As Degroot et al. (2021) demonstrated recently, a lot of scholarship in the field of the “history of climate and society” (HCS) is being produced by scholars from different disciplines. The same applies to an ever increasing amount of scholarship dedicated to volcanic eruptions, as becomes apparent with interdisciplinary collaborations such as the VICS-group, which is part of the PAGES (Past Global Changes) network.
Often in interdisciplinary collaborations between natural scientists and historians, difficulties arise when dealing with the specifics of the sources of both disciplines. We define sources as original material from the past. Regarding the archives of nature, this can include volcanic material captured in ice cores (e.g., ash, tephra), pollen found in lake sediments, or evidence gleaned from tree rings to name a few. The archives of society can be documents written during the time in question such as chronicles, newspaper articles, administrative literature, poetry, ego-documents, weather records, etc., but also inscriptions referring to a certain event (e.g., flood marks on bridges, dikes, buildings, etc.), and epitaphs. Every type of source has its particular strengths and weaknesses. With the “Interdisciplinary Nature-Induced Disaster Index” (INID-index), we want to offer scientists a tool to assess the quality of historical sources when analyzing nature-induced disasters, such as historical volcanic eruptions, from an interdisciplinary perspective.
The starting point of the index is a “signal” in the archives of nature (e.g., ice cores). This indication is verified by further analyses following the single steps on the index. Each step of the analysis is evaluated in terms of its significance. If the analysis is clearly in favor of the natural event, the value of the index increases by 1. If additional effects contributing to the event can be identified, the value is increased by 1 as well. If no evidence for a naturally caused effect can be identified, the value decreases by 1. In any case of uncertainty, the value remains neutral at 0.
Once a signal or peak has been identified in the archives of nature, the next step is to date the event. Statistical uncertainties must be addressed clearly in this case and need to be discussed and evaluated in terms of +1/0/–1. Additional paleoclimatological evidence (e.g., tree-ring data) should be used to determine whether more data is available that may support the initial information. Evidence from the archives of society (e.g., chronicles) is often used to calibrate studies and finetune the dating of specific events. Historical evidence, however, is challenging. A series of research steps need to be taken in order to verify its validity.
It should be considered that historical documents are available in different forms. There are plain translations of manuscripts, publications offering the original language and a translation, and critical editions informing readers about the status of the manuscripts (some parts might have been inserted years after the initial text had been written), varying handwritings in the manuscripts (which would indicate different authors), etc. Finally, there are the original documents themselves. For example, only a critical edition will indicate whether a report of extreme weather is authentic or simply a copy taken from a different source from a different region. Also, inserts from authors having worked on the given document years after it was initially written will only be described in critical editions or become obvious when looking at the original archival material. In some cases, critical editions or the original documents are not available. Some manuscripts did not stand the test of time and are lost forever. Some documents are still waiting to be critically edited, and some might have been transcribed in the past but destroyed during military conflicts or fires for example. In the case of the latter, a critical edition will therefore never be realized. This needs to be evaluated when looking for historical evidence. Only when original material or critical editions are available and analyzed by a specialist can a positive value be applied to the index (+1). Because of the uncertainty that lies in non-critical transcriptions, their value is neutral (0). This is also the case when using translations of critical editions without consulting the editions themselves. Translations without any information concerning critical editions should be excluded from interdisciplinary research (–1). In order to avoid negative values, an expert should be consulted here.
After a source has been identified, more historical material is needed that meets the aforementioned quality standards. This may help to identify the regional scope of a natural impact. Besides the availability of historical sources and their critical editions, however, it is necessary to evaluate their authors and their credibility. What is known about the author(s)? What was their intellectual background (what knowledge was available to them), at which location were the authors writing the text, and/or what was the quality of the information they could possibly rely on? An educated abbot writing a letter to his king informing him about poor provisions in his monastery due to a bad harvest related to unfavorable weather conditions is a considerably more trustworthy informant than an anonymous monk writing about some misery in a distant region he heard about. These questions can contribute to a deeper understanding of the environmental circumstances possibly related to an event spotted in the archives of nature (INID-index 0–5).
In order to distinguish between reliable and less reliable sources, historians look for “contemporary” evidence. “Contemporary” means that the author and/or the historical document, as well as the events described, belong to the same time period in the past (at best, the text was written by an eyewitness the same year, month, and day). Historical writings often depend on one another and so do the pieces of information they provide. Through clerical connections, for example, information written in one monastery could well be transferred to another, and accounts written in one region could be copied in another so that the same information can be found in two historical sources from different regions. This information is not independent—it is simply a copy. By evaluating and considering these factors in an interdisciplinary context, the findings in the archives of nature can be combined with dense, corroborative historical evidence. These evaluations can lead to an increase in the likelihood of a nature-induced disaster for a given incident (INID-index 5–7).
