Climate and the Ancient World: Beyond Present Concerns to Complications, Where Details Matter

Abstract

Current modern attention and concern about (human-driven) climate change has prompted much focus on the historical/archaeological relevance and role of (natural) climate change in the past. The topic is both relevant and important—and especially those short(er)-term events that perhaps helped trigger historically substantive change episodes. But, at the same time, initial, somewhat naïve enthusiasm has now run headlong into the limitations of the available data sources before the early modern era, and the many complications of establishing actual causal associations. These need to be, first, closely defined in terms of timing and effects, and then also, second, established as relevant to the specific human societies/civilizations and contexts in question. This paper seeks to highlight the need for appropriate care and rigorous method when seeking to associate climate and environmental events with the available ancient historical and archaeological evidence, and investigates three illustrative, problematic, cases from the Classical Mediterranean world.

1. Introduction: Avoiding “Black-Box Determinism” and “Suck-In and Smear”—Details Matter

Many studies which attempt to bring climate to bear as explanation or causal agent in history or archaeology struggle to avoid the fundamental challenge of first establishing that the climatic or environmental evidence they seek to associate with some historical/archaeological process or change is in fact (i) closely (temporally especially, and spatially) associated, and (since mere correlation is not causation) (ii) that there are plausible and relevant causal connections (e.g., [1]). As we move back into the past, this usually becomes more challenging: we must be very careful not to be tempted to try to associate different approximately dated events and evidence with something that is precisely dated (suck-in), while also being on our guard against the problems of the smearing of approximately dated evidence over a period of time (whether from the ranges from radiocarbon dating or from history) such that either we lose actual single, separate, specific events, or we end up creating false periods of change from what are really unconnected elements [2]. The further necessary step is to avoid what might be termed “black-box determinism”. A chain of plausible causality needs to be investigated and established. Human societies and their political structures (e.g., cities, states, empires) are highly complex and usually resilient within and among their several dimensions [3]. Demonstration that climate or environmental episodes caused “collapse” are often problematic when critically examined [3,4,5]. As Butzer [3] (p. 3638) observes with reference to western Europe since ca. 1200 CE, “This more recent historical experience … suggests that environmental or economic disasters do not necessarily lead to social breakdown or collapse”.

This essay seeks to do three things:

• Discuss some of the issues involved with bringing current climate change concerns into studies of the past;
• Address what types of climate changes are plausibly relevant to historically forcing episodes (inherently short-term), versus those that may form the general background varying from longer-term historical trajectory to evolutionary context (important but an entirely distinct category);
• Investigate three Classical cases wherein claims have been made for the relevance of climate or environmental effects (affecting climate) but where, upon examination, the actual details do not support the claim or raise questions about causal primacy and the relevant complexity involved.

2. From the Present to Past and a Necessary Separation

Increasingly modern politics and the world of the earlier 21st century CE, even in democratic countries, is divided/polarized and noisy (e.g., [6,7]), with the USA, formerly perceived as the “leader of the free world” post-WWII (e.g., [8]), now an exemplum of the challenge (e.g., [9]). This division spills over into, is reflected in, and shapes, approaches and academic responses across many fields from literature (e.g., [10]) through—and even centrally involving—climate change and any effective responses thereto (e.g., [11,12,13]), and again particularly in the USA (e.g., [14,15,16,17]). We have plenty of noise but relatively little clear signal on most big topics, and, in particular, on how to address current climate change.

Today, unless a climate-denier, the threat to lives and livelihoods from substantial and rapid climate change around the globe is regularly observed and stated—this modern climate change is, according to the vast majority (>99%) of scientists, primarily human-caused [18]. It is the collateral damage from the massive human transformation of the Earth since the beginning of the Holocene and especially agriculture [19], and, notably, the dramatically increased human activities and their discharges into the planet’s atmosphere, oceans, and total landscape following the mature Industrial Revolution (e.g., from ca. 1850 CE, as illustrated by greenhouse gas emissions [20]). It is in turn matched by an exponential increase in human population associated with this same technology-driven era of human change across recent centuries [21], and the associated and now widespread human presence/pressure (>65–75% of the planet already in two 2016 reviews [22,23]), habitat fragmentation (e.g., [24]), and endemic pollution from human-created materials like (micro-)plastics, everywhere ([25] frighteningly, the vast amount of micro-plastics actually attested on the seabed are but the tip of the iceberg [26]), including within our own bodies [27].

We are well into the fourth decade since the Intergovernmental Panel on Climate Change (IPCC) (https://www.ipcc.ch, accessed on 30 March 2025) was created in 1988, and, starting with the First IPCC Assessment Report of 1990, there have now been six IPCC Assessment Reports through 2023. Across the history of these increasingly sophisticated reports backed by more and more data and the work of a large international scientific community, the picture is grim (e.g., [28]). Guzman [29] reviews and provides numerous examples of the major threats to the world: for instance, in Asia, more than a billion people “face the risk of severe water shortages” (p. 117) because of impending climate change (i.e., more than 12% of the total world population)—and by 2016 it could be estimated that water scarcity affected 58% of global population in the 2000s [30]. Wallace-Wells [31] articulates the existential threat we all face. Climate change is and should be a focus of serious attention (irrespective of any politics) as a common threat to humanity’s future. One could go on and on. And this almost appears to be a fundamental problem itself. As in “the Boy who cried wolf” (one of Aesop’s Fables [32]), the repetition and dramatic scare-mongering around climate seems to inure. Thus, the “U.N. climate chief says two years to save the planet” (10 April 2024, Reuters: https://www.reuters.com/world/un-climate-chief-says-two-years-save-planet-2024-04-10/, accessed on 30 March 2025) becomes almost self-defeating, as for many people climate change seems detached from fundamental quotidian concerns. Climate change only ranks 11th of top problems facing the nation (USA) in a 23 May 2024 Pew Research center study (https://www.pewresearch.org/politics/2024/05/23/top-problems-facing-the-u-s/, accessed on 30 March 2025).

But, despite the impressive inability of the contemporary world as a whole to grasp these problems and to coordinate seriously effective responses to all our modern human-driven climate change and associated pollution and increasingly existential threats, there has been a marked “climate turn” in historical and archaeological scholarship. Is this reassuring or desperate—does this scholarship bring signal or sometimes more “noise”? From a relatively niche concern started in modern times by scholars like Lamb [33,34], in recent decades numerous academics writing about the past have increasingly focused on climate change as the lens or explanation for changes observed in the historical or archaeological records (e.g., [35,36,37,38,39,40,41,42,43,44,45]). Of course, this is not surprising: prehistory/history is always really about the present and is inherently political. Political and intellectual fashions in the present have shaped and continue to shape our interpretations of the past (e.g., [46,47,48,49]). In the present case, we may even observe the rapidly changing climate of academic fashion, noting how, for example, climate moved to become a/the key factor between the first (2014) and updated and revised (2021) editions of Eric Cline’s 1177 B.C.: The Year Civilization Collapsed [50,51] mirroring the increased attention on climate change as we have moved so far through the 2010s to the early 2020s. The cover of the 2021 revised and updated edition quotes Adam Gopnik: “Astonishing … with eerie relevance”—very much seeking to draw the present–past, past–present connection. However, the concern remains as described and critiqued by Butzer [3] (p. 3633): “Current research in historical collapse suggests a primary fascination with climatic change and environmental degradation as primary agents of change, but at the cost of less attention to the necessary cross-disciplinary integration. Indeed, the recent return to environmentalism is not about a fresh interest in the environment–society interface, but a continuing failure to appreciate the complexity of such inter-relationships”.

