Timing and climate impact of Medieval volcanism are illuminated by lunar eclipses
The Quantitative Wood ANALYSIS. I. The 1783 ce eruption and the timescale of peak cooling associated with the Laki eruption
The study suggests that lunar eclipses happen in about 20 months after an eruption, and that they can be used for dating and for determining the transport of volcanic aerosols. The interval between the first and second lunar eclipses varies from 6 to 3–4 years at some locations. 17,63 The irregular occurrence of total lunar eclipses can therefore prevent the dating of a volcanic eruption if the eclipse occurs outside this 20-month period.
aerosols produce volcanic dust that reduces incoming solar radiation and alter Earth’s surface temperatures.
The wood has a Quantitative Wood ANALYSIS. Edwards et al.100,101 attempted to narrow the period of peak cooling associated with the Laki eruption to late summer 1783 ce using cellular-scale tree-ring proxy measurements. These findings contrast with MXD reconstructions that suggest that the entire 1783 ce summer was exceptionally cold and with tree RW reconstructions that mute the cooling. Such results indicate that QWA data can identify more precisely than tree RW and MXD records the timing of peak cooling following volcanic eruptions within the growing season. Including QWA analyses for the HMP eruptions in the four-step procedure proposed in this study may further refine estimates of eruption timing. Despite promising results, highly resolved QWA is still at an early stage. Further, because highly resolved wood anatomical parameter chronologies are costly and labour-intensive, it is unlikely that an operational (Northern Hemisphere) network of QWA records will be available soon.
Investigating the origins and reliability of lunar eclipses in medieval manuscripts and chronicles: The Middle East, Christianity and South East Asia studied with the eight-millennium catalogue
We extensively reexamined hundreds of annals and chronicles, written in the twelfth and thirteenth centuries ce, searching for references to lunar eclipses. The series of medieval texts edited in the Monumenta Germaniae Historica is one of the more accessible contemporary sources for Europe. Descriptions or observations of lunar eclipses originate from across Europe, namely, Austria, the Czech Republic, England, France, Germany, Iceland, Italy, Sweden and Switzerland. In a few instances, lunar obscurations were observed in the Middle East by Christians engaged in the Crusades. These chronicles were found in the Recueil des historiens des croisades, edited and published in the nineteenth century. Most sources consulted were composed by monks or clerics and, to a lesser extent, by urban laymen. The majority were made up of Latin, but the vernacular was occasionally used. Furthermore, we included observations retrieved from Ukrainian and Russian chronicles in our database51. We looked at the observed lunar eclipses recorded in Arabic chronicles. We did not analyse them further because none of the sources have any information about the colour of the Moon.
Latin translations of Islamic texts were available in Europe from around 11507. The calendar and referenced dates in these texts were meaningless for their original developers. They also described astronomical observations that were made far from northern Europe. Users unfamiliar with the Islamic calendar weren’t helped by early translations.
The solar year was not in line with a year in theJulian calendar. Sacrobosco argued that Easter was being celebrated on the wrong day due to the fact that the accumulated error had reached 11 days. Church leaders were aware of the problem, but feared there would be serious resistance to changing the calendar11. Major religious institutions have their own calendars so the timing of phases of the Moon was out of step.
Historical source criticism and the eight-millennium catalogue of lunar eclipses were used to assess each eclipse observation’s reliability. We made sure that all eclipses reported in historical sources were visible in Europe, Middle East and East Asia using the visibility maps and local circumstances tables provided by both catalogues. Care was taken to identify second-hand reports, that is, those that a given author did not witness but instead drew information on from another source. Frequent duplication occurred in Western and Eastern Christian sources owing to common underlying source materials and the scribal practices of copying, synthesizing and piecemeal updating of annals and chronicles. A table listing all total lunar eclipses records for which information about the colour of the Moon could be retrieved, and providing summary context for the historical sources investigated, is presented in Supplementary Dataset S1. An extended version of Supplementary Dataset S1 containing all descriptions of total lunar eclipses from Europe, the Middle East and East Asia considered in this study (with or without reference to colour), along with the eclipse visibility maps, can be accessed from the Zenodo repository: https://doi.org/10.5281/zenodo.6907654.
Historical observations of lunar eclipse brightness are recognized as a valuable proxy for SAOD following large volcanic eruptions7,16,17,18,19. However, precautions are necessary to use this proxy appropriately:
SAOD time series cover a 56 month time window before the eruption and 52 month after. The time window for SAOD 0.1 varies from model to model, and for eruptions releasing more sulfur than Pinatubo there is substantial uncertainty about the home time of aerosols. Accordingly, for each month within the 56-month time window, we determined the probability for the SAOD = 0.1 threshold to be exceeded on the basis of agreement between time series. If four of the time series indicated exceedance of the threshold in a given month, the probability was very high. Likewise, probability was considered ‘high’ or ‘medium’ if at least three or two, respectively, datasets agreed and as ‘low’ if only one dataset indicated threshold exceedance (step 2.2). A dark lunar eclipse can be seen after an eruption, if the most probable time period isMonth min:Month max. Reciprocally, we can assume that the most probable eruption date falls within the time interval [Month max: Month min] before the date of the dark lunar eclipse (step 2.3).
We used the NVOLC v2 (ref. 7), Sch2015 (ref. 41) and N-TREND2015 (ref. 42) reconstructions to refine the dating of HMP eruptions. The NVOLC v2 reconstruction is composed of 25 tree-ring chronologies (12 MXD and 13 RW chronologies), Sch2015 relies on 15 MXD chronologies distributed across the Northern Hemisphere extratropics, whereas N-TREND2015 is based on a network of 54 records (11 RW, 18 MXD and 25 mixed series composed of RW, MXD and blue intensity records). These reconstructions were selected because they integrate a substantial number of MXD records. We did not use the recently published reconstructions.
It is important to observe totality when the weather is good and not near the horizon or at dawn or dusk.
Observations of lunar eclipses in 1100-1300: evidence for a volcanic haze or a lack of lava?
Reports should be contemporary with the event and preferably by an eyewitness. These conditions are not always met for the medieval sources available (see Supplementary Dataset S1 for more information).
It’s clear in the morning. After the hour of the Snake [9 am–11 am], heavy rain and flooding. People died when houses were swept away. The weather began to clear at the hour of the Horse. The moon was not properly seen during the hour of the Rat.
The work could also further scientists’ understanding of the climate in 1100–1300, which could be key to understanding the development of the exceptionally cold period that followed, called the Little Ice Age.
The authors have reviewed over 400 accounts that documented a number of lunar eclipses. Of those, 37 included information about Moon colouration and darkness; scientists can use both traits to determine whether and to what extent a volcanic haze was present at the time.
History in the Sky: Historical Perspectives on Solar Geoengineering from a Century of Research for a Renaissance of Social and Political Studies
Knowing what happened in the past could help clarify the impact of a proposal called solar geoengineering which would inject particles into the stratosphere to reflect sunlight back into space. Some scientist are of the opinion that such efforts could act as an emergency backstop, cooling the planet and stopping the worst effects of global warming.
For Guillet, who spent a decade poring over historical records in his free time before finishing the project during pandemic lockdowns, the study is also a nod to the medieval observers whose accounts historians have long used to document social and political trends. “If they are able to provide such an accurate record about social and political events, why not about natural events?” he says.