Historical documents—especially chronicles and similar literature—are rarely “fact-reports”. Information in these texts was often adjusted to the purpose of the author(s). Culture and religious belief play an important role here. There is evidence that in some instances writers fabricated natural events, or spun real events in a certain way simply for dramatic purposes (Wozniak 2020). Topoi and narrative motifs are frequently used when describing extreme events or social distress (Nünning 2013; Schellbach 2021). Often motifs from the Bible like hordes of locusts or cannibalism indicate cultural patterns of processing certain events rather than historical facts (Stathakopoulos 2016). It is necessary to rule out such aspects when looking for nature-induced disasters.
Vulnerability and resilience studies indicate that a key element in understanding the dimensions of an extreme event is the social structures it collides with. In order to overcome monocausal explanations, researchers need to distinguish between a nature-induced disaster and a disaster caused by multiple factors. It is therefore crucial to look for other/social stressors that could have contributed to the situation. This could be social and/or political conflicts during the time in question, failing infrastructure, insufficient aid programs, and so on. These factors need to be excluded when arguing for a nature-induced disaster (+1). In order to ascertain how past societies were impacted by and responded to climatic shocks, understanding historic coping strategies is important; this may help identify particular measures related to a certain natural impact (e.g., drought, heavy precipitation, volcanic cooling) or shed light on how people interact(ed) with their environment more generally. By following the index, uncertainties regarding a nature-induced disaster can therefore be limited to a low level (INID-index 7–9).
In the context of measures taken, it should be considered that the way people react to certain events highly depends on how they interpret them (Schenk 2010). An analysis of the connection between the interpretation of a natural event and measures taken may provide a deeper understanding of how past societies responded to climatic shocks. By looking for multiple causes for cascading effects (Pescaroli and Alexander 2015) and by analyzing them precisely, it is possible to evaluate the natural dimensions of a disaster. This not only reduces the uncertainty; it also helps to improve our understanding of extreme natural events and under which circumstances they likely cause the most damage (INID-index 9–11). Since a precise distinction cannot be achieved in every case, the index is designed to ensure a certain flexibility by overlapping values (INID-index 5/7/9).
Eldgjá and Laki will be used as two examples of how to implement the index during research. In addition, an event in ca. 913 is discussed to clarify the contours of the method presented and to highlight desiderata proposed by the index.
Following the INID-index, several aspects can be attributed to the Eldgjá event. The natural event has been identified and dated accurately by paleoclimatology using ice-core data, tree rings, and also historical records from various regions (+5). A substantial number of records are copies from later centuries and they depend on each other. Contemporary and independent manuscripts from Europe, however, remain (Ebert 2021) (+2). They were written in the historical regions of Lotharingia, Saxony, Swabia, and the eastern borders of West Francia as well as northern Italy and the monastery of Montecassino (Italy). While these manuscripts mostly describe hard winters, famines, cattle mortality, comets, or a solar eclipse, four medieval texts, in particular, have been used to date and explain volcanic impacts on climate and society in the tenth century: The Chronicum Scotorum (a compilation of Irish annals from the 1700s based on earlier manuscripts), the Res gestae Saxonicae (“Deeds of the Saxons”) by the Saxon chronicler Widukind of Corvey (written ca. 967–973), the Icelandic Landnámabók (“Book of Settlements,” written in the 1200s), and the Codex Regius that contains the Poetic Edda (written ca. 1270). The Landnámabók and the Codex Regius need to be excluded because of their significant temporal distance. They may refer to a volcanic event, but the information given is too vague to be clearly assigned to Eldgjá. Writing more than 250 years after the eruption, the scribes might have included experiences and reports of other volcanic eruptions in their descriptions (Ebert 2021). Current assumptions based on the Codex Regius about the Christianization of Iceland stimulated by Eldgjá are therefore questionable because the source is not contemporary (see above). The value of the Chronicum Scotorum is ambiguous. Due to its substantial temporal distance, the source can be excluded from the analysis. Studies on the Chronicum Scotorum, however, have at least put its distorted chronology in order so that a dating of phenomena described in the text is possible (McCarthy 1998, 2005). Yet, the account in the Chronicum Scotorum regarding a blood-red sun (see above) cannot be attributed to Eldgjá because the duration of one and a half days is too short and—more important—the subsequent line in the chronicle refers to the siege of the fortification Ailech by “the heathens” (Chronicum Scotorum, a. 939). The blood-red sun could therefore be a topos—a prodigy—indicating that blood will be spilled in the future. The same must be applied to the writing of Widukind of Corvey († after 973). In the second book of the “Deeds of the Saxons,” he writes: “Many people were terrified at the sight of the comets, fearing that there would be a great pestilence or at least a change in ruler since great prodigies were seen before the death of King Henry. For example, the light of the sun could hardly be seen outdoors because of the dark sky, but inside sunlight poured red as blood in through the windows of houses” (eds. Hirsch and Lohmann 1935; trans. Bachrach and Bachrach 2014). King Henry I died in 936. Hence, Widukind refers to an event that occurred not later than this year. According to recent paleoclimatological research, an eruption date for Eldgjá of around 939/940 is likely, which allows for the conclusion that the phenomena that Widukind describes were not connected to the Icelandic eruption. There may have been communication networks from Iceland via England to Saxony, but there is no evidence that supports historical observations connected to the Eldgjá eruption (Wozniak 2020; Ebert 2021). Since there is no evidence for or against the natural signal, the INID-index value is 0.