3. “The Past Is a Foreign Country” [52] (p. 7)—Issues and Questions for Climate History

There are important differences as we combine present and past. Modern climate change and the globally polluted and over-populated world is primarily a human-caused crisis. The vast majority of scientists, at least, more or less know what the causes and vectors comprise, and we have increasingly good records that characterize current changes and circumstances globally versus those of the past several centuries back at least a millennium or two (e.g., [53]). We can, for example, quantify 2023 CE as the hottest year globally over the past 2000 years [54]. Strategic action and resilience/sustainable planning ought to be possible. However, once we move back more than a few centuries, climate and history associations are predominantly about how natural climate changes affected and sometimes shaped human history. In contrast to the present, the people of these times were largely unaware of the processes driving their climate context. At best we have the summary observations of a great thinker, Aristotle, who highlights that rainfall across Greece is diverse, potentially varying greatly from place to place:

“And sometimes drought or rain is widespread and covers a large area of country, sometimes it is only local; for often in the country at large the seasonal rainfall is normal or even above the normal, while in some districts of it there is a drought; at other times, on the other hand, the rainfall in the country at large is meagre, or there is even a tendency to drought, while in a single district the rainfall is abundant in quantity” (Aristotle, Meteorologica 360b5-13, translation of Lee [55]).

But his explanations for why this is so, in terms of the different exhalations (see [56] (pp. 35–50)), fall well short of modern meteorological theory.

In reality, most pre-modern people could only know and use the recent and known past history of their area to construct strategies of resilience; changes beyond those reasonably anticipated from past experience could thus form potential but largely unknown threats (until they occurred). Thus, while past episodes of climate change may offer examples or laboratories of possible scenarios and human responses of relevance as potential analogues for the present and future (e.g., [57]), they are also different.

3.1. Lower to Low Frequency Climate Variations

The telescope of history brings one other key distinction into focus: in broad terms there are two fundamentally distinct forms of climate relevance or effects on history. The first is the lower to low frequency (that is, longer to long-term) cycles and shifts in climate regimes that have shaped the Earth over its history. These longer-term fluctuations range from the centennial scale (argued by some to be associated primarily with solar irradiance cycles [58,59] but cf., e.g., [60]), to the millennial-scale variations like the Dansgaard–Oeschger cycle associated with ocean circulation systems [61]. In addition, there are glacial–interglacial cycles linked with 100,000-year (eccentricity) and 41,000-year (obliquity) orbital forcing cycles (e.g., [62,63]) observed in long ice-core records back over 800,000 years (e.g., [64]—on-going exploration suggests the potential eventually to recover records reaching back beyond 2 million years [65]) and (with less resolution) back many millions of years from isotope records from deep-sea benthic foraminifera recovered from ocean-floor cores (e.g., [66,67]). All these categories of climate variations occur over time periods from much longer than, to vastly beyond, human generations to lifespans. The history of human adaptation and evolution, and human culture, was at least partly driven by these and earlier lower-to-low-frequency climate cycles and changes (e.g., [35,68,69,70]). This type of climate change was background context: humans were not immediately confronted with, or aware of, dramatically changing and challenging circumstances where decisions and strategies were straightaway necessary. Indeed, some such lower-frequency shifts could be to more benign/stable periods that generally favored growth in human societies (like the overall Roman era, e.g., [71]). Altogether, there were merely constant on-going adaptational processes which (with successes and failures) led to longer/long-term evolutionary outcomes.

Here, it is important to mention phenomena such as so-called “mega-droughts” (or similar)—the idea of extraordinary negative episodes lasting many decades, even centuries, and blamed in a very generalizing, reductionist, “black-box” logic for a variety of cases of collapses of civilizations/societies (see, e.g., [44,51,72]—and generally contrast with the critiques of such overly simplistic black-box thinking by, e.g., [3,4]). Yes, paleoclimate archives indicate variations over multi-decade to centennial-type scales. However, such longer-term variations—if not sudden and extreme—that become a new normal for a sustained period, are what humans are very adept at adapting to and, indeed, where climate may enhance/decrease effective environmental circumscription, such patterns of longer-term sustained shifts may play an important background role in driving economic, and sociopolitical evolution or devolution. Egypt offers a case wherein a reasonable argument can be made that periods (multi-decadal to centennial scale) of sustained climate conditions, shifts to periods of more extremes of higher/lower Nile floods versus periods of more stable (non-extreme) conditions, can either encourage or discourage the effective level of environmental circumscription of the agricultural population (and so the basis via resource mobilization, e.g., tax, to socioeconomic and political organization structures), and hence can be associated with the longer-term evolution or devolution of political stability and so economic and socio-political structures [73]. The 4.2 kya episode overall may offer another case with a range of outcomes depending on the local contexts and pressures [74,75]. Furthermore, it is also important to observe that cases of longer-term variations covering a few centuries, like the Little Ice Age, do not represent consistent “bad years”, rather a shift, usually in fact relatively modest, in average conditions. Despite the name and the impression of cold winters and frozen conditions, Lockwood et al. [76] have shown that, during the Little Ice Age period, there were in fact many warm summers and not all winters were especially cold (just cooler winters were more common), and overall, the Little Ice Age represents a climate shift (downturn) that was within the manageable range of many societies. Even in those areas of northern Europe that were most strongly affected, many societies carried on, and indeed some—notably the Dutch Republic—with its more open and inventive social, economic, and political structure found and took advantage of new opportunities managing or benefitting from the different climatic sub-phases from the mid-16th to mid-17th centuries CE (e.g., [77]). Obviously entirely marginal areas may become unsuitable for existing use frameworks over time during such a climate shift (e.g., a swing to much drier conditions making cereal agriculture problematic, or wetter and cooler conditions leading to a change from wheat to more tolerant crops like barley, oats, or rye), and, if too difficult, one adaptive response could be migration (habitat-tracking), while others include alterations in resource emphasis; for example, a move more to pastoralism or other strategies (e.g., [78]).

