the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 08 Sep 2026
A Theory of the Linkage between Volcanoes, Climate Cooling, Malnutrition and Epidemics: Case Study of Justinianic Plague
Abstract. Major volcanic eruptions have long been proposed to have triggered and amplified historical plague pandemics through climatic shocks and harvest failures, but the mechanisms linking volcanic forcing, nutrition and disease remain poorly quantified. We develop a theoretical model of a single region with zero spatial dimensions ("0-D") that couples volcanic forcing, climate, agricultural productivity, food storage, malnutrition and plague epidemiology. The epidemiological component includes reservoir hosts, vectors, human-to-human transmission, disease-driven migration and waning immunity, with transmission and mortality rates dependent on nutritional status.
We first evaluate the model against agricultural and archaeological evidence from England following the 1345 CE eruption, shortly before the Second Plague Epidemic. The simulations reproduce the observed sudden drop in crop output and prolonged deterioration in nutritional conditions, as well as the observed relationship between plague mortality and an isotopic proxy for malnutrition derived from human skeletons. We then apply the model to the Justinianic Plague following the 536 and 540 CE eruptions. Representing three population groups with different vulnerability to harvest failure and malnutrition, the model predicts eventual plague mortality of about 39 %, consistent with historical estimates. Sensitivity experiments show that removing the effect of volcanic forcing on malnutrition substantially reduces predicted plague mortality, whereas excluding disease-driven migration increases it. Doubling eruption intensity produces only a modest increase in total mortality, revealing nonlinear interactions among climate, nutrition and epidemic dynamics.
A reduced phase-space analysis further shows that epidemic feedbacks change from net positive to net negative as susceptible individuals are depleted, with the system transitioning from an initially unstable state towards a quasi-equilibrium attractor. Together, the simulations suggest that volcanic eruptions could have substantially amplified historical plague epidemics through their effects on failure of harvests and food production, malnutrition and human behaviour. The results provide a mechanistic framework linking volcanic climate forcing to epidemic severity and highlight nutrition as a potentially important mediator of climate–disease interactions.
- Preprint
(11583 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 03 Nov 2026)
-
RC1: 'Comment on egusphere-2026-5195', Anonymous Referee #1, 12 Oct 2026
reply
-
AC1: 'Reply on RC1', Vaughan Phillips, 14 Oct 2026
reply
We thank the referee for this positive and encouraging assessment of our study.
We are grateful for the recognition of the potential contribution of our interdisciplinary approach and for the thoughtful observation concerning the limitations of modern understanding of the Black Death.
We appreciate the time and care taken in reviewing our manuscript.
Citation: https://doi.org/10.5194/egusphere-2026-5195-AC1
-
AC1: 'Reply on RC1', Vaughan Phillips, 14 Oct 2026
reply
-
CC1: 'Black Death Historiography: Comment on egusphere-2026-5195', Monica Green, 13 Oct 2026
reply
My congratulations to Dr Phillips and colleagues on this innovative project, which brings modelling to an assessment of possible epidemiological patterns. I write as a specialist on the Black Death, for which I have maintained a publicly accessible interdisciplinary bibliography for the past six years: https://docs.google.com/document/d/1x0D_dwyAwp9xi9sMCW5UvpGfEVH5J2ZA/edit?usp=sharing&ouid=104263744463668175127&rtpof=true&sd=true. Appended to the Second Plague Pandemic section is a comparable bibliography on the Justinianic Plague. Both have sections on Climate.
The reason for publicly sharing the bibliography was to address the deficit in the field of infectious disease history, which by its nature (germs, animals, people, environments) has to draw on a variety of disciplines, yet falls into too many cracks between disciplines in terms of bibliographic resources.
I note this because the present project is framed on the assumption that the Black Death (BD)/Second Plague Pandemic (2PP) is a well-understood and precisely dated phenomenon, one which can be clearly separated from famine episodes. Aspects of the pandemic are well-known and precisely dated, it's true, more so than the Justinianic Plague (JP), because later medieval document production was so much more intense than in the 6th century, and more of it has survived to the present day. But nearly all of that documentation only tells the human part of the story, when we know that plague is really a zoonotic disease. The advent of paleogenomics (which focuses on the bacterium) doesn't resolve that evidentiary bias, since nearly all sampling is done only on human bodies. In other words, we are all dealing with incomplete data.
