstruction shows that these years were also extremely wet in Denmark. 125However, as discussed above, the prevalence of unfavourable weather, or even documented crop failures, cannot be used as proof of a subsistence crisis.Thus, assessing the resilience and vulnerability of the 1310s Danish society might help to evaluate if the extremely wet conditions had the potential to trigger a hunger crisis in the country.
Climate history has been criticized for its determinism, explaining past societal events as resulting directly from weather or climatic conditions. 126In this regard, we propose that by bringing human agency more profoundly into the analysis, for example, by assessing the vulnerability, resilience, and adaptation capacity of the studied societies, we can perhaps avoid some of these monocausal interpretations.Nevertheless, we wish to emphasize that the agency of climate and the agency of humans can be interlinked as, for example, climatic conditions, among other environmental and societal factors, can also influence human vulnerability, resilience, and adaptation capacity.

Outlook
Considering the published studies in the field of climate history for the five Nordic countries and Estonia from the early twentieth century to today, and the research gaps we have identified, we propose ten recommendations for the advancement of the field: (1) increase the number of collaboration among historians and (palaeo)climatologists to ensure that interpretations rest on a thorough and up-to-date knowledge of both past climate variability and human history; (2) increase collaboration across present-day national and language boundaries in research on both documentary-based climate reconstructions and climate-society interactions; (3) publish more of the Nordic climate history research in international peer-reviewed journals to achieve higher visibility and impact outside the Nordic countries and to contribute to the vibrant European research front; (4) practise a more rigorous source criticism when using climate data to understand and interpret past climate variations and the role they have played for human societies; (5)  pay more attention to what distinguishes weather from climate, and attempt to attribute whether certain extreme weather events occur more frequently (or cause more damage) under specific climatic regimes (as during especially cold periods); ( 6) focus more strongly on the local/regional differences of agricultural responses to climate variability to better understand spatio-temporal differences; (7) give more attention to societal vulnerability and resilience to climate-induced adverse effects of food production; (8) pay more attention to the human responses (such as agricultural adaptation) to climate variability instead of a narrower focus on human consequences; (9) expand the research on climatic change-human history linkages to areas other than agricultural and famine history; and (10) provide more university education in climate history to increase awareness about the field and train future historians.

Notes
2018-01272) and conducted the work with this article as a Pro Futura Scientia XIII Fellow funded by the Swedish Collegium for Advanced Study through Riksbankens Jubileumsfond.

Notes on contributors
Heli Huhtamaa (b.1984) has a PhD in History from the University of Eastern Finland and a PhD of Science in Climate Sciences from the University of Bern (both received in 2017).She is currently working at the University of Bern, conducting postdoctoral research with the Institute of History and teaching at the Graduate School of Climate Sciences.Address: Institute of History, University of Bern, Länggassstrasse 49, CH-3012 Bern, Switzerland.[email: heli.huhtamaa@hist.unibe.ch]Fredrik Charpentier Ljungqvist (b.1982)

Figure 1 .
Figure 1.The three main objectives of climate history and related dominant approaches and perspectives on agency.

Figure 2 .
Figure 2. Reconstructed (grey lines) and measured (black lines) June-August northern Fennoscandia and January-April Stockholm mean temperature anomalies, with 31-year moving average (black and white lines, respectively).The temperature data have been standardized over overlapping 1841-1890 period.The lines indicate approximate timing of the 'Medieval Warm Period' (MWP) and the 'Little Ice Age' (LIA), with dashed lines indicating the debated extent of the periods.Source: Matskovsky and Helama, 'Testing long-term summer'; Leijonhufvud et al., 'Five centuries of Stockholm'; Berkeley Earth; Moberg, A. 'Stockholm historical weather observations'.The measurementbased temperature data (Berkeley Earth) is averaged over 63-69°N, 18-24°E.

Figure 3 .
Figure 3. Simplified illustration of the link between the onset and thermal conditions of the growing season, delay of ripening of crops, and frost damage.(a) The growing season starts early, and warm conditions prevail (typical for the mid-twentieth century).(b) The onset of the growing season is delayed, but the growing season is warm.(c) The growing season starts early but the growing season remains cool and/or wet.(d)The onset of the growing season is delayed and the growing season remains cool and/or wet.In the last case the ripening of the crops is delayed to a period when the firsts night frosts commonly occur.The latter three conditions are commonly associated with the LIA in the Nordic countires.
is Associate Professor of History and Associate Professor of Physical Geography at Stockholm University as well as Pro Futura Scientia Fellow at the Swedish Collegium for Advanced Study.Ljungqvist currently leads the interdisciplinary project 'Disentangling socio-political and climatic factors for food insecurity in early modern Europe (c.1500-1800)'.Address: Department of History, Stockholm University, SE-10691, Sweden.[email: fredrik.c.l@historia.su.se]