3.2. Higher to High-Frequency Climate Variations

The second type of climate variations are higher to high frequency over relatively short (decadal) to short, interannual, timescales, often associated with observed climate patterns like the North Atlantic Oscillation [79], the El-Niño-Southern Oscillation [80], the Pacific Decadal Oscillation [81], or Atlantic Multi-decadal Variability [82]. Many of the types of climate anomaly associated with such short-term variations are just that: short, in particular one-year, and not necessarily especially extreme in the longer-term, historically forcing, sense. Let us consider three examples:

(i)

Western Anatolia where Köse et al. [83] reconstruct precipitation from a network of tree-ring (Pinus nigra) sites from 1632 to 1929 CE covering 298 years. While there are 85 years (or 28.5% of years) that exceeded 1 standard deviation (SD) from the mean (43 dry, 42 wet), only 12 years exceeded 2SD from the mean (6 very dry, 6 very wet). Furthermore, most (31 of 43) of the dry or very dry events were just one year, only two cases of very dry events lasted 2 consecutive years, and one case lasted 4 consecutive years. Similarly, the wet events were mostly one year (31 of 42) and only two of the very wet events comprised 2 consecutive years. Thus, in total, consecutive multi-year extremes occurred over just 4% of years.

(ii)

In Medieval Egypt, the available Nile flood records indicate longer and short-term variations (Hassan [84]) (Figure 1A–C)—thus, indicating fairly commonly occurring changes/challenges that people were well aware of. But, looking at the 289 years from 944 to 1232 CE, there were just 22 years where a ≥2 consecutive years of abnormal Nile flooding occurred associated with evidence of catastrophic impacts [85] (Table 2); thus, in about 7.6% of years. Alternatively, assessing flood data more generally from the 7th to 15th centuries CE, Grins [86] (p. 109) observed of 820 floods that just over a quarter (27%) of years saw floods that would not have been recognized as good/propitious, but only up to about 12% of floods would have caused serious alarm (5% destructively high, 7% too low for cultivation). Looking across 7 centuries in the pre-modern second millennium CE, Morris [87] (pp. 6–8) notes six periods of multi-year famines (variously 3–7 years) in Egypt associated with Nile flood irregularities (either too low or too high). Thus, here we see challenges with a frequency of around once a decade (ca. 8–12% of years), but with disaster and especially multi-year crises that were infrequent-to-rare in general (some periods saw heightened risks associated with major explosive volcanism, e.g., [88,89]). Nonetheless, the unusual singular annual dependency of Egypt on the Nile flood made this risk and occasional disaster both severe enough, and it reoccurred sufficiently regularly over time, to form a haunting fear that became ingrained into long-term social memory and practice in Egypt [86,87].

(iii)

India, where from 1871 to 2002 CE Ó Gradá [90] (Table 2) identifies, from examination of, respectively, India, or East Rajasthan, or West Rajasthan, that 20, 21, and 14 extreme droughts occurred and 18, 20, and 17 extreme floods. That is occurrences in between 11% and 16% of the years covered in the respective regions. However, 2-year consecutive droughts (back-to-back as [90] characterizes) are very rare, just once for India and once for West Rajasthan; thus, in 1.5% of the years covered for the respective region. Floods in 2 consecutive years are also much less common, occurring for India just twice, in East Rajasthan five times, and West Rajasthan four times, or in between 3% and 4% of the years covered.

The relevance is that, in many areas of the world, occasional negative-to-extreme 1-year weather events occur and people living in any particular region will expect these with some sort of approximate frequency according to their lived local experience (and that of parents, grand-parents, family, and friends). Events that occur above about once a decade (so around or ≥10% probability per year) were undoubtedly frequent enough that anyone surviving to adulthood likely would have experienced such an event at least once, and those surviving to become parents and those reaching their later 20s to 30s, or older, probably two or more times. Parents, grandparents, family, and friends would add more to such an experience base and collective social memory. Thus, some adaptations to try to be prepared for such known possible challenges may be expected as well as other social strategies to address such anticipated challenges (e.g., [91,92]). For such reasons, across the pre-modern Mediterranean, Halstead [92] (p. 162) reports the aspiration of recent traditional Mediterranean farmers to store enough grain for a second year “so they could ride one total crop failure” (citing, e.g., [93]), while noting that this was not always achieved. And on a wider global basis, we may observe various traditional/ancient strategies to try to adapt to survive “bad years” as relevant to region and society (e.g., [91,92,93,94,95,96,97]).

3.3. Back-to-Back, or Consecutive, Multi-Year Climatic Extremes

In contrast to the anticipated, if feared, known challenges of a “bad year”, the instances of two or more consecutive years of such climatic extremes are much less common in most cases. As in the three cases noted above, two or more such consecutive, or “back-to-back”, extreme bad years only typically occur in a small percentage of years. And rather than once or more a decade, it is perhaps more often once a generation or even less, and only a few times a century, or even less. Such relatively rare events—even if known and feared in social memory (e.g., [86,87]) (and Egypt with its critical annual flood, long-recognized as central to life, and the state, is an unusual case and so the focus of rare early bureaucratic attention—Nilometer records, e.g., [98]; see in general [99] (pp. 26–35))—are much less likely to be anticipated in an everyday practical way, and were unlikely the focus of any regular strategy: they are not really expected (until too late), and are less likely to be part of routine individual or collective adaptation strategies and resilience models for any particular society.

With reference to the ancient Greek world, Garnsey [100] identified this important bifurcation of risk and of human risk adaptation. The comparison of mid-20th century data from 1931–1960 CE for Attica against the approximate plant moisture thresholds for wheat (300 mm per annum), barley (200–250 mm per annum), and dry legumes (350–400 mm per annum), indicated that, in any one year, there was a 71% probability of dry legume failure, a 28% probability of the wheat harvest failing, and a 5.5% probability for barley failure [100] (p. 10). Thus, there was a fairly regular likelihood of some form of annual food insecurity or scarcity and a review of ancient evidence confirms this. Hence, Garnsey [100] (p. 17) observes that “Subsistence crises were common in antiquity”. This was something to expect and to try to plan for—the one bad year strategy. However, [100] (p. 17) then asks how often does such an expected, relatively frequent, instance of food shortage to subsistence crisis “assume the proportions of famine?” This means a subsistence crisis that is not just 1 year but occurs over consecutive years and so overwhelms the adaptative efforts towards resilience via “bad year” economic practices through plausible storage or social strategies. As in the examples above Garnsey [100], (p. 17, Table 3) finds that 2 or more consecutive “bad years” with harvest failures are much less common to rare. In Athens, for wheat, the 1-year probability of 28% becomes only a 7.8% probability for 2 successive years, and for barley the 1-year probability of 9.7% drops for 2 successive years to just 0.3%—thus, less than even once a century. Data from Larisa in northern Greece are similar: for wheat, 28.5% (1 year) to 8.1% (2 years), and for barley 9.7% to 0.9%. Garnsey also looks at Odessa in the Black Sea: wheat 46% (1-year probability of failure) to 21.1% (2-consecutive years probability of failure) and barley 15.6% (1 year) to 2.4% (2 years). Thus, even here where the failure probabilities are higher, a fall back to barley means that a no-cereal situation occurs only maybe 2–3 times a century. “Thus, where drought is an important precipitating cause of food crisis, genuine famines are much rarer than mere shortages” [100] (p. 17).