For both the BD and the JP, the question has arisen as to how much trust we can put into the dating of each pandemic, since the narratives we have inherited have been based solely on human accounts. I note that you cite Bauch and Büntgen's 2025 paper on the 1345 volcanic eruption. Missing from their study was the following piece which had appeared 11 months earlier:
- Green, Monica H., and Nahyan Fancy. “Plague History, Mongol History, and the Processes of Focalisation Leading up to the Black Death: A Response to Brack et al.," Medical History 68 no. 4 (Fall 2024), 411-435, https://doi.org/10.1017/mdh.2024.29.
Adding animal reservoirs and arthropod vectors into our plague narratives is increasingly necessary, if we are to modulate the omissions of human-created accounts. Two recent pieces not mentioned in your draft do this for the JP:
- Brandon McDonald, “Contextualizing Late Antique Natural Disasters: The Elements and Effects of the Late Antique Little Ice Age and the Justinianic Plague,” in: Catastrophes in Context: Disaster and Response in the Roman and Early Byzantine World, ed. Rubina Raja and Andrew Wilson (Turnhout: Brepols, 2026), pp. 263-287.
- Brandon McDonald, Andrew Wilson, and Arietta Papaconstantinou, “The Environmental Context for the Outbreak of the Justinianic Plague in the Mediterranean,” in: Catastrophes in Context: Disaster and Response in the Roman and Early Byzantine World, ed. Rubina Raja and Andrew Wilson (Turnhout: Brepols, 2026), pp. 233-262.
McDonald's work seems important for your own approach, since he raises the question whether the volcano-induced climatic shift is important not for causing famine, but for increasing flea populations. There is an additional question of whether the geographic routes that plague traversed were really fully compatible with the written accounts we have. As I'm sure you know, a genetically basal strain of the JP Y. pestis lineages has been found in Hungary, whose dating is still unclear. (This work has been presented at several conferences, including Leeds 2025: https://www.imc.leeds.ac.uk/imcarchive/2025/sessions/202/.) I recognize that your model assumes that plague is already local and I do think that, because of our prior reliance solely on human-authored accounts, we overestimate plague's "sudden" arrival in certain regions. My point here is simply to say that just as a "simple and sudden importation" model no longer works for the BD, similar considerations may be necessary for the JP, too.
Also, a study whose implications for assessing the presence of plague in different landscapes may be pertinent to your characterizations of human responses to plague is this:
- Sidhu, Ravneet Kaur, Guillem Mas Fiol, Pierre Lê-Bury, Christian E. Demeure, Emelyne Bougit, Rémi Beau, Charlotte Balière, Aurelia Kwasiborski, Valérie Caro, Jennifer Klunk, Daniel J. Salkeld, Ann Carmichael, Nükhet Varlık, Debi Poinar, David J. D. Earn, Benjamin M. Bolker, Jonathan Dushoff, G. Brian Golding, Nicolas Rascovan, Olivier Dussurget, Edward C. Holmes, Javier Pizarro-Cerdá, and Hendrik N. Poinar. “Attenuation of Virulence in Yersinia pestis Across Three Plague Pandemics,” Science 388, No. 6750 (29 May 2025), eadt3880.
If, as they argue, all three plague pandemics were characterized by the (sequential or simultaneous) circulation of lower-virulence strains, can we always assume that human responses will be identical? I believe this gets to the core question you are asking about how climate-inflected famine might be related to patterns of flight and land abandonment during outbreaks of plague. Evidence is mounting that the connections between plague and famine are far more common than we have previously realized. It might be that your "panic threshold" works best in situations where highly virulent strains are circulating, having spilled over into human population centers. If so, traditional historical dating might be perfectly adequate. Lines 490-92 suggest that you have made that part of your model: 'Bubonic plague involves highly visible symptoms, such as buboes and rapid mortality. Both the onset and decline of an outbreak would have been apparent to the population, prompting rapid behavioural responses."