Social stressors are considerable during the time in question—especially in East Francia ruled by Otto I (912–973). In the subsequent years after his coronation in 936 warlike conflicts set in as Otto pursued a new policy that some magnates in his kingdom did not support (Ebert 2021). War, plunder, and looting increased severely. In particular, Saxony, Swabia, and Lotharingia fell victim to these conflicts. Most of the tenth-century sources are located in these very regions (Ebert 2019, 2021; see also the GIS map available at https://arcg.is/rbn5i). This means that the Eldgjá eruption coincided with a period of military conflicts, and political and social instability that severely threatened people’s livelihoods. The disastrous situation was not entirely nature-induced (–1). Rather, the volcanic event contributed to worsening the situation, and was one factor amongst many (+1).
In most climate reconstruction papers reaching back to the first millennium CE, historiography is the primary genre consulted, whereas administrative documents play a secondary role. These sources can cover coping strategies and measures taken in reaction to extreme events. In the case of Eldgjá only two administrative documents offer insights into how people might have reacted to environmental stressors. In the first half of the tenth century, a capitulary (royal decree) dealing with famine in 779 was included in a codex compiled in the regions between Lotharingia and West Francia (Mordek 1995). The capitulary aimed to overcome hunger with spiritual guidance and food aid (i.e., prayers, almsgiving/pecuniary fees, feeding of paupers, and fasting). In the event of social turmoil during the time in question and natural stressors setting in, monastic communities could have found orientation in such a decree and acted accordingly (Ebert 2021). These measures only make sense when interpreting changes in the natural and social environment as divine punishment (+1). In 940, abbess Markswid of the Saxon monastery Schildesche (Germany) ordered her sisters to participate in an annual rogation to pray for a good harvest and an end to adverse weather (Holder-Egger 1888). In a social environment in which Christian belief was not yet totally engrained, instructions of this kind helped to actively involve the population in Christian rituals. Measures such as this, show that extreme events could conceivably lead to the ubiquitous influence of Christianity in everyday life. The historical evidence, however, is very poor. While at least in the case of the capitulary, the dating in the relevant period is certain (+1), the vita of Markswid mentioning the rogation was not written until the thirteenth century and is therefore uncertain—even if it clearly refers to 940 (0).
If the values obtained from the index are added together, the result is an INID-index value of 9. A nature-induced event is likely. This can be confirmed with methods from the natural sciences and historical records. In addition, social stressors and multi-causal effects can be observed, which offer a new perspective on the diverse relation of nature-induced and anthropogenic factors during and after the Eldgjá event. Contemporary coping strategies related to specific societal interpretations of historical events can be observed. Due to the limited quantity and quality of the relevant sources in this case, however, some uncertainty remains for the Eldgjá event as a whole. This might be unsatisfying, but the benefit obtained from the index compensates for this shortcoming: Although the Eldgjá eruption had a high intensity, its impact on western and central European societies was only significant due to an interaction with human-made factors.
Paleoclimatological reconstructions using ice-core data, tree rings, and several historical sources from Europe, North America, North Africa, and Asia are in good agreement about this event (+5) (Hantemirov et al. 2004; D’Arrigo et al. 2011; Sigl et al. 2015; Edwards 2021). The eruption and its effects upon Iceland are well documented by Icelandic eyewitnesses: One example is the Lutheran priest Jón Steingrímsson who kept detailed notes. He lived in Kirkjubæjarklaustur, 32 km from the Laki fissure. These reports give us the eruption’s exact start and end dates as well as its local consequences (Steingrímsson 1998; Steingrímsson 2002; other contemporary Icelandic reports are mentioned in Thordarson 2003; Thordarson et al. 2003). In 1788, Steingrímsson authored a fire treatise (eldrit) about the Laki eruption, which was intended for publication so that all Icelanders “born and yet unborn” would remember the event (Steingrímsson 1998). He also wrote a memoir of his life, most likely in 1788. However, he did not intend for this to be published; it was written exclusively for his descendants, which allows for the assumption that he wrote openly about his struggles. The autobiography was published posthumously in 1916 (Steingrímsson 2002). Both manuscripts survived and are housed at the National and University Library of Iceland. They are also available in English translations.