The end result, compounded by the many difficulties in pre-modern times of storing staple foodstuffs, even the less-perishable, for periods beyond 1–2 years (such as (i) practicalities of suitable space and climate except where very dry, as perhaps Egypt (see [101]), or special one-off oxygen-free cases like storage pits, (ii) unwanted germination, (iii) issues of pests, and (iv) mold and other microorganisms) (e.g., [92] (pp. 162–163), [102,103]), is that, while many ancient and traditional societies had strategies or intentions to attempt to try to cope with anticipated single “bad years”, they did not usually have adaptative mechanisms in place for rare and thus effectively unexpected occurrences of 2 or more consecutive “bad years” and thus famine. Raphael [104] (p. 56) provides an example from the medieval Levant writing that “even well-organized regimes found it hard to cope with long periods (more than two years [my italics]) of food shortage”, and, after a review of granaries and storage [104] (p. 66) concludes that “the medieval Middle East was ill equipped to battle long-term droughts and famines”. Ó Gradá [90] provides other examples where “back-to-back” bad years of extreme weather led to genuine crisis and famine (and this is leaving out the roles of warfare and other forms of human agency in causing or exacerbating famines [105] and the increasing evidence of a general association between climate changes/extremes and conflict, e.g., see discussion in [106]).

One other major difference comparing modern circumstances with those in the past revolves around communication (and so information) and infrastructure. Rapid long-distance communication was a considerably greater challenge in the pre-modern world; typically only the elite and especially the ruling structures of states and especially empires could invest in and support forms of relatively rapid (and, even then, sometimes seasonally or geographically limited or fairly slow—e.g., Middle Bronze Age Assyrian contacts with Anatolia were not possible in the winter when roads were closed for about 4 months: [107] (p. 75)) communications structures (e.g., [108,109]). For most in society, long-distance communication or access (and so information) was limited at best or impossible. Moving large quantities of bulk goods was also a challenge. Transport by water, and especially since sailing ships became available, made bulk transport possible [110,111], but over land this was considerably more challenging until the Industrial Revolution and railways. In what has become the orthodox assessment, Finley [112] (pp. 126–127) argued that the available bulk land transport options (ox power principally) were limited by possible distances and thus costs/practicalities, except for a few state-level efforts (such as using large numbers, e.g., 30, oxen for each marble column drum for a major temple construction), and referred to the passage of Pliny the Younger (Epistles 10.41.2) where, in his letter to the emperor Trajan, Pliny explains that, in contrast with water, transport by cart requires much labor and considerable expense. Finley [112] (p. 126) therefore argued that “towns could not safely outgrow the food production of their own immediate hinterlands unless they had direct access to waterways”. But, as Laurence [113] discusses, it is clear, despite this orthodox view, that roads became key infrastructure in the Roman world notwithstanding the higher terrestrial/road costs (see also [114]). Recent modelling from the distribution of materials recovered from Roman period archaeological sites in lowland Britain by [115] further questions some of this orthodoxy based on literary evidence and Diocletian’s Edict on Maximum Prices [116]. Wiseman et al. [115] find overall that transport by road was (best-fit values) about 1.6–3 times as costly as by river and about four times as costly as by sea, which as an overall heuristic they turn into an approximate 1:3:4 ratio (with 95% confidence intervals of river 1:1-5 and sea 1:1-9).

All this leaves any specific area in the pre-modern world vulnerable if it is hit by an unexpected and rare two or more “bad years”. If a port locus then, with communication, grain, and other transportable food products could potentially be sought, or bought and brought, from elsewhere (and various requests for such relief are known in ancient text sources, e.g., [117] (Table 2)), and might be shipped in if these other areas were not also afflicted (noting the limitation that cereals are much easier to harvest and transport on a larger scale compared with pulse crops: [92] (pp. 103, 105, 114)). An inscription from Cyrene discussed below offers such an illustration of grain from North Africa being provided to cities in Greece at a time of grain shortage (Supplementum epigraphicum Graecum, SEG ix 2, [118] (inscription no. 96), [119]), and such trade—largely maritime—in key staples became central to the support of major urban centers, like Athens and Rome, and a source of considerable profit and state focus [100,120], [121] (pp. 206–257). But addressing the terrestrial settlement landscape more widely, many cities, towns, and regions in the pre-modern period lacked such potential communication and bulk-transport potentials and thus rare to very rare instances of consecutive to even multiple “bad years” likely could or would have overwhelmed storage and resilience practices and led to famine and potential disaster. Here, as discussed by Garnsey [100] (pp. 22–23), although human factors like profiteering often came into play, the limitations of land transport are attested in ancient sources as creating real challenges for inland cities with food crises. In turn, these cases of real crisis, reaching upwards, would subsequently seriously affect regional and inter-regional tax revenues, economic structures, and so on (but by the time communication and information from these became systemically known, highlighting the problem, the original cause was already critical, and likely no effective remedy was immediately possible). In this way, occasional instances of what form effectively extraordinary episodes may prove historically forcing where no help is at hand. An example are three consecutive very dry years in central Anatolia ca. 1198–1196 BCE recognized from tree-ring growth information, within what appears from stable isotope information to be a longer-term shift (centennial-scale) to somewhat drier conditions. This rare confluence not attested for centuries before or after, since coincident, might therefore form a substantive part of an explanation for the apparently sudden end of centuries of the Hittite Empire and its administrative system [122] (and see now further in [123]). Another similar recently investigated case is the role of an unusual three-consecutive years of severe drought 364-366 CE and consequent harvest failures as a likely driving factor behind the massive rebellion known as the “Barbarian Conspiracy” in Roman Britain around 367 CE [106].

4. History, Climate, and Environmental Changes: It’s Complicated

Confronted with data, problems and questions, the usual human (or AI) response is to try to simplify: to reduce “noise” to patterns or signal and then to offer some form of assessment. In general terms, simplification forms a basic cognitive strategy to address an endlessly complex world given limited mental resources (brain or processing power) and available time. And, with a complementary logic, the principle of parsimony (Ockham’s Razor) is regarded as a basic of Philosophy, holding that simpler explanations are superior (and likely to be true) compared with those that are more complex [124]. The identification of (or reducing data to) the simplest pattern(s) or elements—by various means—is regarded as a fundamental of the Gestalt school of psychology and cognition and related approaches, through modern applications in, for example, image analysis, and other complex topics (e.g., [125,126,127,128]). But, despite its central role in all human dealings with the world, and while never perfect, to be robust, simplification must be appropriate. It must, in particular, take into account all the relevant parameters and data—not just (subjectively) selecting out a few observations or impressions and then using these and so creating a biased or even entirely flawed analysis (an obvious example is political debate where various simplifications from the pool of information and misinformation available often offer differing, different, and even contradictory outcomes).