However, the lower virulence strains that Sidhu et al. 2025 document might not have produced such prominent symptoms, nor such rapid mortality. If lower virulence strains are moving through agricultural areas, might lands be abandoned not because people have taken flight because of visible disease, or even because the lands themselves are no longer productive (because of climate impacts), but because the land itself is infested with plague (the process of focalization)? I'm not sure how any of us could answer that question right now from the tools we have. I would, at the very least, suggest engaging with the arguments about "caging" proposed by Borsch and Sabraa, which suggest that the seemingly "inflated" population figures for Muslim cities during the BD may reflect the fact that people fled to the cities because of famine and/or disease in rural areas:
- Borsch, Stuart, and Tarek Sabraa. “Refugees of the Black Death: Quantifying Rural Migration for Plague and Other Environmental Disasters,” Annales de Démographie Historique 2017 N°2, no. 134 (2017), 63–93.
A note on use of evidence from the Great Famine (1315-17): at lines 801-84, you invoke some "archaeological observations ... [that] report evidence for chronic shortages of protein, calcium and vitamin B12" during the period of the Great Famine. Those are inferred shortages based on the assumption that nutritional stress was the main physical hazard being faced in the early 14th century. As the authors themselves indicate (DeWitte and Slavin 2013, p. 44), "The Fisher’s exact test reveals that the differences among the pre-famine, famine, and post-famine groups are not statistically significant for any of the skeletal stress markers considered. The lack of significant results, which may be an artifact of small sample sizes, means that neither of our expectations is supported by the existing data." In fact, this result is not surprising, since it is now likely that mortality during the Great Famine was due to factors beyond nutritional deficit alone.
Whereas climate science/history and its related disciplines are all quite mature in their methods and can draw on deep databases with high confidence of chronological accuracy, that is not true of infectious disease history, which is still charting a path toward viable work between documentary records, bioarchaeology, and genomics.
A final note: on line 290, you cite "Martin 1996" on the topic of land abandonment. No corresponding entry is found in the bibliography.
Again, my congratulations to you and your team. I look forward to being able to include the final version of this study in my bibliographies (which, as you'll note, have already incorporated this draft).
Monica H. Green, Independent Scholar and Fellow, Medieval Academy of America
Citation: https://doi.org/10.5194/egusphere-2026-5195-CC1 -
RC2: 'Comment on egusphere-2026-5195', Anonymous Referee #2, 14 Oct 2026
reply
We have much to learn about sixth-century epidemic disease and climate and so this highly interdisciplinary paper is to be welcomed. Though many have commented previously on these phenomena in the context of the sixth century and the fourteenth, as the authors state: ‘mechanisms linking volcanic forcing, nutrition and disease remain poorly quantified.’ This reviewer very much appreciates the effort made here to move us towards a better understanding of how volcanic climate forcing may have contributed to the emergence of the First Plague Pandemic and how it may have amplified the pandemic’s mortality. I do believe, however, that this paper is not ready for publication. This is a hard paper to review. It is a very interdisciplinary paper and it requires an interdisciplinary review from climate and disease modelers, plague ecologists, historians of the sixth century, and historians of disease. There are aspects of this paper that I cannot comment on. I will restrict myself to areas where I can comment, while also pointing out where and when scholars who do not model for a living might struggle to follow the paper’s logic. What follows is list of questions and remarks that came to mind while reading the 92-page PDF. As the paper draws widely from many fields, these remarks are a bit disjointed.
Three general points:
- the authors might make it clearer how the model they develop can be reliably employed to explain what are very poorly understood phenomena
- the authors would also do well to be more transparent with their use of the historical data
- the authors should consider revisiting some of the relevant historiography
More on these general points:
- how can the model be ‘evaluated’ by considering its output next to what we can understand about effects in England of an eruption in 1345? Are the eruptions of 536, 540 and 1345 comparable? Why can the impact of the climate forcing following the 1345 eruption on crops in England be compared to the impact of 536 and 540 on crops in the Mediterranean region? Do differences in agroecosystems, etc., not greatly matter, nor the partiality of our understanding of sixth-century agriculture? Perhaps these things do not matter, but if that is the case it needs to be made clearer why that is
- on the evaluation test, it is perhaps important to note that Bruce Campbell’s work has not identified hugely significant yield variations in 1345-1347. The harvest failure in the 1340s in England that was very significant started in 1349 (as is briefly noted on p. 7 of the PDF of the paper under review, but neither earlier nor later). I recommend reconsidering Campbell’s “Four Famines and a Pestilence” article (2009). In a 2010 publication (“Nature as Historical Protagonist”), he identifies 1346 as a year in which yields fell ‘below their long-term mean,’ but not ‘by more than 20 per cent.’ 1346 was not a great year to be sure, but it was not 1304, 1315-16 or 1321 or 1349-50. Is that worth pointing out?