As the Laki eruption occurred in the late eighteenth century, many historical sources were produced by contemporaries and have survived up to the present. Some of the numerous sources that make mention of the unusual weather in Europe include newspapers, weather diaries, private correspondence, and scientific publications.
Many articles in dozens of newspapers from different parts of central and western Europe mention unusual weather phenomena. At times, however, different newspapers used the same correspondent or copied from one another, which becomes apparent from the phrasing of individual articles. While newspaper articles can be dated exactly, as most newspapers were printed one to six times per week, the identities of the correspondents responsible for penning the articles are rarely known. Nevertheless, the large number of articles regarding the unusual weather in different regions during the summer of 1783 published by dozens of newspapers over the course of several months make the sources independent. In weather diaries, amateur weather observers and naturalists made observations regularly, often daily. This is the case, for instance, with the weather diary of British naturalist Gilbert White at Selbourne (England), whose “Naturalist’s Journal” manuscript for 1783 and the following years are available at the British Library (White 1783). Several other contemporary naturalists also kept weather diaries, where they wrote about their meteorological observations (Woodforde 1782–1784; Barker 1783; Beroldingen 1783; Christ 1783; and many others). These sources are also contemporary and independent (+2) (Kleemann 2022).
At the time, the Societas Meteorologica Palatina, a meteorological society with its headquarters in Mannheim (Germany), operated a network of more than thirty weather stations across Europe and beyond. The society’s weather observers received standardized instruments and forms with instructions to take note of the instruments’ readings at three specific times per day. After the year’s end, the forms were sent to Mannheim, where they were compiled into annual compilations called Ephemerides. The Ephemerides for 1783 include instrumental data on temperature and pressure and sometimes remarks on meteora, unusual observations that include descriptions of the sun setting in a blood-red color or the Laki haze (Societas Meteorologica Palatina 1783 (1785)). The application of careful scientific methods by many various documentations from different weather stations in different parts of Europe and the collection of large quantities of data rule out topoi (+1).
Throughout the summer of 1783 and the following year(s), several naturalists published their speculations about the source of this unusual weather, particularly the dry fog. Many different theories were in circulation at the time. Some blamed the earthquakes in Calabria in February of that year, which allegedly had caused cracks in the Earth that released the noxious gas, Others suggested that the large-scale introduction of lightning rods deprived the air of its “fertilizing electricity,” which in turn led to the formation of the fog. A few naturalists, as early as the summer of 1783, postulated that an Icelandic volcano, such as Hekla, or the emergence of a new “burning island” off the coast of Iceland was to blame. These scientific explanations were published by contemporary newspapers with an emphasis on calming the readership, which can be seen as a coping strategy. Additionally, newspapers undertook interviews with the elderly who reassuringly stated that similar events in the past were usually followed by good harvests. The newspapers also searched chronicles, which corroborated what the elderly interviewees had said, further comforting their readership (+1) (Kleemann 2019a; 2019b).
While the hot temperatures and Laki haze did affect vegetation in some parts of western Europe, it largely recovered and did not cause a bad harvest or famine that year (Grattan and Pyatt 1994; Stothers 1996). In the same areas, those with pre-existing respiratory diseases did suffer because of the Laki haze. Others complained about sore throats and eyes (Brugmans 1783; Santel 1997; van Swinden 2001). Some regions in England and France experienced increased mortality in the summer and autumn months of 1783. It is, however, inconclusive as to whether these bouts of diseases occurred independently from the Laki haze or were aggravated by it (Durand and Grattan 2001; Grattan, Durand, Taylor 2003; Witham and Oppenheimer 2004; Garnier 2011). Further research is needed to establish which regions in Europe were affected by this increased mortality in the second half of 1783. This may lead to insights about the possible causes of death. Currently, the Laki eruption cannot yet be ruled out as the culprit (0).