Efforts to associate climate or environmental changes with history are very much a case where over-simplification can often seem enticing, but care is necessary to establish whether this is robust. Fundamentally, it is key to remember that correlation is not causation. Coincidence in timing and place may make an association possible, but that is all (and of course failure to securely document such real propinquity undermines any claims for a climate–history association). Establishing whether there is an actual plausible causal association requires detailed investigation and substantive articulation—a point made in various ways and in various contexts several times (e.g., [1,3,4,5,129]).

Both the science side (climate and environmental data and the processes that drive, affect, and propagate these), and the human side (history, politics, society, culture), are inherently hugely complex. Furthermore, from the human side, decisions and actions by individuals and groups will often derive from and involve many different factors, some of which may be largely independent, and some of which may interact. The topic of the relationship between climate and history is a case wherein we must guard carefully against easy but inappropriate simplification, especially as we usually lack much relevant direct testimony. This is not a new statement. Ingram et al. [130,131] provide a critical review of historical climatology up to that point (now more than four decades ago) and they noted then the issue—looking at existing work—of needing much greater critical attention to the sources of information employed in order to build reliable data from which climatic reconstructions are plausible and then whether associations may (or may not) be drawn with human history. But, while in many ways the modern and critical field of climate history begins with the critique of Ingram et al., it remains very much the case that a focus on efforts to robustly link climate with history can often receive insufficient attention.

The idea that climate (and climate change) could be relevant to history is far from new. When Thucydides 1.23 reviews the factors leading to the Peloponnesian war—which he regards as unparalleled in Greek history (as 1.1)—he concludes pointing principally to the rise of Athens to greatness and the associated fear of the Lacedaimonians (1.23.6). En route, however, in 1.23.3, Thucydides refers to various natural events and disasters (a recurring reality in the Mediterranean [132] (pp. 298–341)), finishing with mention of great droughts in some areas with famines and in particular the plague. While ending up identifying a central human agency, the recognition of the potential relevance of an environmental/climate/disease context is striking. But this example also nicely highlights the challenges of moving from noting some disasters or events and speculation to clear and specific historical causality.

In the final part of this essay, I therefore wish to investigate three Mediterranean Classical World examples from recent work which serve to draw out some of the complexities in trying to tie history and climate together without a careful and critical review, both of the evidence used, and how this evidence relates to the relevant history to which a climate association is claimed.

4.1. An Inscription and a Major Drought 330–328 BCE?

In an article usefully reviewing sources, topics, and the limitations of current evidence for environmental–climate history of Classical and Hellenistic Greece, Ruben Post [133] (p. 6 of 12) notes that a high-resolution climate record from dendrochronology:

“would be particularly useful for the study of the Classical and Hellenistic periods given the existence of well-dated written evidence for climatic anomalies, such as an inscription recording a major drought throughout Greece 330–328 BCE” ([133] (p. 6 of 12) citing [118] (inscription no. 96)) (my underlining emphasis).

This “well-dated written evidence … an inscription recording a major drought throughout Greece 330–328 BCE” appears to be important evidence. The inscription in question is from Cyrene in Libya, North Africa, on one face (of four, two others have the sacred Law of Cyrene) of a large white marble stele later reused in Byzantine baths ([118] (no. 96), [119], [134] (pp. 205–207), SEG, ix 2). Investigation proves salutary. Reading the presentation of Rhodes and Osborne [118] (no. 96), the date for the inscription is only estimated and as “c.330–326” (p. 486) and they further qualify the dating of the information in the text by stating “We do not know whether the grain was sent in a single year or more than one year. We do not know in which year or years the grain was sent”. For example, Garnsey [100] (p.159) observes the inscription is “normally placed in the period 330–323 (although 332/1 …recently … suggested by an ingenious argument)”. Thus, the date is not given in the text, and could cover almost a decade and not necessarily the fairly specific date of 330–328 BCE suggested by [133]—although a date likely ca. 330 BCE to the beginning of the 320s BCE seems probable (as [119]).

However, rather more fundamental: where in the text is information allowing the description of an “inscription recording a major drought throughout Greece”? In fact, the text only records:

“Priest: Sosias son of Kallias. These are those to whom the city gave grain during the grain shortage (sitodeia) in Greece. To the Athenians 100,000, to Olympias 60,000 … [and to another 49 (eight repeated) names of cities and two individuals]” [118] (pp. 486–489). Rhodes and Osborne [118] (Figure 2) show a map with all the locations mentioned covering much, but notably not all, of mainland Greece (also [119] (Figure 4.2), Horden and Purcell [132] (p. 73 Map 8))—in particular, most of the Peloponnesian poleis are absent bar those in the northeast (and Elis—which however receives only a modest allotment). While it has been suggested that absences may reflect politics around Macedon (for or against) and/or Cyrene’s network of partners/allies [119] (pp. 73–74), it seems more likely that the majority of the Peloponnese—relatively fertile in agricultural terms—did not require grain [134] (p. 205). The absence of many northern Greek poleis and many in Asia Minor is difficult to explain if there was truly an Aegean region major drought (and in more recent times major basin-wide droughts encompass more or less all of the Aegean (including north Aegean) and western Anatolia: [135] (Figure 5), [122] (Figure 3)). Bresson [119] (p. 74) speculates that perhaps Pontus was responsible for “feeding northern Greece”, but there is no positive evidence for this.

The text enumerates in all 805,000 (we can assume medimnoi) of grain. It mentions a “grain shortage in Greece”. Nowhere does it actually mention the cause and specifically there is no mention of a “major drought”. As Rhodes and Osborne [118] (p. 488) observe, “…this apparently straightforward text is in almost all respects obscure”—much though Bresson [119] seeks to add likely interpretation and context. We do not know whose medimnos measure was used (the Aeginetan/Laconian one was 50% larger than the Attic), which type of grain was sent (e.g., wheat or barley), nor whether this was all sent in one year or over several years. On the last, Bresson [119] (pp. 85–92) makes a reasonable case from linked observations that the sending of the grain did perhaps occur in a single year, and based on parallels, we might assume that Cyrene both gave (allocated) grain to some of the cities or people on the list, but also likely made available (i.e., sold) grain to the sitōnai (public grain commissioners) seeking grain for their cities. The productive potentials of Cyrene and its territory and early harvest date are outlined by [132] (pp. 65–72). But no actual date is given. And no actual reason is stated. As Rhodes and Osborne [118] (pp. 489–490) then observe, the mentions of Olympias, likely the mother of Alexander the Great, and Cleopatra, likely Philip II’s daughter/Alexander the Great’s sister, are both the best dating evidence and indicate a probable political motivation (whether or not any climate or environmentally related factor may have also been involved). Carney [136] (p. 51) makes the case that Olympias and Cleopatra are named because they “were functioning here as heads of state”. For Cleopatra this could either be as regent from 334 BCE when her husband and uncle (Alexander I of Epirus) went to fight in Italy, or perhaps best from her period as regent and ruler following his death in 331 BCE. Hence, widest likely dates are between ca. 335 (after Cleopatra’s marriage) and probably 334 BCE to 324/323 BCE (with the political conflict in Cyrene from 324 BCE leading to its takeover by Ptolemy I in 321 BCE [137] (pp. 92–93), and/or death of Alexander the Great in 323 BCE, setting a terminus ante quem), and more likely, since Cleopatra is named, ca. 330 BCE or a few years later.