- it would seem that the model considers harvest data from a decade following the 1345 eruption (p. 77 of the PDF), but why is that? Does doing so associate the extremely poor harvests of 1349-50 with the eruption of 1345? If it does, and it might not, what is the basis for assuming the climate forcing of the eruption would cause such a poor harvest four and five years after it took place? What of the role of plague mortality on the harvests of 1349-50? Campbell (2009, p. 12) stresses that any inclement weather then ‘was greatly exacerbated by the heavy mortality of workers and managers’
- about food shortages, also note that it has been argued that the Samalas eruption, which comes up in the paper at times, aggravated, but did not initiate the food shortages reported in England or Europe more generally around the time of the eruption. From Campbell’s Great Transition (pp. 57 and 396), Samalas ‘clearly cannot have caused the extreme weather of 1256 and early 1257 or the disastrous northern European harvest failures of 1256 and probably 1257;’ ‘any volcanic forcing of the climate must have been felt between 1258 and 1261; years, intriguingly which do not stand out as particularly extreme or unusual in the paleo-environmental record;’ Samalas ‘did not necessarily have proportionately substantial and enduring environmental and human impacts.’ Does a more recent publications question this thinking?
- the caption of Figure 1a brings up issues that pop up elsewhere in the paper about the interpretation and use of written evidence for epidemics. How ‘well-documented’ or ‘major’ is the epidemic of c.43 BCE? The plague diagnosis is not secure, as the authors note, but why does that matter? Please let the readers know
- should Figure 1 and the corresponding text also not give readers an idea of how many epidemics do not chronologically align with volcanic eruptions? Most documented epidemics do not align with major volcanic eruptions. Plague outbreaks are near constant through much of the second plague pandemic (~1335-1850). The vast majority of them, including some of the best documented and the clearly demographically significant, do not line up temporally with a volcanic eruption. I think the paper would benefit from being clearer about this issue. It is said (on page 5 of the PDF) that ‘epidemics generally have always occurred throughout history irrespective of eruptions’ but I do think this needs to be much more emphasized
- it is as well said on page 5 that ‘the selected epidemics are disproportionately represented among the higher-mortality events in the historical epidemic record.’ I am sorry to say that this is not a provable statement. Many of the clearly large-scale and demographically severe epidemics of the second plague pandemic are not selected here. Not to press too far with this, but the mortality of some of the plagues included is unknown. How many people died in the Athenian Plague, 100,000? We have no idea. Why is mention not given here to the Antonine or Cyprianic Plagues which many have argued killed a lot of people, far more than the Athenian?
- about the epidemics mentioned, I am not sure many disease historians would agree that epidemics in Russia between 1230-32 were plague. How non-contemporary is the source used for this epidemic?
- perhaps more important for the argument: about the onset of fourteenth-century second pandemic Yersinia pestis, the authors cite (p. 5 of the PDF) Cui et al 2013 via Fell et al 2020. It would be better to focus on Syprou et al 2022 (the discussion under the heading ‘Divergence Time’), a paper that is cited in the following sentence but for a slightly different purpose. In short, we cannot say with confidence on the basis of current data that the Y. pestis responsible for the Black Death had a ‘genetic origin’ in the mid 1250s, that is, around the time of the Samalas eruption, or even within a few decades of the 1250s. The genetic origin is likely later, very likely in the early fourteenth century (as per Spyrou et al 2022 and subsequent aDNA papers)
- are several major eruptions not missing from Figure 1a, particularly between 536/40 and 1230, like those of 574, 682, 939, but also before 536, like 266? Or why are they not included?