The Laki haze—then known as “the dry fog,” “haze,” or “mist”—was blamed for several of the observed unusual phenomena in nature, such as a seeming increase in the number of thunderstorms (with frequent lightning strikes that injured and killed people) and the unusually hot summer. Some contemporaries suggested there was a connection between the eruption and the severe winter(s) that followed. Modern natural scientists confirm some of these assumptions, at least for a connection between the eruption and the subsequent cold winter(s) (+1) (Franklin 1785; Hochadel 2009; Zambri 2019a; 2019b). During the summer of 1783, contemporaries in Europe were unaware of the eruption in Iceland, instead the Calabrian earthquakes were discussed as a potential cause of the dry fog. Many coping strategies were developed: A newspaper report in the Königlich-Privilegirte Zeitung, a Berlin-based newspaper, from 24 July 1783 suggested that vegetables affected by the fog should be washed properly before consumption. The diligent smoking of tobacco was also recommended, as was keeping livestock inside. These recommendations were written by an anonymous correspondent from Hanau (Germany) in a newspaper. From the records there is no indication that suggestions like those mentioned above were widely adopted. (Kleemann 2022).
When calculating the values from the index, the result is an INID-index value of 10. Therefore, a nature-induced event is likely from both the perspective of interpreting paleoclimatological data and historical records. While it is not yet clear beyond a reasonable doubt whether the Laki eruption was responsible for the peak in mortality in 1783, the dry fog and other unusual phenomena definitely captivated onlookers and naturalists alike. They collectively processed the various observable events that included a sulfuric-smelling dry fog, which triggered respiratory problems, sore throats and eyes in some regions, the withering of vegetation, “blood-red” sunsets, many thunderstorms, and a colder-than-usual period following an unusually warm summer. Contemporaries developed many different interpretations of the unusual weather and numerous coping strategies.
3.3. Volcanic Induced Harsh Winter Around 913 CE?
In this chapter we aim to contrast the previous analyses by comparing the INID-index results with a case notably different toEldgjá and Laki. A peak in SO4 sediment in the GISP2 ice core dated to ca. 913 has been assumed to be connected to an Icelandic eruption—probably Katla (Zielinski et al. 1995) (+1). A precise dating or localization of the eruption, however, has yet to be published (0). Dendrochronological reconstructions of the summer months (JJA) do not indicate a significant temperature decrease (Luterbacher et al. 2016) and a wet summer can only be identified in Ireland in 912 where dry summers followed in the years after—beginning in 914 (Cook et al. 2015). Regarding the INID-index, additional studies from the natural sciences are available, however, their results do not support a definite nature-induced disaster (0). Harsh winters were documented in Ireland and along the river Rhine in 912–914 (McCormick et al. 2007; Wozniak 2020) by various, contemporary, and independent writers (+3). The Irish Annals of Ulster even mention “a dark and rainy year” in 912 and 913. It is not clear whether this report refers to both years or one report is simply a duplicate of the other. The observations could well be influenced by volcanic aerosols in the atmosphere, but there is no similar report from another region that supports this assumption. The annals document many deaths of various individuals in 913, and so, the “dark and rainy year” could also be interpreted as metaphorically dark and rainy, given the events mentioned (0). From a dendrochronological perspective, however, this account matches the wet summer reconstructions in Ireland in 912 (see above) (+1). The only reference that may explain the darkness in the Irish annals is a solar eclipse documented in Córdoba (Spain) on 17 June 912 (Wozniak 2020), but according to the NASA Catalog of Solar Eclipses it is unlikely that this event was visible in Ireland (–1). Adémar of Chabannes (989–1034), a chronicler from Aquitaine (southwest France), describes a severe famine in ca. 913, but he attributes the widespread hunger and punishment to the rule of Alduinus, a French count who had unlawfully acquired a relic (Wozniak 2020). Hence, a link between the famine and environmental circumstances is uncertain in this case (0). Hunger is not reported elsewhere and social stressors during the time in question are mainly documented in Lotharingia, Saxony, Bavaria, and Alemannia—regions not affected by famine. A connection between a harsh winter, limited food provisions, and social conflicts contributing to the limited food supply is not possible in the case of 913 (–1).
In short: There are indications of a volcanic induced winter, but there is a high uncertainty regarding a nature-induced disaster so the index value is 3. Due to the poor evidence in the archives of nature and society it is difficult to evaluate how much the event spotted in the GISP2 ice core contributed to the winters of 912/913 and 913/914. The impact on society, it seems, was not substantial as there are no reports of vast mortality in multiple regions in Europe. The chronicle of Adémar of Chabannes, however, could be a case for further investigation. Micro studies on the historical region of Aquitaine and its environmental conditions could prove whether Adémar really uses the famine as a topos or if there is more evidence for a natural impact and its cultural consequences—a desideratum highlighted by following the steps of the INID-index and contributing to the understanding of past interrelations of humans and nature.