It may be noted that some other sources indicate grain shortages and/or grain gifts 330/329 and 328/327 BCE, and these may have climatic associations, although as Rhodes and Osborne [118] (p. 490) further observe, it seems likely that warfare and other strife at these times may have also contributed (and, as the study of Manning et al. [88] illustrates looking at Egypt 305–30 BCE—a very little later—with higher resolution historical and climate-proxy data series, both issues are often intertwined and thus co-correlated). For example, we hear of production deficits and restrictions on grain exports in Egypt, and profiteering by its governor, Kleomenes, at the beginning of the 320s BCE [119] (p. 77, 84). Such actions may well have affected wider grain supplies also. Indeed, more generally, given the timeframe involved, it might be suspected that an obvious elephant in the room is being missed in considerations, since we may assume that Alexander the Great’s (and associated allies’ including Alexander I) many and enormous military activities across the region and then beyond 336–323 BCE both created many upheavals across the entire eastern Mediterranean region and their own major provisioning–supply needs (as Duncan-Jones [138] (p. 104) highlights in his review of [100]; for detailed analysis of feeding and supply of Alexander’s army, and the implications for local crops and hence massive local to regional disruptions, see [139]). As a slightly later example, considering the grain shortage in the Peloponnese ca. 191–188 BCE, Post [134] (pp. 200–202) considers the possible relevance of various military activities and the requirements to provision armies, and, having reviewed the evidence, observes for example that “The period from 191–189 BC was thus one in which the Roman state’s demand for provisions was exceptional, evidently straining its grain supply system”. Likely much greater mobilizations and regional demands applied in the late 330s to early 320s BCE.

Returning to the Cyrene inscription itself, there is no actual evidence to relate the Cyrene gifts to any of these other events—but the mentions of Olympias and Cleopatra do point to the relevance of Macedonia. In all, reduced grain supplies which, among other factors could be because of drought or other environmental causes, enormous and many military activities in this period, allied with political considerations—and particularly relations with the ascendant Macedonians—may singly or together be behind this inscription, but we really can only guess, and the actual date is simply not known. Thus, despite the frequent mention of this inscription as supporting a period of famine (already extrapolating from a sitodeia) ca. 330 BCE (e.g., [140] (p. 99)), the cause of this sitodeia, and whether there was in fact a drought (or other environmental issue), and the exact date, are unknown.

While we at present lack high-resolution climatic data for the Aegean for this time period, more recent evidence from the last millennium raises two further and somewhat contradictory issues. Analyzing annually resolved tree-ring evidence for drought in the Mediterranean over the last 900 years, [135] (p. 2065) observe that their “Figures 5 and 6 … show evidence for antiphasing behavior in the eastern end of the basin, where there is a tendency for Libya, Egypt, and the southern Levant to be wet or near normal when Greece and Anatolia are in drought”. They relate this circumstance possibly to the North Atlantic Oscillation (NAO) [135] (pp. 2063, 2066, Figures 3 and 7). This observation could be used to argue that, as stated in the inscription, Cyrene was able to supply grain at a time Greece experienced drought because the area around Cyrene did not. But, there is also an obvious complication! Egypt (at least northern Egypt) and Cyrene (north Libya) appear usually to experience typically similar conditions probably as linked with the NAO [135] (pp. 2063–2066, Figures 3–7). Hence, when part of the Cyrene discussion links to stated problems with grain supply in Egypt 330/329 and 328/327 BCE, this seems less plausible since typically Egypt should have shared the Cyrene climatic circumstance (both positive or both negative regarding likely precipitation and grain harvest). As in all aspects of a review of this notable but difficult inscription, therefore, the totality of the available evidence is less than clear-cut regarding the cause or causes of the sitodeia mentioned and the exact date of these circumstances.

4.2. Aristotle and a Mycenaean Drought?

Discussing the period around 1200 BCE and the topic of societal collapse, Kaniewski et al. [141] (pp. 370–371) write:

“There are no written sources with direct information on climate for this period except Aristotle’s statement about the Mycenaean drought around 1200 BCE”. (my underling emphasis).

The name Aristotle of course draws immediate attention. The above quoted sentence ends with a footnote in [141] (footnote 25). This refers the reader not to Aristotle directly but to an article of 30 years earlier by Neumann [142]. Moving to this article, [142] (p. 441) begins by noting that:

“Aristotle’s pronouncement has a measure of relevance to the ‘Mycenaean drought (1200 B.C.)’ problem which led to an interdisciplinary controversy among some classical archaeologists and historians as well as meteorologists. In fact, it was this problem that aroused part of the stimulus to undertake the literature search”.

Neumann [142] (p. 445) addresses the text of Aristotle (Meteorologica 1.14) using the Loeb translation of Lee [55] (dated 1952—[142] gives the date as 1962). He introduces the argument of Aristotle that slow changes—well beyond any one human lifetime—in landscape (relating to wetness/dryness) occur such that (352a7–16):

“… places that formerly enjoyed a good climate deteriorate and grow too dry. This has happened in Greece to the land about Argos and Mycenae. In the time of the Trojan War Argos was marshy and able to support few inhabitants only, while Mycenae was good land and therefore the more famous. Now the opposite is the case for the reason given above: for Mycenae has become unproductive and completely dry, while the Argive land that was once marshy and unproductive is now under cultivation”.

Aristotle is talking in 1.14 about land–sea changes and the role of rivers. Hence, Aristotle states “The same parts of the earth are not always moist or dry, but change their character according to the appearance or failure of rivers” (351a19-20). As [56] (p. 172) summarizes: “According to Aristotle, districts of the earth undergo periodic senescence and rejuvenation according to the amount of wet exhalation present in the ground. A district in its youth is moist and rich in vapors”. Aristotle in particular sees a progression that starts from new land that is too wet. Egypt is his example of this, writing: “But this [too wet] land changes in its turn and in time becomes thriving. For as places dry they improve, and places that formerly enjoyed a good climate deteriorate and grow too dry” (352a6-8). Aristotle observes that these processes and the changes take place slowly over long time periods, much longer than human lives, and given that people die in wars, because of pestilence or plague, etc., there is thus no record of these changes preserved by humans, and people thus forget (351b8-22).

Aristotle employs rivers as the linking element in a complex argument, and one against his opponents (the Ionian “entropists” as [56] characterizes them), in which he contrasts but seems causally to associate changes in wet/dry areas:

“The various species of the genus of moisture need not in their own right have a causal connection among themselves. Rather, the compelling reason for positing a causal connection between vapor, rivers, and the sea is the polemic against the entropists. The apparently paradoxical, but clearly intended, result of Aristotle’s scheme is that one district by becoming wet can make a neighboring district dry. In two distinct dynamics the presence of one opposite in one place causes the other opposite to appear elsewhere” [56] (p. 173).