- I am not certain most plague historians would say the First Plague Pandemic ‘spanned’ ‘central and southern Asia’ (p. 7). We do not have the data we would need to state that. That ‘it spread throughout Europe’ is also not supported by the data available (the written and paleogenetic evidence do not cover all European regions). That said, paleogenetics have now many times located Y. pestis beyond the boundaries of the Roman Empire (contra what is said on p. 7 of the PDF; see Keller et al PNAS 2019). In addition to Scandinavia, the authors could also list Ireland, Italy, Greece, all of northern Africa, etc, as places where paleogenetics have not yet recorded the First Plague Pandemic.
- about the First Plague Pandemic, what (see p. 9 of the PDF) about the Irish annals reference to disease c.544 suggests that epidemic was ‘major’?
- that Gildas’ text might not actually relate to the events of 530s or 540s should be made clear to readers, as the debate about the text’s composition continues
- the paper takes an almost catastrophist perspective to sixth-century plague. On the basis of the evidence available that is one perspective of several that can be adopted depending on how one wants to employ all the evidence available. The ongoing debates about this make that clear. The authors might let readers know that too
- on another topic, reference is made a few times to Dean et al’s 2018 study on lice and plague (on pp. 63, 67-8 of the PDF). That study was questioned by Park et al 2018 (PNAS) that highlighted Dean et al’s ‘inconsistent modeling choices and reliance on strong assumptions.’ Does this matter for the paper and should this be addressed in the text?
- on another topic, not all ‘major volcanic eruptions’ were followed by ‘acute’ ‘food stress’ in all regions. The Millennium Eruption and even the Samalas eruption show that to be the case, but others too. The tone here is a bit catastrophist here too
- about food shortages, some historians will really struggle with the claims made about the famines that the model predicts after the 536/540 eruptions. This scale of disaster is simply not evidenced in the source material
- the same can be said about the projected plague mortality. Might the model inputs (about things like population density and distribution, plague transmission mechanics, the weight given to malnutrition, etc.) that account for this output be briefly summarized in a table early on in the paper so historians can decide for themselves if they wish to follow the conclusion? The conclusion (p. 52 of the PDF) suggests the result cannot be applied to both major cities and also rural communities, and even that the ‘geographical applicability of the present model is uncertain.’ I think that that needs to emphasized in the abstract and introduction
- related to this, the very important role (about 3/4s of all deaths (p. 34 of the PDF) and more than 2/3rds of infections (p. 53 of the PDF)) ascribed to pneumonic plague transmission (person-to-person via respiratory droplets) is fascinating but seems to be out of sync with well documented outbreaks of pneumonic plague, published analyses of pneumonic plague’s epidemiology, and also the written evidence available. It seems key papers on pneumonic plague are not cited, such as Gani and Leach EID 2004, Kool and Weinstein Clinical Infectious Diseases 2005, Nishiura et al Journal of Epidemiology 2006, and Hinkley et al Epidemiology and Infection 2012
While other such points could be raised and revisions are required, the interdisciplinary nature of the paper is to be really praised and the authors congratulated for their collaboration. It does seem that assumptions feed into assumptions on a few occasions, but I suppose that is to be expected with a modelling paper that models what are for the most part poorly understood phenomena. That said, more transparency is needed in the text about the choices made and another reading of the literature discussed above is perhaps needed. I recommend the editors seek reviewers from the impressive range of fields this paper engages with.
Citation: https://doi.org/10.5194/egusphere-2026-5195-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 426 | 198 | 59 | 683 | 48 | 51 |
- HTML: 426
- PDF: 198
- XML: 59
- Total: 683
- BibTeX: 48
- EndNote: 51
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
I think this a deeply impressive and very important study which draws upon quantitative techniques to model how the interaction between climate change and bubonic plague in the sixth century may have served increase the societal impact of the latter.
The article is carefully considered, the historical literature and primary sources are judiciously handled, and the modelling is carefully done. Extensive use is made of comparative evidence pertaining to the later medieval Black Death, but this is justified by the authors to good effect. It is possible to argue that the authors perhaps overstate how much we currently really know about the Black Death, but ceteris paribus their approach strikes me as perfectly valid.
The argumentation and presentation are also a model of clarity. It is an important contribution to an important field.