It is in this context of slow and effectively unobserved (by humans) long-term change that Aristotle comes to the Greek case when he tells us the information quoted above from his Meteorologica 352a7-16 comparing Argos and Mycenae contrasting the past with the present stating that in his own time “Now the opposite is the case for the reason given above: for Mycenae has become unproductive and completely dry, while the Argive land that was once marshy and unproductive is now under cultivation”.. (my underling emphasis)

On the question of the specific reason, see the text emphasized (underlined) above, Aristotle is not in fact explicit. The implication appears to be that the relevant local river, the Inachus, is the cause of the change (also [56] (p. 174)). On the question of the timing the text is clear. Aristotle is saying that, between the time of the Trojan War and his own time (so classical period and in particular the 4th century BCE), formerly moist Mycenae has become dry and the formerly swampy Argos has become arable (352a9-14). Yes, the majority of classical sources place the Trojan War somewhere around 1200 BCE, give or take a margin of a century or two. For example, the first historical source available to us, Herodotos (2.145), although demonstrably poorly informed on historical events and dates before about 700 BCE [143], nonetheless indicates a date for the Trojan War somewhat more than 800 years before his own time (mid-later 5th century BCE), but less than 900 years earlier (placed earlier and linked with Heracles instead), and so perhaps a date somewhere ca. 1300 BCE to mid-13th century BCE. Subsequent more detailed chronographic efforts, such as Eratosthenes especially and his date of 1184/1183 BCE, and others (see [143] (pp. 142–143), [144] (esp. pp. 246, 249–250)) led to dates mainly between the 14th and the 12th centuries BCE. Thus, Aristotle (384–322 BCE: [145]) does indeed tell us that, between ca. 1200 ± 100 BCE (give or take) and the mid-later 4th century BCE, a period of ca. 750–950 or so years, formerly moist Mycenae became dry and the formerly swampy Argos became arable. But this was, according to the theory he expounds in Meteorologica 1.14, a gradual change imperceptible to people at the time. Therefore, there is absolutely no evidence or statement by Aristotle about a “Mycenaean drought around 1200 BCE” as claimed by Kaniewski et al. [141] from Neumann [142].

Neumann [142] (p. 445) comments that “Aristotle’s … important statement appears to have passed unnoticed by those who participated in the controversy over the ‘Mycenaean drought’ problem”. An obvious response and explanation is that Aristotle says nothing about any Mycenaean drought at all. And when Neumann [142] (p. 446) then goes on to write “that Aristotle’s statement concerning Mycenae ‘confirms’ our picture that the period from about 1200 was relatively warm and dry in Greece…” we entirely leave Aristotle’s actual text behind. Aristotle specifically states that, around the time of the Trojan War, Mycenae was good land—he does not say warm and dry. It only became (too) dry later sometime across the 750–950 or so years between the Trojan War and Aristotle’s time (and the reverse for Argive territory).

4.3. Okmok Volcanic Eruption in 43 BCE and the “Fall” of the Roman Republic

McConnell et al. [146] argue that the climatic effects of the massive Okmok volcanic eruption in Alaska in 43 BCE, during what was already a cooler period, created substantial effects (very cold years in 43 and 42 BCE) that played a part—the authors note accounts of unusual climate, crop failures, famine, disease, and unrest—in what they describe as the “fall” of the Roman Republic, and, not satisfied with just this, they include also the end of Ptolemaic Egypt (last ruler Cleopatra VII). The abstract of the article ends stating:

“While it is difficult to establish direct causal linkages to thinly documented historical events, the wet and very cold conditions from this massive eruption on the opposite side of Earth probably resulted in crop failures, famine, and disease, exacerbating social unrest and contributing to political realignments throughout the Mediterranean region at this critical juncture of Western civilization. [146] (p. 15443)”

This study is the product of 20 authors in total, including several major figures in ice-core and climate science as well as some well-known ancient historians. The idea of a volcanic association is not entirely new. Mention in ancient sources of an eruption of Etna in Sicily at this time had already prompted suggestions that some of the unusual occurrences noted and linked with the assassination of Julius Caesar had a possible volcanic association [147]. However, this new work identifies Okmok as a dramatically larger, and thus much more climatically effective, volcanic eruption and main would-be culprit, and the new ice-core data are high-resolution, with the events identified dated accurately to the year. In order to further interpret scale and effects, the Okmok eruption ice-core signal is linked also with absolutely dated tree-ring evidence as well as speleothem data, and climate impacts are estimated via computer modelling employing the Community Earth System Model ([148], and subsequent versions, see: https://www.cesm.ucar.edu, accessed on 30 March 2025). This is a state-of-the-art fully coupled global climate model. The data and modelling show that the Okmok eruption likely caused substantial cooling with the authors stating: “The CESM simulations suggest that the Okmok II eruption in early 43 BCE resulted in 0.7 to 7.4 °C seasonal cooling in specific regions of southern Europe and northern Africa, with cooling especially pronounced during summer and autumn” [146] (p. 15447). The authors on the same page note a variety of ancient sources that mention cold or stormy weather, issues with agriculture, and instances of famine, and so on. The challenge is whether a connection can be drawn with wider Roman history and all the issues noted by the various historical sources.

To begin, the scale of the effects postulated and their spatial reference both need care—if we examine the reconstruction in [146] (Figure 4A,B). The reconstructed temperature anomalies affecting most of Italy (including around Rome) and several of the main grain supply areas are in the more modest −2.0 to −2.5 °C range and only some in the −3.0 to −2.5 °C range and few in the more dramatic change ranges. The authors caution on the challenges of accurate precipitation reconstructions, but, nonetheless, the more strongly affected areas are the very west Mediterranean and parts of the Aegean, Anatolia, and the northeast Mediterranean, whereas much of Libya and Egypt and almost all of Italy and Sicily are in the smaller two categories of positive precipitation increase (0–3 mm and 3–6 mm per month).

Given the paper’s title and dramatic claims, and the not surprising substantial media coverage, it attracted criticism from some quarters. Strunz and Braeckel [149] (p. 32207) complain that McConnell et al. [146] does not provide the “detailed analysis of integrated socioenvironmental mechanisms [that] would be indispensable to overcome ‘black-box determinism.’” They conclude that “a superficial correlation between climatic and social events cannot substantiate the purported effects”, noting Butzer’s [3] (p. 3633) critique of the “continuing failure to appreciate the complexity of [human–nature] interrelationships”. McConnell et al. [150] (p. 32209) respond to the criticism and point to the large climate episode they document and model arguing that:

“It is only logical to conclude that such an extreme climate event—including the second- and eighth-coldest years of the past 2500 y at the start of the fourth-coldest decade—had a significant effect on food production and society during this already tumultuous, critical juncture of antiquity”.

However, it is also true that this is a classic case where one needs to carefully investigate whether correlation is actually causation. Indeed, if we consider some of the historical events that [146] (Figure 2A) and [150] (Figure 1) identify and propose to associate with the Okmok eruption, then there are complications. Let us examine:

• McConnell et al. [146] (Figure 2A) and the improved McConnell et al. [150] (Figure 1)—see https://www.pnas.org/doi/full/10.1073/pnas.2019906117, accessed on 30 March 2025—note “food scarcity in Rome” as starting specifically from 43 BCE—more than implying that it was caused by the eruption at the same time. However, if we examine the ancient source material, there are indications of shortages of food in Rome 49–43 BCE (e.g., [100] (pp. 201–202)), i.e., before the Okmok eruption in 43 BCE, and these are stated as caused by military and political threats. And these military and political threats and food shortages continue through to 36 BCE. Garnsey summarizes from statements in several contemporary ancient sources ([100] (p. 202 n. 15) taking evidence and quoted text from, respectively, Cicero, Epistulae ad Atticum, 7.9.2 and 4; 9.9.4; Epistulae ad familiars 14.7.3; Appian, Bella civilia 2.48; Cassius Dio 41.16.1):

“The civil war that broke out in 49 between Caesar and Pompey put the inhabitants of Rome at risk. Blockade of the city was more or less inevitable, as Cicero had hinted in a letter of December 50. In March 49 Cicero reported to Atticus Pompey’s ‘first plan’: it was ‘to throttle Rome and Italy and starve them, then to lay waste and burn the country, and not to keep hands off the riches of the wealthy’. A few days later Cicero was gloomily predicting a terrible war, ‘ushered in by famine’. He was aware that a large fleet was being prepared ‘to cut off the supplies of Italy and blockade the grain-producing provinces’. In June he advised Terentia and Julia to head for the family estate at Arpinum if grain became more expensive. When Caesar arrived back in the city from Spain, he found the people ‘starving’. He had grain brought in from the islands and distributed”.

Thus, Okmok perhaps exacerbated a situation ongoing for some years, but did not start it nor necessarily form the prime cause.

2.

In the more elaborated work by McConnell et al. [150] (Figure 1—https://www.pnas.org/doi/full/10.1073/pnas.2019906117, accessed on 30 March 2025), perhaps aware of some such complications, McConnell et al. add underneath the “Food scarcity in Rome” red bar a second red bar from 40 to 36 BCE labelled “Sextus Pompeius blockaded Italian Peninsula, limiting grain imports”. This is placed a few (three) years after the Okmok-43 BCE start shown by them for the “Food scarcity in Rome”. However, in fact, Sextus Pompeius controlled the relevant seaways and started a blockade of Rome in 43 BCE!—i.e., this, rather than Okmok, may specifically explain or cause the “Food scarcity in Rome”—even if the Okmok climate impact worsened the effects. To quote Garnsey [100] (p. 202): “From late in 43 to 36 Sextus Pompeius was entrenched in Sicily (and his lieutenants in Sardinia and Africa), and able to exploit his naval superiority to cut off shipments of grain to Rome. By 42 many were dying in the city”. Thus, Sextus Pompeius caused “Food scarcity in Rome” and not just from 40 BCE but from 43 BCE. Key to this continuing from 43 BCE onwards and to beyond 38 BCE was first a victory by Sextus Pompeius over the Octavianic fleet off the promontory of Scyllaeum in 42 BCE (Cassius Dio 48.18-19; Appian Bella civilia 4.85) and second a storm and effective victory in the same area in 38 BCE (Cassius Dio 48.47-48; Appian, Bella civilia 5.88). Neither of these circumstances and events can be regarded as caused by Okmok, nor the intervening short-lived pact of Misenum. More generally, this overlooking of the critical challenge posed by Sextus Pompeius conforms to the pattern in ancient history of rather down-grading what was a major and almost central/critical role of Sextus Pompeius. This is all very much an historical outcome of the eventual success of Octavian/Augustus in largely writing Sextus Pompeius out of history (after the eventual victory in 36 BCE by Octavian and Anthony), and hence the usual scholarly treatment of Sextus Pompeius as a sideline and relatively minor figure. In fact: Sextus Pompeius was likely a key part of the existential threat to the Triumvirs as they sought to establish control of the Roman World over a number of years (e.g., [151]), and much more of a threat to Rome and its food supply and much else than the Okmok volcanic eruption. If indeed there was a “fall” of the Roman Republic, it was a relatively long and highly complex set of processes and protagonists, with a prolonged denouement 49–31 BCE, but beginning in earnest well before, at least in the contest between Marius and Sulla in the 80s BCE. It is difficult to detect any actual, let alone significant, role for the Okmok volcanic eruption in these human affairs spread across an empire (e.g., [152]).

3.

Overall, the work of McConnell et al. [146,150] provides an impressive assembly of high-resolution data. But, even so, some of the fundamental historically associating, and historical-narrative forcing, information that they list and invoke as explained by and so linking with the Okmok volcanic eruption involve significant question marks over the relevant exact temporal placements and causative associations. It is also good to remember Garnsey’s [100] (p. 205) conclusion from a review of later Republican Roman history: “Flood, pestilence and harvest failure play minor roles in the ‘famine narratives’ of late Republican Rome”. Although he was thinking about Etna 44–42 BCE and possible short-term effects (and not yet Okmok), Sallares [153] (p. 20) nonetheless captures the bigger picture historical canvas when he concluded: “However volcanoes did not have major long-term effects on the climate or the economy in classical antiquity”.

5. Conclusions

The relationships between changing climate and human history are complex and numerous depending on timescales, contexts, and the many interconnections between human and ecological factors. The object of the investigation also requires care: what often is characterized as “collapse”—and which then needs a cause (and is this climate?)—can instead better be interpreted as resilience via transformation. The Late Bronze Age to Iron Age transition is a case in point. Although some specific cases of rapid historical change may relate to dramatic immediate multi-year climate shocks in potentially vulnerable regions (e.g., [122]), other areas of the East Mediterranean experienced differing climate contexts and were situated in differing geographic/ecological, economic, social, and political contexts, and this overall regional transition occurred across a considerable time-span [117]. Hence, Broodbank [154] (pp. 460–461) argues that “climate played no significant part…This transformation began and ended with human actions”. As in this case, rarely is there a binary situation, rather shades of nuance and difference at many scales.

Two general categories for analysis emerge. There are (i) general longer-term climate variations and plausible associations of these with human history entwined with other social, economic, political, and ideational issues, but also (ii) some circumstances wherein immediate shock episodes of back-to-back or multi-year years of climate extremes can trigger crisis and thus force potential disaster or change. In both cases, the specifics of a climate association beyond general background factor need close argument. However, this is especially pressing for the second category. If we wish to try to establish a particular causal association between a climate episode and potentially historically forcing event, like famine—and so move climate forward from one of several elements of the context for human decision-making to the primary concern—then it is critical to establish a tight temporal, spatial, and causal connection, not just allege a general approximate coincidence. As reviewed in some cases above, all too often the lack of careful focus on details is a weakness in work trying to propose climate as explaining history.